I? I), _ . V V I . V V . V V . V . I V . o .I V. a . . V I I. I I . V V . V . . . . . . V V. V V. I . . . I V . . . I . V . . . I . . V . V . . I . V . V V V I I V . . V. . . .. . . I I . V V . . . V. . V I V . I . . . . . V.V V V . V. VI .. .. V . . V VI I . . . . . .V . . V V . I . .V I V IV. I V . I V V. I . I .V V . IV . V I I V . .. . I . I VI V. V V V I V . V I . . V . V V I V . \V. V. IV . . I I I V V . . V . V I V I V V V V V . . V V I V I V . . V I . . V . . I. I . I V . I . V 1 U.- V . . V V I. .. . V . V V . . V V .. . I V . . . V . I . I I I I I I. Q I V I .I I I I . V. . I VVV I V I I V I I I I I V I . VI I V I I. V .I I V I II I . I . V . V I V . V V . . I I V I I I I I . V V . . VV . IV . . . . I . V . V . V . I.I . .. V . I . I. . . . I V. . V V I V V I V .I . V I . I .I I I. .. . . V V V V . . V . . . . . . .V _ . . .V I I I .V . V I I . V . V . . I . I . V V V V . . V V V I V V . . V I .I I .V I V I V V I V . V V V V . V V I I.II . . V . .I V V .V I x I I . . I I I . I V II I I I I I I V V I V . V .. V V V I . V I I IIV . V . I . . . . I .. .V I . V . V I I... V V I . . .. I II V I . I I . . I: I I . . . I . V . . . V V . . . I . . . . . 1 VI . . . . . V V I . I V . . . . V . V IV _ V. I V V V .V V I . V I V V I I .IV. . . . I . I . I I. V . I V . V .V . . V V I. IV. I . . V. V . . . .. . . V I II \ | V I V V . .V I V V . . V . . V I V V . I I . . . V V . . I I I I - I . I I V II V I ‘ I V I I I I I I II. .I I I V I I . I I I I . .I . I I . . . . I I . V . . I I I V V . V . I . . I . V . . . . . I V . . V I . . . V . . I . I . IV . . V V. I . I. V V I. II V I . .I . V I V . . . I V V . I V V .- .V . . . I . .. I ...V .... V .x V. V . V . V V V V V V V V V . V . V I . I V V 5 V . V I . . . V V . . I . V I . I . V. I . V V V. .VI . . I I . V . . . V V . V V .I V . ..V V . I. I V V V I . . . .I V . II. .V I . . V .V I V . . .V V . . . V V V I V . .V. V . IV . .. V. . . . I. . I V I . . V I V . I . . I. V V V V I .I V V I V . . II V I .0 VI .II I . V . . . V . V V V . I V V V V V. .. . . V V. .V .V I V . I V .. V I V . V . .... . s .I I a .I I I I . I V . . . I V . .I . V . V V . . V . . V V I V V .V V V I .. . . V I . . I. . I V . I V V V . V V V V V I . . V V V V. . I .V. I . .. . V I i. I I V . I ..V I . V. V V I . .. V . . V. I I . V I. V IV V V V I V V . . VlV.... . V V I I . VVI V ..I V . V. ..I I V I I. . V I . .V V . . . I . V . I V V V V V V V V . _ . V V . . I V I V. . V I V V V V . IV V . I . . II. I I V I V. . V V . I . . . . V V V V I . I . V V.. . . V. . V I . V \ I . V I . I V V I I . I V V .I V . . . V I I .. V V IV.. V . V V .. . .. . . . . V . I .V I V I I V . . I V ..V . V . V I V . I. V . I V . V. . V V V V . I V . . .V V I V V. I V V . V V .II .V. .V . . V I V . I I I . V . V V . . .VI VV. .I . . I . . I . I. . VI. . . . . C I I I I I I I I. I | I V I V . V .V. V V V V .V V V . I V .. I .VIVI V. IVI V V V V I V . V . . V V V V . V . I V I V I I I I I . I. V V . ‘ . .I V . I I . V I V . I. V . I I V V .I I V . .V V I I . V I I V V. I. I V ..I I V . I . V . V . I. V . . . . . IV. I. I‘V V .I I I. I I . I V .I . V .I V V . V I . . . V I V . V V . V . . V . .. .V V V .V II. I V. I V V I VII .I V V V. V V V . ..V I V . . V I . I. V I . I O I II I II I I. . I . . I . V . . I V V V I V I V. . I. . V V . . II . V V . I I . I V V . . V V .. . V . . V I V V . I.. V. I . V . I V V V. I . I . . V V I V V V V I I V . I V I V V V V . I.I . . ... I . . I V . V . V I . I .. . V V V I. . V V. . . V V V V V .V V V IV._ V... I.IVV V . V V V . . I I . V . V I. V V . V I I I V . ‘ II o I V V V . I. VVVIV I V . I . .. I V . V I V... VI I I . V . I V V V V . I . . V V . V . I . . I I V. V . . .V . . V V . V V . I V . V.. . V V I. . V I . V V V V I V V I I V V . VVV .. I . I V . . IV . . V I V . I V . . V I I. I. I . I . . V .. I'V I. .V . V .V V V V V V V V I . . V I . I . . . V V .I... . I I. .\I I V . I . V V V I. I V I I I I V I ' . I V .V V V I V V . . V V m .V V. V V V I . I. . . II.. V O n I l I D n h a V . V V . V I I I.V V V I V. V I . . . .II . V V V.I I. V . V. V V V V .. V V V . . V V I . I . . . V I . V... . . . V . . .. . V . V . I V I . . I . V V . . . V V V . I . . I . .I I . I .V . I I I . . I I . V . V. V I V V I I V V . .V. V . V I I V V I I V . . . V V . V _ . V . VIVV V . V. I V . .. . I V V V I I I I . I I V . IVII I. I I 0. I I I I II I II I O I I . . . .V . . V . I I . . . I I V V . . .. . . V I V . I I. .. . I I I . V I I I . V V . . V . I. I I V I . . I . V . V . V V. I . II . V I I .. V . I V .I . I V V V I I . . . V I V . V V V . V V . .V . V . . V V. V . V V . V. I . V . V V I .V V V V V V I V . I. V. V I . V . I I . . V. I I Q . V I . V. V I V . . V . . . . V .V V . I .V . A V . . .I V V V V V V . V . . V . . V . I I .V . . . . V . . . I . . I . . V I . I I I I I . . V . I . V . V . . V. . . . . V. . . . I V . IIVV . . V I I . .V V V. .. I I. I V . I . V . . . .V V V V V. V V . V . I . I . .. V . V . . . V V V V . I . . I . V . . . V I V V V . I V. I . . V .l V V . V . . V. . . . . V I V . V I .I I I . V V . V . V . V V. . V V I I V I . V V I .I. I V I . . . . V . V .. V V I V I VVV .. V V V I V I I. . I I V V V . . .V . . . I . . I V I I . . V. V V V .I V I I V V . V . .. . . V .V .. . V. V. I . V .. V V .V . V I . V V I I . V. I . . .. V . V I VIII I... I I V V I I V V I I V . . V V . . . . . . V V . . V . . V V V I I I . V V. I . I . . I I . I V V .I I V V . I I I I V I . V V . V . . I . I I V V V . . V . . V. . V a V . V . . I . I V . . I . . V V . V. . . V .V . V . I I. I I V . V . . V V V V . . . I V I . I V V V . . V V I V . .. . V V V . . V I . I . V . . V .. . . V I . .V V I I. . . . I . V I . V. I. V . I V I . . . V. V V I V. . V V . . V I V V V VI. V . I . . . V . . V . I I . I . V . V V . V . V V V . I I . I V V . . V V I. . V .V . I V . . V. . I ..V . V . V V V V V . . . V . . V . . . . . r I. V I . V. . .V I . I . V V V VI . .I . . . V V . . . V V V . . V. I I. . I I . V . I . I . .. . V V V V V . I I I I . V V V V . V . . . V . . l . .l I V V I. I I . . I V V V I . I . . V V . I I i I I a V I o I V I . V I. . V I. V . I . . . V V V I V V V I I V . V I. V . . V. I . V V I I . V V . . . V I . . I . .I . V I.. . . I. V V . . .V . . V I I .V . I V V V I V . . .. I . V V . V . V . V I V .V I I . . .V . .. . I I I V . I . I V . V V V . . V .. V . I V V. V I I V . n V V . . V V . V . V . . .. . V I . V V . I V V . I . I II I V . I . . V V I V IIV . ‘ I I V . . . VV . V I V .. I. V . . I . . V . I I . V . I V I . I I o . I I O I I 4 . V . . .I I . I I I I I I I I I V . 0 V . V V .. V . ... I . . V . V I V . . . . I . . . V I . . V I V V. . . . I V . . . V .. I .4 . . V . . . V . . . I I I D II I V I . . I: .. . . . . . V . V V. . . . I I . . . V V I V I . . . I V V V I . V I . V I I I I . V V . V v I O V I I I I V . I . I l V I V . I V I I I I . V V V V V V I . . V V V .V I V . I . V. I I. I I . V V I V . I I V .. I . .V V I . V IV V . I I . I I V V V I V . I. .V . V . I V I . . V V V. V .V . . . . . . I I V I . . V . . . V . . I . . . . V . I V . . V V V . . V . V V .. . V I . . V V V I . . I . I V V V I . V V V . . . V V . . V . . I V I . I . I V V V I I V .I I Q I V I I . . . V I V V VV V V . . V . VI . . . I V . V .VVI . . V . .. . . V I . . V . V V I V V . I V .V V . . . I I I V. I . . V . I . V . . . V V . . V V I V . V V V I V V I V V . I I I I I I I I V . V V V V V V V . .I V . V V I . . . I . I I . V V V . .. . I I . . . I I I . I V . . I V V I I I . . O .I . . . .. . . V. . I . I O V V . . . V o I I I I V I I I I . . I . V . V . V I . \ V I V V . . I V . I . . . . V I V V V . V . I I . V V . V V V I . . V I V V l I . I I I I . . . I V . . V I . . . V . V . I . .. I . . I . V . . V V . V . V I .. V V . I . V . . . I . 'IIIIIIII I I I I I I I I [III meme: ESR SITY LIBRARI Illlllllllllllllllllllllll llll‘ll Will 3 129300 This is to certify that the dissertation entitled Therapeutic Alliance, Patient Object Relations and Outcome in Psychotherapy presented by Gary John Gunther has been accepted towards fulfillment of the requirements for PhD degreein Psychology «a. flu Major professor Date \\ Lug! qi MS U is an Affirmative Acn'on/Eq ual Opportunity Institution 0-12771 _...__. _ 4 _—-—_—-—- ___,_____.-.——-— LiERARY Michigan state University 1 1...- PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE “ MSU Is An Affirmative Action/Equal Opportunity Inflitution CW ”3'94 IF'“ \ has... L.3_3_ hee-. “~~~.- THERAPEUTIC ALLIANCE, PATIENT OBJECT RELATIONS AND OUTCOME IN PSYCHOTHERAPY BY Gary John Gunther A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1991 Patie" a IQI n / H ’ abolaZLe/O‘ZC- ZZCIEESLZC-" ‘ 1 . . 50:2.61' ‘ ‘ es A an-.- engined the .V‘u p :eeuveures' Y“ . a fi ‘ &.‘ FI‘ L‘II‘Q ; " l ‘ b He‘a4l0“c ‘l Q :A'». ‘«4\Q ‘ v“ Q’ (sh ‘ycuoe.a 5..- n O a .: “4“;6’.‘ a ‘H‘ P kwf‘r “'7 ‘a '-v-.'_ 3““: “‘7 ‘ . a yc‘ f"(fl.‘ ‘A.\“ v- fl.’ I u ,_ "EC'M. \‘.' 'I’e ‘ I 0 ‘1 & ’7r"_’/‘5 ("’l Ki ABSTRACT THERAPEUTIC ALLIANCE, PATIENT OBJECT RELATIONS AND OUTCOME IN PSYCHOTHERAPY BY Gary John Gunther Patient and therapist factors which moderate the alliance/outcome relationship have proven to be increasingly important in the understanding of the complexities inherent in psychotherapy. This study examined the interactive effects of patient early— therapy object relations on the therapeutic alliance and psychotherapy outcome. Object relations refers to representations of self and other, and the interactional dynamics between them, as well as the interpersonal manifestations of these inferred psychic structures. It was hypothesized that object relations- —especially for patients with lower levels of object relations—~would interact with the therapeutic alliance, moderating its relationship to outcome. Forty-one outpatient psychotherapy cases were rated for (a) level of early-therapy object relations based upon rating of interpersonal episodes using the (Object Relations Scale, and (b) therapeutic alliance lacrcss psychotherapy sessions using ratings from the (lalifornia Psychotherapy Alliance Scales (CALPAS). The principal findings were: (a) Object relations is predictive of the initial, but not the middle or final therape LZCESECCLVE O.‘eCt C. 1 .‘OIa" a A VA. jec: I l‘. u hr- q C e relati theta; rela: ‘..'.€‘.’ (3:38;: L C ~r02 e; ‘“ t. . H .e e9: ‘ 9“~ 5“.“8 O; « Q'fiw I: :H‘as:0 i “fl ‘ ..b n: “v- 1. - H‘ I he“ Ct;ve c A‘ =VK~e Cymfi‘ 1”H»Or “‘\"Ca“* h i. {'e Q. ~ h: “s *k Endese f ‘ ‘ h «.4 "2“ ?~ 5?: Q~~*Ce Gary John Gunther therapeutic alliance; (b) Object relations exerts an interactive effect across treatment: patients with low object relations have lower levels of initial, but not final, therapeutic alliance, while patients with high object relations establish and maintain a high level of therapeutic alliance throughout treatment; (c) Initial therapeutic alliance is predictive of middle and final therapeutic alliance; (d) Initial therapeutic alliance is not predictive of outcome, while the final therapeutic alliance is broadly correlated with outcome; (e) There was an overall agreement for ratings of quality of process and outcome for psychotherapy from the perspective of the therapist, patient and in terms of symptom reduction; (f) Neither object relations nor overall therapeutic alliance was jpredictive of outcome: while 71% of the patients showed some symptom reduction, there were no consistent nor significant patterns as to which process variables may have accounted for the improvement. The implications ‘of these findings for future research and clinical practice are discussed. ~. . H\ ‘ ...- Q Ebuxn cecaae :0: 23¢“- i ck. EE'S. Men Ralf? - .ze ACKNOWLEDGEMENTS This study used data collected over the last decade for the Michigan State University Psychotherapy Research Project, under the direction of Norman Abeles. I want to thank the patients and therapists who have participated by the anonymous sharing of psychotherapy data, as well as the various research coordinators who have collected and organized the data. I want to thank George Ankuta, Tim Eaton, Janice Gutfreund and Michelle Klee for giving generously of their time and clinical expertise as raters of the object relations and therapeutic alliance scales. Janice Gutfreund has been a consistent and indispensable research associate over the last few years. I want to thank her for her help and support which goes back to the very beginning of this project. I also want to thank my dissertation committee. Gershen Kaufman provided helpful feedback on several of the conceptual issues in this study through his broad ‘understanding of object relations theory. Robert (Jaldwell was extremely helpful in sharing his knowledge (of methodological and statistical issues. He helped me iv in developing zezltodologica. has been n it ' I .P" "'V" ”7' We- Wu uU-.Ou\. . J ‘ CLIECC KILLEI . Raina thue:.;C 'offi v ‘~--~h> .. ‘ . i baud-Ab “‘e .O 4 5 Chat .-.:. in need, ‘fighfi of psyc “W! 'e'-2~"~" and g ~r ‘-I :1 '-~- . c.c:'s a‘.’a;‘ a” ‘ ~ c.‘ . « fig“ ‘ bct’V-H :0 K96“ y \g; ““~ .ensia. ‘ .h k .. he Q‘\“ V‘chates ’- u-s “-J ke~~ . ~‘*~;ed a 1 4 Drah‘:‘ ~'*L O“ ‘ HE" F :41 a I 5‘ C:‘-.h ‘l‘. «‘6‘ § “ tie _ ‘e59aw ‘ r; ’ J‘ in developing a very solid and efficacious methodological foundation to this study. Bertram Karon has been an important and inspirational figure throughout my graduate education, both in terms of his direct influence as a teacher and also through his academic "offspring" of professors and supervisors. He taught me to respect patients—-especially those who are most in need, to understand the broad and lasting impact of psychoanalytic psychotherapy, and to develop a healthy and strong identity as a clinical psychologist. Norman Abeles, as both guidance and dissertation committee chairperson, has consistently provided direction in my graduate education and the development of a professional identity. He has functioned as a realistic and practical force to help temper and guide me throughout the dissertation; he was always available, extremely patient and kind, and helped to keep me on track. Dr. Abeles has been very influential in my academic and professional identity—— he emanates that mixture of pride and humility, he has exemplified a healthy balance of the scientist— practitioner paradigm and helped me to feel proud to be a clinical psychologist. Michael Teixeira, while not directly involved in 'the research, has helped me to understand the concepts (If object relations and the therapeutic alliance in a very deep anc invaluaox 5:; 21; graduate e. zac‘erlies the : r “.‘:A ’ an s s b..‘b ‘ .h ~ “ F‘H . do. “y' "5 .1 ~98: h " is . fl n .D.. ‘. e L o; r ‘ ' a; ' . -“e ‘ongrter: 36-;ef 1 life (' gratem; f0? ‘ :0 03* c.. 0-,, y. s ‘hbu‘e very deep and abiding way. In addition to his invaluable support and encouragement throughout much of my graduate education, he has taught me, in a special way, about the "alliance" and "real relationship“ aspects of psychotherapy, that is, the connection which underlies the work of therapy and of all relationships. For this I am grateful. Finally, I want to thank Kathy Kowalski, who has been with me during the most difficult and demanding times of this research. She has made sacrifices toward the "long-term good" and shown an understanding of and belief in me that really made the difference. I am grateful for her patience and love, and I look forward to our future together. vi . « .. .N. ‘9 1 Q 5. e‘ e .. I.‘ pg p» ...‘ av ad 0‘ fi \ hw§ we st 0‘. I. i 0‘ ‘ uV Q -. ‘ u. \ e r. . no w. an. ...... .. .31. 4 .C.e. S . . 3 Ce T . . S S A: . .. .. a“ e a C... t 3.. .C 7, ... n18 .....l 1.. a e e e «We .. 2w St. . Hy Ca 0 Der“ C ‘Hu 0 C )ha. ‘5 C hh~ v a v.1ha. I e D. a A... h S S I h . Hi . a p e «c y o I. & a e o It $ L» .s CE .3 .fm .fimnflmfliawe .0000 OOO.\A.~ P. ‘ a. n u u .R u "U i M. '7.“ I NC A e MM - H “Rd “c “an “iv. “gt “U‘ D; an \ “Us “Q “U H5 u u 1 . e s a .e 1.»! PI PV. v f s c. O 0 MW W n O S "Nee .nH Kid 5 Q G S Qi‘ 3 Q5 SRAV VMO‘IQAJvVedevn" QM vu'.‘ "VJ ~u..£ ‘H .5 .05 VA 0 e V § Y A vNe the Thin N: "Nee “Nee Y ‘ S n: as“ .e .n... .d MIMI ... -~ u u fie . o a o.- o A\~ . VOW an. I?“ ‘V \...H TABLE OF CONTENTS List of Tables. List of Figures Introduction. Therapeutic Alliance Early Therapeutic Alliance Formation. Contribution of the Therapeutic Alliance to Psychotherapy Outcome Other Studies . . . . . . California Psychotherapy Alliance Scale (CALPAS) . Object Relations Measures. Object Relations Scale. Empirical Research. . Interpersonal Checklist (ICL) Summary and Objectives Hypotheses Method. Data Therapists . General Procedure. Subjects Instruments. . . Data Sampling and Rating Raters Results . . . . Hypothesis Hypothesis Hypothesis Hypothesis pronJH- Discussion. . . Hypothesis 1 Hypothesis 2 and 3 . . . Implications for Research and. Practice Summary. References. vii ix 10 12 14 29 31 36 44 48 52 56 67 77 77 78 80 82 83 96 99 101 109 114 127 144 147 147 156 164 167 169 Appendices llll. C... C. T. v. c v. e «C 5.. r. «V e a I. .n“ v. P. e P. e. w. ”W e» e; as. a. I. A” .fiy n u. v. R. 0 e. 0 9A 3 C D be. Appendices A. Literature Review: Patient and Therapist Variables Associated With Psychotherapy Process, Outcome the Therapeutic Alliance. . . . . . Patient Variables. Ego— Strength. Motivation. Object Relations. Symptom Severity. Therapists' Contribution . Therapist Personal Factors. Training, Style, and Interventions Therapist Technical Factors Therapist Relational Factors. B. Therapeutic Alliance Scale (CALPAS) C. Object Relations Scale (OR Scale) D. Interpersonal Checklist (ICL) E. Outcome Measures: SCL- 90- . . . . . . Post— —Therapy Client Questionnaire. Post— —Therapy Therapist Questionnaire viii 189 189 190 195 197 206 210 211 217 234 265 273 275 278 279 281 286 T." ' 43-8 1, $ m'J’U ITO ('D r' '1 T—U.’ .'< U) .‘< .'< < (I) (D (U (D (I) r") (I) 0) 0) v '0) '— .‘l' O 0) C) :1 (T) r). 0! ’ l o . ' 1 D m F) : i C) O. '1 O ‘. (9 J r) C) '1 l 9 U7 0 .‘J‘ O (7 {I (D (I) ('7 (n *4 ‘ (D '1 f' r1 :1 l i f ."1' I (7 !"i(‘fll '1 (I) .'7'Ih(OO 0m F7 HF)! 'I'JOU) I" ? .'1(‘D(D '1 "1 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 10. 11. 12. 13. 14. 15. 16. LIST OF TABLES Potential Interactions among between Therapeutic Alliance, Object Relations across Sessions . . . . . Means and Correlations for Object Relations Scale . Mean Scores for the CALPAS by Session Mean Scores for the CALPAS by Session and Level of Complexity Mean Scores for Pre— and Post-Therapy SCL-90—R . . . . . Correlations between Therapeutic Alliance and Global Severity Index (GSI). Correlations between Therapeutic Alliance and Psychotherapy Outcome Using Different Outcome Calculation Formulas. Correlations between Post Therapy Client Questionnaire (PQCL) and CALPAS, SCL-90—R, PQTH. . . . . . . . . . . Correlations between Post Therapy Therapist Questionnaire (PQTH) and CALPAS, SCL— 90- R, PQCL. . . Mean Scores for Pre- and. Post— —Therapy SCL- 90- R by Level of Complexity Correlations between Pre- and Post— —Therapy SCL- 90— R and Therapeutic Alliance by Level of Complexity . Therapeutic Alliance, Therapy SCL— 90— R. . Correlations between Therapeutic Alliance and Psychotherapy Outcome by Complexity Level Using Different Outcome Calculation Formulas. . . . . Correlations between Post Therapy Client Questionnaire (PQCL) and CALPAS, SCL— 90— R, PQTH by Complexity Level. . . . . . . . Correlations between Post Therapy Therapist Questionnaire (PQTH) and CALPAS, SCL-90-R, PQCL by Complexity Level. . . . Correlations between the Interpersonal Checklist (ICL), OR Scale and Therapeutic Alliance by Complexity. ix Complexity and Post- 70 102 107 111 115 117 121 124 126 129 131 133 138 142 143 146 LIST OF FIGURES Figure 1. The Basic Interactional Model. . . . . . 68 Figure 2. Means for Initial and Final Therapeutic Alliance by Level of Complexity. . . . . . 110 The SILLC', ‘1: Ffi ". A .' aag‘a..\.e ..ab ‘ P: 5988 and 0| 9“ V ‘ 'V,AY\ \ Pfir 'f1 INTRODUCT ION The study of various aspects of the therapeutic alliance has been at the center of recent psychotherapy process and outcome research. Two general qualities about the therapeutic alliance seem to have contributed to this phenomenon. First, it has been a robust variable, consistently accounting for a meaningful amount of variance in a variety of experimental paradigms (e.g., Frieswyk, Allen, Colson, Coyne, Gabbard, Horwitz & Newsom, 1986) Second, the notion of alliance or working relationship is so fundamental to most forms of psychotherapy that it possesses an almost universal applicability. Bordin (1979) stressed the generalizability of the therapeutic alliance for a wide variety of psychotherapies and viewed the alliance as a principal factor in therapeutic outcome. Luborsky (1984) stated that the therapeutic alliance is essential in the establishment of the therapeutic process and decisive in influencing the eventual outcome of psychotherapy. Although the concept originated in and has been more closely identified with psychoanalytic and psychodynamic areas, divergent sources of re- :biquitous cc~ expirical Lit: The gene: alLiance :ew deal with the :esis:ance th: to as the 'agé between the w CT sources of research and practice have made it a ubiquitous concept in both the theoretical and empirical literature. The general conceptualization of the therapeutic alliance grew out of Freud's (1912, 1937) attempts to deal with the vicissitudes of transference and resistance through the development of what he referred to as the "analytic pact". He made distinctions between the working alliance aspects, that is, the "ego ‘which is our collaborator" (Freud, 1916-17) and the transference aspects of the relationship. Freud (1913) asserted that both rapport-—charged by its unconscious similarity to significant early relationships--which is the basis of transference, and the therapist's technical activities--necessary to guide the nascent relationship through resistances and other conditions ‘which threaten the establishment of a therapeutic relationship--contributed to the patient—therapist collaboration. He noted the importance of developing a positive transference early in treatment to provide the impetus and create the foundation for the therapeutic *work.that would follow. Its importance is made clear ‘when Freud (1916-17) states: ...that a transference is present in the patient from the beginning of treatment and for a while is the most powerful motive in its advance. We see no trace as it 0; analysis 3:: as Diches Freud we. Fro: a Q: PngreSSes m '\ .-_ HE‘ ‘ ' h k“kJ‘h 3“ e‘, . - “‘C:r a~~ “59» L‘Vy 3 no trace of it and need not bother with it so long as it operates in favour of the joint work of analysis. (p. 443) But as Dickes (1975) indicates: Freud was quite aware that transference could not only function as the motive force for the patient's cooperation but transference could be used in the service of resistance. At such a time, transference becomes an obstacle; it does not serve the cooperative relationship and thus is no longer part of the defined alliances. (p. 3) From a psychoanalytic perspective, as therapy progresses into its inevitable stages of resistances, the rational aspects or working alliance components of the relationship gain increasing importance; Dickes (1975) believes that Freud undoubtedly understood the “importance of rational factors in the therapeutic partnership and of the part in this relationship played by sensible motivations and the relatively healthy or mature portion of the ego" (p. 5). Sterba (1934) developed the focus upon the alliance aspects of the therapeutic relationship 'through elucidating the process of the formation of the «abserving ego within the patient. This called for a dissociation enabling the reactions. S .-.—' ; fithn‘s Ca: . 4 dissociation of one aspect of the "contemplative ego," enabling the patient to reflect on his/her transference reactions. Sterba stated: This capacity of the ego for dissociation gives the analyst the chance, by means of his interpretations, to effect an alliance with the ego against the powerful forces of instinct and repression (p. 120). From the outset the patient is called upon to “cooperate" with the analyst against something in himself. The use of the word "we" always means that the analyst is trying to draw that part of the ego over to his side and to place it in opposition to the other part which in the transference is cathected or influenced from the side of the unconscious. (p. 122) This facilitated a separation between those aspects of the patient's reaction which were reality— focused and those which were not. The more realistic aspects allowed the patient to identify with the treatment goals of the therapist. This led to the consideration of therapeutic matters from the therapist's viewpoint and culminated in an identification with the therapist (Friedman, 1969). Freud (1933) recognized the fundamental importance of the observing ego when he proposed that, in successful 6- Q r ‘V- tree .2" " a cab, iron Zetzel ( used the ter: alliance i::e - . ‘ I WWOQ“... V .u.G...4-e ant: .8- ersni . ESQ " ; _.' .vaS 0‘ t. 6‘, ‘He‘5$' ' “..l; ‘ «leg "1 Dow .. Q‘s 0f ,_ gnu 5 treatment, the patient must be able to observe him or herself from another person's perspective. Zetzel (1956) further developed this concept and used the terms working alliance and therapeutic alliance interchangeably. Moving away from the infantile and irrational aspects of transference, she emphasized the collaborative, reasonable qualities of the relationship that aided in the development of a partnership. She noted that analysis of the regressive aspects of the transference can occur only in the context of some degree of mature ego functioning, in particular evidenced by the maintenance of a functional therapeutic alliance. Greenson (1965, 1967; Greenson & Wexler, 1969) differentiated three overlapping components of the therapeutic alliance: transference, the working alliance and the real relationship. Like Sterba's (1934) notion of the collaborative, dissociated aspect of the patient's ego, Greenson's working alliance emphasized the non—neurotic, cognitive aspects of the patient's ego; these aspects of the patient's ego identified with the therapist as well as the mutual goals of therapy. Perhaps as much as any writer since Freud, Greenson elucidated the notion of the "real" relationship, that is, the realistic (non-transference) aspects of the therapist-patient relationship. This is based upon : I it ntificati: :herapiSC- E work (Greensc | carefully dif reactions to enpiricaily—r external woril components of three conponei real relations interact with P v- ‘ ~ . examnle, 1 nqa; 9C.-e:t pe :12: 1 'Y‘H: ' bk LN“; 1n 9- .. \r V“ H§ adia‘nce. ‘ h “e tats~. ‘4 .2“ _ ““Ct*oh; a‘*ZC) d a I 6 >ased upon the patient's growing perception of and Ldentification with a consistent, genuine and involved :herapist. He recommended, particularly in his later work (Greenson & Wexler, 1969), that the therapist :arefully differentiate the patient's realistic reactions to the therapist (more accurate and empirically-based perceptions of the therapist and external world) from more primitive, transferential :omponents of the therapeutic relationship. These :hree components (transference, working alliance and real relationship), while conceptually distinct, -nteract with each other in the therapeutic situation. Por example, the therapist's “real" stance with the >atient permits an identification with the therapist which in turn becomes the basis for the working alliance. Dickes (1975) proposed that there are important listinctions between the therapeutic alliance (the >roader rapport that includes the sum of beneficial lSPGCtS of the therapeutic relationship) and the working alliance (a more circumscribed concept dealing with the mature aspects of the patients's ego functioning). He stated that the therapeutic alliance 'encompasses within its conceptual boundaries such factors as the transference, negative as well as >ositive, the real relationship, the working alliance itself, and 0) factors and r. e‘! L" (D I: O '1 :R‘ .1. '3 0 0’ We . beds 61‘ «a?» 7 itself, and all other factors inclusive of irrational factors and nonrational factors" (p. 6). Phe working alliance, for Dickes, is comprised of: ...such aspects of a reasonably mature ego as the ability to form reasonably reliable object relationships, and some ability to delay gratification coupled with suitable aim inhibition and sublimation of sexual and aggressive impulses. In addition, a certain amount of psychological mindedness is necessary, as is the intelligence to grasp on and reflect upon the views implied in the confrontations and interpretations offered by the therapist. (p. 5) Frieswyk, Colson, & Allen (1984) have suggested :hat many of the current definitions of the therapeutic alliance, which encompass various aspects of patient and therapist activity, are too global and have :herefore recommended——because of both conceptual and research implications--a more circumscribed definition: We believe that such conceptualizations of the alliance are too broad and are subject to diverse interpretations thereby contributing to theoretical controversy and impeding systematic research...we propose that the therapeutic alliance active C 3) :r 31 r; -' (D r: I p- D I) r? all).rfi '? C All?! 8 alliance be more narrowly defined as the patient's active collaboration in the work of psychotherapy or psychoanalysis. This definition enables us to distinguish the alliance from various parts of the patient's experience of the relationship (especially transference) and to separate the alliance from issues of technique. (Frieswyk et al., 1984, pp.460-461) Bordin (1979) has been a central figure in the area of the therapeutic alliance literature especially in terms of attempting to bring some conceptual clarity and operationalization to the concept. He listed four propositions to provide a conceptual framework to aid in the convergence of different approaches: 1) All genres of psychotherapy have embedded working alliances and can be differentiated most meaningfully in terms of the kind of working alliance each requires. 2) The effectiveness of a therapy is a function in part, if not entirely, of the strength of the working alliance. 3) Different approaches to psychotherapy are marked by the differences in the demands they make on the patient. 4) The strength of the working alliance is a function of the closeness of fit between the demands of a particular kind of working alliance and the personal characteristics of the patient and therapist. (Bordin, 1979, p.253) Borcin aspects of ti all alliance: includes (a) of tasks, ani vary a great psychodynar i .; experience tc, behavior cine: the activitie the purpose c of the H ...EI‘a' ~. a 9 Bordin (1979) elaborated the working alliance aspects of the therapeutic alliance by proposing that all alliances possess three general components. This includes (a) an agreement on goals, (b) an assignment of tasks, and (c) the development of bonds. Goals may vary a great deal from therapy to therapy (e.g., from a psychodynamic emphasis on the understanding of internal experience to a behavioral emphasis on concrete behavior changes), having far-reaching consequences for the activities and outcome of the therapy. In general, the purpose of goals is to denote the desired outcome of the therapeutic interaction. The need for a mutual agreement between the therapist and patient that these are reasonable and agreeable goals to pursue is central. This ties in with the tasks of psychotherapy, e.g., how active or passive the therapist or patient will be, the emphasis upon internal versus external phenomena, and so on, will be determined by the nature of the therapeutic tasks. Bordin states that some basic level of trust marks all types of psychotherapy, but when attention is directed toward the deeper and more personal aspects of experience, possibly greater trust and attachment are required. Overall, the bond refers to the optimal interpersonal relationship between therapist and patient, and to those factors that inhibit or facilitate its development. ' ' 1 . .- Wnie tne 1m developed ‘1‘. aS’QEC’LS Oi t. Baily £0175 C ‘ 1 bozd (Bord; Therapeutic Alliance As discussed in the introduction, the quality of the therapeutic relationship is seen as a central factor in the successful outcome of psychotherapy. While the notion of the therapeutic alliance was developed within a psychoanalytic framework, core aspects of the alliance may be generalized to include many forms of psychotherapy; definitions from this broader perspective usually include: (1) agreement on goals, assignment of tasks, and the development of a bond (Bordin, 1979); (2) a conceptualization of the therapeutic alliance as comprised of a transference, real relationship and a working alliance (Greenson, 1965); and (3) and objective behaviors and inferred attitudes on the part of the therapist and the patient which allow them to realistically collaborate in a relationship which is founded on mutual trust and respect, and a commitment to the therapeutic task. A positive therapeutic alliance requires the careful and ongoing attention of the therapist. The therapist must be therapeutically and affectively attuned to the NOTE: See "Appendix A" for a literature review on patient and therapist variables associated with psycho- therapy process, outcome and the therapeutic alliance. 10 'l moment-tome:- As research ': factors sect " and respect : sufficient t - Roth r, it 1 collaborative PSYChothera-Qs, reach a succe number Of con as well as r". 6;, v- “ . cnelapet-tic c T: A ‘s . ) tue 1504.101 EEQCC: .4! 0: pSYChOpa t, moment—to-moment state of the dyadic relationship. As research has indicated (discussed in therapist factors section——Appendix A), mutual positive regard and respect between patient and therapist are not sufficient to insure a successful therapeutic outcome. Rather, it is through the development of a true collaborative atmosphere, where the various tasks of psychotherapy are addressed, that psychotherapy may reach a successful outcome. This process depends on a number of complex therapist and patient contributions, as well as the therapist-patient interactions via the therapeutic alliance. In the review of patient contributions to the therapeutic relationship (Appendix A) the following factors were identified: level of ego- functioning, object relations, motivation and severity of psychopathology. The following were identified in the therapist factors section (Appendix A): personal factors, level of training and style, technical interventions and therapist relational factors. This section will focus on the effects of the interaction of therapist and patient contributions to the therapeutic alliance as they pertain to psychotherapy process and outcome. 11 writers in t: 1975; Marzia Of a therape" psflhOthetag; 'i'CI‘k and pOS_ State .t-he 71 decelgping :- Q‘QCOZe is Q. Some 8:; 12 Earlv Therapeutic Alliance Formation. Several writers in the field (e.g., Horwitz, 1975; Luborsky, 1976; Marziali, 1984) have stressed that the formation of a therapeutic alliance in the beginning of psychotherapy is crucial for subsequent therapeutic work and positive outcome. Hartley and Strupp (1983) state "the most important phase of therapy for developing the therapeutic alliance and predicting outcome is the initial phase (p. 29)“. Some studies have specifically investigated the importance of the initial sessions upon the development and/or outcome of treatment, although not always directly within the conceptual framework of the therapeutic alliance. Ryan and Cicchetti (1985) measured various patient pre—therapy variables (e.g., object relations, psychological mindedness, hope for success, etc.) as predictors of readiness for therapeutic alliance in the initial therapy interview. About 40% of the variance was attributable to pre— therapy factors, the largest portion of which, in both the expressive (psychological freedom) and collaborative (quality of alliance) dimensions, was object relations. Lehrke (1977) noted that the factor which most contributed to early alliance development ‘was the indirect structuring of a safe and supportive treatment situation. Secondly, rather than the therapist's comunicatic; capacity for best predict- thtari and ' predictive c5 Ltpressmns patient 5 oe: engagm i H ‘1 13 therapist's specific interventions or the patient's communications, it was the patient's pre-therapy capacity for object relations and ego—functioning which best predicted therapeutic alliance development. Bottari and Rappaport (1983) investigated the predictive capacity of patient and therapist initial impressions in psychotherapy and found that the patient's perception of the therapist's affective state and therapeutic style during the first session significantly correlated with: (a) therapist and patient reports of symptom change in the next four sessions, (b) the length of therapy, and (c) a linear decrease in patient symptoms after eight sessions. The authors emphasized the critical importance of events occurring in the initial sessions with regard to the subsequent development of therapy. Crowder (1972) reported that patients who were rated as less passive resistant in the first three sessions had better therapeutic outcomes. Also, successful therapists engaged in more alliance enhancing and alliance inhibiting behaviors, and were more competitive and less passive—resistant than unsuccessful therapists, implying the benefit of greater therapist activity early on in therapy. Saltzman, Leutgart, Roth, Creaser, and Howard (1976) investigated therapeutic relationship formation and found that successful therapies we caring, warn“ disruptions . relationship in poor outc. 1 “tbs \.' 19 80; 14 therapies were characterized by greater mutual respect, caring, warmth and understanding. They noted that disruptions and dissatisfaction in the treatment relationship were evident as early as the third session in poor outcome cases. Contribution of the Therapeutic Alliance to Psychotherapy Outcome. There is an evolving research literature elucidating the relationship of the therapeutic alliance to successful psychotherapy. Perhaps the earliest evidence which supported this link arose from the longitudinal psychotherapy study at the Menninger Clinic (Horwitz, 1974). Horwitz noted that the greatest change in supportive and expressive psychotherapies was connected to the development of the therapeutic alliance and he conceptualized this as foundational to the change process. The following is a brief review of the major therapeutic alliance research projects. Penn Psychotherapy Research Project. These studies were some of the first comprehensive research projects to empirically investigate the therapeutic alliance as a predictive component of therapeutic outcome (Luborsky & colleagues—-Luborsky, Mintz, et al., 1980; Luborsky, Crits—Cristoph, Alexander, Hargolis, s Mintz, Labor Crits—CriSto finding was 1 therapy Outcg ‘ A alliance (Mo were identif. the Patient « helpful with the Patient ] word heip t] process; :ke‘ 15 Margolis, & Cohen, 1983; Luborsky & Spence, 1978; Mintz, Luborsky, & Cristoph, 1979; Morgan, Luborsky, Grits-Cristoph, Curtis, & Solomon, 1982). The major finding was that the variable most predictive of therapy outcome was the quality of the therapeutic alliance (Morgan et al., 1982). Two forms of alliance were identified by Luborsky (1976): Type 1 alliance-— the patient experiences the therapist as warm, caring, helpful with the patient as the recipient; also that the patient believes that the therapist or therapy would help the patient overcome their problems; Type 2 alliance--the patient is involved in a collaborative process; there is a sense of shared responsibility with the patient's contribution being equally valued as the therapist, indicating the rudiments of internalization of the therapeutic process. Luborsky (1976) believed the latter alliance type was a more robust and central factor for long-term therapeutic gain and that this probably occurred later in therapy. The research findings indicated that the type 1 alliance tended to increase over the course of therapy while the type 2 tended to be a fairly stable process across therapy. Both types of alliances were found to be positively related to successful therapeutic outcomes. In subsequent analyses, Luborsky compared the ten most and ten least improved cases from a larger pool of subjects. r- for 25 or at and eight le found that s the low inpr type 1 helni: Two of the t the end of t' Tg‘.‘ ye ' ' 5:, oenellts fron ' .. :: 08.811125 preg 16 subjects. After examining cases of those patients seen for 25 or more sessions, a subsample of the seven most and eight least improved cases was generated. They found that six of the seven high improvers and none of the low improvers showed signs of the development of type 1 helping alliances as early as the fifth session. Two of the high improvers developed type 2 alliances by the end of therapy while none of the low improvers had. Luborsky suggested that the best predictor of later benefits from therapy may be the signs of early benefits present in therapy sessions. Utilizing two different methods of measuring helping alliance, Luborsky et al. (1983) found a positive relationship between higher scores on helping alliance (counting signs method and global rating method) and psychotherapy outcome. Vanderbilt Studies. A major contribution of these studies is the development and refinement of the Vanderbilt Therapeutic Alliance Scale (VTAS) and its related scales (VPPS and VNIS) (Strupp & colleagues—- Hartley, 1978; Hartley & Strupp, 1983). The use of multidimensional scales, which considered the components of the therapeutic alliance, allowed for a more fine—grained analysis of patient and therapist contribution to the therapeutic alliance. The VTAS scale consis literature 1." dimensions: . . :.J tn rapist l.. JI respons;b;t; ‘ ' I +H. ‘7‘ *V‘“w 5.3-5 ‘A;S a- 5A.. . L14 - . «his in tn- ‘ treatment 5 e; 1 directly prec the therapep.‘ negative tree The Va :1 arose fI'Om t O'Maliey, 19 divided in . 17 scale consists of 44 items derived from the dynamic literature which were factor analyzed into six dimensions: positive climate, patient resistance, therapist intrusiveness, patient motivation, patient responsibility and patient anxiety. Research using this instrument (Hartley & Strupp, 1983) identified shifts in the therapeutic alliance during early treatment sessions. While therapeutic alliance did not directly predict outcome, the pattern of the level of the therapeutic alliance differentiated positive and negative treatment outcomes. The Vanderbilt Psychotherapy Process Scale (VPPS) arose from the same research efforts (Suh, Strupp & O'Malley, 1986). There are 80 Likert type items divided into the following patient and therapist scales: patient participation, patient hostility, patient psychic distress, patient exploration, patient dependency, therapist exploration, therapist warmth and friendliness, and negative therapist attitude. These subscales were combined into three process dimensions: exploratory processes, patient involvement and therapist-offered relationship. Gomes—Schwartz (1978) found the patient involvement scale (this included patient active involvement and patient negative affect toward therapist) of the VPPS was the best predictor of improvement rates (rated by judges and therapists) as well as targ O'Malley, Sn findings not consistent r.- t’ne associat session and p an attent- ‘ and Silbersit. Slbjects 1m, .| Ler V’ev used, The Q‘ the e“d es .1 ~..e‘ap; Undone 1:0. .. , halfe I‘os ‘verw re 18 well as target symptom change (rated by therapists). O'Malley, Suh, and Strupp (1983) corroborated these findings noting that patient involvement was the most consistent predictor of overall outcome; specifically, the association was almost nonexistent in the first session and increased steadily to the third session. In an attempt to replicate the VPPS studies, Windholz and Silbershatz (1988) looked at VPPS ratings of the middle psychotherapy session (i.e., eighth) in 38 subjects involved in brief (i.e., 16 session) therapy. A briefer version (i.e., 44 items) of the VPPS was used. The treatment population was more diverse and the therapists more exclusively psychoanalytic than previous studies. Outcome ratings were made from all three perspectives and, as before, the results indicated that therapist ratings of patient involvement and therapist-offered relationship were correlated with outcome. However, as Windholz and Silbershatz point out, the question still remains as to why VPPS ratings have not correlated with patient's and independent observer's ratings. The third therapeutic alliance measure developed at Vanderbilt is the Vanderbilt Negative Indicators Scale (VNIS), which assesses various impediments to the development of a viable therapeutic alliance (the specifics of this scale were discussed in an earlier section n r research w'ni predictive o detailed exa noted a high negativistic These fin 1: literature ( SUPPOrtec’ 1 Patient “.1: Connec3i03 extremely C alliances 19 section on negative effects). This scale arose from research which considered patient profiles that were predictive of poor outcome. Strupp (1980a), in a detailed examination of cases from this population, noted a higher degree of hostile, resistant and negativistic patient behavior in poor outcome cases. These findings are convergent with the related literature on patient variables, and they are also supported in the Langley Porter studies. Specifically, patients who had a limited capacity for interpersonal connection were typically unable to develop (or it was extremely difficult to develop) helpful therapeutic alliances in their treatment. Langley Porter Institute. In more recent years, a number of important studies have come from the Langley Porter Institute (Horowitz, Marmar, Weiss, Dewitt, & Rosenbaum, 1984; Marmar, Horowitz, Weiss, & Marziali, 1986; Marziali, 1984a, 1984b; Marziali, Marmar, & Krupnick, 1981) including the development of a widely— used therapeutic alliance measure. Marziali et al. (1981), building upon previous therapeutic alliance scales developed by the Penn and Vanderbilt groups, attempted to integrate some of the essential aspects of 'the therapeutic alliance and focused upon the affective (and.attitudinal aspects of the alliance, rather than emphasizing responses. Hm therapis develop and Hm patient. therapist cc that they a: Expeding tn. System (TAR scal‘95) w'ni therapist C C0ntrib Ctic A 20 emphasizing specific therapist actions and patient responses. This allowed for a general examination of the therapist and emphasized the therapist's efforts to develop and sustain a positive working alliance with the patient. This is reflected in items such as "the therapist conveys a "we" bond with the patient, a sense that they are in a joint struggle against what is impeding the patient". The Therapeutic Alliance Rating System (TARS) consists of 41 items (on six—point Likert scales) which are divided into four subscales: therapist positive contribution, therapist negative contribution, patient positive contribution and patient negative contribution. A therapist total contribution and patient total contribution score are obtained by collapsing the appropriate subscales. Therapist and patient contributions were not examined as a dyadic unit because of previous research indicating that these factors should be considered as separate units; for example, Comes-Schwartz (1978) pointed out that it was necessary to consider the patient's contribution independent of the therapist's in order to realize a significant relationship with the treatment outcome. In the pilot study using this instrument (Marziali et al., 1981) five good outcome and five bad outcome cases were drawn from the larger sample of 25 patients. When the two groups were compared across the four subscales an! significant contribution those who we: sustain a p; (Marziali, 1 therapeutic (iYSEpenc‘eg-t ‘Hr: 3 ‘ u.-.e._ ps‘Irch '1". “‘e find C‘ ‘ F L. I“ UE‘R. 21 subscales and total contribution scores, the only significant relationship to outcome was the patient's contribution. Patients who were able to develop and sustain a positive alliance had better outcomes than those who were unable to. In a subsequent study (Marziali, 1984) the TARS was used to evaluate the therapeutic alliance from three perspectives (independent rater, therapist and patient) in a more heterogenous clinical population of 42 patients in brief psychotherapy (20 sessions). Marziali found that patient and therapist positive contributions were significantly associated with treatment outcome across all three rating perspectives. The patient's negative contribution was correlated with negative outcome, while therapist's negative contribution was unrelated to outcome. Further, there was evidence of these relationships as early as the first or third sessions. Horowitz, Marmar, Weiss, Dewitt, and Rosenbaum (1984) investigated the relationship between process, outcome and therapeutic alliance in a sample of 52 bereaved patients involved in short—term dynamic psychotherapy. The findings indicated only a slight relationship between patient negative contribution and symptom change. Concluding that no patient or therapist characteristic viewed in isolation was strongly connected with outcome, Horowitz and colleagues conducted a: data. This h interactions particular, account the on therapegz IELatiOEShigfi variables) to C0150n, Coy: Hrs: “ ‘: leak-LS as.a‘ys ~.Qd.¢¢te:anc‘ s.pp0rtiv€ pOoryy “(‘0‘ 0th th $ negatl‘.’ t 305%.. “1Ve h 5‘. Se‘vewr 7‘ a 22 conducted an hierarchical regression analysis on the data. This allowed for the exploration of interactional characteristics of process variables; in particular, this was accomplished through taking into account the impact of patient predispositional factors on therapeutic alliance and outcome. The interactional relationship (process variables and predispositional variables) was predictive of outcome. Frieswyk, Allen, Colson, Coyne, Gabbard, Horwitz & Newsom (1986) have summarized some of the key findings which emerged from this analysis: (a) The therapeutic alliance is fostered by the exploration of the negative transference with highly motivated patients; (b) The therapeutic alliance could be undermined by the expression of negative transference in more poorly motivated patients; (c) The therapeutic alliance and treatment outcome was positively affected by the maintenance of a consistently positive attitude in a supportively oriented treatment approach for more poorly motivated patients; and (d) Negative effects in both the therapeutic alliance and treatment outcome could be induced by an avoidance to adequately explore negative transference in otherwise motivated and positive patients. Several other studies have been conducted by researchers at Langley Porter which have expanded the direction- or studies rep: from the 1:24 I Horowitz, Mt- Marrnar, Weig subscales of? based sCaleS Involvement,- Pa'tien: C01,. {8) Patient ThESe f original foo 23 direction of the research outlined above; the following studies represent a cross-section of recent research from the institute. Using the same database as Horowitz, Marmar, Weiss, Dewitt and Rosenbaum (1984), Marmar, Weiss and Gaston (1987) factor analyzed the subscales of the TARS and arrived at five empirically- based scales: (a) Therapist Understanding and Involvement; (b) Therapist Negative Contribution (c) Patient Commitment; (d) Patient Working Capacity; and (e) Patient Hostile Resistance. These factors represent an elaboration of the original four factors, in particular, the bifurcation of patient positive contribution into two patient factors (this will be discussed in more detail at the end of this section). The findings of Marmar, Weiss, and Gaston (1987) indicated that a number of patient pre-treatment variables (e.g., educational level, motivation for psychotherapy, pre-treatment interpersonal functioning, etc.) correlated with therapeutic alliance. The therapeutic alliance positive contribution scales were negatively correlated to stressful life events and to therapist interventions which addressed the patient's resistance. The authors explained this last point as a function of the patient reacting with higher patient hostile resistance in response to the therapist pointing out patient avoidance s interprets: toward the relationshi - c “‘* 9 .- vp- % * kg..- up.“ .. s, scores were patient ' s . address the 24 avoidance of material. Further, therapist interpretations which linked problematic feelings toward the therapist with the patient's early relationships was reflected in lower scores in patient commitment and patient working capacity. These lower scores were understood to be a function of the patient's reaction to the therapist's efforts to address the emerging negative transference, and are in keeping with previous findings (e.g., Foreman & Marmar, 1985; Malan, 1976). Gaston, Marmar, Thompson and Gallagher (1988) examined the relationship between therapeutic alliance and patient pre-treatment characteristics in a sample of 60 elderly patients involved in brief cognitive, behavioral and dynamic therapies. Using a revision of the TARS, the California Psychotherapy Alliance Scales (CALPAS; Marmar & Gaston, 1989), Gaston, Marmar, Thompson, and Gallagher (1988) found that lower patient contributions (i.e., patient working capacity and commitment) were associated with higher degrees of patient defensiveness. These findings are congruent with Foreman and Marmar's (1985) results and prompted the authors to conclude that the therapist must diligently focus on the patient's resistance and avoidant strategies early on in psychotherapy to facilitate an active collaboration. The availability of environment therapeutic patient pr alliance Wi pproac‘nes 94“"3': ant: “nth 25 environmental support was positively related to therapeutic alliance. The overall relationship between patient pre-therapy characteristics and the quality of alliance was similar across all three treatment approaches as was overall rate of improvement. Gaston, .Marmar and Ring (1988), following Foreman and Marmar's (1985) approach, investigated five cases of patients 'with initially poor therapeutic alliances (as measured by the CALPAS) who were seen in brief cognitive therapy for depression. Three of the cases went on to develop improved alliances and satisfactory outcomes, while the other two cases did not indicate improvements in process or outcome. Gaston et a1. (1988) found that, for patients whose alliance did not improve, there was a tendency to avoid dealing with interpersonal problems. Also, therapists tended to focus on non- interpersonal material, leading the authors to suggest that: Therapist and patient avoidant contributions to the task was reinforced through a reciprocal pattern of defensiveness against problematic relationship themes. (Gaston, Marmar & Ring, 1988; pp. 12-13) While the improved group also began treatment with a problematic therapeutic alliance, they were able to overcome ‘ their con: add: ss a1 patient's and worki: evidenced discussinc a 26 overcome the initial resistance and move on to discuss their concerns; furthermore, therapists tended to address affective and cognitive aspects of the patient's interpersonal problems. Patient commitment and working capacity was higher for this group, evidenced by a "reciprocal pattern of communication at discussing relevant problematic issues“ (Gaston, Marmar & Ring, 1988; p. 12). Further analysis revealed that these two groups differed with regard to pre—treatment levels of avoidant strategies to deal with depression. Some important findings and general conclusions should be highlighted from the Langley Porter studies. First, these studies tend to support previous findings: (a) patient's and therapist's ratings support the relationship of patient contributions to the alliance to therapeutic outcome (e.g., Luborsky & Auerbach, 1985; Marziali, 1984); and (b) patient pre-treatment characteristics influence the therapeutic alliance and outcome (e.g., Moras & Strupp, 1982; Piper, deCarufel, & Szkrumelak, 1985). Second, as Foreman and Marmar (1985) proposed, the importance of identifying and addressing the patient's initial avoidant strategies or defensiveness, and/or the therapist's avoidance of problematic transferences, as it affects the early therapeutic alliance formation, was shown to be a critical factor in accounting for some of the equivocal findings in e~7-Pi1'ically the origina; such as the conceptual?“ Horowitz! V. :19 use of T. de“. 27 findings in the literature. Third, the five empirically derived factors represent a refinement of the original TARS, and are compatible with other scales such as the VTAS (where five of the six scales conceptually correspond). As demonstrated (e.g., Horowitz, Marmar, Weiss, Dewitt, & Rosenbaum, 1984), the use of more sophisticated statistical analyses (e.g., hierarchical multiple regression) have begun to address the complexity and interactional nature of the alliance and allows for more subtle statistical relationships to emerge. A particularly promising avenue has been the delineation of the two patient factors. Marmar, Gaston, Gallagher and Thompson (1989) have stated: Taken together, these findings suggest that the patient's commitment to treatment and capacity to participate in treatment are factors generic to the change process in different patient populations and in diverse therapies. (p. 20) Marmar et a1. (1989) proceeded to discuss the important clinical differences between the two factors: The distinction between commitment and working capacity appears to be meaningful. There are patients who may be capable of engaging in the work of dynamic therapy with the capacity to self— 28 observe, self-reflect...who, for other reasons, may not strongly value treatment or be willing to make the necessary sacrifices. Conversely, there are patients who are highly committed to dynamic therapy, value the treatment, attend regularly, and yet have considerable difficulty engaging in the working tasks of dynamic therapy. The emergence of these two patient positive dimensions supports the theoretical distinction between the therapeutic climate, which may reflect primary patient commitment, and the therapeutic alliance, which implies both commitment and working capacity (Greenson, 1965; Horowitz, Marmar, Weiss, Dewitt, & Rosenbaum, 1984; Langs, 1975; Marmar, Horowitz, Weiss, & Marziali, 1986). These authors described instances where a positive climate in the relationship was based on a misalliance in which there was a mutual avoidance of engaging in treatment strategies...and other instances where the patient was in the throws (sic) of intense hostile feelings towards the therapist, yet maintained a strong commitment to understand the origins of angry reactions. (pp.20-21) Thus, in addition to the differentiation of patient contribution into patient working capacity and patient con: factors anc therapist-p: patient fat“ (1989) expha athiValenCel mindedges S their Prom. factors. toward the Aomote the is 1 . . «'19 C0nvergent' afists” . Eiffel-ext ‘5 .ad hree v Nesures) : 29 patient commitment, the distinction between these factors and a positive emotional climate in the therapist—patient interaction has been made. The patient factors appear to reflect some combination of the four factors of the patient's contribution to the therapeutic process (discussed earlier) which the patient brings to therapy. The positive emotional climate is similar to Freud's (1912) notion of an unobjectionable positive transference. Marmar et al. (1989) emphasize that tolerating frustration, ambivalence, and a capacity for psychological mindedness and dealing with one's own contribution to their problems, are central aspects of these patient factors. In contrast, a friendly, positive attitude toward the therapist is most likely insufficient to promote therapeutic change, especially when the patient is lacking in the patient factors. Other Studies. Finally, given the number of therapeutic alliance measures which exist and have been employed in various studies, the matter of the convergent and divergent validity of these scales arises. Tichenor and Hill (1989) compared four different therapeutic alliance measures (one of which had three versions, for a total of six different measures) in order to determine the correlations u ; between t..e. the Califor: Vanderbilt T Helping All _. | Of these me: I I a . rater re“c. ‘ ' Pacient Sin correla 9-: LIOn WAI‘O wEre '1: high intern.- l: . . re“ablllty the Other ClientlbaSe orre} ‘ L8 “’1 that t“e pen allia 30 between these various measures of treatment alliance; the California Psychotherapy Alliance Scales (CALPAS), Vanderbilt Therapeutic Alliance Scale (VTAS), Penn Helping Alliance Scale (Penn), and Working Alliance Inventory (WAI—O, C, T) were the measures used. Further, they attempted to assess the validity of each of these measures with a particular interest in inter— rater reliability on observer-rated therapeutic alliance measures. Using a small sample size of eight patients in short-term treatment, they found a strong correlation among three of the therapeutic alliance :measures--all observer rated. The CALPAS, VTAS and WAI-O were found to be equivalent as measures of alliance; the authors conclude: "All three measures had high internal consistency, high inter-rater reliability, and high correlations with other measures of working alliances (Tichenor & Hill [1989]; p. 198)." The Penn measure did not significantly correlate with the other measures. Also, the therapist-based and client-based ratings of the WAI were not found to correlate with the other measures. The authors suggest that the perspective from which the therapeutic alliance is measured may be a critical factor. The Cal (ALPAS). ‘ (Marnar & Gd the TARS and Ef'j. “ ‘tiOns .es, .«Esentlh liarc « e ( p The C 31 The California Psychotherapy Alliance Scales (CALPAS). The most recent version of the CALPAS (Marmar & Gaston, 1989) reflects the transformations of the TARS and the conceptual understanding arising from factor analyses. Because different definitions of the therapeutic alliance have emphasized different aspects of the alliance, Gaston (1990) suggested that the therapeutic alliance was best seen as multidimensional. Basically, these dimensions include: (a) the therapeutic alliance--this is compatible with Freud's (1912) and Zetzel's (1956) positive affective relationship, and the attachment of the patient to the therapist; (b) working alliance-—this is in keeping ‘with Sterba's (1934) and Greenson's (1965) definition ‘which emphasizes the patient's capacity to effectively and purposefully work in psychotherapy; (c) therapist capacity for empathy and understanding-~this dimension underscores the therapist's empathic and technical abilities; and (d) agreement on goals and tasks-—this is from Bordin's (1979) perspective on the collaborative aspects between therapist and patient. Gaston (1990) concludes—-"These four alliance definitions are complementary and compatible, each representing a relatively independent dimension of alliance (p. 148)." The current version of the CALPAS has integrated | these dimer. 9 capacity LP patient C0175. consensuS [I involvenent ALPASnthe judge—rated discussed it present stt rated on a there is 0:14 scales. Til scale. 32 these dimensions into five scales: (a) patient working capacity [PWC]; (b) patient hostile resistance ; (c) patient commitment [PTC]; (d) working strategy consensus [WSC]; and (e) Therapist understanding and involvement [TUI]. Of the three versions of the CALPAS--therapist-rated, patient-rated and clinical judge—rated, only the clinical judge—rated form will be discussed here (as it will be the scale used in the present study). The CALPAS-R consists of 30 items rated on a 7 point degree of occurrence scale. Also, there is one general rating for each of the four scales. The following is a brief explanation of each scale. The patient workinq capacity reflects Sterba's (1934) idea of patient's need to be able to work in analysis. The formation of the therapeutic alliance is dependent on the patient's capacity to work actively and intentionally in therapy. This requires that the patient self—disclose clinically significant material and respond to therapist's interventions in a manner that promotes a deepening of the therapeutic process. This scale is comprised of six positive (PWC) and six negative (patient hostile resistance - PHR) items. The subcomponents are: self-disclosure and self-reflection on important themes, exploration of one's own contribution to problems, ability to experience strong affects wit comments, c thenes, wor of distanci an easy anci The ne has its ori transferenc( 1393897611330 grounds: :0 efforts W1] sacrificeS. basic truss COZTllete t The w Bo: in's ( §Etie3t fix tasks a“ as such iI'r-n. . lytlcit (205790;: *2. CHE Cha 33 affects without decompensation, use of therapist's comments, deepening of the exploration of important themes, working toward problem resolution, an absence of distancing attitudes, and an absence of the wish for an easy and quick cure. The next scale is patient commitment. This factor has its origins in the idea of unobjectionable positive transference (Freud, 1912), but includes the patient's perseverance to remain committed during difficult periods of the therapy. It reflects the degree of the patient's attachment, on both rational and emotional grounds, to the therapy and therapist. The subcomponents of this scale are: confidence that efforts will lead to change, willingness to make sacrifices, view of therapy as an important experience, basic trust in therapist and therapy, taking part in therapy despite painful moments, and commitment to complete the therapeutic process. The working strategy consensus scale is based upon Bordin's (1979) assertion that the therapist and patient must have an agreement on major therapeutic tasks and goals for a therapeutic success to occur and, as such, this scale reflects the degree of explicit or implicit agreement on how therapy should proceed. The subcomponents include: sharing compatible ideas about the change process, similar ideas about how to proceed in therapy, goals of ti. and do not | The 15‘ inwolsmrent erpathic "r Ho. components aCCEptance Patient's 5 aCtively p: an timing ‘ ~I€dt me:t f, 34 in therapy, therapist's understanding of patient's goals of therapy, therapist and patient work jointly and do not work at cross purposes. The last scale-—therapist understanding and involvement--has two subcomponents: therapist's empathic understanding of the patient's problems and therapist's active interventions which indicate a genuine concern in helping the patient. The specific components of the scale include: non-judgmental acceptance of the patient, understanding of the patient's subjective experience and psychic distress, actively pursuing the patient's core conflicts, tact and timing of interventions, therapist does not misuse treatment for personal needs, and commitment to helping patient in overcoming his or her problems. In summary, many of the studies are in support of a significant relationship between the therapeutic alliance and psychotherapy outcome. The alliance appears to be established early on and is a predictor—- though not in a direct fashion-—of the course of the therapy. Depending on the experimental paradigm involved, the relative importance of patient factors and/or therapist factors varies. Factors including patient commitment and working capacity, motivation, ego—strength, object relations capacity, and therapist understanding and/or errors in technique have all been inplicated types of so ' 4- different 0) different Dr, between all (A) complex an analyses; ’9. critical d; re.ationsni 35 implicated as important mediating variables. Different types of scales have been developed to measure different aspects of the therapeutic alliance from different perspectives. Overall, the relationship between alliance and outcome has been shown to be more complex and in need of more sophisticated statistical analyses; further, mediating factors appear to be a critical dimension in the elucidation of this complex relationship. relations 1 0f psychot‘: internal rl perspectiv. considerec relations TAT reSpo; nature of and the af representa im’01‘v'ed i manta; Ops repreSepte aspeCts 0f repreSEZtE 1 MW a dc W83: "1 e“, Measurement of Object and Interpersonal Relations In the literature review (see Appendix A), object relations was discussed specifically within the context of psychotherapy research. Approaching the study of internal representational processes from a different perspective, there is a body of research which has considered the measurement of the quality of object relations through dreams, early memories, Rorschach and TAT responses. These studies have concentrated on the nature of the subject's self and other representations, and the affective processes associated with these representations. The representational processes involved in object relations have to do with various mental operations which organize internal representations of self and others. The distinguishing aspects of the patient's store of mental representations are believed to correspond to points along a developmental continuum (Urist, 1980). Leigh, Westen, Barends, Mendel, and Byers (1989a) note that: “A key assumption of object relations approaches is that individual differences may reflect developmental differences; that is, adults differ in their tendency to represent people in developmentally mature or immature ways (p. 2)". 36 issue relat of se bound more. bad a Geller“ C0350! 37 There are certain general developmental aspects or issues that seem critical to the quality of object relations. First, there is an evolving differentiation of self and other representations, and various boundaries between inner and outer. Second, representations are hypothesized to increase in complexity, dimensionality and integration with maturation. Third, while early experiences tend to be internalized with more primitive, univalent affects, more mature representations should integrate good and bad aspects of significant others. Thus, the development of these qualities allow for the consolidation of complex and realistic representations of self and other. Diverse psychoanalytic and developmental theories underlay the theoretical basis of the object relations scales in use today. There is a compatibility between the major scales in that they are typically unidimensional scales organized according to a developmental continuum; the following is a brief overview of some key studies. Mayman and his colleagues at the University of Michigan (Krohn & Mayman, 1974; Mayman, 1967, 1968; Urist, 1977; etc.) have pursued research investigating various thematic and content areas of object representations revealed in projective and interview data. Operating within an object relations and ego psycholo object I characte enter :- Tr“? .elneyl "Qy—n KG ‘\ “la: A w ., \‘al C) [1‘ 38 psychology approach, they explored the relationship of object representations to levels of psychopathology, character structure and the ability of a patient to enter into a psychotherapy relationship (Blatt & Lerner, 1983). In one of his early investigations, Mayman (1967) examined patients' level of psychopathology and related dimensions (e.g., severity of symptoms, ego strength, defenses, motivation, etc.) based upon inferences from Rorschach responses as a measure of the patient's “relationship potential“. He found that these dimensions were significantly related to independent interviewer's assessments of patients (this included 12 pre—treatment variables and Health- Sickness ratings; patients were from the Menninger Psychotherapy Research Project [e.g., Kernberg, Burstein, Coyne, Appelbaum, Horwitz, & Voth, 1972). Mayman (1968) next considered the use of early memories as a measure of object relations. Because of the screen function of early memories, Mayman suggested that they might be treated like manifest dream content. As such, he reasoned that they could provide information about defensive structure and internalized object representations. Mayman constructed a scale based upon psychosexual ego modes of development and organized them in terms of prototypical interpersonal themes found in early memories. While this was not a p Q n” .U. ..l d‘. o'n be. . .1 39 formal empirical study, the several case examples adapted well to the scale, leading Mayman’to conclude that it was a useful source of data on the differentiation and richness of internal object representations. Krohn and Mayman (1974) developed The Object Representation Scale for Dreams based, in part, upon Kernberg's (1975) treatment of early object relations and applied it to a battery of measures for 24 outpatients. The patients' dreams, early memories, Rorschach responses, health—sickness rating and therapist/supervisor ratings of object relations were used. There were significant correlations and interaction effects among these five variables accounting for a total of 38% of the variance. Krohn and Mayman conclude: This study suggests that the modal object representational score of a patient's dreams will constitute the best predictor of the therapist's evaluation which is based on months of contact with the patient. (p. 463) The fundamental conclusion to be drawn from this study is that level of object representation appears to be a salient, consistent, researchable personality dimension that expresses itself through a relatively diverse set of avenues, anc R U. .l Q «\~ e as . O L .. .5 an... s . flu .1 .C y ..L r . n u e In Us 40 ranging from a realm as private as dream life to one as interpersonal as psychotherapy. (p. 464) Urist (1977) developed an object relations scale which "focuses on the developmental progression toward separation-individuation, with particular emphasis given to the issue of autonomy vis—a-vis the self, and...others (p.4)". Forty inpatients provided Rorschach responses, completed an autobiography, and were rated by staff on their actual interpersonal functioning. The findings indicated a consistently high correlation among all measures. Urist concluded (a) that there is support for the enduring aspects of the patients's relational capacity across different measures, (b) the ratings of the Rorschach successfully assessed the mutuality of autonomy between self and other, and (c) that it was the patient's range of responses (i.e., their repertoire of healthy and unhealthy responses) that was the best predictor. Ryan and Bell (1984) collected early memories from over 60 hospitalized psychotic patients across the course of their treatment and rated these memories on the Ryan Object Relations Scale, which is a 20 point unidimensional scale with a developmental continuum very similar to those scales described above. They found an increase in the developmental level of object re re CO :5: C "J l dIlC 41 relations from a psychotic state at admission to a more recovered state upon follow-up. There was a correlation between higher level of object relations scores and fewer re-hospitalizations. The object relations measure was not related to other outcome measures such as symptom severity. Concluding an overview of the University of Michigan studies, Blatt and Lerner (1983) state: In summary, the research findings of Ryan, Krohn and Urist, based on the theoretical conceptual— izations and research methodology articulated by Mayman, provide further support for the importance of assessing object representations in clinical research and practice. Object representations assessed with several instruments (e.g., early memories, dreams and the Rorschach) have significant relationships with independent assessments of the capacity to enter into and benefit from psychotherapy. (pp. 209-210) Blatt and his colleagues (e.g., Blatt, Brenneis, Schmick, & Glick, 1976; Blatt & Ritzler, 1974) from Yale University have developed a line of research related to Mayman and colleagues. The Yale research, however, has focused on the structural or formal dimensions of object representations. Integrating 42 cognitive, developmental and psychoanalytic research, this group sees the development of cognitive schemata, especially concerning self and other representations, as a principal aspect of the interpersonal process (Blatt & Lerner, 1983). Basic to this approach is the differentiation of boundaries on several dimensions including--between self and object, inside and outside, and object and mental representations. Much of this research was concerned with the boundary impairment associated with severe psychological disturbances. Various studies from this group (e.g., Blatt, Brenneis, Schimek, & Glick, 1976; Blatt & Ritzler, 1974) have used the Developmental Analysis of the Concept of the Object scale to assess the differentiation, articulation and integration of object representations of different data including Rorschach human figure responses, TAT protocols, interview data, manifest dream content, etc. (Blatt & Lerner, 1983). Research using this scale has shown consistent differences in the structural aspects of object representations with various diagnostic groups (for a review, see Blatt & Lerner, 1983). For example, Blatt and Ritzler (1974) found that the degree of disturbance in boundary articulation, a fundamental aspect of object relations, as measured by thought disorders seen in Rorschach responses was significantly related to the quality of 43 ego intactness; poorly articulated boundaries were more likely to be found in more severely disturbed patients whose ego functions are less intact. Lerner and St. Peter (1984) used the same scale to rate the quality of Rorschach human figure responses for patients with neurotic, borderline, and schizophrenic disorders in order to examine differences in developmental level of object relations. Responses were classified as to their degree of perceptual accuracy based upon response form level. The ratings of 70 Rorschach protocols indicated that schizophrenics are differentiated from the other diagnostic groups based upon lower developmental level (in all three categories) and lower level of perceptual accuracy of human figure responses. An interesting finding is that inpatient borderlines had the highest levels of differentiation, articulation and integration, as well as the highest number, of inaccurate responses and malevolent responses. In contrast, outpatients with borderline diagnoses had higher developmental levels of accurate responses and quasi-human responses. The authors note that there is an increasing relationship between the quality of the subject's reality testing (i.e., response accuracy), developmental level of the concept of the object (3 scales of developmental index) and patient psychopathology (diagnostic continuum from 0Cr 44 schizophrenic, inpatient borderline, outpatient borderline and neurotic). The Object Relations Scale (OR Scale). Westen and colleagues (Westen, Barends, Leigh, Mendel, & Silbert, 1988a; Westen, Huebner, Boekamp, Lifton, & Silverman, 1989; Westen, Ludolph, Lerner, Ruffins & Wiss, 1990) integrating the object relations research of Mayman and Blatt described above, have developed a multidimensional object relations scale which considers several distinct but interdependent cognitive and affective processes hypothesized to be associated with object representations (Westen, Ludolph, Lerner, Ruffins, & Wiff, 1990). The measure has been applied to normal and clinical samples and the scales have successfully differentiated borderline adolescents and adults from normal and other diagnostic groups, as well as differentiating certain general personality types. The scales have been adapted for various types of data including TAT and Rorschach protocols, self/other descriptions, early memories and excerpts from psychotherapy sessions. The OR scale consists of four subscales which assess four interdependent aspects of object relations. Each scale has a five point rating system where level one is associated with the lowest developmental level and level five with the highest (exc is br 45 (exception: complexity has a seven point scale). A brief description follows: Complexity and Differentiation of Representation. This scale measures the extent to which others are seen as differentiated and complex persons. While there are diverse theories concerning object relations, there appears to be a consensus about three general observations of the development of object representations: increasing differentiation, complexity, and the integration of ambivalent representations. There is a substantive body of literature from the areas of infant research, developmental and cognitive psychology which is supportive of this general framework of the development of self and object representations (Westen, Huebner, Boekamp, Liftin, & Silverman, 1989). At the lowest level of this scale the person has difficulty differentiating aspects of self from others. At intermediate levels there is differentiation of one's own from others' perspective, but persons are seen in simplistic, unidimensional terms. At the highest level individuals are able to experience self and others in terms of complex, subjective, and psychologically oriented factors. fix 2. r . n u no. r ,. us :1 v 1 C. l. 1.. 1‘ . . l . S e r . r“ e 0 a f 8 n3 0 .d 0 D R. an tt X a p. o e 0. 0 1. v . b n n a K d t a C O l. .. p... $4.. 9 m a D S 46 Affect-Tone of Relationship. The second scale assesses the degree to which the participant expects or perceives relationships to be destructive and dangerous or benevolent and enhancing. This scale considers the underlying affective quality associated with object representations. Arising from research conducted on borderline personality organization, various writers have discussed the implications of a malevolent affective charge to internal representations (e.g., Kernberg, 1975). In such persons, frustrating, aggressive or depriving early relationships are thought to be internalized in non-integrated or pathological ways leading to a generally malevolent or threatening coloring of the internalized object world. Some of the object relations studies reviewed above have supported the observation that borderlines tend to have more malevolent responses in a variety of test situations. .At the lowest level of this scale, the person experiences others and the world as dangerous and malevolent. At the highest level, the person has an abundant range of affective expectations which are benevolent and hopeful. Capacity for Emotional Investment. The third scale measures the degree to which the participant sees others as relatively differentiated and whether or not 11:1 IT.C'\-’l b . .. -Ohc (Wes 47 they are viewed as primarily need—gratifying or as ends in themselves. There is a general agreement in the literature of the developmental trend of the person moving from experiencing objects as a means to an end toward experiencing objects as ends in themselves (Westen, Barends, Leigh, Mendel, & Silbert, 1988a). Westen et al. (1988a) outline three processes involved in the capacity to emotionally invest in others: (a) the capacity to regulate emotional investment in terms of rapidity, intensity and consistency of investment; (b) investment in specific others for their personal attributes; and (c) a trend toward investment in values and morals which transcend one's immediate wishes and impulses. There is research in the cognitive, developmental and moral development areas which is supportive of this (see Westen et al., 1988 for a review). The scale uses a three step model of rating developmental level. At the most basic level others are experienced in terms of their gratifying or comforting functions almost exclusively. At the intermediate level others become valued as ends in themselves; the internalization of values and ideals is apparent, and guilt and shame result when these values are not maintained. The highest level is characterized by a striving for autonomous selfhood, and a genuine and enduring investment in others and society according to ac deep, EEC-1'8 b h 48 to accepted norms. The individual is able to invest in deep, committed relationships where the other is appreciated as an unique individual. Understanding of Social Causalipy. This scale considers the logic, complexity, accuracy and psychological mindedness of the subject's description of interpersonal interactions. (Note: this scale will not be used in the present study; the reader is referred to Westen, Huebner, Boekamp, Lifton, & Silverman [1989a,b] for further information). Empirical Research Involving the OR Scale. A series of studies have been undertaken by Westen and colleagues to investigate the construct, discriminant and convergent validity of the OR Scales (Leigh, Westen, Barends, Mendel, & Byers, 1989; Westen, Barends, Leigh, Mendel, & Silbert, 1988a; Westen, Huebner, Boekamp, Lifton, & Silverman, 1989). Strong preliminary support has been gathered for the scale's capacity to discriminate between certain diagnoses and personality types. With regard to the scale's application to psychotherapy process data, an empirical study was recently conducted which assessed changes in patients' object representations over the course of brief psychotherapy (Schneider, 1990). Westen, re: 03 be: Co: 49 Ludolph, Lerner, Ruffins and Wiss (1990) reported inter—rater reliabilities for the OR Scale for TAT data: Affect-Tone .93, Complexity .91, Emotional Investment .87, and Social Causality .95. Interview data using the Complexity scale had an inter-rater reliability of .91 (Leigh, Westen, Barends & Mendel, 1989). Other studies using TAT data have achieved similar reliabilities (Westen, 1990, 1991; Westen, Huebner, Boekamp, Lifton & Silverman, 1989). A discussion of the major validation studies follow. Westen, Ludolph et al. (1990) attempted to differentiate borderline patients from others through the examination of the quality of object relations as measured by TAT responses in a female adolescent population of normal, psychiatric (various non— borderline diagnoses) controls and borderlines. TAT responses were analyzed on all four dimensions of the OR scale. The findings clearly differentiated borderline patients from normals and psychiatric controls on the basis of their object relations. Further, the scales were able to elaborate several distinctive qualities of borderline object relations. Borderlines, in contrast to normals and psychiatric controls, showed more malevolent representations, lower capacity for emotional investment in others and less accurate attributions about others' motives. 50 Interestingly, borderlines showed the highest level of complexity of object relations; this was often colored by malevolent and inaccurate attributions about people (in keeping with the findings of Lerner & St. Peter, 1984). The authors concluded that while patients with borderline diagnoses may produce more pathological or distorted representations, these representations are not simple nor unidimensional, guiding explanations away from a simple deficit etiological model. Westen, Huebner, et a1. (1989) compared natural science and clinical psychology graduate students on measures of complexity of representations and understanding of social cognition using TAT stories. The stories were scored on three dimensions--complexity of representations and understanding of social causality (inter—rater reliabilities were above .90). For discriminant validity the affect-tone of relationships scale was used as this was hypothesized not to distinguish the two groups. The findings indicated that both complexity and social causality reliably discriminated between natural science and clinical psychology graduate students, while affect— tone did not, with clinical psychology students scoring significantly higher on both measures. These relationships remained even when cognitive functioning (verbal GRE's) were covaried with the scales. The F). f). de 51 authors concluded that these scales are individual difference measures which capture differences in developmental levels. Leigh, Westen, Barends, Mendel and Byers (1989) developed and evaluated two measures of complexity of representations—-one measure was used to assess TAT stories while the second assessed narrative data (semi— structured interview with self and other descriptions). These two scales, hypothesized to correlate with each other, were also compared with a similar measure of complexity form Blatt's object relations scales to test convergent validity. Discriminant validity was evaluated by use of a Blatt subscale for ambivalence and malevolence/benevolence which, because it measures affective dimensions of representations, was hypothesized to not be correlated with the measures of complexity. The findings were largely supportive of the hypotheses-—the two complexity measures correlated well with each other (p = .42, p (.0001) and with the Blatt measure of complexity, and maintained high internal consistency, but it did not correlate with the affective measure. Inter—rater reliability was very high for both the TAT and the narrative data. The authors concluded that this study offers preliminary support for these measures which can be applied to a broad array of data. 52 The Interpersonal Checklist. The Interpersonal Checklist (ICL; Leary, 1957) is a questionnaire listing 128 behavioral adjectives which correspond to 8 (or 16, depending on the scoring approach) interpersonal styles. These items are analyzed into four quadrants based upon two bimodal scales: Dominance—Submission and Affiliation-Hostility (quadrants 1 and 2 are dominant, quadrants 3 and 4 are submissive; quadrants 1 and 4 are affiliative, quadrants 2 and 3 are hostile). Previous studies have indicated that these two dimensions are a fundamental and significant way to describe interpersonal behavior and they account for a preponderance of the variance in interpersonal interactions (see Hurley, 1980, for a review). Adams (1964) suggested that the affiliation-hostility dimension delineates the extent of acceptance or rejection of others; while the dominance-submissive dimension describes the degree of acceptance or rejection of self. As such, Hurley (1980) has named these dimensions ARO (acceptance/rejection of others) and ARS (acceptance/rejection of self). Low scores on either dimension tend to offset the implications of high scores on the other dimension; and low scores on both dimensions indicate a relatively destructive interpersonal style. SEC: rela 53 In a recent study (Filak, Abeles & Norquist, 1986), patient's pre—therapy ICL score, in particular, the affiliative stance, was used to predict psychotherapy outcome (discussed in greater detail elsewhere). The findings indicated strong support for the relationship between an high affiliative score and successful outcome, while a hostile stance had a poor relationship to positive therapy outcome. Crowder (1972) used the ICL to examine the relationship between therapist and client interpersonal behavior in psychotherapy. Based upon the Freedman, Leary, Ossorio, and Coffey (1951) diagnosis schema, four general interpersonal styles (derived from the four quadrants) were used to categorize interactional styles of transference and countertransference configurations in psychotherapy (supportive-interpretive, support- seeking, hostile-competitive, and passive-resistive). This schema was successful in differentiating successful and unsuccessful therapeutic interactions. The ICL will be used in the present study to assess patients' pre-therapy interpersonal attitudes on the affiliative versus hostile dimension (acceptance/ rejection of others; ARO). In the Filak et a1. (1986) study, the findings indicated that ARO, but not ARS, was an important predictor of therapeutic outcome. 54 Almost twice as many (21 of 29) successful patients, as opposed to unsuccessful patients (10 of 26), had an affiliative stance with no significant differences noted on the ARC dimension. Filak et al. (1986) concluded that while attitudes toward the self (ARS) may be affected short-term psychotherapy (i.e., mean = 24 sessions), interpersonal attitudes toward others (ARO) are not as easily changed. It may be the case that attitudes toward the self are more easily modified via internalization of the therapist's accepting stance, while basic therapeutic of others are more deeply embedded and based upon archaic, and hence less mutable, representations. In conclusion, the studies from the University of Michigan, Yale University, as well as Westen and colleagues, and research utilizing the Interpersonal Checklist (ICL) are supportive of the enduring, measurable and predictive qualities of self and object representations and the affective processes associated with them. Research findings indicate that object representations (and interpersonal attitudes) can predict certain aspects of the process of psychotherapy, differentiate diagnostic groups, are related to the development of various aspects of ego functions, and can change or mature over the course of 55 treatment. The concept of object relations as a predictive variable and model of internal representation of interpersonal processes appears to be a relevant and robust research variable. Because the present study will investigate the effect of pre- therapy interpersonal attitudes on the development of the therapeutic alliance and its relationship to outcome, the ICL will be included to complement the OR scale as a measure of patient pre-therapy factors. Since the two measures differ (i.e., the ICL purports to measure interpersonal attitudes determined by self— report while the OR scale assesses inferred internal processes of object relations judged via ratings of early psychotherapy sessions), convergent, divergent and construct validity will be considered through the comparison of the two measures. 56 Summary and Objectives Research findings indicate a significant relationship between the quality of the therapeutic alliance and outcome in a variety of psychotherapies. Gaston (1990) states "Empirical evidence supports the direct association between alliance and outcome“ (p. 148). While there is a general consensus as to the association between the therapeutic alliance and psychotherapy outcome, this unanimity decreases as soon as attempts are made to specify specific patient and therapist factors which contribute to this relationship. Gaston (1990) underscores this point: “...there has been less emphasis on defining its (therapeutic alliance) specific components, and on empirically investigating the roles played by the alliance in psychotherapy (p. 143)”. Studies differ as to the amount of variance which is attributed to various patient and therapist variables, and whether positive factors or negative factors are more predictive. Further, several studies indicate that the relationship between therapeutic alliance and outcome is neither direct nor simple (e.g., Gaston, 1990; Horowitz, Marmar, Weiss, DeWitt & Rosenbaum, 1984). More sophisticated statistical w“. M 57 procedures which have investigated multi—factorial interactions have revealed the importance of several other factors as significant mediating variables. Such factors as therapist errors in technique (Foreman & Marmar, 1985; Sachs, 1983), patient problem avoidance (Marmar et al., 1989), phase of treatment (Klee, 1986; Morgan, Luborsky, Crits-Cristoph, Curtis, & Solomon, 1982), and patient predispositional factors--including object relations (Marziali, 1984b; Moras & Strupp, 1982; Ryan & Cicchetti, 1985), and motivation (Horowitz et al., 1984) have exerted a significant interactive effect on alliance and outcome. Several writers have discussed the importance of patient predispositional qualities, in particular, the patient's internalized object representations or interpersonal capabilities, in relation to the development of the therapeutic alliance (e.g., Frieswyk, Allen, Colson, Coyne, Gabbard, Horwitz, & Newsom, 1986; Horowitz et al., 1984; Strupp, 1974). For example, Klee's (1990) findings support: "...the assertion that patients possess a core capacity for relatedness, reflected in pretreatment measures and in interpersonal behavior from the first session of psychotherapy (p. 13)". There are a limited number of empirical studies which have been, with some qualifications, supportive 58 of this relationship (see Appendix A for a review). There are, however, some research issues which should be addressed. One difficulty is how object or interpersonal relations is defined. While the constructs of object relations and interpersonal relations overlap, the possible range of phenomena falling under their purview is considerable when one moves between the poles of inferred internal psychic processes on the one hand and patterns of potentially observable interpersonal abilities on the other. Urist (1980) suggests that: "The confusion (in assessment) represents a failure to distinguish conceptually between external relationships in the real world and the internal images or mental representations of those relationships (p. 828)". Thus, some studies have failed to delineate adequately the construct and have minimized the implications of the differences between object and interpersonal relations. Perhaps the most obvious limitation of many of these studies are the dependent measures employed to assess object or interpersonal relations. The degree to which the validity and reliability of certain measures have been researched has also been of concern. In some studies the measure has been inadequate to capture the complexity of this dynamic construct. For example, one study (Gaston, de CC «C S 59 Marmar, Thompson and Gallagher, 1988) which examined the therapeutic alliance with elderly patients used the Young Loneliness Inventory; items such as “I cannot seem to communicate", "I've always felt lonely" and "I've felt different from other people" from the inventory, ostensibly "permitted assessment of the patients's interpersonal functioning" (p. 485). In another study, (Horowitz, Marmar, Weiss, DeWitt & Rosenbaum, 1984) the measure was more sophisticated and congruous with an assessment of object relations (a "developmental level of self“ measure which followed developmental stages) but the rating procedure, construct (Horowitz, 1979) and test validity have been researched and validated in a limited manner. More recently, Ryan developed the Ryan Object Relations Scale (Ryan & Bell, 1984). This is a well—validated scale whose development has been informed by the major object relations measurement paradigms (see Appendix A for a review). However, the 20 point scale attempts to collapse a number of variables organized around developmental level and degree of psychopathology. Such an approach does not allow for examination of the separate components involved, nor does it differentiate between aspects of object relations as distinct from maturational or diagnostic factors. Instead, a more thorough process of assessing object relations in terms V . 1 «-1 t by D; ..C. s; 60 of the discrete qualities of various self and other representations is recommended (Urist, 1980). This approach also seems to follow from the proposal made by Westen, Ludolph, Lerner, Ruffins and Wiss (1990) stating that object relations "consists of a number of distinct but interdependent cognitive and affective processes“ (p. 5). Many of these concerns have been addressed by Westen (e.g., 1990) in the current version of the OR Scales—-the scales are multidimensional, reliable and valid, and appear to be conceptually sound. The OR Scales measure four interdependent aspects of (the representational processes associated with) interpersonal relations. There has been a scarcity of research which has investigated specific aspects of how object relations affect the quality of the alliance. Further, it is only more recently that the OR Scales have been applied to the psychotherapy process. In addition, the ICL, an interpersonal attitude questionnaire, will be used as an adjunct to the OR scale to assess validity issues. Patient early-therapy object relations are postulated to exert an interactive effect between the therapeutic alliance and psychotherapy outcome. There appear to be three general ways in which this occurs: (a) entering into a viable therapeutic alliance is predicated upon a requisite minimal level of patient a}. e we. COT. . ‘Y‘v‘ ‘6..'I. B; S. v. at RU; 61 capacity for object relatedness; (b) the major patterns and qualities of the patient's object relations are predictive of the attitudes and transferences the patient brings to bear in the therapeutic relationship; and (c) depending on how these attitudes are expressed, they can either help or hinder the work of therapy, contributing significantly to an interactive sequence of patient and therapist responses which are an important process component. While the major emphasis of this study concerns the general relationship between patient early-therapy object relations, the therapeutic alliance and psychotherapy outcome, an object relations theoretical framework can be helpful to generate speculations about more specific relationships between object relations and therapeutic alliance. There are a number of possible inter—relationships between the specific subscales of the OR Scale, the CALPAS and the outcome measures, but the major factors under consideration for the present study are: (a) OR scale-—three of the four OR scales (AT--affect tone, EI—-emotional commitment, and C—-complexity) and (b) TA scale—-the three patient subscales of the CALPAS (PWC-—patient working capacity, patient hostile resistance and patient commitment). The following is a brief conceptual overview of these factors. 62 Affect tone (AT) may be viewed as a general indicator of the affective coloring of the patient's internalized object relations and interpersonal expectancies. These expectancies range from hostile/malevolent to benign/enriching perceptions of the world. Developmentally, this parallels preoedipal issues, in particular the capacity to integrate positive and negative aspects of self and others. Also, the basic conflict or developmental task involved in the AT scale corresponds with trust versus mistrust. Problems with these sorts of psychostructural issues-- trusting, integration—-are typically associated with borderline personality organization. In terms of the particular dimension of the therapeutic alliance, AT ties in with transference--the positive (or negative) affective attachment to the psychotherapist. Emotional investment (EI) seems to be an indicator of ego/super-ego maturity as reflected in object relationships. At the lower end, the level of maturity is egocentric and need gratifying in orientation; it is concerned with mutual need gratification and conventional morality toward the middle of the scale; and a striving toward deep commitment and moral ideals at the high end. Developmentally, this seems to parallel the move from the pleasure principle to the reality principle. The move from hedonism toward 63 internalized super—ego anxiety and internalized conventions corresponds to oedipal issues, with the focus on neurotic concerns. EI appears to coincide with the working alliance aspects of the therapeutic alliance. Complexity (C) appears to deal with the most developmentally early processes. As such, it considers boundary development--from the most rudimentary differentiation of self from other to the most sophisticated where integrative and inferential processes are involved. At the lower end of the scale, the person may have trouble differentiating their thoughts and emotions from those of others; and, subjective and enduring states in others are not recognized. At the high end of the complexity scale, others are perceived as complex, psychological, clearly differentiated persons with unique and enduring traits, allowing inferences about the motives and internal states of others. Because complexity is perhaps the most fundamental of all the scales, there is some overlap with each of the CALPAS subscales. Aside from clear boundary definition, the capacity to note and tolerate differences, empathize with others and infer more complex internal states, all seem related to aspects of the scales. 64 With regard to the patient subscales of the CALPAS, Marmar, Weiss and Gaston (1989b) proposed that the patient commitment (PTC) scale should reflect the therapeutic climate and note the patient's transference and positive attachment to the therapist. The PTC dimension is congruent with the affect tone (AT) scale of the OR scale. However, the six items of this scale can also be interpreted as indicative of level of maturity of ego-functioning in that it requires commitment and tolerance. As such, the scale would be conceptually more compatible with the emotional investment (EI) scale. From a third viewpoint, Gaston (1990) explains that PTC can also be affected by a therapeutic misalliance (Langs, 1974) where the patient and the therapist unconsciously collude in avoidance of conflictual material: negative feelings and conflictual material may be split off and dissociated from the therapeutic process. There are, then, at least three interpretations of PTC in its relationship to the OR scales. Also, as discussed above, the complexity (C) scale deals with fundamental aspects of object relational processes and is expected to be related to all of the CALPAS subscales. Each of these will be considered in an attempt to clarify which explanations and theories seem to fit best. ’4. '11 65 Patient working capacity (PWC) coincides with the working alliance dimension of the therapeutic alliance. The positive scale seems indicative of the patient working actively and deeply within the relationship and therapeutic process. From this perspective, it seems congruent with E1. However, PHR scale items appear to be more analogous to the AT scale as directly indicative of negative transference. Again, the major interrelationships will be investigated to learn more about the conceptual compatibility of the scales. Another CALPAS scale also merits consideration as it monitors, more directly, the collaborative aspects of the therapeutic alliance. The working strategy consensus (WSC) scale assesses the degree to which the patient and therapist agree on the goals and tasks of psychotherapy, based upon Bordin's (1979) theories about the alliance. This scale is postulated to be related to AT in that it requires that the patient have positive interpersonal expectancies of the therapeutic relationship. WSC corresponds with EI in that it requires a level of mutuality, commitment and acknowledging the other in more than a need gratifying manner. The following is an example of how the OR scales might be predictive of specific alliance scenarios unfolding. If the patient has a low AT score (e.g., tea tna n6- ‘0‘!- an 66 mean = 1 or 2), this might be seen as an indication that he or she tends to expect hostile and malevolent interactions with others. This may manifest itself in an evolving negative transference toward the therapist and therapy (e.g., mistrustful attitude, hostile attributions toward the therapist, power struggles, difficulty participating in the work of therapy, etc.). This would affect the development of the therapeutic alliance and be reflected in the PHR subscale, as this most directly corresponds with the negative transference. A lower WSC score might also be expected as the development of the working alliance would most likely be compromised by the patient's attitude. Conversely, if the patient has a high AT score (i.e., mean = 4 or 5), this would be an indication that their interpersonal expectancies are of a more benign and enriching nature. The patient would be a cooperative participant in the development of the alliance. The patient's involvement will be evidenced in a variety of ways and this should register on each of the CALPAS patient subscales: a clear commitment (PTC), an ability to work with the therapist (WSC), and an unfolding of the patient's capacity to open up, experience strong emotions, and gain insight (PWC). elm +1. D :..E A. O C . C» out: HYPOTHESES The present study will attempt to undertake a more thorough consideration of the measurement of object relations, and its relationship to the therapeutic alliance and outcome. This will be accomplished through the use of a multidimensional object relations measure (OR Scale) and a measure of interpersonal attitude (ICL). Interactions between object relations (OR) and therapeutic alliance (TA), as it pertains to outcome, will be examined on both a general and specific level. In addition to the hypothesized general relationship between patient early therapy object relations and therapeutic alliance, various relationships between subscales of the OR Scale and the CALPAS will be examined in order to investigate some of the specific components of the interaction. A graphical depiction of this basic model is illustrated in Figure l. The overall purpose of the present study is to further establish the empirical relationship between the therapeutic alliance and psychotherapy outcome, and to begin to elucidate some of the patient factors which affect this relationship. 67 ,_. h. {'13 Bee. 1 1.?: \ 68 Figure 1 Basic Interactional Model 0R <———> TA X > O OR x TA OR = OBJECT RELATIONS TA = THERAPEUTIC ALLIANCE 0 = OUTCOME 1.Process Variable Hypotheses Because the relationship between the therapeutic alliance and therapeutic outcome is neither direct nor simple, the factors——in the present study the focus is on patient factors--which moderate this relationship require further investigation. The first set of hypotheses concern the therapeutic process; specifically, they examine the relationship between patient early-therapy object relations and the therapeutic alliance. Principal Process Hypotheses: Hypothesis 1A. It is predicted that there will be a positive relationship between level of early~therapy object relations and the quality of the initial therapeutic alliance (rating of sessions one and 69 three). It is predicted that this positive relation- ship will continue through the course of therapy (i.e., an initial and overall relationship). Specifically, positive correlations between subscales of the OR scales (affect tone-—AT, emotional investment--EI, and complexity—-C) and CALPAS patient scales (patient working capacity--PWC, patient hostile resistance--PHR, and patient commitment--PTC) is predicted. There will also be a positive relationship between the OR scales and the working strategy consensus (WSC) scale. Several specific positive relationships between the three OR scales and the four CALPAS scales, across time points in the therapy process (i.e., session one, session three, middle and end of therapy) are also expected; the potential interactions among these subscales are illustrated in Table 1. Hypothesis 18. It is predicted that level of early—therapy object relations, as measured by ratings on the three OR Scales, will exert an interactive effect upon the therapeutic alliance across sessions (i.e., initially and overall). Specifically, object relations is hypothesized to differentiate initial levels of therapeutic alliance, and to moderate the relationship between the overall alliance. It is predicted that subjects with lower levels of object 70 Table 1 Potential Interactions Between Therapeutic AllianceL Object Relations and Session CALPAS OR Scale 3’ '6 E1 C EEC Session 1 Session 2 Middle End Session 1 Session 2 Middle End Session 1 Session 2 Middle End Session 1 Session 2 Middle End pg; Session 1 Session 2 Middle End "U :1: $0 'U *3 O 2 m 0 -— —-—-—-—P—h-—-- —- —- P- -—————--—-—- —— -- relations will have lower levels of initial therapeutic alliance. The strength of the difference between high and low levels of object relations will be less decrease over across sessions (i.e., low object relations will begin to 'catch up' with high object relations). High levels of object relations will consistently be related to higher levels of therapeutic alliance over the course of psychotherapy. 71 Exploratory Process Hypotheses Hypothesis 1C. Patient scores on the three Object Relations Scale-—affect tone (AT), emotional investment (EI) and complexity (C)--scale will be positively correlated with scores on the patient working capacity (PWC) scale, patient hostile resistance (PHR), patient commitment (PTC) and, to a lesser extent, positively correlated with working strategy consensus (WSC). In terms of the strength of association for the three OR scales, affect tone, then emotional investment, then complexity will be correlated to the therapeutic alliance. The specific hypotheses concerning the relationship between object relations and therapeutic alliance are as follows: Hypothesis 1C—1. The OR scales will be positively correlated with the patient positive scales——PWC and PTC . Hypothesis 1c—;. The OR scales will be negatively correlated to the PHR scale. Hypothesis 1C—3. The OR scales will be positively correlated with the WSC scale. Sign; c 4“l Skip torn. 72 Hypothesis 1D. The greatest within-group variance is predicted to occur among patients with relatively lower object relations and therapeutic alliance scores. It is suspected that deficits and/or conflicts which are associated with patients' internalized object representations will contribute to more tenuous, variable and idiosyncratic responses to the therapeutic situation, thus disrupting the therapeutic alliance. There will be less within-group variance for patients with higher object relations and alliance scores. 2. Outcome Variable Hypotheses The therapeutic alliance is seen as a multidimensional and interactive process which is affiliated with many of the core components of the change processes in psychotherapy. As such, it is seen as a predictor of the processes which lead to positive psychotherapy outcome. The following hypotheses attempt to elucidate some aspects of that relationship. Principal Outcome Hypotheses: Hypothesis 2A. It is predicted that there will be a general “treatment effect" for the subjects in this study. That is, there will be a statistically significant reduction of patient—reported symptomatology when pre— and post-therapy SCL—90—R SCO‘IBS IH4 73 scores are compared. Hypothesis 2B. It is predicted that (a) there will be a positive relationship between the quality of the initial therapeutic alliance (sessions one and three) and outcome; (b) there will be a relationship between overall therapeutic alliance (beginning, middle and termination phase session) and outcome. Specifically, the subscales of CALPAS (the patient scales--PWC, PHR and PTC)--from the initial alliance rating, as well as the overall therapeutic alliance rating——will be positively correlated with the outcome measures . Exploratory Outcome Hypotheses: Hypothesis 2Q will examine how the way in which symptom change is calculated affects the measurement of the relationship between therapeutic alliance and psychotherapy outcome. It is predicted that those calculation formulas which most take into account the amount of symptom change relative to initial level of symptomatology will be more highly correlated with the therapeutic alliance. Specifically, in order of the strength of the relationship, it is predicted that (a) residual gain, (b) standardized score change, (c) raw gain (pre-post difference scores), and, lastly, (d) raw post-therapy scores, will be correlated with the 74 therapeutic alliance. Hypothesis 2D. It is predicted that the patient's appraisal of certain dimensions of the therapeutic process-—measured via items from the Post Therapy Client Questionnaire (PQCL)—-will be correlated with the therapeutic alliance and other outcome measures which have been measured from other perspectives (i.e., symptom level, therapist evaluation). Also, level of object relations will exert an interactive effect on these relationships. Hypothesis 2E examines the relationship between the therapist's evaluation of the patient, and the psychotherapy process and other process/outcome measures. Specifically, it is hypothesized that there will be a positive correlation between items taken from the Post Therapy Therapist Questionnaire (PQTH) and (a) therapeutic alliance, (b) patient satisfaction and (c) symptom reduction. Again, level of object relations is predicted to interact with these factors. . In t- 9.. C. a». 3.. O .l\ 75 3. Interaction of Object RelationsL,Therapeutic Alliance and Outcome. Hypothesis 3A. There will be an interaction between level of object relations and each of the main process and outcome measures: therapeutic alliance, symptom change, client appraisal and therapist evaluation of psychotherapy. Overall, object relations is predicted to exert a (a) stronger effect for the low object relations group, and (b) be more associated with dynamic measures of outcome (PQTH, PQCL) than with symptom level or symptom change (SCL—90—R). 4. Additional Hypotheses The Interpersonal Checklist (ICL) will be used as an adjunct to the OR scale. This is, in essence, a measurement validation study. These two measures are hypothesized to assess, from different perspectives, and via different measurement techniques, complimentary aspects of the same core process of capacity for object relatedness (i.e., the ICL is a measure of interpersonal behavior assessed via self-report and the OR scales concentrates on representational processes indicative of interpersonal relationships via clinical judge ratings of relationship episodes). 76 Hypothesis 4. It is hypothesized that (a) there will be a significant positive relationship between the ICL and OR scales; and (b) the ICL will be correlated, in a pattern similar to OR, with the therapeutic alliance. . .nqqgfirehuussdswnhhaflhuuspQ. METHOD The purpose of the study will be to examine the relationship between patient and therapist factors that contribute to positive and negative treatment outcome via the therapeutic alliance. There are three major objectives: (a) to examine patient interpersonal factors that promote or inhibit the development of a therapeutic alliance; specifically, to study how patient early-therapy object relations are associated with the therapeutic alliance and therapeutic outcome; (b) to evaluate therapist, patient and interactional factors within the context of the therapeutic alliance; and (c) to assess the relationship between therapeutic alliance and psychotherapy outcome. The Michigan State University Psychological Clinic Psychotherapy Research Project served as the source of data for the present study. The data were collected over a seven year period (1978-1986) for the purpose of ongoing psychotherapy research at the Michigan State University Psychological Clinic. The clinic, a training and research agency of the MSU Department of Psychology, serves primarily non—student members of the 77 ,_. '.2" ' O 4. a? a—O -‘ — o .7; -. 1... ‘ ‘5. 'P ‘b— *6 fi 1.: up 9 O - - .- ...- U 1!- ..._ u— o - — o- —- '2- -—.. - ‘— - a mu. 0.. AU; .l. 78 community, providing psychological services, based upon an income-adjusted fee schedule, for adults, children and families. Ongoing data collection was begun to create a data base of psychotherapy process and outcome - en- ..p .. 4 a- .‘0- T. — -' - '0 - c - c. .9 .- o , .4— -. 1" c . ..- 7 - - .. 14 a . — «.4 - . a. . information from the clinic to be used in various research projects. The information consists of pre— and post-therapy client and therapist measures and audiotapes from selected psychotherapy sessions. During this seven year period there were over 200 cases for which at least some portion of research data were collected. Of these, approximately 115 cases included all necessary pre— and post—therapy measures. Cases that were at least ten sessions in duration and included the requisite number of audiotaped psychotherapy sessions comprised the final data base—— consisting of 41 cases-—used for analysis in the present study. Therapists Therapists participating in the MSU Psychological Clinic Psychotherapy Research Project were clinical psychology graduate students involved in the clinic practicum and internship program. At the beginning of each academic year practicum students and interns are given a general orientation concerning the overall psychotherapy research project and informed consent is 79 obtained at this point. The therapists' level of experience ranges from the first year of clinical practicum (second year in clinical program) to several years of experience and nearing completion of the doctoral degree to a few highly experienced Ph.D.'s. The therapeutic orientation was varied, but, in order of frequency, basic forms of interpersonal, eclectic, psychodynamic, and cognitive-behavioral perspectives were represented. From one perspective, the average level of experience is comparable to many community mental health centers (although supervision and various didactic support is typically more rigorous within a clinical program). From another perspective, the present study contrasts with many of the research projects which have studied the therapeutic alliance because the latter have used highly experienced clinicians with a more circumscribed range of psychopathology and diagnoses in their patient sample (e.g., Hartley & Strupp, 1983; Horowitz, Marmar, Weiss, Dewitt & Rosenbaum, 1984; Marziali, Marmar & Krupnick, 1981; Mintz, Luborsky & Cristoph, 1979). It will be important to compare the data of the present study with the research literature to consider the generalizability of the effects of therapeutic alliance on treatment outcome (as factors such as the severity of disturbance of the clinical population and level of 80 therapist training affect the therapeutic process). General Procedure During the patient's initial intake contact with the clinic they were informed of the Psychotherapy Research Project and for those patients who wished to participate in the project (and were deemed appropriate) informed consent was obtained. Patients were asked to complete the pre-therapy research packet (consisting of the written consent form, the Hopkins Symptom Checklist [SCL—90-R], the Interpersonal Check List [ICL] and a demographic information sheet) and return it to their therapist at the first ongoing psychotherapy session. Intake therapists completed a research packet consisting of a pre—therapy information sheet and a clinician's version of the Hopkins Symptom Checklist. Audiotapes of the entire psychotherapy session were obtained for the first, third and every fifth session thereafter, as well as the termination session. If one of the predetermined sessions was unavailable for audiotaping the subsequent session was substituted. A stamped, pre-addressed post-therapy research packet (consisting of the Hopkins Symptom Checklist [SCL—90-R], the Interpersonal Check List [ICL] and a post-therapy client questionnaire) was given to patie psycl to tl clier the t ccrpe that 11181133 IESea; the 3. Quest. the c. clie.. 81 patients upon completion of planned termination of psychotherapy, with instructions to mail the forms back to the Psychological Clinic upon completion. The client was provided with a financial compensation for the time invested in completing the questionnaires (the compensation was based upon a percentage of the amount that the patient paid for psychotherapy: the reimbursement was 10% of the amount paid up to a maximum of $50.00). If the termination was unanticipated or the therapist forgot to forward the forms, the post—therapy research packet was mailed to the patient directly, the instructions being the same. Therapists were given the therapist post—therapy research packet consisting of the clinician's form of the SCL—90-R and a post—therapy therapist questionnaire. The confidentiality and anonymity of the clients' research records was explained to the clients with the following qualifications: only the research coordinator and future researchers would have access to the information (all others--including the therapist——would not). This was done to ensure that patient self-reports could be honest and to keep therapists blind to research data and patient reports. All patient data were assigned a research code to insure anonymity. - - ---0- <— -. .- . —'. v . ‘ , — .4. ' run, .- ‘- _. 4 .1 . 4A ' - ~ .— ..— ‘—4 o c c. "L -'- a" -4 _. - I 4-4 - u -> sele crit 82 Subjects Forty-one cases from the tape library were selected for use in the present study based upon the criteria that (a) all necessary data were available on the subject and (b) the sample was representative of the total project data base in terms of age, sex and duration of therapy. The mean number of psychotherapy sessions was 30 (SQ 14.66), distributed as follows: 10- 20 sessions 16 (33%), 21-30 sessions = 14 (31%), 31- 6 (12%), 41—50 sessions = 9 (18%), 51 40 sessions sessions and above = 3 (6%). The mean age of the participants was 30 (pp 8.0), with a range of 20 to 57 years. Gender composition was 73% female and 27% male. The average educational level was 15.5 years (SQ 2.3) with a range of 11 to 20 years. The average annual income was $11,000 (SQ 8.2) with a range of $2,000 to $35,000. Half of the participants were single, one quarter were married and one quarter were separated or divorced. The occupations of the subjects varied, but half were in clerical or unskilled worker positions, about a third were in management and related positions and the remainder were students, homemakers or skilled laborers. All subjects were white. While no formal diagnostic assessments were conducted, all subjects were deemed appropriate for 83 outpatient psychotherapy. Scores from the SCL-90-R indicate that the mean Global Severity Index (GSI) for subjects at the beginning of psychotherapy was 1.27 (S2 = .63) and at post-therapy was .83 (SQ = .46). These scores are very similar to the normative data which Derogatis (Derogatis, Rickels, and Rock, 1976) compiled on over 1000 psychiatric outpatients. Instruments (1) Therapeutic Alliance. The most recent revision of the California Psychotherapy Alliance Scales (CALPAS; Marmar and Gaston, 1989) was used as the primary measure of therapeutic alliance. This multidimensional scale assessed patient, therapist and interactional contributions to the therapeutic process. The current version has integrated these aspects into five scales: (a) patient working capacity (PWC); (b) patient hostile resistance; (c) patient commitment (PTC); (d) therapist understanding and involvement (TUI); and (e) working strategies consensus (WSC). There are three versions of the CALPAS--therapist-rated, patient-rated and clinical judge—rated. Only the clinical judge-rated form (CALPAS-R) will be used in the present study. The CALPAS-R consists of 30 items rated on a 7 point degree of occurrence scale. Previous studies using the CALPAS and its predecessors (e.g., TARS, Marziali, 1984; measu- dSsES 84 CALTARS, Marmar, Weiss, & Gaston, 1989b) have shown the therapeutic alliance measure to have a high degree of internal consistency and inter-rater reliability (ranging from .73 to .85 on the subscales; Horowitz, Marmar, Weiss, DeWitt & Rosenbaum, 1984; Marziali, Marmar & Krupnick, 1981; Marmar et al., 1989b). In a recent study, inter-rater reliability on the CALPAS was as follows: TUI = .69, WSC = .71, PWC = .80, and PTC = .69 (L. Gaston, personal communication, March 20, 1989). Independent judges listened to and rated audiotaped excerpts from psychotherapy sessions. This provided a psychotherapy process measure as well as baseline information concerning the patient's initial level of therapeutic alliance. (2) Object Relations/Interpersonal Relations. Two measures were used: (A) Object Relations Scale: This scale, which assesses patient object relations on multiple dimensions based upon independent judge ratings, was used to provide a measure of patient pre-therapy object relations. The Object Relations and Social Cognition Scale (hereafter OR Scale) (Westen, Barends, Leigh, Mendel, & Silbert, 1988a) consists of four scales which assess four interdependent aspects of object relations. Each scale has a five point rating system (except for 'lll.‘. —-- ' .— _ c o u. a ' c - ‘ J .- - -- —- — o- v .a - . ’ .e 4— _ - 85 Complexity which has a seven point rating scale) where level one is associated with the lowest developmental level and level five with the highest. The four scales are described briefly: (a) Complexity and Differentiation of Representation [C]--this scale measures the extent to which others are seen as differentiated and complex persons; (b) Affect—Tone of Relationship [AT]——this scale assesses the degree to which the subject expects or perceives relationships to be destructive and dangerous or benevolent and enhancing; (c) Capacity for Emotional Investment [EI]—— this scale measures the degree to which the subject sees others as differentiated and whether or not they are viewed as primarily need—gratifying or as ends in themselves; and (d) Understanding of Social Causality-- this scale considers the logic, complexity, accuracy and psychological mindedness of the subject's description of interpersonal interactions (note: this scale was not used in the present study). The OR Scale has been used, as a whole or in parts, in a number of recent studies (e.g., Leigh, Westen, Barends, Mendel & Byers, 1989; Westen, Barends, Leigh, Mendel, & Silbert, 1988a; Westen, Huebner, Boekamp, Lifton, & Silverman, 1989) and has been found to possess a high degree of construct validity and inter—rater reliability; overall ratings have ranged from .72 to .94 for the four 4 a -\u 86 scales. A recent psychotherapy process study achieved inter-rater reliabilities of .82 to .91 across the scales (Schneider, 1990). Independent judges rated tape excerpts of interpersonal episodes gathered from the initial ongoing psychotherapy interviews (see data sampling section for an elaboration). The OR scale will be used in the present study to rate taped excerpts from early psychotherapy sessions. Brief descriptions of relationships, called interpersonal episodes, will be used as the basic unit of measurement. An interpersonal episode refers to aspects of the therapy session where the patient offers a self-description, or a narrative of his or her relationships with others or with the therapist. Each interpersonal episode will be separately rated on several dimensions. The actual subject of the episode (i.e., self, other, therapist) is noted. Then the degree of “completeness“ of the episode is rated. This comes from Luborsky's (1984, 1985) CCRT rating scheme and refers to the elaboration of the subject's wish and response, and the response of the other. Next, each of three (the social cognition scale will not be used in the present study) OR scales is separately rated (the rating process is described in detail in the methodology section). '! W!" 1'"? «1.16 *‘li'flflllli l’ ',6 cl-u'lI-vl'i'lmirt‘ri r‘ 87 The following is an example of the rating process for an interpersonal episode: Example #1 (Subject 8025, RE#10; A 22 y.o. depressed female) I'I still feel threatened by my mother. I was talking to her on the phone last night. I started getting really anxious and pissed because she wouldn't hang up. She was wanting me to come and visit her, and also she wanted to come and visit me. And, she was whining and acting like a baby; and wanting me to mother her and, just, it was just a really demanding situation. So I still have those feelings". In terms of the object of this discussion, it is clear that it is her mother-—a specific object (OS). The degree of completeness is around a 3.5. The wish is to not feel threatened or intruded upon. The response of the self is anger and anxiety, and the response of the other is intrusiveness and an emotional demandingness. Affect tone (AT) rates a 2. This is because there is a predominantly negative affect tone but not overwhelmingly so (anger, anxiety, intrusion) and generally negative expectancies. There is an air of disappointment. Emotional investment (EI) rates a 3. This is because there is a sense of an orientation toward the needs of self and others, a rudimentary sense of guilt and obligation within a conventional relationship. It may be toward a 2 rating because the patient is also responding out of obedience and anxiety. 88 Complexity (C) is more equivocal. A justifiable rating can be made for a 3 or, perhaps, a 4. There is adequate self—other differentiation; indeed the preservation of boundaries is what the patient is anxious about. There is an elaboration of the patient's internal states and some rudimentary attributions about the other and a sense of these being enduring traits. It may toward a 4 in the sense that the patient's own affective state is described in a fairly differentiated manner. (B) Interpersonal Checklist (ICL): This self— report measure was used to provide a measure of patient pre—therapy interpersonal relations. The ICL (Leary, 1957) was used as an adjunct to the object relations measures and was considered more exploratory in nature. The ICL is a questionnaire listing 128 behavioral adjectives which correspond to 8 (or 16, depending on the scoring approach) interpersonal styles. These items are placed into four quadrants based upon two bimodal scales: Acceptance-Rejection of Self (ARS) and Acceptance-Rejection of Others (ARO) (quadrants 1 and 2 are dominant, quadrants 3 and 4 are submissive; quadrants 1 and 4 are affiliative, quadrants 2 and 3 are hostile) (see the section on object relations for an elaboration). The ICL will be used in the present study (a) to help validate the OR Scales. Because l - a- <— .. A - ’- - 4.. f'. ., - a 4‘ u— up o ,. an i..- 1’ .- " .- .- -— ,4 89 these scales measure similar constructs from differing conceptual (object relations versus interpersonal relations) and measurement perspectives (self—report versus judge—rated), a comparative score will be considered; and (b) to cross—reference the OR Scales with regard to the therapeutic alliance. (3) Outcome Measures. A post-therapy client questionnaire, a post—therapy therapist questionnaire and a patient symptom checklist were used to measure outcome. (A) The Client Post Therapy Questionnaire (PQCL; Appendix E) is from Strupp, Lessler and Fox (1969) and consists of 56 items that assess beliefs concerning the effectiveness and outcome of therapy; each item has a nine point scale ranging from ”a great deal" to “to some extent“ to "not at all". The level of success regarding psychotherapy outcome from the client's perspective was judged from the client's ratings on selected questions. Correlations and content were examined for the individual items of the PQCL. The most relevant items were selected and subjected to a four, five and six factor principle components factor analysis. A four factor solution was judged to fit the chosen items best and the factors were conceptualized in the following manner: (1) CHANGE--the patient's 90 appraisal of change over the course of their therapy. This is a 7 item scale (questions 1, 2, 3, 4, 11, 15, 26) in which the patient is asked to appraise their further need for therapy, satisfaction with and benefit from treatment, symptom change and overall change, and whether external or personal factors led to termination; (2) PROCESS—-the degree to which the therapy was an intense, emotional and psycho- dynamically-oriented process. This includes five items (questions 19, 20, 41, 43, 44) which assess how much of the therapy the patient remembers, emphasis on feelings, intense and painful experience, therapist was neutral and therapist discussed psychoanalytic ideas; (3) EMPATHY--rating of the therapist's empathic abilities. There are five items (questions 27, 28, 33, 37, 51) evaluating the patient's perception of the therapist with regard to his or her level of attention and understanding, therapist saw patient as valuable and worthwhile human being, and wish to genuinely help patient; and (4) STYLE--assessing the therapist's technical style. There are five items (questions 34, 38, 39, 46, 50) concerning the therapist's integrity as a person, activity level, emphasis on childhood experience, fully accepting and warm in their style toward the patient. 91 The reliability of these factors was assessed and alpha coefficients were calculated for the four factors. The factors are considered to be relatively independent as they correlate from p = —.04 to p = .44 with each other. The alpha coefficient are as follows: CHANGE = .70; PROCESS = .62; EMPATHY =.80; STYLE = .74; based upon the new “scale", an overall alpha of .74. (B) The Post-Therapy Therapist Questionnaire (PQTH; Appendix E) includes ratings for the ten scales from the SCL-90-R therapist version as well as 23 items (note: some items had "before" and "after" therapy ratings) from Strupp, Lessler and Fox (1969) dealing with the therapist's subjective appraisal of the effectiveness of therapy. The level of success for psychotherapy outcome from the therapist's perspective was based upon the therapist's ratings on selected questions. The PQTH was subjected to the same factor development procedure as the PQCL, described above. The factor analysis produced an acceptable solution with the following factors: (1) SYMPTOM-—the therapist's rating of the patient's overall change due to psychotherapy. There are five items (questions 9, 12, 13, 17, 18) which consider the amount of symptom change, of change in personality, overall success of the therapy, amount of improvement expected and overall working relationship with the patient; (2) ANXIETY—- 92 level of anxiety and defensiveness. This scale include six items (questions 1, 2, 4, 10, 23) which assess the patient's level of anxiety at the end of therapy, defensiveness before and after therapy, counter- transference after therapy and an overall characterization of the therapist's experience with this patient; (3) STRUCTURE—-rating of more psychostructural or characterological traits. This factor has five items (questions 3, 5, 6) which concern the patient's overall adjustment afterwards, their ego- strength before and after therapy, and overall level of disturbance before and after therapy; (4) LIKING—-the therapist's subjective evaluation of their experience with this patient. This factor is composed of four items (questions 7, 11, 15, 40) assessing the therapist's emotional investment in the patient, enjoyment working with this “kind" of patient, personal liking for patient, and patient's warmth toward therapist). The factors seemed to be measuring reasonably orthogonal data with factor inter-correlations ranging from: p = .05 to p = .46 with each other. The alpha coefficients were as follows: SYMPTOM = .89; ANXIETY = .81; STRUCTURE = .85 LIKING = .82; yielding an overall (new) scale alpha of .88. (C) The (Symptom Checklist (SCL-90-R; Derogatis, .9” ‘ I pro Hy. 93 1977) client form (Appendix E), is comprised of ninety problem statements (this measure was used as both a pre- and post-measure). The statement of problems constitute nine symptom dimensions as follows: (a) somatization, (b) obsessive compulsiveness, (c) interpersonal sensitivity, (d) depression, (e) anxiety, (f) hostility, (g) phobic anxiety, (h) paranoid ideation and (i) psychoticism; and three global measures of psychopathology: the Global Severity Index, the Positive Symptom Distress Index, and the Positive Symptom Total. The Global Severity Index (GSI) was used in the present study. Clients' checked those statements, ranking the level of distress on a five point scale, that were indicative of current problems experienced at the beginning of therapy (pre-therapy symptomatology) and at termination (post-therapy symptomatology). Individual items within each symptom dimension were summed and averaged yielding a level of pathology score for each of the nine dimensions, as well as a Global Severity Index rating (GSI) noted above. The SCL-90—R, a widely used measure of psychological symptomatology, has been shown to possess adequate internal consistency and test-retest reliability on the symptom subscales (from .74 for psychoticism to .90 on depression), and on the global severity index (Derogatis, Rickels, and Rock, 1976). ‘ ”H -04 (I) f \C) 94 The SCL-90-R has successfully differentiated psychiatric outpatients from normals (e.g., Derogatis, 1977). Concurrent validity has been evaluated in several studies, one of which used the Middlesex Hospital Questionnaire, and correlations from .36 to .74 on similar symptom dimensions, as well as a correlation of .92 for equivalent global symptom scales, were achieved (Boleloucky and Horvath, 1974). The construct validity of the SCL—90-R has been assessed by Derogatis and Cleary (1977) using a varimax factor analysis on a sample of over 1000 psychiatric outpatients. The researchers found a significant correspondence between the empirically derived and theoretically based structure of the symptom dimensions. Calculation of Outcome. In order to circumvent some of the problems inherent in measuring outcome from a simple post-therapy symptom level or pre- to post— therapy symptom change score, the type of calculation used to measure outcome was considered. It was hypothesized that those formulas which are most sensitive to symptom reduction relative to initial symptom level would be better able to pick up meaningful overall symptom reduction in the sample. Four formulas were employed: (a) residual gain scores 95 (RSCL; Mintz, Luborsky and Christoph, 1979); (b) change in pre- to post-therapy standardized scores (ZCHG); (c) pre— to post-therapy difference scores (DIF); and (d) post-therapy symptom level (TGSI). The residual gain score has been employed in recent studies in an attempt to minimize some of the inherent bias toward pre-test levels when considering outcome. Cronbach and Furby (1970) note that the post— test score may be regarded as a deviation from the regression line of the pre-test—on—post-test. The residual gain score partials out that part of the post— test data which can be predicted from the pre—test data, in order to get a more meaningful measurement of outcome. The basic formula is as follows: 2 POST ‘ (z PRE) (r PRE, POST) The other outcome calculation formulas are also intended to compensate for the bias toward initial symptom level, though in a more direct manner. The ZCHG formula measures the change in pre— to post—test symptom level based upon standardized scores, while the DIF formula is a simple raw gain score. .C aru— /I\ r. 96 Data Sampling and Rating (1) Therapeutic Alliance. The general procedure for rating of the therapeutic alliance was as follows: 20 minute tape excerpts were taken randomly from the early (first and third), middle (40—60% through) and late (BO-90% through) phases of psychotherapy; also, the excerpts were randomly taken from the beginning, middle or end third of the psychotherapy session. A total of four excerpts (1 hour 20 minutes) were rated for each subject. Excerpts from all subjects will be placed in random order on a master tape, identified only by segment number. The training procedure for the CALPAS was as follows. Raters received instruction on the scale and became familiar with the coding manual; both raters have had previous experience with the scale and the therapeutic alliance literature. A pilot study was conducted, using 20 minute segments of psychotherapy sessions, and training continued until ratings achieved reliabilities of .70 or above for each of the four CALPAS subscales. The rating schedule was constructed such that the initial and final set of segments were rated by both raters, and then approximately 60% of the remaining segments overlapped, with a periodic recalibration occurring approximately every quartile over the course of rating the total of 192 segments. 97 Inter—rater reliabilities were calculated using a fixed-effects method of intraclass correlation coefficient (Shrout & Fleiss, 1979). (2) Object Relations. Two baseline measures were be developed: (1) Raters listened to tape excerpts gathered from early psychotherapy sessions. These excerpts were comprised of from 10 to 15 interpersonal episodes. The interpersonal episode constituted the basic unit of measurement to assess the patient's level of object relations. An interpersonal episode refers to aspects of the therapy session where the patient offers a narrative of his or her relationships with others or with the therapist. This approach was taken from Luborsky's (1984, 1985) Core Conflictual Relationship Theme (CCRT) method in which relationship episodes are the basic component. Three basic criteria were used to identify episodes: (a) patient description of a relationship, (b) identification of the object with which the relationship occurs, and (c) rating the episode in terms of its completeness (Luborsky, 1985). The interpersonal episodes were gathered, using the criteria mentioned above, from tape recordings of the first few sessions of the patient's psychotherapy. The actual episodes were compiled by the experimenter listening to the first ongoing psychotherapy session, 98 collecting interpersonal episodes which meet the above criteria, and continuing until approximately 15 interpersonal episodes were identified, and/or until the end of the third taped session. The interpersonal episodes were placed contiguously on a tape identified by episode number and patient research code number. Validity of these episodes were based upon inter-rater reliability of the extent to which the segments met the "completeness" criteria of an interpersonal episode. The training procedure began with raters reading the Manual for Codinq Object Relations and Social Cpgnition from Interview Data (Westen, Barends, Leigh, Mendel & Silbert, 1988a) and becoming familiar with the specific scoring criteria for each of the three scales. This very thorough manual also included a number of practice exercises (which the raters completed). Pilot ratings were done to assess the level of inter-rater reliability and attempts were made to keep reliabilities at the .70 level or higher. As with the therapeutic alliance scale, inter-rater reliabilities were calculated using a fixed-effects method of intraclass correlation coefficient (Shrout & Fleiss, 1979). Training sessions between the experimenter and the two raters were conducted to identify problem areas and to improve reliability. Periodic checkups were conducted and recalibration was implemented as needed. £1 any 5. r. .me «Q Du 99 The actual rating procedure was as follows. Raters first listened to a particular subject's entire set of interpersonal episodes, rating the level of completeness of each interpersonal episode and noting the "object" of the episode (i.e., self, other-- specific or general, or therapist). Next the rater rated each episode on the affect tone and emotional investment scale; the rater then listened to the tape once again and rated each episode on the complexity scale. These three ratings provided the patient pre- therapy level of object relations. (2) Ratings based upon the affiliation—hostility (ARO) scale of the ICL were used, in addition to the object relations ratings, to provide a measurement of patient pre—therapy object relations. As discussed earlier, this also functioned as a validation study of the measurement of patient pre-therapy interpersonal factors. Raters Ratings for therapeutic alliance (CALPAS) and for object relations rating (OR Scale) were obtained from two separate teams of judges. Each team was comprised of two persons connected with the MSU clinical psychology doctoral program with at least four years of supervised clinical experience (the CALPAS ratings were do: po: Sca 1111 be I8! “:1 I'd? 100 done by two males, one of whom is three years postgraduate, the other is a doctoral intern; the OR Scale ratings were done by two females, one is two years postgraduate, the other has completed a two year internship). Attempts were made to have parity on crucial variables between judges--all four judges have a strong psychodynamic orientation in part derived from the clinical program and the internship; all four have been involved to a significant extent in psychotherapy research, and all are very familiar with the therapeutic alliance literature (three of the four raters have published in this area). RESULTS Reliability of Measures Object Relations Scale (OR Scale). The reliability of the OR Scale was assessed (SPSS/PC+ reliability) and alpha coefficients were calculated for the three subscales (AT--affect tone, EI--emotional investment, C--complexity; alpha = .64). The three scales correlated moderately with each other (see Table 2), indicating a reasonable amount of independence. The alphas and inter—scale correlations are similar to those reported by Westen, Barends, Leigh, Mendel and Silbert (1988a). Examination of the score frequencies indicates that the distribution was skewed and not normal, and that there was a restriction of range in the rating scores. Forty—six percent of the interpersonal episodes were rated by both raters (i.e., 209 of 441 interpersonal episodes). The intraclass correlation coefficient recommended by Shrout and Fliess (1979) was used; this ICC formula calculates an estimated reliability of ratings for one rater based upon a two- way random—effects ANOVA. The ICC's for the OR sub- scales are as follows: AT = .49, E1 = .38 and C = .57. 101 102 Table 2 Meapg and Correlations for Object Relations Scale Scale Mean SD Affect Tone (AT) 2.79 .34 Emotional Investment (EI) 2.79 .26 Complexity (C) 3.62 .67 Correlations AT EI C AT 1.000 .5580** .4877** E1 1.000 .4202* C 1.000 fl=41 1-TAILED SIGNIFICANCE = *.01 **.001 CHEN +F n UKA LL In distri ind: .. iCa P ‘ange 103 there inter—rater reliabilities are considerably lower than those reported by Westen, Barends, Leigh, Mendel and Silbert (1988a). However, the complexity scale met the criteria proposed by Jacobson, Follette and Revenstorf (1984), suggesting that the data may be interpreted with caution. The reliabilities ratings were low given the amount of time and effort devoted to rater training and previous inter-rater reliabilities of the object relations practice sets. Two possible contributing factors may be suggested. First, there is a considerable restriction of range in the frequency distributions of the three subscales. As Lahey, Downey and Sall (1983) have advised, a restriction in variability can diminish intraclass correlation coefficients. Second, this is a relatively unique adaptation of this scale (i.e., to these particular type of data); this issue will be considered in the discussion section. In addition, a percentage agreement score was calculated on the same, mutually judged cases; the exact agreements were: AT=60%, EI=67% and C=39%. CALPAS. Examination of the frequency distributions (total and by subscale) of the CALPAS indicate that there is a significant restriction of range and that the distribution is kurtotic and non- were Solut the b. organs PTC), 104 normal. The measurement model of the therapeutic alliance factor was evaluated by testing the reliabilities of the five subscales and the overall CALPAS. Because there was a statistically significant sessions effect for each of the subscales (i.e., between the first, third, middle and final sessions), reliabilities were first calculated for each session. However, because there was little variation from session to session, reliabilities averaged across sessions were used. The averaged subscale alpha coefficients are as follows: Patient Working Capacity (PWC) = .91; Patient Hostile Resistance (PHR) = .70; Patient Commitment (PTC) = .87; Working Strategy Consensus (WSC) = .86; Therapist Understanding and Involvement (TUI) = .90; and overall CALPAS alpha = .89. The subscale correlations, item—factor and item— total reliabilities are very high, suggesting a lack of independence between the subscales. Consequently, a factor analysis was conducted on the thirty CALPAS items and two, three, four and five factor solutions were calculated. Aside from the existing five factor solution, a three factor solution appeared to provide the best fit. In this solution the three factors were arranged as follows: patient positive scales (PWC and PTC), patient negative scale (PHR) and therapist neg CW 156 C01 est 105 negative scale (TUI). The alliance scale (WSC) overlapped with each of the factors and could not be used. However, the three factor solution yielded lower correlations with other variables indicating that some fine-grained aspects of the relationship between the variables may have been lost. As a result, because the original five factor solution correlates more strongly with a wide range of variables in this study, a decision was made to use the original five factor solution. The rating scheme for this study utilized 37 of the total 164 CALPAS session segments (i.e., 23%) rated by both judges. Because of this, inter-rater reliability was examined via the intraclass correlation coefficient (ICC). The formula used by the ICC estimates the reliability of a single session rating by one judge utilizing a two—way random-effects ANOVA. An overall inter-rater reliability was calculated based upon a weighted average of the combined ratings of the judges. The weighted averages were: PWC=.42, PHR=.48, PTC=.36. The Spearman-Brown Prophecy formula was then used to calculate inter-rater reliabilities at the treatment level (i.e., the average of four sessions): PWC=.74, PHR=.79, PTC=.69. The inter-rater reliabilities were somewhat lower than those found in previous studies (e.g., Marmar and Gaston, 1989). The acrc , :0: pet; COHEc 106 individual session ratings ranged from .25 to .87. Some possible explanations for this can be advanced. Table 2 indicates a considerable restriction of range which can be associated with decreased ICC (Lahey, Downey and Sall, 1983). Because the ratings were done over a long period of time, the possibility of rater drift seems likely; this appears to be reflected in the diminished reliabilities for the final session. Because of this, inferences made based upon the CALPAS ratings of the final session must be viewed with some caution. The therapeutic alliance subscale means compared across session indicated a significant sessions effect for some of the subscales. Although there were significant differences found for third and middle sessions as well, only the initial and final therapeutic alliance sessions were examined in this study. Paired sample t—tests indicated that the patient working capacity (PWC) subscale showed a trend toward significance (M = 5.16, 5.41) t = (df = 40) —1.83, p = .07; for patient hostile resistance (PHR), a significant change was found: (M = 1.81, 1.53) t = (df = 40) 2.80, p = .008; no significant difference for patient commitment (PTC) or for working strategy consensus (WSC) was found. r ‘3‘“ Q 1 PM In PVC-p PHP ('F DTC-F 107 Table 3 Mean Scores for CALPAS bv Session .— - -- 1— vi. .- ' -.- ’ . o — .- -.- - '- -.- -. w - - Scale PWC PHR PTC WSC TUI Humv- «mu SESSION MEAN SD MEAN SD MEAN SD MEAN SD MEAN SD FIRST 5.17 .82 1.81 .54 5.16 .65 5.24 .76 5.22 .83 THIRD 4.92 .79 1.81 .67 4.74 .80 4.76 .78 5.13 .72 MIDDLE 5.41 .56 1.39 .39 5.27 .56 5.35 .58 5.51 .63 FINAL 5.42 .54 1.53 .56 5.30 .59 5.44 .52 5.47 .61 TOTAL 5.23 .46 1.64 .35 5.11 .44 5.20 .47 5.33 .55 PVC - Patient Working Capacity USC - Working Strategy Consensus PHR - Patient Hostile Resistance TUI - Therapist Understanding and Involvement PTC - Patient Commitment Hypotheses Hypothesis 1A examined the relationship between object relations (OR) and the initial therapeutic alliance (TA). There was a significant relationship between certain subscales of the object relations measures and the first session therapeutic alliance; there was no relationship for the third session. A consistent relationship between the object relations subscales and the CALPAS patient working capacity (PWC) subscale was seen (Affect Tone [AT] 3 = .30, p = .028; Emotional 1an r‘. 1' 57.1 Co~ Becé ‘¥L " 7 «e ‘at 108 Investment [EI] ; = .27, p = .04; and Complexity [C] r = .55, p = .0001). Further, the C subscale was significantly related to all five CALPAS subscales (PWC ; = .55, p = .0001; PHR ; = -.26, p = .04; PTC g = .44, p = .002; WSC ; = .42, p = .003; TUI ; = .25, p = .06). The prediction that the relationship between OR and TA would continue across session was not supported. The relationship diminished by the third session and remained nonsignificant. It should be mentioned that the original choice for statistical analysis of the hypotheses under consideration was multiple regression. It was believed that a multiple regression analysis would allow for the most sophisticated and accurate examination of the interaction among therapeutic alliance, object relations and psychotherapy outcome. Several different combinations of the therapeutic alliance and object relations subscales, as well as different outcome calculation formulas, were utilized in the multiple regression formulas (stepwise method), but no statistical significance was achieved when using the comprehensive interactive model described earlier. Because of this the focus of the statistical analyses shifted to a more direct examination of the relationships between specific variables. 109 Hvoothesis 13 stated that OR would exert an interaction effect on therapeutic alliance across sessions, with a stronger interaction occurring for lower object relations. This relationship was examined by dichotomizing the object relations scale into low and high complexity groups (M um = 3.09; M mm,==4.13) according to a mean split. As Table 4 shows, object relations differentiated the level of initial therapeutic alliance for three of the four CALPAS scales (PWC, PTC, WSC), but there were no significant differences between low and high complexity levels on CALPAS scales at final therapeutic alliance. While subjects with higher complexity had higher levels of initial therapeutic alliance and subjects with lower object relations had lower levels of initial alliance, there was no significant difference observed between the complexity groups for the final therapeutic alliance. Figure 2 depicts the relationship between object relations and alliance (initial and final). V: W... 1:12 Figure 2 110 Means for Initial and Final Therapeutic Alliance by Obiect Relations Level (Complexity) Mean Level of Therapeutic Alliance Across Sessions High .1, as. Complexity Low Complexity I l I I Initial Final Session Session Therapy Session Table 4 Mean Scores for CALPAS by 111 Session and Level of Complexity M SD df t p FIRST SESSION PWC LOW 4.81 .72 39 —2.96 .005 High 5.50 .79 PHR LOW 1.91 .43 39 1.16 .25 High 1.72 .63 PTC Low 4.95 .47 39 -2.18 .04 High 5.35 .74 WSC LOW 4.99 .76 39 -2.21 .03 High 5.49 .69 FINAL SESSION PWC Low 5.39 .51 39 —.29 .77 High 5.44 .57 PHR Low 1.45 .47 39 —.99 .33 High 1.62 .64 PTC Low 5.34 .47 39 .43 .67 High 5.26 .70 WSC Low 5.43 .47 39 .09 .93 High 5.45 .57 Low = 20; 9 HIGH = 21 PRC - Patient working Capacity PHR - Patient Hostile Resistance PTC - Patient Commitment NSC - working Strategy Consensus Hygothese' afiect to complexit pantie a? relatior alliance limita: {Blatio were nc ObjECt hYpoth. m A" othes be “egat 112 Hypotheses 1C examined the three OR Scale variables—— affect tone (AT), emotional investment (EI), and complexity (C)—-and their relationship with the thera— peutic alliance. Significant relationships were noted with the initial session PWC (AT ; = .30, p = .028; EI ; = .27, p = .04). There were no significant relationship for the third, middle or final therapeutic alliance sessions. Because of the statistical limitations noted earlier and the absent of relationship with other variables, the AT and EI scales were not considered further. This shifted the focus of the OR Scale to complexity (C). As indicated earlier, there was a significant relationship between C and all five CALPAS subscales at the initial session only. All hypotheses were confirmed, in the correct order, with PWC having the strongest relationship. Hypotheses 1C—1 stated that the patient positive scales (PWC, PTC) would be the most strongly correlated with object relations among the CALPAS subscales. This hypothesis was confirmed (PWC ; = .55, p = .0001; PTC ; = .44, p = .002). Hypothesis 1C:2 predicted that object relations would be negatively correlated with patient hostile resist- ance. This was supported (PHR ; = —.26, p = .04). Risothe sorrel: ‘viO‘C $111 supper objeC' posit 113 Mgpothesis 1C-3 stated that object relations would be correlated the "general" therapeutic alliance (i.e., working strategy consensus, WSC); this hypothesis was supported (WSC ; = .42, p = .003). Overall, these findings suggested that level of object relations was meaningfully related to patient positive and negative contributions to the therapeutic alliance for the initial session, but not for the final session. Negative contributions were less strongly correlated than predicted, while the working strategy consensus was more strongly correlated than predicted. Hypothesis 1D proposed that the greatest amount of within-group variance would be found in the low object relations group. Examination of the Patient Working Capacity (PWC) for the initial alliance rating (M um = 4.80, SQ = .72; M mm,= 5.50, S2 = .79) and the final alliance rating (M um = 5.39, S = .51; M mm,==5.44, S2 = .57) was not supportive of this hypothesis as there was greater variance in the high complexity group at the beginning and end of treatment. Hypothesis 2 considered the relationship between therapeutic alliance and outcome. The multidimensional aspects of psychotherapy outcome were addressed by an evaluation from multiple perspectives--symptom E129 "tre wide the: the: 90-R (GSI overé Addit Showe faile 114 reduction, patient appraisal and therapist evaluation of outcome. Hypothesis 2A predicted that there would be an overall "treatment effect". There was a significant and widespread reduction of symptoms from pre- to post- therapy for 71% of the subjects. Table 5 summarizes the means for pre— and post-therapy scores on the SCL- 90-R. As a summary measure, the Global Severity Index (GSI) of the SCL-90—R indicated that there was an overall reduction in symptoms at post-therapy (M.mw = 1.27; M mm.= .83; t_= [df = 40] 3.99, p < .001). Additionally, seven of the nine SCL—90—R subscales also showed significant symptom reduction at post-therapy at the .005 level. Two scales——hostility and paranoia-- failed to reach statistical significance. Table 5 115 Mean Scores for Pre— and Post—Therapy SCL—90-R Pre—Therapy Post—Therapy Scale Mean SD Mean SD Global Sever. Index(GSI) 1.27 .63 .83 .46 Somatization (SOM) .80 .71 .43 .38 Obsessive-Comp. (OBS) 1.57 .82 1.07 .64 Interpers. Sens. (INT) 1.67 .85 1.15 .64 Depression (DEP) 2.00 .92 1.36 .82 Anxiety (ANX) 1.55 .89 .88 .59 Hostility (HOS) 1.05 .84 .66 .62 Phobic Anxiety (PHO) .70 .75 .45 .56 Paranoia (PAR) 1.15 .83 .95 .66 Psychoticism (PSY) .91 .67 .54 .38 41 I2 u 116 Hypothesis 2B. This hypothesis examined the relationship between the therapeutic alliance-—initial and overall-— and psychotherapy outcome. Beginning with the examination of the most basic relationship of therapeutic alliance and outcome, an analysis between therapeutic alliance and Global Severity Index (GSI) was conducted. The prediction was that level of initial and overall therapeutic alliance would be positively correlated with symptom reduction at outcome. As Table 6 illustrates, there was no simple, direct relationship between therapeutic alliance and the post-therapy Global Severity Index (nor between alliance and pre—therapy symptoms). One relationship was significant (WSC ; = -.26, p < .05), implying that lower pre-therapy symptoms were associated with the working alliance. Since the basic hypothetical model predicted an interactional relationship, the focus will turn to a more fine—grained analysis of the intricacies of the relationship between psychotherapy process and outcome in the section dealing with the interaction between object relations, therapeutic alliance and outcome (hypothesis 3A). 117 Table 6 Correlations Between Therapeutic Alliance and Global Severity Index (GSI) Initial Session Final Session PreGSI TGSI PreGSI TGSI PWC —.07 -.19 PWC —.04 .22 PHR .11 .06 PHR .13 —.05 PTC —.21 —.21 PTC —.08 .06 WSC —.31* —.09 WSC .05 .19 M=41;*=p<.05) PWC - Patient Working Capacity PTC - Patient Commitment PHR - Patient Hostile Resistance WSC — Working Strategy Consensus 118 Hypothesis 2C. It was predicted that the way in which symptom change was calculated would have an impact on the statistical relationship between therapeutic alliance and outcome. The hypotheses stated that the formulas which were the most sensitive to symptom change relative to initial symptom level might best measure the alliance/outcome relationship. In order of strength of relationship, the predictions were as follows: The most predictive outcome calculation formula was hypothesized to be the residual gain score (RSCL) which is purported to adjust the post—therapy score for the portion contributed by the pre—therapy score. The next outcome calculation formula was the difference between standardized pre- and post-therapy scores (ZCHG). The third most predictive formula was raw gain (DIF) which simply measured the pre— minus post—therapy raw score differences. It was hypothesized that the least predictive outcome calculation formula would be poet—therapy symptom level (TGSI). Table 7 summarizes the correlations between the i115-135431 and final CALPAS subscales and the outcome calcrullation formulas noted above. 1“Tor the initial therapeutic alliance, the residual gain Outcome calculation formula (RSCL) was correlated Wit; h the therapeutic alliance for the working strategy 119 consensus (WSC) scale (p = .29, p <.05). The pre- minus post-therapy standardized scores (ZCHG), was correlated with the initial therapeutic alliance for the WSC scale (p = .40, p (.01). The raw gain outcome calculation formula (DIF) was correlated with three of the four initial therapeutic alliance scales (PHR p .30, p <.05; PTC p = —.33, p <.05; and WSC p = .47, p < .01). The outcome calculation formula hypothesis formula was not confirmed for the initial therapeutic a]. liance. The residual gain score (RSCL) was significantly correlated with the final therapeutic alliance (PWC p -.28, p <.05; WSC _1; = .30, p .40, p (.01; PHR p - The pre- minus post—therapy standardized score < .05). = .33, p difference (ZCHG) was correlated with PWC (_r; < . 05) and PHR (p = -.27, p <.05). The pre- minus Poet-therapy difference score (DIF) was correlated with PWC (p = -.26, p <.05). Finally, there were no Significant correlations between post—therapy symptom level (TGSI) and either the initial or final therapeutic alliance. Hypothesis 2C was only partially confirmed. The general findings were: (a) initial therapeutic alliance and Outcome were not correlated for the major scales; only a peripheral scale (WSC) was correlated; (b) there is a fairly pervasive and consistent relationship 120 between final therapeutic alliance and outcome, depending on the way in which outcome is calculated; (c) the order of strength of the correlations with final alliance was as follows: residual gain score (RSCL), pre- to post-therapy standardized score differences (ZCHG), then pre- to post—therapy (difference scores (DIF). Thus, the prediction that (Jutcome calculation formulas which were more sensitive t1) the relative level of symptom reduction--with the above qualifications in mind-—would be more highly cxorrelated with the therapeutic alliance was confirmed for the final alliance only. Table 7 Correlations Between Therapeutic Alliance and Psychotherapy 121 Outcome Using Different Calculation Formulas PHC - Patient Working Capacity i’liR - Patient Hostile Resistance PTC — Patient Commitment ‘JESC - Working Strategy Consensus Alliance RSCL ZCHG DIF TGSI INITIAL PWC .16 .18 -.23 —.19 PHR -.15 .18 .30* .06 PTC .14 .25 —.33* -.21 WSC .29* .40** .47** —.09 FINAL PWC .40** .33* -.26* .22 .PHR .28* .27* .16 —.04 PTC .24 .22 .21 .06 VVSC .30* .19 .16 .19 M = 41 *p < .05 **p < .01 RSCL - Residual gain ZCHG - Standardized change DIF - Pre/post difference TGSI - Post-therapy symptoms 122 It was predicted that the patient's Hypothesis 2D. appraisal of certain dimensions of psychotherapy would be related to the overall process and other outcome measures. This hypothesis was examined via the correl— ations between selected items from the Post Therapy Client Questionnaire (PQCL) and (a) the therapeutic alliance, (b) post-therapy symptom level, and (c) the Post Therapy Therapist Questionnaire (PQTH). Four factors were developed were developed from this 57 item scale (see Appendix E): (a) CHANGE—-the patient's appraisal of change over the course of psychotherapy; (b) PROCESS--how dynamically oriented treatment process was; (c) EMPATHY——perception of the therapist's empathy; and (d) STYLE--impression of the therapist's technical style and activity level (Note: the items are scored such that a low score indicates a £308 itive response). A matrix which lists the correlations among PQCL items, PQCL factors, therapeutic alliance, and other outcome measures is summarized in Table 8. There was a set of significant relationships between the initial therapeutic alliance and EMPATHY (we r = -.48. p < «- 01; PTC p = -.44, p < .01). For the final t herapeutic alliance, EMPATHY was correlated with a - p t‘lent hostile resistance (PHR p = .44, p < .05). The r6 was a relationship between STYLE and the final 123 - —.37, p < .01; PTC therapeutic alliance (PWC p p = -.52, p < .01). The patient's appraisal of CHANGE was correlated with the post-therapy Global Severity < .05), as well as three of the Index (TGSI p = .31,p EMPATHY was correlated with TGSI nine SCL—90—R scales. _r_’ = .27, p = (.05) and three of the nine SCL—90—R There were several correlations between the scales. Scales patient and therapist scales (PQCL and PQTH). measuring symptom change and therapeutic success (CHANGE, SYMPTOM) were correlated (p = -.42, p (.01). The patient's appraisal of the therapist's empathy was correlated with the therapist's evaluation of overall ego strength (p = —.33, p <.05). All four PQCL scales correlated with ANXIETY for the therapist scale-~from r = .29 to p = —.35, indicating that (From the therapist's perspective) as patient anxiety level decreased, appraisal of change, the therapeutic process and the therapist's style became more positive. There was qualified support for hypothesis 2D in that (a) the Patient's appraisal of therapist empathy was correlated with initial therapeutic alliance and patient appraisal of therapist style was correlated with final alliance ratings; (b) patient perception of change was c"D-‘E‘Il’i‘e.‘Lated with symptom reduction; (C) and there was congruence between the patient's and therapist's 9V a l u ation of symptom change. 124 Table 8 Correlatiops Between the Post Therapy Client Questionnaire (PQCL) and CALPASL SCL-90-R CHANGE PROCE SS EMPATHY STYLE Alliance PWCl -.05 .01 -.48** -.03 PHRl .17 —.18 .24 .19 PTCl -.15 .13 -.44** .06 PWC4 .11 .09 —.04 -.37** PHR4 .08 —.04 .32* .23 PTC4 —.12 .02 -.20 -—.52** P TH SYMPTOM -—.42** —.18 —.22 -.12 .AJQXIETY -.35* -.32* -.34* -.29* STRUCT —.25 -.08 -.33* .03 LIKING .19 -.08 .05 .01 SCL-90—R T—SOM .14 .03 .06 .08 T—OBS .32* .08 .10 .01 T- INT .19 .02 .36** .11 T—DEP .35* .04 .13 .06 T—ANX .35* .01 .11 .04 T—HOS .10 .06 .08 .04 T—PHO .22 .07 .20 .13 T—PAR .25 .09 .51** .23 T—PSY .25 .18 .45** .03 we 'I'GS I .31* .05 .27* -.02 RSCL .35* -.03 —.03 —.02 DIF -.23 .01 .01 .16 \ M = 41 * = p (.05 ** = p (.01 PHC — Patient Working Capacity SON - Somatization I’liR - Patient Hostile Resistance OBS - Obsessive-Comp. l’irc - Patient Commitment INT - Interpers. Sens. “SC - Working Strategy Consensus DEP - Depression ANX - Anxiety HOS - Hostility PHO - Phobic Anxiety PAR - Paranoia PSY - Psychoticism RSCL - Residual gain ZCHG - Standardized change DIF - Pre/post difference TGSI - Post-therapy symptoms 125 Hypothesis 2E examined the relationship between the therapist's evaluation of the patient, psychotherapy process and outcome. A positive correlation was predicted between the Post Therapy Therapist Quest— ionnaire (PQTH) and the therapeutic alliance, and other outcome measures (symptom level, PQCL). As with the PQCL, relevant items were taken from the PQTH and, via a scale development procedure, factor analyzed into :flour factors (see Method section). The factors which eunerged were: (a) SYMPTOM--the therapist's evaluation cxf the patient's progress, symptom and personality (b) ANXIETY--rating the patient's level of cilange; (c) STRUCTURE alnxiety before and at the end of therapy; -—measurement of so-called psycho-structural factors (ego-strength, capacity for insight, and overall acflgjustment; and (d) LIKING——the therapist's feelings arlci countertransference concerning the patient. Table 9 reviews the correlations between the PQTH fulci the process and outcome measures. The relationship .bet:vveen the PQTH factors and the therapeutic alliance indicates that the STRUCTURE factor was significantly In addition, three correlated (PHR _; = -.28, p <.05). out: <>f a possible 36 correlations were significantly COL"IZ'C‘eeJated. The therapist's evaluation of the pat‘jL-eent's level of anxiety was correlated with global post‘therapy symptoms (TGSI E = -.27, Q < .05). 126 Table 9 Correlations Between the Post Therapy Therapist Questionnaire (PQTH) and CALPAS! SCL-90-RLyPQCL SYMPTOM ANXIETY STRUCT LIKING Alliance PWCl .09 —.01 .06 -.O3 PHRl -.16 .14 -.28* —.11 PTCl .20 -.01 .11 .12 PWC4 —.13 —.10 -.19 .09 PHR4 .09 —.17 —.25 —.08 PTC4 --.13 .10 -.04 .19 P CL CHANGE —.42* —.35* —.25 .19 PROCESS -.18 -.32* -.09 .07 EMPATHY -.22 -.34* —.33* —.05 SCPYLE .12 -.29* .03 .01 SCL—90-R fP-SOM -.22 -.11 .05 -.26* T—OBS -.19 -.17 —.15 -.09 fP—-INT —.11 —.18 —.15 -.07 T—DEP -.34* —.11 -.21 —.07 T—ANX -.16 -.18 —.10 —.04 T—HOS —.04 -.24 -.04 —.16 T—PHO -.01 -.17 -.07 —.08 T—PAR —.23 -.36* —.22 -.16 T—PSY -.24 -.25 -.18 -.08 Gem TGSI -.22 —.27* —.16 —.09 RSCL -.24 -.23 —.08 .01 DIF .16 -.09 -.16 -.06 \ H = 41 * = p <.05 PWC - Patient Working Capacity 50" - Somatization PHR - Patient Hostile Resistance OBS - Obsessive-Comp. PTC - Patient Commitment INT - Interpers. Sens. WSC - Working Strategy Consensus DEP - Depression ANX - Anxiety HOS - Hostility PHO — Phobic Anxiety PAR - Paranoia PSY - Psychoticism IRSCL - Residual gain ZCHG - Standardized change DIF - Pre/post difference TGSI - Post-therapy symptoms 127 Hypothesis 3A focused on the effect of object relations as an important moderator variable. For several of the central hypotheses, it was predicted that the dichotomization of variables according to low and high complexity would permit a more fine-grained analysis of the relationship and potential interactions between process and outcome measures. The effect of level of complexity was first considered by re-examining outcome (i.e., SCL—90-R). A review of Table 10 shows that for the low complexity group, the Global Severity Index (GSI) and eight of nine subscales (hostility was not significantly reduced) indicated a significant symptom reduction at post—therapy. For the high complexity group, the GSI and five of nine subscales (obsessive-compulsive, interpersonal sensitivity, phobic anxiety and paranoia were non—significant) showed significant pre- to post- therapy symptom reduction. However, group mean .2 -tests indicated that the low and high group did not Significantly differ from each other for (a) pre- therapy (PreGSI M Low = 1.39, M HIGH = 1.15, p (39) = 1. 22, p,= .23), (b) post—therapy (TGSI M um = .89, M me“ = .78, p (39) = .76, p = .45), nor (c) in the amount of pre- to post-therapy symptom reduction (DIF M mw ==.59, M mm,= .44, p (39) = .66, p = .51). The exétmination of differences between high and low 128 complexity groups across subscales indicated that only two of the nine subscales differed at pre—therapy (obsessive—compulsive, interpersonal sensitivity), none at post therapy, and none when comparing pre- to post- therapy symptom reduction. Overall, the dichot— omization according to level of complexity did not provide any additional information beyond that which was learned earlier. 129 Table 10 Mean Scores for Pre— and Post—TherapyfiSCL—90—R by Level of Complexity Low Complexity Pre Post Difference GSI 1.39 .89 .50 SOM .80 .45 .34 OBS 1.84 1.20 .63 INT 1.92 1.20 .71 DEP 2.13 1.43 .70 ANX 1.61 .94 .67 HOS .97 .70 .28 PHO .78 .44 .34 PAR 1.36 1.00 .36 PSY 1.10 .62 .47 High Complexity Pre Post Difference GSI 1.15 .78 .37 SOM .81 .40 .41 OBS 1.32 .94 .38 INT 1.43 1.10 .33 DEP 1.87 1.30 .58 ANX 1.49 .81 .68 HOS 1.13 .63 .50 PHO .62 .46 .16 PAR .96 .91 .05 PSY .73 .46 .27 glow = 20; EHIGH = 21 GSI - Global Severity Index SON - Somatization OBS - Obsessive-Comp. INT - Interpers. Sensitivity DEP - Depression ANX - Anxiety HOS - Hostility PHO - Phobic Anxiety PAR - Paranoia PSY - Psychoticism 130 As table 11 illustrates, there were no significant correlations for either the initial or final thera- peutic alliance (patient working capacity [PWC]) and the SCL-90-R when the complexity dimension was considered. There were a few significant correlations (Low complexity—-post—therapy anxiety and final PWC; high complexity-—initial PWC and pre-therapy paranoia, psychoticism, post—therapy psychoticism, and final PWC and pre-therapy anxiety), and there was a trend toward significance for a final PWC/TGSI correlation. However, no consistent or interpretable pattern was established. 131 Table 11 Correlations Between Pre— and Post—Therapy SCL—90—R and Therapeutic Alliance (PWC) by Level of Complexity Initial Alliance Final Alliance PreSCL PostSCL PreSCL PostSCL Low Complexity GSI —.06 —.03 .09 .35 m==nm SOM -.10 .03 —.16 .28 OBS —.26 -.04 —.10 -.05 INT -.09 -.04 -.26 .31 DEP .03 .07 .12 .33 ANX -.06 .03 .28 .39* HOS .08 .31 -.01 .38 PHO —.21 .04 .14 .33 PAR .01 -.30 .17 .24 PSY .19 -.13 .16 .22 Initial Alliance Final Alliance Pre Post Pre Post High Complexity GSI .01 —.25 22 .13 SOM .06 -.10 .09 .22 OBS -.08 —.17 -.29 .21 INT -.15 -.34 -.10 .06 DEP .15 —.10 -.04 .21 ANX .12 —.08 -.37* .12 HOS .27 —.10 .01 .15 PHO .03 -.24 —.26 .02 PAR -.39* -.45* -.27 .01 PSY -.24 -.36* -.27 .05 p_um = 20; p mm,= 21 * = p (.05 ** = p (.01 GSI SOM OBS INT DEP ANX HOS Global Severity Index Somatization Obsessive-Comp. Interpers. Sensitivity Depression Anxiety Hostility PHO Phobic Anxiety PAR Paranoia PSY - Psychoticism 132 A Chi Square analysis indicated that object relations exerted an interactive effect when process and outcome are mutually examined. Table 12 examines the interaction of object relations level (i.e., complexity), therapeutic alliance, and level of outcome (post-therapy SCL-90-R scores). Object relations showed a trend toward exerting an interactive effect on the initial therapeutic alliance and the relationship between therapeutic alliance and psychotherapy outcome (X2== 2.75, p =.09). For the low OR (i.e, complexity) group, more subjects begin (i.e., initial therapeutic alliance) in the low therapeutic alliance cells (15 of 20 8'5); whereas the high OR group tends to load up in the high therapeutic alliance cells (16 of 21 8'8). While the hypothesis is not supported for the initial therapeutic alliance, the final therapeutic alliance indicates a strong relationship for the moderating effect of object relations on the therapeutic alliance- -«outcome relationship. For the low OR group there is a Cflnange toward higher therapeutic alliance by the end of tllerapy (X2 = 9.73, p;.001). While high OR S's loaded HE) in the high therapeutic alliance cells initially, lfllere was no significant distinction for therapeutic aLiliance by the final session. 133 Table 12 Therapeutic Alliance, Complexitprand Post-Therapy SCL-90-R Initial Therapeutic Alliance Low Object Relations High Object Relations Outcome Outcome LOW HIGH Row LOW HIGH Row Total Total PWC LOW 9 6 15 LOW 4 1 5 HIGH 3 2 5 HIGH 6 10 16 Column 12 8 20 10 11 20 Total Variable Pearson Value Df Significance PWC Low OR .0000 1 1.0000 PWC High OR 2.7586 1 .0967 (PWC — Patient working capacity) Final Therapeutic Alliance Low Obiect Relations High Obiect Relations Outcome Outcome LOW HIGH Row LOW HIGH Row Total Total PWC LOW 2 7 9 LOW 4 6 10 HIGH 10 1 11 HIGH 6 5 11 Column 12 8 20 10 11 21 Total Variable Pearson Value Df Significance PWC Low OR 9.7306 1 .0018 PWC High OR .4443 1 .5050 134 Table 13 illustrates the relationship between differing outcome calculation formulas and the initial and final therapeutic alliance by level of complexity. Dichotomization did allow for more specificity as to the nature of the association among variables--in most cases (6 of 8) the correlation was with the low complexity group, with only very weak correlations for the high complexity group on corresponding variables. For the initial therapeutic alliance, there were no statistically significant correlations with the residual gain score (RSCL) outcome calculation formula. For change in standardized scores (ZCHG), there was an additional correlational relationship between the low complexity group and PTC (p = .42, p <.05) as well as with WSC (p = .42, p (.05). High complexity was not correlated with ZCHG. The pre- to post-therapy difference score (DIF) manifested a pattern similar to the previous analysis--for the high complexity group there were correlations in the wrong direction with PHR (r = .42, p <.05) and WSC (p = -.56, p (.01). Low complexity was correlated with PWC (p = .42, p <.05). There were no statistically significant relationships for the TGSI calculation formula. The same overall relationship remained for the final alliance and outcome. For the low complexity group, RSCL was related with PWC (p = .57, p (.01) and 135 WSC(p = .48, p (.05). Low complexity was not correlated with either the DIF or TGSI outcome calculation formula. In the high complexity group, final therapeutic alliance and RSCL were correlated on the PHR variable (p = -.44, p (.05). There were no statistically significant correlations with the ZCHG or TGSI outcome calculation formula. The ZCHG formula was the most predictive for the relationship between the initial therapeutic alliance and outcome (as the DIF formula had a problem with directionality). Also, the correlations were primarily with the low complexity group. For the final therapeutic alliance, the RSCL outcome calculation formula was the most predictive; ZCHG was next, while DIF and TGSI were not significantly correlated with the alliance. There were correlations for both the low and high cxnnplexity groups with outcome; the correlations were never on the same CALPAS scale nor with the same outcome calculation formula. The hypothesis that the diChotomization of the complexity variable would help elUCLidate the process-relationship was partially confirmed. It was noted that the significant relationships tended to load up on either the low or high; complexity groups singularly, typically preserving the :same pattern of relationship seen in the previous 136 data analysis. In order to examine for the presence of an interaction, a test of the difference between the correlations for the low and high complexity groups was conducted. The formula used was based upon the assumption that these two groups are independent samples. Briefly, the computational formula for the statistical test (Shavelson, 1988) is as follows: 2 — z m (observed) = z — z w r1 The correlation coefficients from the two independent samples, 1‘, and r'2, are transformed to Fisher's 2's. This comparison was applied to all data sets where complexity was dichotomized (Tables 10 through 16). When the low and high complexity groups' correlations were transformed to z scores and compared, virtually none of the correlations achieved enough significant difference to support the conclusion that this was an interaction. The one significant new pattern which did emerge was that the correlations tended to load up heavily on the low complexity group. There was some support, again, for the impact of the tYpe of outcome calculation formula employed; this was 137 more evident for the final therapeutic alliance. Overall, while there were some minor shifts and changes noted, the same fundamental pattern emphasizing the relationship between the final therapeutic alliance and outcome remained as existed with the non-bifurcated sample. 138 Table 13 Correlations Between Therapeutic Alliance and Psychotherapy Outcome Using Different Calculation Formulas by Complexity Alliance RSCL ZCHG DIF TGSI INITIAL P319 Low .25 .22 -.20 —.03 High .03 .03 —.21 —.25 BE Low -.21 -.10 .07 —.27 High .09 .18 .42* .22 2.9. Low .29 .42* .42* .07 High .01 .09 .27 —.24 HS_C Low .22 .46* -.31 .07 High .17 .09 —.56** -.05 FINAL 21E Low .57** .37* -.29 .35 High .28 .30 —.24 .13 BE Low -.09 -.03 .02 .18 High -.44* —.47* .30 -.16 BT_C_ Low .17 .04 .01 -.09 High .30 .35 —.36* .13 FEE Low .48* .27 -.35 .23 High .28 .15 .30 .25 pLum = 20; p mm,= 21 * = p <.05 ** = p <.01 PWC - Patient Working Capacity PHR - Patient Hostile Resistance PTC - Patient Commitment WSC - Working Strategy Consensus RSCL - Residual gain ZCHG — Standardized change DIF - Pre/post difference TGSI - Post-therapy symptoms 139 Tables 14 and 15 refer to the dichotomized versions of the Post Therapy Client Questionnaire (PQCL) and the Post Therapy Therapist Questionnaire (PQTH) examined earlier. A similar general pattern emerged where (a) the same simple correlational patterns between alliance and outcome measures were typically preserved, and (b) the statistical significance tended to load up on either the high or the low complexity groups in a singular fashion. There were, however, some interesting patterns which emerged. For the PQCL (Table 14) some complicated and paradoxical patterns emerged due to the dichotomization of the complexity variable. For example, some relationships shifted from positive to negative (e.g., the initial patient working capacity (PWCl) and EMPATHY: p = -.48, p (.01; dichotomized: p um = -.45 and p mm,==.55, p <.05), suggesting that there was potential for an interaction embedded in the non- bifurcated data. Overall, a pattern between initial therapeutic alliance and PQCL factors emerged where significant correlations tended to load up on the high complexity group, and the number of variables which were significantly correlated with the therapeutic alliance doubled. The final therapeutic alliance indicated a pattern of improvement and positive evaluation of the therapist's abilities associated with 140 a positive alliance. For the low complexity group, higher final therapeutic alliance was associated with lpgp perceived change and therapeutic benefit. There was a strong and pervasive relationship, for the low complexity group, between higher post-therapy symptom level and lack of perceived improvement, non-dynamic therapeutic process and low ratings of the therapist's abilities. There were no statistically significant relationships for the low complexity group. For the PQTH, the very minimal relationship to the alliance is unchanged (Table 15). For the initial therapeutic alliance, for the high complexity group, there is a correlation between patient hostile resistance and therapist evaluation of ego strength (STRUCTURE p = .45, p <.05) as well as anxiety level (ANXIETY p = .44, p <.05). For the low complexity group, patient's with higher levels of hostile resistance were rated as less pleasant to work with and generating more counter—transference (PHR and LIKING p = -.49, p < .05). There were no statistically significant relationship in the high complexity group for the final therapeutic alliance. Therapist and patient evaluations, again, tended to show a pattern of congruence. The same general pattern was observed, but with over 70% of the correlations occurring in the low complexity group. 141 The patient's perception of change is associated with the therapist's evaluation of symptom reduction, ego strength and anxiety level. Lower levels of the patient's perception of the therapist's empathic capabilities are associated with the therapist's rating the patient's ego strength and defensiveness more negatively. In summary, while the correlational patterns were somewhat complex, there was support of hypothesis 2D, 2E and, to a lesser extent, for the presence of an interaction (hypothesis 3A). The statistical significance tended to load up on either the low or high complexity groups. However, when the amount of difference between the low and high groups' correlations was examined, it was not sufficient to suggest the presence of an interaction. 142 Table 14 CorrelationspBetween the Post Therapy Client Questionnaire (PQCL) and CALPAS, PQTH and SCL—90-R by Level of Complexity CHANGE PROCESS EMPATHY STYLE Low High Low High Low High Low High PWCl .04 -.31 -.11 .24 -.45* .55** -.11 .04 PHRl —.10 .49* -.25 .20 .04 .44* .04 .27 PTCl .04 .45* .16 .19 -.24 .61** .24 -.15 PWC4 043* -026 025 006 012 -020 005 -054** PHR4 .05 .16 .05 .07 .37* .34 .19 .27 PTC4 .30 -.49* .08 .04 —.15 -.25 —.24 —.64** SYMPT -.58**-.15 —.13 -.29 -.24 -.19 .21 .02 ANXIET -.48* -.21 -.35 -.35 -.54* -.09 .01 —.43* STRUCT -.48* .14 .02 .16 -.46* -.08 .07 .04 LIKING .21 .18 .01 .13 .11 .04 .15 .06 TGSI .50* .14 .43* -.34 .50* .04 .05 -.07 RSCL .58** .09 .42* -.44* .16 .06 .04 -.12 DIF -.10 —.10 -.23 .23 .11 -.15 .01 .28 Q-um = 20; p mm1= 21 * = p <.05 ** = p (.01 PWC - Patient Working Capacity PHR - Patient Hostile Resistance PTC - Patient Commitment RSCL - Residual gain ZCHG - Standardized change DIF - Pre/post difference 143 Table 15 Correlations Between the Post Therapy Therapist Questionnaire (PQTH) and CALPAS, PQCL and SCL-90—R by Level of Complexity SYMPTOM ANXIETY STRUCT LIKING Low High Low High Low High Low High PWCl .08 .11 .12 .02 -.23 .25 -.28 .03 PHRl —.22 —.15 .12 .01 .05 -.45* -.49* .01 PTC1 .21 -.11 .21 .03 -.24 .35 .16 .09 PWC4 -.14 —.19 -.20 .01 -.53** .17 .09 .01 PHR4 .16 .11 -.44* .06 —.16 -.34 .07 .07 PTC4 -.27 -.11 .21 .03 —.24 .06 .07 .14 CHANGE -.58**-.15 -.48* -.25 —.48* .13 .21 .18 PROCES -014 -029 -035 -036 002 -016 001 -012 EMPATH -.24 —.19 —.54** .09 —.46* .08 -.07 .03 STYLE —.25 —.02 -.15 —.43* .10 .04 .15 .06 TGSI -.28 —.17 -.46* .07 —.27 .05 -.16 .07 RSCL —.42* .02 -.34 .04 -.32 .20 -.08 .05 DIF .26 .04 .09 —.30 .06 -.38* .06 .22 Q-um = 20; p mm.= 21 * = p <.05 ** = p (.01 PWC - Patient Working Capacity PHR - Patient Hostile Resistance PTC - Patient Commitment RSCL - Residual gain ZCHG - Standardized change DIF - Pre/post difference 144 HvDothesis 4 served as a measurement validation of the OR Scale by comparing the Interpersonal Checklist (ICL), a conceptually parallel self—report, interpersonal attitude questionnaire, with the judge— rated OR Scale. The acceptance-rejection of others (ARO) dimension of the ICL was hypothesized to be positively correlated with the OR Scales. Both pre- therapy (PARO) and post—therapy (TARO) mean scores for the ICL were calculated for the sample and then correlated with the OR Scales, which is summarized in Table 16. Additionally, correlations with the initial and final therapeutic alliance were included to cross- reference the OR Scales. There were no statistically significant relationships between either pre-therapy (PARO) or post—therapy (TARO) ICL scores and the three OR Scales. The TARO was correlated with the initial therapeutic alliance. When the complexity variable was dichotomized into low and high groups, statistically significant relationships were established, loading heavily on the high complexity group. The complexity variable was significantly correlated with both pre— and post-therapy scores of the ICL (PARO, TARO). Additionally, the correlations between the ICL and the therapeutic alliance broadened and became stronger. There were no statistically significant relationships with the final alliance. For the low complexity group, 145 only the PHR scale was related to the TARO. Only indirect support could be established for hypothesis 4. There was no direct relationship between the OR Scales (AT, EI, C) and the ICL at pre— or post— therapy (PARO, TARO). When dichotomized into low and high complexity groups, a significant correlation was noted within the high complexity group, between alliance and the ICL. Also, there was a relationship between the ICL and the initial therapeutic alliance for the high complexity group. There is some equivocal support to suggest that the OR Scales and the ICL are measuring significantly overlapping variables and that, when complexity is high, higher level object relations are associated with higher levels for both pre-and post—therapy interpersonal attitudes, as well as higher levels of initial therapeutic alliance. 146 Table 16 Correlations between OR Scales, Interpersonal Checklist (ICL)! and Therapeutic Alliance by Level of Complexity PARO TARO OR Scales Full Low High Full Low High AT .13 .08 .04 .04 .04 .15 El .13 —.17 .11 .04 -.31 .33 C .13 .01 .49** .13 .06 .43* Initial Alliance Full Low High Full Low High PWC .09 .02 .20 .15 .03 .40* PHR -.20 -.21 -.27 -.36*-.10 -.36* PTC .15 .12 .25 .26* .06 .40* WSC .06 .04 .23 .09 .08 .29 Final Alliance Full Low High Full Low High PWC .13 .02 .21 .03 .01 .04 PHR -.25 -.21 -.26 -.19 -.17 -.20 PTC .15 .12 .16 .05 .14 .01 WSC .04 .04 .05 .08 -.10 .06 EFULL : 41; i-I-LOW = 20; 11 HIGH = 21 9 .05 ** = <.01 PARC - Pre-Therapy Interpersonal Checklist TARO - Post-Therapy Interpersonal Checklist AT - Affect Tone El - Emotional Investment C - Complexity PWC - Patient Working Capacity PHR - Patient Hostile Resistance PTC - Patient Commitment WSC - Working Strategy Consensus DISCUSSION This study examined the effects of object relations upon the relationship between therapeutic alliance and psychotherapy outcome. Three principal questions were addressed: (1) What is the nature of the relationship between object relations and therapeutic alliance? (2) What is the nature of the relationship between therapeutic alliance and outcome in psychotherapy? and (3) Do object relations, therapeutic alliance and outcome interact? Hypothesis 1. The series of predictions subsumed under hypothesis 1 (1A—1D), were concerned with the relationship between therapeutic alliance and object relations. The hypotheses were partially confirmed; there were two central findings: (a) object relations (i.e., the complexity subscale) and therapeutic alliance were significantly correlated at the initial session of psychotherapy. Once the treatment process began, the relationship diminished and was not reestablished; and (b) when the sample was dichotomized into low and high object relations groups 147 148 (i.e., complexity levels), the low complexity group had a significantly lower mean score for the initial therapeutic alliance. However, the score increased across sessions so that for the last half of treatment (i.e., middle and final alliance ratings) the low complexity group had the same level of therapeutic alliance as the high complexity group. Only one of the three OR Scales was significantly related to therapeutic alliance. While the affect—tone [AT] and emotional investment [EI] scales were expected to be more conceptually related to the variables under study, only the complexity [C] scale showed a consistent relationship to therapeutic alliance and other factors. This was unexpected not only because the complexity scale involved factors thought to be less central to the hypothesis, but also because it was argued earlier that a measure which assessed multiple dimensions of object relations would have more validity. The complexity scale is considered to be a more basic and general object relations measure, focusing on the degree to which the subject differentiates self and others in terms of subjective, complex and psychologically-oriented factors. For this study it may be best viewed as a measure of the patient's psychological mindedness, interpersonal sophistication and perhaps as a gross indicator of the 149 intricacies of the his or her internalized object representations. Three methodological concerns are important in understanding the limitations of the OR Scales in the present study: (a) there was a marked restriction of range in the data. In contrast to the AT and EI scales, only the C scale--perhaps, in part, because it was a seven point scale which called for more extreme responses--had a relatively large amount of variance; (b) the inter—rater reliabilities were low, due, in part, to the restriction of range; and (c) there were some limitations as to adaptability of the OR Scale to the data. Although the scale has been applied to a broad array of data, fewer studies have considered psychotherapy or interview data. The data used for the interpersonal episodes were taken directly from ongoing psychotherapeutic sessions, as opposed to structured interviews (Leigh, Westen, Barends, & Mendel, 1989; Ryan, 1973) or a pre-therapy test battery (Westen, 1990). The interpersonal episodes varied greatly in terms of their richness, depth, frequency, interpersonal object of focus, and subject matter. The mean number of episodes per session ranged from none to eight; the mean number of total episodes per subject was eleven. Thus, it seems reasonable to conclude that the data may not have been sufficiently rich, varied or 150 complex to take full advantage of the OR Scales as they were originally established. A strong association between object relations and the initial therapeutic alliance was established. Patients with lower levels of complexity began treatment with fewer positive contributions (patient working capacity [PWC], patient commitment [PTC]), somewhat more negative contributions (patient hostile resistance [PHR]) and a less effective overall working alliance (working strategy consensus [WSC]), relative to the high complexity group. This finding is in keeping with other studies that examined the relationship between patient pre-therapy interpersonal (or object) relations and therapeutic alliance (e.g., Klee, Abeles & Muller, 1990; Horowitz, Marmar, Weiss, Dewitt & Rosenbaum, 1984; Marziali, 1984; Moras & Strupp, 1982; Ryan, 1973). The (initially lower level of) therapeutic alliance for the low complexity group increased such that parity was achieved with the high complexity group's level of therapeutic alliance by the middle of psychotherapy. For patients in the high complexity group, the level of therapeutic alliance remained steady over the course of psychotherapy. The direct influence of object relations upon therapeutic alliance seemed to dissipate once the psychotherapeutic process 151 was set in motion. This change over time (initial to final therapeutic alliance) between the low and high complexity groups was the strongest evidence for an interaction effect of object relations. The consensus for therapeutic alliance research is that the alliance is typically established very early in the treatment process (Hartley, 1978; Horowitz, Marmar, Weiss, Dewitt, & Rosenbaum, 1984; Klee, Abeles, & Muller, 1990; Luborsky, 1976; Marziali, 1984; Marziali, 1984). An important finding in the present study is that while the alliance is quickly established for patients who comprise the high complexity group, some amount of time and therapeutic interaction is necessary for the formation of the alliance in the low complexity group. This pattern of differential development is in keeping with the findings of Ryan (Lehrke, 1977; Ryan, 1973; Ryan & Cicchetti, 1985) concerning the initial alliance, and of other studies where object relations were predictive of therapeutic alliance (Marziali, 1984b; Moras & Strupp, 1982). It may well be that highly psychologically minded patients enter psychotherapy with the expectation of an intense interpersonal experience with the therapist, while those who are less psychologically oriented will require some guidance and effort to build an alliance. In contrast to several studies of therapeutic 152 alliance where the alliance was treated as a more or less unitary phenomenon (i.e., therapeutic alliance ratings were either taken from one time point or collapsed across sessions; e.g., Eaton, 1987; Gaston, Marmar, Thompson & Gallagher, 1988; Marziali, 1984a), the present study attempted to consider the formation and developmental course of the therapeutic alliance over the course of psychotherapy. The significance of early alliance formation is augmented by introduction of the object relations variable which mediated the formation of therapeutic alliance. It took the low complexity group several sessions (potentially up to the mid-point of treatment) to develop a working level of alliance comparable to the high complexity group. This finding has specific meaning for the area of therapeutic alliance research since many studies have utilized a time-limited or brief psychotherapy model (e.g., Gaston et al. [1988], 16 to 20 sessions; Gomes— Scwhartz [1978], M = 18 sessions; Horowitz, Marmar, Weiss, Dewitt & Rosenbaum [1984], 12 sessions; Marmar, Weiss & Gaston [1989b], 12 sessions; Windholz & Silbershatz [1988], 16 sessions), while the present study had a mean of 30 sessions (SD = 14.66), with one third of cases being substantially longer. This longer treatment period may have allowed for a qualitatively different or more involved interaction to occur, one 153 which may not have been evident in studies utilizing a briefer treatment, between patient object relations and therapeutic alliance (across sessions). This suggests that object relations limitations exerted a diminishing role in its effect on the alliance and treatment outcome. This may help to explain some of the equivocal findings in therapeutic alliance literature. It may be the case that not only does therapeutic alliance fluctuate over the course of treatment (e.g., Hartley & Strupp, 1978; Klee, 1986), but, more specifically, that the alliance may have its own developmental course (Ticho, Appelbaum, Binstock & Appelbaum, 1971), and, for patients with limitations in object relational processes, the working alliance aspects of therapy may take some time beyond the initial phase (Greenson, 1967). Because psychotherapy is such a uniquely relational medium, it appears that an impairment in the patient's object relations has a significant impact on the formation of therapeutic alliance at the beginning of the therapeutic process. In an attempt to understand the differential development of therapeutic alliance (i.e., when the interactive effects of object relations are considered), it appears that the low complexity group may have entered treatment with a relational limitation or a handicap which impeded the 154 development of therapeutic alliance, and were able to overcome the handicap over time. In contrast, the high group may have entered the treatment process with the potential for a high level of alliance virtually ready- made. This may offer a partial interpretation of what then happened over the course of psychotherapy in the present study: the low group, perhaps due to the therapeutic effects of the alliance, eventually ”compensates" for its impairment, while the high group has little room to improve, implying that a ceiling effect may be in operation. The present study contained two seemingly paradoxical patterns with regard to the course of therapeutic alliance. First, the mean scores for therapeutic alliance for the third session were significantly lower and not correlated with the other sessions. The reasons for this are unclear, although lower inter—rater reliabilities may offer a partial explanation. Another possibility is that therapeutic alliance ratings beyond the initial treatment session may more accurately represent the therapeutic process in vivo. For example, patient hostile resistance as well as therapist negative contributions tend to increase after the initial therapeutic alliance rating, toward the midpoint of therapy, as the therapist begins to confront and focus on the patient's patterns of 155 conflict (e.g., Gaston, Marmar, Thompson & Gallagher, 1988; Hartley & Strupp, 1983; Klee, Abeles & Muller, 1990; Marmar, Weiss & Gaston, 1989b). Second, the initial, middle and final alliance showed a steady increase across sessions and a relatively consistent correlational pattern (among sessions) for the low complexity group. The mean alliance ratings for the high complexity group remained virtually unchanged across sessions but the sessions were not significantly correlated. One speculation is that the pattern for the low complexity group indicates a more unified, linear process of alliance formation, while the high complexity group, perhaps reflected in the greater within-group variance, may have an alliance pattern which is more heterogeneous and flexible. Again, low inter-rater reliabilities must be considered. The patterns of differential alliance formation add weight to the argument that the basic course of the alliance, from the beginning of therapy to the end, was dissimilar for these two groups. However, conclusions are limited to the statement that something different happened for the low and high complexity groups; the nature of that process is unclear. Thus, for hypothesis 1, object relations is predictive only in terms of ypppp the patient enters treatment and not necessarily how the process unfolds 156 or what exactly happens to the alliance over the course of psychotherapy. Hypothesispgpand 3. The second set of hypotheses predicted a positive relationship between therapeutic alliance and psychotherapy outcome. There were two principal predictions: (a) an overall treatment effect; and (b) a simple relationship between therapeutic alliance and post—therapy symptom level; there were three exploratory hypotheses: (c) measurement of symptom reduction utilizing different outcome calculation formulas; and (d) and (e) measurement of outcome from multiple perspectives (patient appraisal and therapist evaluation, in addition to symptom reduction). Also, because hypotheses 3--the examination of the interactive effects of object relations with therapeutic alliance/outcome relationship--is interwoven with the set of second hypotheses, both hypotheses will be reviewed together in this section. First, it was predicted that there would be an overall ”treatment effect", defined, in this case, as an overall pre- to post-therapy reduction in symptom level. There was significant symptom reduction; a majority of the sample (71%) experienced a decrease in pre- to post—therapy SCL-90—R scores. Mean scores on 157 both the Global Severity Index (TGSI) and a majority of the SCL—90-R subscales (seven of nine scales) showed a statistically significant reduction in pre— to post— therapy symptoms. This finding is also significant when placed in the context of the research literature (e.g., Horowitz et al., 1984; Marziali, 1984) since psychotherapy in the present study was conducted largely by supervised advanced graduate students with a heterogenous and perhaps more disturbed clinical population. Hypothesis 2B and 2C examined the relationship between therapeutic alliance and outcome (i.e., symptom change), based upon different outcome calculation formulas. There was no consistent relationship between the initial therapeutic alliance and outcome, regardless of how outcome was calculated. The only support came from a peripheral variable (working strategy consensus [WSC]), which was significantly correlated with outcome. There was support for the relationship between fippl alliance and outcome. The outcome calculation formula hypotheses were confirmed: formulas which took into account the relative change in symptoms from pre— to post—therapy had the strongest final therapeutic alliance/ psychotherapy outcome correlations. This is in keeping with previous research (e.g., Eckert, 1986; Mintz, 158 Luborsky, & Cristoph, 1979) where the residual gain score, by correcting for the influence of the pre- therapy score on the final outcome score, allowed for a stronger relationship between the independent variables and outcome. In contrast, outcome formulas utilizing a simple raw gain (i.e., pre- minus post-test) were less strongly correlated with alliance. The research literature is generally supports the relationship between therapeutic alliance and outcome, although there are areas which require clarification. While some studies show a direct relationship, other studies have noted that the relationship between alliance and outcome is significant only when patient “predispositional factors" have been considered. In the absence of any significant relationship between the object relations variable and pre— or post- therapy symptom levels, this relational limitation was specific to the interpersonal, and not the symptomatic, dimension. There were no significant differences between the low and high complexity groups for pre- therapy or post—therapy symptom levels. In fact, data regarding complexity showed a consistent relationship with patient symptom level from pre— to post-therapy (PreGSI p = —.31, p < .05; TGSI p = -.30, p < .05). This contradicts Eaton, Abeles, and Gutfreund (1988) finding that pre—therapy symptoms were a significant 159 predictor of psychotherapy outcome. The lack of an interaction between object relations level and symptoms is supportive of an argument developed earlier. The conclusion was that psychostructural factors, such as object relations, belong to a different realm of psychological functioning and, consequently, a different dimension of psychotherapy process than that of symptoms. This is especially true for dynamically- oriented psychotherapy where inferred internal processes such as intrapsychic conflict, quality of object and interpersonal relations, ego and super-ego processes are at the center of the treatment process. In the present study, even when patient object relations were included, no consistent significant relationships evolved between the initial alliance and outcome. An examination of some aspects of the psychotherapeutic process may help to shed some light on the this phenomenon. Several other studies have failed to note a direct relationship between therapeutic alliance and psychotherapy outcome (e.g., Eaton et al., 1988; Gomes— Schwartz, 1978; Horowitz, Marmar, Weiss, Dewitt & Rosenbaum, 1984; Klee, Abeles & Muller, 1990; Marziali, 1984a). Specifically, there has been a lack of a direct (process/outcome) relationship when outcome is measured by patient self-reported symptom change on 160 scales such as the SCL-90 or the MMPI. However, one explanation for the lack of consistent empirical support for the predictive power of therapeutic alliance concerning patient symptom change is that the process/outcome relationship is not direct and must be considered in the context of other variables. The present study introduced the interactional variable of patient early—therapy object relations to attempt to address this concern. Some studies have considered experimental paradigms which included multiple variables and noted a process/outcome correlation when object relations was examined along with other patient predispositional variables, such as motivation for treatment (e.g., Horowitz et al., 1984; Ryan, 1973). Furthermore, symptom change might be considered only one, albeit important, dimension of patient change and/or psychotherapy outcome. Multiple dimensions from different perspectives can be considered in addressing the outcome question (e.g., Schaffer, 1983). For example, the present study considered, in addition to symptom change, patient and therapist rated evaluations of the treatment process and therapeutic benefit. Some studies have noted that while therapeutic alliance was not related directly to symptom change, it could be related to dynamic change. For example, Marziali (1984a) established that judge-rated therapeutic .2. .Hne no.3. ~he.’ A '0, .923. z... I. ..a. .... o. .1 .r. .n. ._..... ..e. O C , a. .~ .1... .3 . :4. . r. a . ..._ a... 3. a. L..- .2 .m z e w. L. _— 25...; .2.» fl 2.. s :4 ._. .A t L. I .1 .s... t 161 alliance was related to clinical evaluations of "dynamic outcome", but not to patient rated symptom change. Ryan and Bell (1984) found a decrease in the number of hospitalizations and improvement in the quality of object representations, but not symptom level, over the course of psychotherapy for psychotic patients. Schneider (1990) reported changes in OR Scale ratings from the beginning to the end of treatment, but a non-significant relationship of process variables to symptom change for outpatients in brief psychotherapy. The conclusion which can be drawn from these findings is that there is a need for the measurement of psychotherapy outcome from a multidimensional perspective (e.g., Klee, 1986; Schaffer, 1982). Hypotheses 2D and 2E considered the relationship of the (a) patient's appraisal of the therapeutic process and outgome (PQCL), and (b) the therapist's evaluation of the treatment process and the patient's change (PQTH), to therapeutic alliance and psychotherapy outcome. There were a number of significant findings, a majority of which were associated with the low complexity group. Only minimal correlations between the PQCL factors and therapeutic alliance were found. Patient ratings of the therapist's empathic and technical abilities were 162 related to lower levels of patient hostile resistance in the alliance scales. However, the PQCL factors which were more pivotal to the hypothesis (i.e. symptom change) were not consistently related to the alliance. The relationship between the therapist's evaluation (PQTH) and therapeutic alliance was extremely limited. Only a few of the many potential combinations of variables were significantly related. The virtual lack of association between the PQTH and therapeutic alliance seems to be related to the fact that the items which comprise the four PQTH factors focus primarily on the evaluation of the functioning of the patient and the therapist's feelings about the patient, and not conceptually based in the framework of therapeutic alliance. In terms of psychotherapy outcome, there was a strong relationship between the low complexity groups' appraisal of change, the degree to which patient's saw their treatment as dynamic and intensive, therapist empathy, and post-therapy symptom level; both the SCL— 90-R Global Severity Index and the subscales were highly correlated to these client factors. For the high complexity group only factors dealing with the therapeutic process were related to lower post—therapy symptom levels. However, the PQCL factors also assess a broader dimension of the treatment process and 163 therapy outcome than the symptomatic aspect, and link it to the other variables involved in this study. For the patient then, there is a clear and direct relationship between satisfaction with and perceived benefit from therapy, positive impressions of the therapist and therapeutic process and overall pre— to post-therapy reduction in symptoms. There is also a clear relationship (for the low complexity group) between the therapist's evaluation (PQTH) of various aspects of the patient and psychotherapy process and the patient's evaluation of the therapist and treatment process. There were, for the low complexity group, a set of significant correlations between the PQCL and PQTH factors, indicating that there is a meaningful consensus between the patient and therapist in the evaluation of the patient's functioning, the therapeutic process and psychotherapy outcome. These findings suggest a congruence not only between the patient and therapist's evaluation of symptom reduction and satisfaction with therapy, but also with some aspects of both the treatment process and what has been referred to in the present study as the psychostructural dimensions of the patient. In theory, this addresses the concern raised above regarding the need for a multidimensional approach to 164 matters of psychotherapy process and outcome. However, for the present study, despite this congruence, no factor or set of factors was/were successful at predicting the overall course of alliance or the outcome (regardless of perspective). Implications for Research and Clinical Practice. Some of the limitations of the present study have been alluded to or discussed earlier. Factors such as limited power connected with the small sample size, inter-rater reliability concerns, restricted ranges and non-normal distributions, and some questions about the adaptability of scales to the data have been mentioned. The heterogeneous clinical population, and the level of experience and diversity of the therapists has also been discussed. While it has been argued that therapeutic alliance is a nonspecific variable (e.g., Jones, Cummings & Horowitz, 1988), there is a clear need for further delineation of the specific components of the alliance and a greater understanding of how the alliance affects the therapeutic process and related patient and therapist variables. Also, measures which allow for more precise and less inferential rating of the alliance can help ameliorate problems in training and reliability among raters. 165 The need for research designs which take into account multiple patient and therapist variables becomes more apparent as the field of psychotherapy research unfolds. The investigation of the complex interplay and sequence of therapeutic events between the patient and therapist over the course of psychotherapy is an obvious focus of future research. Of particular importance is the study of how the therapist deals with patients exhibiting limitations in object relations and initially low therapeutic alliances, and how the patient responds to those efforts to be engaged in the alliance (e.g., Gaston, Marmar & Ring, 1988; Sachs, 1983). In the area of measuring psychotherapy outcome, there is a need to move toward the consideration of multiple perspectives of ratings and multiple dimensions of patient change and improvement (e.g., Klee, 1986; Marziali, 1984a; Schaffer, 1982, 1983). In addition to a decrease in patient symptoms, factors such as improved interpersonal functioning, changes in inferred internal structures or processes such as object relations (e.g., Schneider, 1990; Ryan & Bell; 1984), and overall outcome from a clinical evaluation of dynamic outcome (e.g., Malan, 1976) should be considered important criteria indicating therapeutic movement. 166 There are a number of avenues open concerning the study of the effects of patient object relations upon psychotherapy process and outcome. Because object relations refers to inferred mental representations, it follows that changes in object relations from the beginning to end of treatment (e.g., Schneider, 1990; Ryan & Bell, 1984) could provide helpful information on the interpersonal dimension of psychotherapy. The use of more highly structured and tightly controlled object relations data--whether it be interview data, process recordings, early memories, TAT responses, etc.--wou1d be extremely helpful in uncovering the potential richness and improving the validity of object relations measures. Also, because object relations is only one aspect of the psychostructural dimensions discussed in this study, the consideration of conceptually related variables (patient motivation, other ego functions, developmental history, etc.) would help to address the need for an adequately complex model. There are two primary clinical implications from the present study. First, the centrality of the interpersonal aspects of psychotherapy cannot be overstated. Whether one is considering patient symptom level, therapeutic process, therapist interventions, therapeutic alliance, object relations, etc., the critical aspect to note is the clear relational 167 emphasis. The relational dimension of the psychotherapy process is a fundamental and universal way to think about and study treatment. The results of the present study suggest that the alliance is of particular importance to patients with object relations limitations, thus directing the therapist's technical focus to this relational dimension when working with interpersonally limited or difficult patients. Second, the finding that therapeutic alliance may take more time to form for patients with relational limitations is particularly relevant for short—term psychotherapy; the therapist may need to make a concerted effort to engage the patient in the alliance. Summary. The present study attempted to elucidate some of the aspects of the relationship between patient early-therapy object relations, therapeutic alliance over the course of treatment and psychotherapy outcome. It was proposed that an emphasis upon an interactional model which included patient predispositional variables might help to clarify some of the equivocal findings of therapeutic alliance research. While patient object relations did prove to exert an interactive effect on the psychotherapy process/outcome relationship, many essential questions remain as to the specific processes involved in that relationship. 168 The findings of the present study suggest that object relations does exert an interactional influence upon therapeutic alliance, although its effects appear to attenuate over time. While object relations was predictive of the formation of therapeutic alliance, it was not predictive of the final alliance nor of psychotherapy outcome. The final, but not the initial therapeutic alliance was predictive of outcome in terms of symptom change. Despite a clear decrease in symptoms, patient and therapist concurrence in appraisal of therapeutic improvement, positive "dynamic“ therapy outcome, no explicit patterns emerged which linked the therapeutic process (i.e., therapeutic alliance) with psychotherapy outcome. The results of the present research support the need to continue to pursue more sophisticated research models which consider patient predispositional variables within the context of more specific process measures and multidimensional outcome measures, hopefully assisting researchers and clinicians to better understand the complex interactions between the patient and therapist over the course of treatment. LI ST OF REFERENCES LI ST OF REFERENCES Abeles, N. (1985). Ego psychology in 1985. Paper presented at the Missouri Psychological Association Meeting, Columbia, Missouri, October. Adams, H. (1964). "Mental illness" or interpersonal behavior?. American Psychologist, 12, 191-197. Ashby, J., Ford, D., Guerney, B, & Guerney, L. (1957). Effects of a reflective and leading type of psychotherapy. Psychological Monographs, 11, 24. Auerbach, A., & Johnson, M. (1977). Research on the therapist's level of experience. In A. Gurman and A. in (Eds.), Effective psychotherapy: A handbook of research (pp. 84—102). New York: Pergamon Press. Bachrach, H., & Leaf, L. (1978). Analyzabilty: A systematic review of the clinical and quantitative literature. Journal of the American Psychoanalytic Association, pp, 881-920. Baekeland, F., & Lundwall, L. (1975). Dropping out of treatment: A critical review. Psychological Bulletin, p2, 738-783. Baker, E. (1979). The relationship between locus of control and psychotherapy: A review of the literature. Psychotherapy, lg, 351—362. Barrett-Lennard, G. (1962). Dimensions of therapist response as causal factors in therapeutic change. Psychological Monographs, 4;. Barron, F. (1953). An ego-strength scale which predicts response to psychotherapy. Journal of Consulting Psychology, 11, 327-333. Bellak, L. (1984). Basic aspects of ego function assessment. In L. Bellak and L. Goldsmith (Eds.) 1M; broad scope of ego function assessment (pp. 6-19). New York: John Wiley and Son. 169 170 Bergin, A. (1971). The evaluation of therapeutic outcomes. In A. Bergin and S. Garfield (Eds.), Handbook of psychotherapy and behavior change (pp. 217-270). New York: Wiley. Bergin, A. & Soloman, S. (1970). Personality and performance correlates of empathic understanding in psychotherapy. In J. Hart & T. Tomlinson (Eds.), May directions in client—centered therapy (pp. 223—236). Boston: Houghton Mifflin. Berzins, J. (1977). Therapist-patient matching. In A. Gurman and A. Razin (Eds.), Effectiveppsychotherapy: A handbook of research (pp. 222-251). New York: Pergamon Press. Beutler, L. (1981). Convergence in counseling and psychotherapy: A current look. Clinical Psychology Review, 1, 79—101. Beutler, L., Johnson, D., Neville, C., Jr., Elkins, D., & Jobs, A. (1975). Attitude similarity and therapist credibility as predictors of attitude change and improvement in psychotherapy. Journal of Consulting and Clinical Psycholpgy, 4;, 90-91. Beutler, L., Pollack, S., &, Jobs, A. (1978). "Acceptance" values and therapeutic change. Journal of Consulting_ and Clinical Psyphology, 4p, 198—199. Beutler, L., Crago, M., & Arizmendi, T. (1986). Therapist variables in psychotherapy process and outcome. In S. Garfield & A. Bergin (Eds.) Handbook of psychotherapy and behavior change (3rd edition) (pp. 257-310). New York: John Wiley and Sons. Bibring, E. (1954). Psychoanalysis and dynamic psychotherapies. Journal of the American Psychoanalytic Apsociation, 2, 745-770. Blatt, S., Brenneis, C., Schimek, J., & Glick, M. (1976). Normal developmental and psychopathological impairment in the concept of the object on the Rorschach. Journal of Abnormal Psychology, pp, 36-373. Blatt, S., & Lerner, H. (1983). Investigations in the psychoanalytic theory of object relations and object representations. In J. Masling (Ed.), Empirical studies of psychoanalytical theoriesprol. 11. Hillsdale, New Jersey: The Analytic Press. 171 Blatt, S., & Erlich, H. (1982). Levels of resistance in the psychotherapeutic process. In P. Wachtel (Ed.), Resistance: Psychodynamic and behavioral approaches (pp.69-91). New York: Plenum Press. Blatt, S., & Ritzler, B. (1974). Thought disorder and boundary disturbances in psychosis. Journal of Consulting and Clinical Psychology, 2;, 370—381. Boleloucky, Z. & Horvath, M. (1974). The SOL-90 rating scale: First experience with the Czech version in healthy male scientific workers. Act. Nerv. Super., 1p, 115-116. Bordin, E. (1955). Ambiguity as a therapeutic variable, Journal of Consulting Psychology. 19, 9-15. Bordin, E. (1979). The generalization of the psychoanalytic concept of the working alliance. Psychotherapy, 1Q, 252- 260. Bottari, M., & Rappaport, H. (1983). The relationship of patient and therapist-reported experiences of the initial sessions to outcome: an initial investigation. Psychotherapy, 29, 355-357. Brown, J. (1987). A review of meta-analyses conducted on psychotherapy outcome research. Clinical Ppychology Review, 1, pp. 1-23. Buckley, P., Karasu, T., & Charles, E. (1981). Psychotherapists view their personal therapy. Psychotherapy, 18, 299-305. Cox, W. (1978). Where are the A-B therapists, 1970-1975. Psychotherapy, 1;, 108-121. Crits-Cristoph, P., Cooper, A., & Luborsky, L. (1988). The accuracy of therapists' interpretations and the outcome of dynamic psychotherapy. Journal of Consulting and Clinical Psychology, pp, 490—495. Crowder, J. (1972). Relationship between therapist and client interpersonal behaviors and psychotherapy outcome. Journal of Counseling Psychology, lg, 68-75. Cronbach, L.J., & Furby, L. (1970). How should we measure change-—or should we? Psyphological Bulletin, 14, 68-80. Davanloo, H. (1979). Techniques of short-term dynamic psychotherapy. Psychiatric Clinics of North America, a, 11-22. 172 Derogatis, L.R. (1977; 1983). The SCL—90 scoring, administration and procedures, manuals 1 and II (for the revised version). Baltimore: Johns Hopkins University School of Medicine, Clinical Psychometrics Research Unit. Derogatis, L.R., & Cleary, P. (1977a). Confirmation of the dimensional structure of the SCL-90: A study in construct validation. Journal of Consulting and Clinical Psychology, 22, 981-989. Derogatis, L., Rickels, K., & Rock, A. (1976). The SCL-90 and the MMPI: A step in the validation of a new self- report scale. British Journal of Psychiatpy, 128, 280- 289. Dickes, R. (1975). Technical considerations of the therapeutic and working alliance. International Journal of Psychoanalytic Psychotherapy, 4, 1-24. Duckro, P., Beal, D., & George, C. (1979). Research on the effects of disconfirmed client role expectations in psychotherapy: A critical review. Psychological Bulletin, 24, 260-275. Eaton, T. (1987). Therapeutic alliance and negative indicators in outpatient psychotherapy. Unpublished doctoral dissertation, Michigan State University. Eaton, T., Abeles, N., & Gutfreund, M. J. (1988). Therapeutic alliance and outcome: Impact of treatment length and pretreatment symptomatology. Psychotherapy, 2;, 536-542. Eckart, P.A. (1986). Client—perceived therapist positive regard and psychotherapy outcome. Unpublished doctoral dissertation, Michigan State University. Filak, J. (1985). Object representations in Clients' verbal discourse: an exploratory study. Unpublished doctoral dissertation, Michigan State University. Filak, J., Abeles, N., & Norquist, S. (1986). Clients' pretherapy interpersonal attitudes and psychotherapy outcome. Professional Psychology: Research and Practice, 21, 217-222. Finell, J. (1987). A challenge to psychoanalysis: A review of the negative therapeutic reaction. Psychoanalytic Review, 14, 487—515. 173 Forsyth, N., & Forsyth, D. (1982). Internality, controllability, and the effectiveness of attributional interpretations in counseling. Journal of Counseling Psychology, 22, 140-150. Frank, J. (1982). Therapeutic components shared by all psychotherapies, in J. Harvey and M. Parks (Eds.), 1M2 master lecture seriespIVol I): Psychotherapy research and behavior change. Washington, D.C.: American Psychological Association. Frank, J. (1973). Persuasion and healipg (2nd edition). Baltimore: Johns Hopkins University Press. Freedman, M., Leary, T., Ossorio, A., & Coffey, H. (1951). The interpersonal dimensions of personality. Journal of Personality, 29, 143-162. Freud, S. (1900). The interpretation of dreams. Standard Edition, Vol.5 4 & 5. London: Hogarth Press (pp. 1-610). Freud, S. (1912). The dynamics of transference. Standard Edition, Vol. 12. London: Hogarth Press (pp. 97-109). Freud, S. (1916-17). Transference. Introductory lectures on psychoanalysis. New York: Norton (pp.431-447). Freud, S. (1918). From the history of an infantile neurosis. Standard Edition, Vol. 17. London: Hogarth Press (pp. 1-123). Freud, S. (1923). The ego and the id. Standard Edition, Vol. 19. London: Hogarth Press (pp. 3-66). Freud, S. (1926). Inhibitions, symptoms, and anxiety. Standard Edition, Vol. 20. London: Hogarth Press (pp. 75-175). Freud, S. (1933). New introductory lectures on psychoanalysis. Standard Edition, Vol. 22. London: Hogarth Press (pp. 1-183). Freud, S. (1937). Analysis terminable and interminable. Standard Edition, Vol. 23. London: Hogarth Press (pp. 209-255). Friedman, L. (1969). The therapeutic alliance. International Journal of Psychoanalysis, pp, 139-153. Frieswyk, S., Colson, D., & Allen, J. (1984). Conceptualizing the therapeutic alliance from a psychoanalytic perspective. Psychotherapy, 21, 460-464. 174 Frieswyk, S., Allen, J., Colson, D., Coyne, L., Gabbard, G., Horwitz, & Newsom, G. (1986). Therapeutic alliance: it's place as a process and outcome variable in dynamic psychotherapy research. Journal of Consulting and Clinical Psychology, 24, 32—38. Gabbard, G., Horwitz, L., Frieswyk, S., Allen, J., Colson, D., Newsom, G., & Coyne, L. (1988). The effect of therapist interventions on the therapeutic alliance with borderline patients. Journal of the American Psychoanalytic Association, 697-733. Garb, H. (1988). Clinical judgement, clinical training, and professional experience. Psychological Bulletin, 105, 387—396. Garduk, E., & Haggard, E. (1972). Immediate effects on patients of psychoanalytic interpretations. Psychological Issues (Monograph 28), 1, 3-83. Garfield, S. (1986). Research on client variables in psychotherapy. In L. Greenberg and W. Pinsof (Eds.), 1M9 pgychotherapeutic process: A research handbook (pp. 213- 256). New York: The Guilford Press. Garfield, S., & Bergin, A. (1971). Personal therapy, outcome, and some therapist variables. Psychotherapy, 8, 231. Gaston, L. (1990). The concept of the alliance and its role in psychotherapy: Theoretical and empirical considerations. Psychotherapy, 21, 143-153. Gaston, L., Marmar, C., Thompson, L., & Gallagher, D. (1988). Relationship of patient pretreatment characteristics to the therapeutic alliance in diverse psychotherapies. Journal of Consulting and Clinical Psychology, 99, 483-489. Gaston, L., Marmar, C., & Ring, J. (1988, June). Engaging the difficult patient in cognitive therapy. Paper presented at the annual meeting of the Society for Psychotherapy Research, Santa Fe, NM. Gendlin, E. (1962). Experiencing and the creation of meaning. New York: The Free Press of Glencoe. Gendlin, E., Beebe, J., Cassens, J., Klien, M., & Oberlander, M. (1968). Focusing ability in psychotherapy, personality and creativity. In J. Shlein (Ed.) Research in psychotherapy, Vol 3. Washington, D.C.: American Psychological Association. 175 Gill, M., & Hoffman, I. (1986). A method for studying the analysis of aspects of the patient's experience of the relationship in psychoanalysis and psychotherapy. Journal of the American Psychoanalytic Association, 19, 137—167. Goldstein, A., & Simonson, N. (1971) Social psychological approaches to psychotherapy. In A. Bergin & S. Garfield (Eds.), Handbook of psychotherapy_and behavior change (pp. 154-195). New York: Wiley. Gomes-Schwartz, B. (1978a). Psychotherapy process variables distinguishing the "inherently helpful“ person from the professional psychologist. Journal of Consulting and Clinical Psychology, 4p, 196-197. Gomes-Schwartz, B. (1978b). Effective ingredients in psychotherapy: prediction of outcome from process variables. Journal of Consulting and Clinical Psychology, 49, 1023-1035. Greenson, R. (1965). The working alliance and the transference neurosis. In Explorations in psychoanalysis (pp. 199—264). New York: International Universities Press. Greenson, R. (1967). The Technique and Practice pf Ppyphoana1ysis. New York: International Universities Press. Greenson, R., & Wexler, M. (1969). The non-transference in the psychoanalytic situation. International Journal of Psycho-analysis, pp, 27-40. Grigg, A. (1961). Client response to counselors at different levels of experience. Journal of Counseling Paychology, g, 217-222. Grigg, A., & Goodstein, L. (1957). The use of clients as judges of the counselor's performance. Journal of Counseling_Psychology, 4, 31-36. Gurman, A. (1977). The patient's perception of the therapeutic relationship. In A. Gurman and A. Razin (Eds.), Effective psychotherapy: A handbook of research (pp. 503-543). New York: Pergamon Press. Hadley, S., & Strupp, H. (1976). Contemporary views of negative effects in psychotherapy. Archives of General Psychiatry, 11, 1291—1302. 176 Hartley, D. (1978). Therapeutic alliance and the success of brief individual psychotherapy. Unpublished Doctoral Dissertation, Vanderbilt University. Hartley, D. (1985). Research on the therapeutic alliance in psychotherapy. In R. Hales and A. Frances (Eds.) in Psychiatric Update: Annual review, Volume 4, (pp. 532- 549). Washington, D.C.: American Psychiatric Press. Hartley, D., & Strupp, H. (1983). The therapeutic alliance: Its relationship to outcome in brief psychotherapy. In J. Masling (Ed.), Empirical studies of psychoanalytical theoriesgLVol. 1). Hillsdale, New Jersey: The Analytic Press. Horner, A. (1987). Object relations and transference reactions. In D. Milman & G. Goldman (Eds.) Technigues of working with resistance. Northvale, NJ: Jason Aronson, Inc. Horowitz, M. J. (1979). States of mind. New York: Plenum Press. Horowitz, M., & Marmar, C. (1985). The therapeutic alliance with difficult patients. In R. Hales and A. Frances (Eds.) in Psychiatry update: Annual review, Volume 4, (pp. 573-585). Washington, D.C.: American Psychiatric Press. Horowitz, M., Marmar, C., Weiss, D., DeWitt, K., & Rosenbaum, R. (1984). Brief psychotherapy of bereavement reactions. Archives of General Psychiatry, 41, 438-448. Horvath, A., & Greenberg, L. (1986). The development of the Working Alliance Inventory. In L. Greenberg and W. Pinsof (Eds.), The psychotherapeutic process: A research handbook (pp. 529-556). New York: The Guilford Press. Horwitz, L. (1974). Clinical prediction in psychotherapy. New York: Jason Aronson. Hoyt, M. (1980). Therapist and patient actions in "good“ psychotherapy sessions. Archives of General Psychiatry, 21, 159-161. Hoyt, M., Marmar, C., Horowitz, M., & Alvarez, W. (1981). The Therapist Action Scale and the Patient Action Scale: instruments for the assessment of activities during dynamic psychotherapy. Psychotherapy: Theory, Research, and Practice, 12, 109—116. 177 Hoyt, M., Xenakis S., Marmar, C., & Horowitz M. (1983). Therapists' actions that influence their perception of "good" psychotherapy session. The Journal of Nervous and Mental Disease, 111, 400-404. Hurley, J. (1980). Two interpersonal dimensions relevant to group and family therapy. In L. Wolberg and M. Aronson (Eds.) Group and family therapypll980). New York: Brunner-Mazel. Jacobson, N.S., Follette, W.C., & Revenstorf, D. (1984). Psychotherapy outcome research: Methods for reporting variability and evaluating clinical significance. Behavior Therapy, 1;, 336-352. Jones, E., Cummings, J., & Horowitz, M. (1988). Another look at the nonspecific hypothesis of therapeutic effectiveness. Journal of Consulting_and Clinical Psychology, pp, 48-55. Karon, B., & VandenBos, G. (1981). Psychotherapy of schizophrenia: treatment of choice. New York: Aronson. Keithly, L., Samples, S., & Strupp, H. (1980). Patient motivation as a predictor of process and outcome in psychotherapy. Psychotherapy and Psychosomatics, 22, 87- 97. Kernberg, O. (1975). Borderline conditions and pathological narcissism. New York: Aronson. Kernberg, O., Burstein, E., Coyne, L., Appelbaum, A., Horwitz, L., & Voth, H. (1972). Psychotherapy and psychoanalysis: Final report of the Menninger Foundation's Psychotherapy Research Project. Bulletin of the Menninger Clinic, 1p, 1-276. Kiesler, D. (1973). The process of psychotherapy. Chicago: Aldine. Klee, M. (1986). Psychotherapy outcome and the course of the therapeutic alliance. Unpublished doctoral dissertation, Michigan State University. Klee, M., Abeles, N., & Muller, R. (1990). Therapeutic alliance: Early indicators, course, and outcome. Psychotherapy, 21, 166—174. Klein, M. (1957). Envy and gratitude. In Envy and gratitude and other works 1946-1963. London: Hogarth Press. 178 Klopfer, B., Kirkner, F., Wisham, W., & Baker, G. (1951). Rorschach prognosis rating scale. Journal of Projective Techniques, 1;, 425-428. Kohut, H. (1977). The restoration of the self. New York: International Universities Press. Kraemer, H.C. (1981). Coping strategies in psychiatric clinical research. Journal of Consulting and Clinical Psychology, 42, 309-319. Krohn, A., & Mayman, M. (1974). Object representations in dreams and projective tests. Bulletin of the Menninger Clinic, 12, 445—466. LaCrosse, M. (1980). Perceived counselor social influence and counseling outcomes: Validity of the Counselor Rating Form. Journal of Counseling Psychology, 21, 320-327. Lafferty, P., Beutler, L., & Crago, M. (1989). Differences between more or less effective psychotherapists: A study of selected therapist variables. Journal of Consulting and Clinical Psychology, 21, 76—80. Lahey, M.A., Downey, R.G., Sall, F.E. (1983). Intraclass correlation coefficients: There's more than meets the eye. Psychological Bulletin, 2;, 586-595. Lambert, M. (1983). The patient-therapist relationship as a factor in psychotherapy outcome: An introduction. In M. Lambert (Ed.). Psychotherapy and patient relations (pp. 1-35). Illinois: Dow Jones-Irwin. Lambert, M. (1983). The patient-therapist relationship: A summary. In M. Lambert (Ed.). Psychotherapy and patent relations (pp. 290-293). Illinois: Dow Jones-Irwin. Lambert, M., & Bergin, A. (1983). Therapist characteristics and their contribution to psychotherapy outcome. In C. Walker (Ed.), The handbook of clinical psychology (Vol. 1, pp. 205-241). Homewood, IL: Dow Jones-Irwin. Lambert, M., Bergin, A., & Collins, J. (1977). Therapist- induced deterioration in psychotherapy. In A. Gurman and A. Razin (Eds.), Effective psychotherapy: A handbook of research (pp. 452-480). New York: Pergamon Press. Lambert, M., Shapiro, D., & Bergin, A. (1986). The effectiveness of psychotherapy. In S. Garfield & A. Bergin (Eds.) Handbook of psychotherapy and behavior change (3rd Eda) (pp. 157-211). New York: Wiley and Sons. 179 Langs, R. (1973). The technigue of psychoanalytic psychotherapy, Vol. 1. New York: Aronson. Langs, R. (1974). The technique of psychoanalytic psychotherapy, Vol. 2. New York: Aronson. Langs, R. (1982). Psychotherapy: a basic text. New York: Aronson. LaTorre, R. (1977). Pretherapy role-induction procedures. Canadian Psychological Review, 12, 308—321. Leary, T. (1957). Interpersonal diagnosis of personality. New York: Ronald Press. Lehrke, S. (1977). Working alliance in early psychotherapy. Unpublished doctoral dissertation. University of Florida. Leigh, J., Westen, D., Barends, A., Mendel, M., & Byers, S. (1989). The assessment of complexity of representations of people using TAT and interview data. Unpublished manuscript, University of Michigan. Lerner, H., & St. Peter, S. (1984) Patterns of object relations in neurotic, borderline and schizophrenic patients. Psychiatry, 41, 77-92. Loewald, H. (1960). On the therapeutic action of psycho— analysis. International Journal of Psychoanalysis, 12, 16—33. Luborsky, L. (1976). Helping alliances in psychotherapy. In J. Claghorn (Ed.), Successful psychotherapy (pp. 92- 116). New York: Brunner—Mazel. Luborsky, L. (1984). Principles ofppsychoanalytic psychotherapy. New York: Basic Books. Luborsky, L. (1985). The core conflictual relationship theme (CCRT) method: Guide to scoring and rationale. Unpublished manuscript, University of Pennsylvania. Luborsky, L., & Auerbach, A. (1985). The therapeutic relationship in psychodynamic psychotherapy: The research evidence and its meaning for practice. In R. Hales and A. Frances (Eds.) in Psychiatry update: Annual reviewL Volume 4, (pp. 550-561). Washington, D.C.: American Psychiatric Press. 180 Luborsky, L., Bachrach, H., Graff, H., Pulver, S., & Cristoph, P. (1979). Preconditions and consequences of transference interpretations. Journal of Nervous and Mental Disease, 141, 391-40. Luborsky, L., Crits-Cristoph, P., Alexander, L., Margolis, M., & Cohen, M. (1983). Two helping alliance methods for predicting outcomes of psychotherapy. Journal of Nervous and Mental Disorders, 171, 480—491. Luborsky, L., Crits-Cristoph, P., Mintz, J., & Auerbach, A. (1988) Who will benefit from psychotherapy? New York: Basic Books. Luborsky, L, Crits-Cristoph, P., McLellen, A., Woody, G., Piper, W., Liberman, B., Imber, S., & Pilkonis, P. (1986). Do therapists vary much in their success? American Journal of Orthopsychiatry, pp, 501-512. Luborsky, L., Mintz, J., Auerbach, A., Cristoph, P., Bachrach, H., Todd, T., Johnson, M., Cohen, M., & O'Brien, C. (1980). Predicting the outcome of psychotherapy: findings of the Penn Psychotherapy Project. Archives of General Psychiatry, 11, 471-481. Luborsky, L., & Spence, D. (1978). Quantitative research on psychoanalytic therapy. In S. Garfield and A. Bergin (Eds.) Handbopkpof psychotherapy and behavior change (pp. 191-232). New York: John Wiley and Sons. Malan, D. (1963). A study of brief psychotherapy. London: Tavistock. Republished by Plenum Press, New York, 1975. Malan, D. (1976). Toward the validation of dynamic psychotherapy. New York: Plenum Medical Book Company. Marmar, C., & Gaston, L. (1989). Manual of California Psychotherapy Alliance Scales. Unpublished manuscript. Marmar, C., Gaston, L., Gallagher, D., & Thompson, L. (1989a). Therapeutic alliance and outcome in behavioral, cognitive, and brief dynamic psychotherapy in late-life depression. Journal of Nervous and Mental Disease, 111, 464-472. Marmar, C., Horowitz, M., Weiss, D., & Marziali, E. (1986). The development of the Therapeutic Alliance Rating System. In L. Greenberg and W. Pinsof (Eds.), 1gp psychotherapeutic process: A research handbook (pp. 367— 390). New York: The Guilford Press. 181 Marmar, C., Weiss, D., & Gaston, L. (1987). Towards the validation of the Therapeutic Alliance Rating System. Unpublished manuscript. Marmar, C. R., Weiss, D. S., & Gaston, L. (1989b). Towards the validation of the California Therapeutic Alliance Rating System. Journal of Consulting and Clinical Psychology, 1: 46-52. Marziali, E. (1984a) Three viewpoints of the therapeutic alliance: similarities, differences, and associations with psychotherapy outcome. Journal of Nervous and Mental Disorder, 112, 417-423. Marziali, E. (1984b). Prediction of outcome of brief psychotherapy from therapist interpretive intervention. Archives of General Psychiatry, 41, 301—304. Marziali, E., Marmar, C., & Krupnick, J. (1981). Therapeutic alliance scales: development and relationship to psychotherapy outcome. American Journal of Psychiatry, 12§J 361-364. Matarazzo, R., Wiens, A., & Saslow, G. (1966). Experimentation in the teaching and learning of psychotherapy. In L. Gottschalk and A. Auerbach (Eds.), Methods of research in psychotherapy. New York: Appelton, Century and Crofts. Mayman, M. (1967). Object-representations and object— relationships in Rorschach responses. Journal of Projective Technigues, 11, 117-25. Mayman, M. (1968). Early memories and character structure. Journal of Projective Technigues, 22, 303-316. McCarron, L., & Appel, V. (1971). Categories of therapist verbalizations and patient-therapist autonomic response. Journal of Consplting and Clinical Psychology, 11, 123- 134. iMeisels, M. (1988). A critique and redefinition of the concept of resistance. Unpublished manuscript. Mintz, J. (1977). The role of the therapist in assessing psychotherapy outcome. In A. Gurman and A. Razin (Eds.), Effective psychotherapy: A handbookyof research (pp. 590-601). New York: Pergamon Press. 182 Mintz, J., Luborsky, L, & Auerbach. A. (1971). Dimensions of psychotherapy: a factor analytic study of psychotherapy session. Journal of Consulting and Clinical Psychology, 35, 106-120. Mintz, J., Luborsky, L., & Cristoph, P. (1979). Measuring the outcomes of psychotherapy: findings of the Penn Psychotherapy Project. Journal of Consulting and Clinical Psychology, 41, 319-334. Mitchell, K., & Hall, L. (1971). Frequency and type of confrontation over time within the first therapy interview. Journal of Consulting and Clinical Psychology, 11, 437-442. Moras, K., & Strupp, H. (1982). Pretherapy interpersonal relations, patients' alliance, and outcome in brief therapy, Archives of General Psychiatry, 11, 405—409. Morgan, R., Luborsky, L., Grits-Cristoph, P., Curtis, H., & Solomon, J. (1982). Predicting the outcomes of psychotherapy by the Penn Helping Alliance Rating Method. Archives of General Psychiatry, 11, 397-402. Morgan, R., Luborsky, L., Crits-Cristoph, P., Curtis, H., & Solomon, J. (1982). Predicting the outcomes of psychotherapy by the Penn Helping Alliance Rating Method. Archives of General Psychiatry, 11, 397—402. Mullen, J., & Abeles, N. (1971). Relationship of liking, empathy and therapist's experience to outcome of psychotherapy. Journal of Counseling Psychology, 11, 39- 43. O'Malley, S., Foley, S., Rounsaville, B., Watkins, J. Sotsky, S., Imber, S., & Elkin, I. (1988). Therapist competence and patient outcome in interpersonal psychotherapy of depression. Journal of Consulting and Clinical Psychology, 11, 496-501. O'Malley, S., Suh, C., & Strupp, H. (1983). The Vanderbilt Psychotherapy Process Scale: a report on the scale development and a process-outcome study. Journal of Consulting and Clinical Psychology, 11, 581-586. Orlinsky, D., & Howard, K. (1975). Varieties of psychotherapeutic experience. New York: Teachers' College Press. 183 Orlinsky, D., & Howard, K. (1980). Gender and psycho- therapeutic outcome. In A. Brodsky and R. Hare—Mustin (Eds.), Women and psychotherapy, (pp. 3-34). New York: Guilford. Orlinsky, D., & Howard, K. (1986). Process and outcome in psychotherapy. In S. Garfield & A. Bergin (Eds.) Handbook of psychotherapy and behavior change (3rd Ed.) (pp. 311- 381). New York: Wiley and Sons. Ornston, P., Cicchetti, D., Levine, J., & Fierman, L. (1968). Some parameters of verbal behavior that reliably differentiate novice from experienced therapists. Journal of Abnormal Psychology, 11, 240-244. Parloff, M., London, P., & Wolfe, B. (1986). Individual , psychotherapy and behavior change. In Annual Review of Paychology, 31, 321-49. Parloff, M., Waskow, I., & Wolfe, B. (1978). Research on therapist variables in relation to process and outcome. In S. Garfield and A. Bergin (Eds.) Handbopk of psychotherapy and behavior change (pp. 233-282). New York: John Wiley and Sons. Patterson, G., & Forgatch, M. (in press). Therapist behavior as a determiner for client noncooperation: A paradox for the behavior modifier. Journal of Consulting and Clinical Psychology. Piper, W., deCarufel, F., & Szkrumelak, M. (1985). Patient predictors of process and outcome in short—term individual psychotherapy. Journal of Nervous and Mental Disease, 111J 726-733. Razin, A. (1977). The A—B variable: Still promising after twenty years? In A. Gurman and A. Razin (Eds.), Effective psychotherapy: A handbook of research (pp. 291— 324). New York: Pergamon Press. Rice, L. (1965). Therapist's style of participation and case outcome. Journal of Consulting Psychology, 21, 155- 160. Rogers, C. (1957). The necessary and sufficient conditions of therapeutic personality change. Journal of Consulting and Clinical Psycho1ggy, 21, 95-103. Rosenbaum, R., & Horowitz, M. (1983). Motivation for psychotherapy: a factorial and conceptual analysis. Psychotherapy: Theory, Research, and Practice, 29, 346- 355 O 184 Rosenfeld, H. (1975). Negative therapeutic reaction. In P. Giovacchini (Ed.), Tactics and techniques in psychoana1ytic,psychotherapy, (Volume 2: Countertransference), (pp. 217—228). New York: Aronson. Ryan, E. (1973). The capacity of the patient to enter an elementary therapeutic relationship in the initial psychotherapy interview. Unpublished doctoral dissertation. University of Michigan. Ryan, E., & Bell, M. (1984). Changes in object relations from psychosis to recovery. Journal of Abnormal Psychology, 11, 209-215. Ryan, E., & Cicchetti, D. (1985). Predicting the quality of the alliance in the initial psychotherapy interview. Journal of Nervous and Mental Disease, 173, 717-725. Sachs, J. (1983). Negative factors in brief psychotherapy: an empirical assessment. Journal of Consulting and Clinical Psychology, 11, 557-564. Saltzman, C., Leutgert, M., Roth, C., Creaser, J., & Howard, L. (1976). Formation of a therapeutic relationship: Experiences during the initial phase of psychotherapy as predictors of treatment duration and outcome. Journal of Consulting and Clinical Psychology, 44, 546-555. Schaffer, N. (1982). Multidimensional measures of therapist behavior as predictors of outcome. Psychological Bulletin, 12, 670-681. Schaffer, N. (1983). Methodological issues of measuring the skillfulness of therapeutic techniques. Psychotherapy: Theory, Research, and Practice,2g, 486-493. Schneider, E. L. (1990). The effect of brief psychotherapy on the level of the patient's object relations. Unpublished doctoral dissertation, New York University. Schuller, R., Crits-Cristoph, P., Connolly, M. (1988). The resistance scale: Background and psychometric properties. Manuscript submitted for publication. Shapiro, D. A., & Shapiro, D. (1982a). Meta-analysis of comparative outcome research: A replication and refinement. Paychological Bulletin, 12, 581-604. Shavelson, R. J. (1988). Statistical reasoning for the behavioral sciences (2nd edition). Boston: Allyn & Bacon. 185 Shrout, P. E., & Fleiss, J. L. (1979). Intraclass cor- relations: Uses in assessing rate reliability. Psychological Bulletin, 11, 420-428. Sifneos, P. (1978). Motivation for change: A prognostic guide for successful psychotherapy. Paychotherapy and Psychosomatics, 21, 293-298. Sifneos, P. (1979). Short-term dynamic psychotherapy: Evaluation and technique. New York: Plenum Press. Silberschatz, G., Fretter, P.B., & Curtis, J.T. (1986). How do interpretations influence the process of psychotherapy. Journal of Consulting and Clinical Psychology, 14, 646-652. Sloane, R.B., Cristol, A.H., Pepernik, M.C., & Staples, F.R. (1970). Role preparation and expectation of improvement in psychotherapy. Journal of Nervous and Mental Diseases, 111, 18-26. Smith, M., Glass, G., & Miller, T. (1980). The benefits of psychotherapy. Baltimore: Johns Hopkins Press. Sterba, R. (1934). The fate of the ego in analytic therapy, International Journal of Psychoanalysis, 11, 117-126. Stiles, W., Shapiro, D., & Elliot, R. (1986). Are all psychotherapies equivalent? American Psychologist, 41, 165-180. Strong, S. (1968). Counseling: An interpersonal influence process. Journal of Counseling Psychology, 11, 215—224. Strupp, H. (1974). On the basic ingredients of psychotherapy. Psychotherapy and Psychosomatics, 260. 4, 249— Strupp, H. (1977). A reformulation of the dynamics of the therapist's contribution. In A. Gurman and A. Razin (Eds.), Effective psychotherapy: A handbook of research (pp. 1-22). New York: Pergamon Press. Strupp, H. (1980a). Success and failure in time-limited psychotherapy: a systematic comparison of two cases (comparison 2). Archives of General Psychiatry, 11, 708- 716. Strupp, H. (1980b). Success and failure in time-limited psychotherapy: further evidence (comparison 4). Archives of General Psychiatry, 11, 947-954. 186 Strupp, H. (1982). The outcome problem in psychotherapy: contemporary perspectives. In J. Harvey and M. Parks (Eds.), Master Lecture Series, Volume I: Psychotherapy research and behavior change (pp. 43-71). Washington, D.C.: American Psychological Association. Strupp, H. (1986). Psychotherapy: Research, practice, and public policy (how to avoid dead ends). American Psychologist, 41, 120—130. Strupp, H., & Binder, J. (1984). Psychotherapy in a new key. New York: Basic Books, Inc., Publishers. Strupp, H., & Hadley, S. (1979). Specific versus nonspecific factors in psychotherapy: a controlled study of outcome. Archives of General Psychiatry, 11, 1125- 1136. Strupp. H., Keithly, L., Moras, K., Samples, S., Sandell, J., & Waterhouse, G. (1980). Toward a measurement of negative effects in psychotherapy. Paper read before the annual meeting of the Society for Psychotherapy Research, Pacific Grove, CA. Strupp, H., Lessler, R., & Fox, K. (1969). Patients view their psychotherapy. Baltimore: Johns Hopkins University Press. Suh, C., Strupp, H., & O'Malley, S. (1986). The Vanderbilt Process Measures: The Psychotherapy Process Scale (VPPS) and the Negative Indicators Scale (VNIS). In L. Greenberg and W. Pinsof (Eds.), The psychotherapeutic process: A research handbook (pp. 285-324). New York: The Guilford Press. Sundland, D. (1977). Theoretical orientations of psychotherapists, in A. Gurman and A. Razin (Eds.), Effective psychotherapy: A handbook of research (pp. 189- 219). New York: Pergamon Press. Sundland, D., & Barker, E. (1962). The orientations of psychotherapists. Journal of Consulting Psychology, 21, 201-212. Taube, C., Burns, B., & Kessler, L. (1984). Patients of psychiatrists and psychologists in office-based practice: 1980. American Psychologist, 11, 1435-1447. Tichenor, V., & Hill, C. (1989). A comparison of six measures of working alliance. Psychotherapy, 29, 195-199. 187 Ticho, E., Appelbaum, A., Binstock, W. A., & Appelbaum, S. (1971). Panel on recent advances in psychoanalysis and psychotherapy. Bulletin of the Menninger Clinic, 11, 447-460. Thoits, P. (1985). Negative outcome: The influence of factors outside therapy. In D. Mays & C. Franks (Eds.), Above all do no harm: Negative outcome in psychotherapy and what to do about it. New York: Springer. Tollington, H. (1973). Initial expectations and outcome. British Journal of Medical Psychology, 4a, 251. Tracey, T., & Ray, P. (1984). Stages of successful time- limited counseling. Journal of Counseling Psychology, 11, 13—27. Truax, C., Wargo, D., Frank, J., Imber, S., Battle, C., Hoehn-Saric, Nash, E., & Stone, A. (1966). Therapist empathy, genuineness, and warmth and patient therapeutic outcome. Journal of Consulting and Clinical Psychology, 19, 395-401. Truax, C., Fine, H., Moravec, J., & Mills, W. (1968). Effects of therapist persuasive potency in individual psychotherapy. Journal of Clinical Psychology, 24, 359- 362. Urist, J. (1977). The Rorschach test and the assessment of object relations. Journal of Personality Assessment, 41, 1—9. Urist, J. (1980) Object relations, in R. Woody (ed.) Encyclopedia of clinical assessment (pp. 821-833). San Francisco: Jossey-Bass Publishers VandenBos, G., & Karon, B. (1971). Pathogenesis: A new therapist personality dimension related to therapeutic effectiveness. Journal of Personality Assessment, 11, 252-260. Westen, D. (1990). Towards a revised theory of borderline object relations: Contributions of empirical research. International Journal of Psychoanalysis, 11, 661-693. Westen, D. (1991) Clinical assessment of object relations using the TAT. Journal of Personality Assessment, 11, 56- 74. 188 Westen, D., Barends, A., Leigh, J., Mendel, M., & Silbert, D. (1988a). Manual for coding dimensions of object relations and social cognition from interview data. Unpublished manuscript, Department of Psychology, University of Michigan. Westen, D., Barends, A., Leigh, J., Mendel, M., & Silbert, D. (1988b). Assessing object relations and social cognition from interview data. Unpublished manuscript, Department of Psychology, University of Michigan. Westen, D., Ludolph, P., Lerner, H., Ruffins, S., & Wiss, C. F. (1990). Object relations in borderline adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 21, 338-348. Westen, D., Huebner, D., Boekamp, J., Lifton, N., & Silverman, M. (1989). Assessing complexity of representations of people and understanding of social causality: A comparison of natural science and clinical psychology graduate students. Unpublished manuscript, University of Michigan. Wenegrat, A. (1976). Linguistic variables of therapist speech and accurate empathy ratings. Psychotherapy, 11, 30-33. Windholz, M., & Silbershatz, G. (1988). Vanderbilt psychotherapy process scale: a replication with adult outpatients. Journal of Consulting and Clinical Psychology, 11, 56-60. Windholz, M., Weiss, D., & Horowitz, M. (1985). An empirical study of the natural history of time-limited psychotherapy for stress response syndromes. Psychotherapy, 22, 547-554. Wogan, M. (1970). Effects of therapist-patient personality variables on therapeutic outcome. Journal of Consulting and Clinical Psychology, 11, 356-361. Wolff, W. (1956). Contemporary therapists examine themselves. Springfield, IL: Thomas. Zetzel, E. (1956). Current concepts of transference. International Journal of Psychoanalysis, 11, 369—378. APPENDICES APPENDIX A : Literature Review Literature Review: Patient and Therapist Variables Associated with Psychotherapy Process, Outcome and the Therapeutic Alliance Patient Variables Before reviewing patient factors, the various requirements inherent in the therapeutic situation for the patient should be noted. Bordin (1979) has summarized the basic dimensions of demands placed on the patient entering therapy; these include (a) the degree of self-observation and self-disclosure; (b) the range and complexity of cognitive processing required; (c) the extent of dependence on or adherence to what the therapist does; (d) the amount of ambiguity inherent in the therapy; and (e) the amount of reliance on non-verbalized material. An overview of the patient qualities needed to deal with these therapeutic demands follows. Probably the most thoroughly developed literature concerning the critical patient variables involved in the patient's capacity to engage in and benefit from psychotherapy--in particular, the therapeutic alliance--has been advanced by psychodynamically and psychoanalytically- 189 190 oriented writers. Based upon a review of the literature by the present author, the four general patient qualities consistently represented in the literature are ego-strength or mature ego-functioning, motivation, a capacity for object relatedness and symptom severity. Ego-Strength While ego-strength (or ego-functioning) is often used in an indeterminate manner, several more specific notions, which can be subsumed under ego-strength, can be elaborated to give some utility to the concept (one qualification--object relations, usually considered one facet of ego-functioning will be discussed separately because of its central nature to this study). Bellak (1984), who has written widely in the area of ego-functioning, outlines 12 major ego functions. Briefly, these include: reality testing, judgment, sense of reality of the world and of the self, regulation and control of drives, affects and impulses, object relations/interpersonal relations, thought processes, adaptive regression in service of the ego, defensive functioning, stimulus barrier, autonomous functioning, synthetic-integrative functioning, and mastery-competence. The following ego functions seem especially pertinent to the present discussion: object relations, regulation of drives, defensive functioning and autonomous functioning; these dimensions of ego functioning 191 will be integrated into the following discussion. Gaston, Marmar, Thompson, and Gallagher (1988) propose that "Probably the most important determinant of the patient contribution to the therapeutic alliance is the mature functioning of certain aspects of the patient ego" (p. 483). Dickes (1975) and Greenson (1969), among others, emphasize such areas of ego-functioning in the development of the therapeutic alliance as (a) the ability to distance oneself from experience temporarily, (b) the capacity to comprehend and reflect on the therapist's interventions, (c) the ability to communicate both cognitive and affective experience, and (d) the capability of shifting between mature, logical reality contact and a temporary regression into one's inner experience and fantasies. Summarizing the Menninger Psychotherapy Research Project, Frieswyk, Colson, and Allen (1984) specify six interrelated aspects of the patient's collaboration in the therapeutic relationship: (a) the patient brings meaningful material to the session, (b) the patient strives to openly share and express his or her emotions, (c) the patient works on therapeutic tasks, (d) the patient attempts to use the therapist's interventions to help understand and change their problems, (e) the patient implements what he or she has learned in therapy to their daily life, and (e) the patient begins to adopt a "therapist-like" attitude toward self in his or her identification with the therapist. 192 Langs (1973) has emphasized that the patient must be able to accept his or her responsibility of being the prime communicator in the treatment, and he or she must be able to withstand some treatment-induced short—term deprivation in the service of long-term treatment goals. Frustration tolerance, delay of gratification, bypassing short-term for long-term goals and enduring some level of uncertainty and deprivation are frequently mentioned as necessary ego- functions. Hartley and Strupp (1983) indicate that the patient's capacity to recognize and overcome personal problems, a desire to cooperate with the therapeutic tasks and the capacity to tolerate the helplessness, dependency and ambiguity that can accompany psychotherapy as critical patient factors. Kernberg, Burstein, Coyne, Appelbaum, Horwitz, and Voth (1972) noted that clinical appraisals of patient ego- strength correlated significantly with global outcome measures. There are several studies which used the Rorschach Prognostic Rating Scale (RPRS; Klopfer, Kirkner, Wisham, & Baker, 1951) to examine the relationship between ego-strength and therapeutic outcome, but the results remain equivocal. While there are more studies suggesting a positive relationship between ego-strength and outcome, as measured by the RPRS, than negative or nonsignificant findings, there is considerable variability in group scores and several measurement problems have been noted (e.g., 193 prediction rates are about 60%--no higher than predictions made without any measures; see Garfield, 1986, for a discussion). Barron (1953) developed the Ego Strength Scale (ES) based upon MMPI items which showed an initial relationship between ego-strength and therapy outcome. Luborsky, Crits-Cristoph, Mintz, and Auerbach (1988) reviewed nine studies using the ES scale and indicate that about 45% show positive relationships, compared with 45% not significant and 10% negative relationships between ego strength and outcome. Garfield (1986), in his review of similar studies, concludes that subsequent studies have not replicated Barron's initial findings. Hartley and Strupp (1983) also suggest that the therapeutic alliance is directly affected by the quality of the patient's self—exploration. The self-exploration process is characterized by open reporting of rich and diverse inner experience, an ongoing capacity to make self- observations and generalize these observations, and the ability to alternate between direct experiencing and cognitive processing of the experience. Several of the therapeutic alliance studies found self-exploration to be an important component of the patient’s contribution to the therapeutic alliance. For example, Comes-Schwartz (1978) and O'Malley, Suh, and Strupp (1983), using slightly differing experimental paradigms, found that greater self- exploration and self—examination of emotions and personal 194 experiences were positively related to psychotherapy outcome when measured from the perspective of a clinical observer. Orlinsky and Howard (1986) summarized the findings of some 37 studies and noted that a majority of the studies (70%) were non-significant. This was, in part, related to the use of the Truax scale in many of the older studies. The above- mentioned studies (which indicated a positive relationship) used a therapist perspective and more sophisticated measures of process and outcome. Luborsky and Auerbach (1985) discuss the importance of the patient's ability for "experiencing“ (Gendlin, 1962; Gendlin, Beebe, Cassens, Klien, and Oberlander, 1968) as one of the key patient contributions in successful psychotherapy. Experiencing has been described as the client's inward attention, including the felt experience of one's phenomenological field with an emphasis upon personally felt meaning of the experience. Luborsky, Crits- Cristoph, Mintz and Auerbach (1988) reviewed eleven studies which examined experiencing and found that there was some support for experiencing as a predictor of better outcomes (37% positive, 18 negative, 45% not significant). Orlinsky and Howard (1986), using a roughly analogous but somewhat broader term--patient self-relatedness--concluded that there was a strong relationship between this variable and process and outcome in psychotherapy. This concept includes patient exploration, defensiveness versus openness and the overall 195 capacity to absorb therapeutic interventions. Over 80% of the 12 studies reviewed indicated a positive relationship to psychotherapy outcome. It would then be expected that the presence of obstacles to the self-exploration process--for example, resistant attitudes, defensive avoidance of inner experience, externalization of conflict, excessive dependency, etc.--would be considered threats, which have originated with the patient, to the working alliance. '1 Empirical studies concerning patient resistances will be examined in an upcoming section. Sifneos (1978) developed a list of patient prognostic indicators for short-term psychotherapy that included psychological mindedness, active participation, curiosity to understand oneself, willingness to change and experiment, honest communication, realistic goals and a willingness to make sacrifices for long-term gains in psychotherapy. Motivation Closely connected with the general concept of ego- functioning is patient motivation. Many writers have included this variable as a rather self-evident patient attribute involved in successful psychotherapy but, as Rosenbaum and Horowitz (1983) have discussed, the concept of patient motivation is quite global and subject to a variety of interpretations. In an attempt to clarify 196 conceptualizations of patient motivation they constructed a 36 item raters measure. A factor analysis revealed four factors--patient active engagement, psychological mindedness, incentive-mediated willingness to sacrifice and positive valuation of therapy; the factor structure was further consolidated to emphasize active engagement and psychological mindedness. Rosenbaum and Horowitz (1983) concluded: The results indicate that motivation needs to be treated as a multidimensional construct. Distinctions need to be made between factors pertaining to patient motivation and those which refer to patients' suitability for a particular treatment. Motivation should not be considered as a static quality, but as a fluctuating interaction of multiple factors. (p. 346) Malan (1976b) ranked patient initial motivation as the most important patient characteristic of his psychotherapy studies. The working definition of motivation used emphasized motivation for insight based on a psychodynamic understanding of problems. Over 80% of the patients who scored highest on psychotherapy outcome also had high initial motivation scores. Keithly, Samples, and Strupp (1980) found a strong concurrence between therapists' and outside raters' ratings of global changes and ratings of 197 patient pre-therapy motivation. Sifneos (1979) also considers patient motivation to be a central prognostic predictor. Orlinsky and Howard (1986) reviewed seven studies which investigated patient motivation and therapeutic outcome. Support for a relationship is present when ratings were done by the therapist or outside observer, but the only study using patient ratings failed to show a significant relationship. Similarly, Luborsky, Crits- Cristoph, Mintz, and Auerbach (1988) reviewed eight studies and concluded that greater patient motivation, especially “need for change", for psychotherapy was positively related to outcome. Patient engagement, one of factors identified by Rosenbaum and Horowitz (1983), was reviewed by Orlinsky and Howard (1986). They indicated that over 70% of the findings were supportive of a relationship between patient ”role—engagement" and outcome, predominately when therapist's evaluated outcomes. Obiect Relations The third general patient factor has to do with the quality of object relations. Many writers in the field have stressed the fundamental importance of the quality of the patient's object relations and see it as almost synonymous with the capacity to enter into and benefit from the therapeutic alliance. Moras and Strupp (1982) note in their review of the psychoanalytic literature that the patient's 198 pre-therapy capacity for interpersonal relations has been repeatedly emphasized as an important screening variable and it is seen as a critical factor in the outcome of psychotherapy. Davanloo (1979) used the quality of the patient's relationships (e.g., the presence of at least one meaningful relationship) as an important criterion for treatment suitability. Bachrach and Leaf (1978), in their review of "analyzability" in the psychoanalytic literature, concluded that the quality of patient object relations was consistently related to the degree to which patients derive benefits from psychoanalysis. Strupp (1974) stated that one of the core determinants of the patient's capacity to form a working alliance is the quality of their previous experiences with significant others. Frieswyk and colleagues (Frieswyk, Colson, & Allen, 1984; Freiswyk, Allen, Colson, Coyne, Gabbard, Horwitz & Newsom, 1986) state that the patient's capacity to see the therapist as a good object is a critical determinant toward establishing the therapeutic alliance. Bordin (1975) suggests that seeing the therapist as a good object is "intimately related to hopeful and trustful states and dispositions" (p. 258). There are both intrapsychic and interpersonal components to one's object relational capacity; thus, a conceptual distinction should be made between the interrelated concepts of object relations and interpersonal relations. Object relations refers to internal 199 representations of relationships, in particular, significant early relationships. Horner (1987) defines object relations as: The term object relations refers to the nature of self and object representations, and the dynamic interplay between them. The mental structures become manifest in interpersonal relationships, particularly in the transference. (p. 227) The distinction between internal and external representations is elaborated by Urist (1980): In the experiencing of “real" relationships between self and others in the external world, the individual processes and registers present experience in the context of the ways in which past experience has been organized. In attempting to "make sense" out of interpersonal experience, the individual brings to bear psychological processes whose function it is to register and organize mental representations of the self and others. These processes...take on the task of organizing an affectively charged interweaving of psychic content, deriving from more fantasy than from any veridical perception of the external world...object relations refers, then, to...intrapsychic functioning, not social 200 behaviors per se. (p. 821) In contrast to object relations, interpersonal relations refers to those affects, cognitions, attitudes, and skills--translated into observable behavior--that enable or hinder one's capacity to engage in an interpersonal relationship with another person. The distinction between object and interpersonal relations is, in many ways, more conceptual than actual since there is a fluid continuum of connection between these two sets of processes. The internalized aspects of significant early relationships-- object relations--colors all subsequent interpersonal relationships, in particular, close relationships. A prime example of this is the therapeutic relationship. Loewald (1960) elucidates the connection between the object relations and the therapeutic alliance: The patient tends to make this potentially new relationship into an old one. On the other hand, to the extent to which the patient develops a positive transference (not in the sense of transference as resistance, but in the sense in which 'transference' carries the whole process of an analysis) he keeps this potentiality of a new object-relationship alive through all the various stages of resistance. (p. 17) 201 Given the parallels between early childhood relationships and the therapeutic relationship (e.g., regression, transference, dependency, unequal status with unilateral emphasis, etc.), it follows that the therapeutic alliance is vulnerable to the same sorts of stresses and conflicts that are inherent in close relationships; the alliance is affected by the quality or level of maturity of the patient's object relations. It is a logical extension to suggest that any notable disturbance in the patient's early significant relationships (which in turn affect the patient's capacity for object relatedness) should ultimately influence the patient's capacity to engage in and benefit from the therapeutic alliance. For a vast majority of patients, the disruptions experienced in early relationships will be comparably minor and the actual implications for the treatment process may be slight. However, for those patients who have experienced severe disruptions or deprivation in their early significant relationships (and thus affecting their overall internal representation of object relations), the potential implications for the therapeutic alliance are pivotal and far-reaching. Gaston, Marmar, Thompson, and Gallagher (1988) have emphasized that "aspects of the patient ego...as it is reflected by the quality of interpersonal functioning“ as a critical determinant in the patient's contribution to the therapeutic alliance. They elaborate: 202 The quality of the patient's interpersonal functioning is seen as central to the development of the therapeutic alliance. In brief psychotherapy, for patients to work through problems with the therapist, they must be capable, prior to treatment, of forming and sustaining trusting and mutual relationships. Otherwise, the work must focus on developing the patient capacity to relate to others, and the task of addressing behaviors, cognitions, or intrapsychic conflicts that contribute to symptom formation is often deferred. (p.483) Researchers in this area are divided in their opinion of the capacity of patients assigned to the more severe diagnostic categories--in particular, borderline, narcissistic and schizophrenic disorders--to engage in the therapeutic alliance and derive benefits from psychotherapy. For example, some authors (e.g., Greenson, 1967; Langs, 1973, 1974) emphasize that the establishment of a rational working alliance, based upon some degree of mature object relations, is necessary for therapeutic progress. Other authors (e.g., Karon and VandenBos, 1981; Kernberg, 1975; Kohut, 1977) argue that limitations in object relatedness or psychostructural disturbances do not necessarily preclude the development of a therapeutic alliance, but instead 203 require modifications in the therapeutic approach which take into account the developmental level of the patient. This leads to a primary conclusion that the quality of the patient's object relations has an impact on the formation of the therapeutic alliance and, subsequently, the treatment process and outcome. For the purposes of the present study, limitations in the patient's capacity for object relations are likely be manifested in subtler ways such as defensiveness, negative alliances and transferences and difficulties in establishing a therapeutic alliance; and such limitations will be reflected ultimately in the course and outcome of psychotherapy. Empirical findings, though few in number and limited by their methodology, have been overwhelmingly supportive of the connection between the quality of a patient's object and/or interpersonal relations and the therapeutic alliance as well as outcome. Ryan (1973) and subsequent studies using Ryan's paradigm (Lehrke, 1977; Ryan and Cicchetti, 1985) found that patient pre-treatment variables, in particular patient pre-therapy object relations and patient hopefulness, accounted for a significant amount of variance (up to 30%) in the formation of the alliance. The strongest relationship was between object relations and two dimensions of patient participation in the initial interview: specifically, personal freedom (the balance between psychic flexibility and integration) and quality of alliance 204 (willingness of the patient to give consideration to the therapist's intervention). Moras and Strupp (1982) concluded that patients who were rated as having more positive interpersonal relationships upon entering therapy made greater positive contributions to the therapeutic alliance and, to a moderate degree, had better treatment outcomes; they noted that up to a 25% of the variance of the patient's collaborative activity in sessions could be accounted for by pre-therapy assessment of interpersonal factors. Horowitz, Marmar, Weiss, Dewitt, and Rosenbaum (1984) discovered a complex, interactive relationship between therapy process variables and "developmental level of self-concept“, an object relations measure (as well as “patient motivation for dynamic therapy"); this interaction was also predictive of outcome. Marziali (1984b) found a relationship between patient self-reports of social adjustment and alliance measurements from three perspectives (patient, therapist and rater). She noted that "the patient's capacity to in engage in adaptive social interactions is intrinsically associated with the patient's responses to the therapeutic relationship" (p. 422). Two studies of interest to the present discussion considered patient interpersonal factors using the Leary circumplex model (Crowder, 1972; Filak, Abeles, & Norquist, 1986). While the basic research questions and design were different, both found support for the relationship between 205 patient interpersonal factors and process and outcome in therapy. In the Filak, Abels and Norquist (1986) study, 72% of patients whose pre-therapy ICL score was characterized as "affiliative" had successful outcomes, whereas only 38% of patients whose pre—therapy ICL score was characterized as “hostile" had successful psychotherapy outcomes. Crowder (1972) found an interaction between the phase of therapy and therapist-patient complementarity in successful therapy. In the early phase of psychotherapy, successful patients manifested less transference reactions as indicated by their greater hostile-competitive, and less passive-resistant and support-seeking in comparison to unsuccessful patients, who showed a greater frequency of transference reactions. By the later phase of treatment, transference reactions no longer differentiated successful from unsuccessful patients. Piper, deCarufel, and Szkrumelak (1985) found that the best predictor of process and outcome was the patient's object choice (in addition to the defensive style of the patient), a measure which assessed the quality of relationships with important persons. Recent research from the Langley Porter Institute (e.g., Gaston, Marmar, Thompson, & Gallagher, 1988; Marmar, Horowitz, Weiss, & Marziali, 1986; Marmar, Weiss, & Gaston, 1987) has shown that patient pre-therapy interpersonal factors have been consistently associated with specific patient variables contributing to the development of the therapeutic alliance in a variety of patient 206 populations and therapy approaches. The one exception to this conclusion concerns psychotherapy with elderly patients, where Gaston, Marmar, Thompson and Gallagher (1988) failed to see a relationship between the patient's quality of pre—treatment interpersonal functioning and therapeutic alliance contribution. It should be noted, however, that pre—treatment interpersonal functioning was assessed via the Young Loneliness Inventory, a measure which appears to have limitations regarding its capacity to measure object or interpersonal relations. Finally, Luborsky, Crits-Cristoph, Mintz, and Auerbach (1988) reviewed eight studies (several were over 20 years old) and found extremely strong support for the predictive significance of patient pre-treatment object relations. Symptom Severity The impact of patient symptom severity (pre-treatment symptomatology) upon the formation of the therapeutic alliance and treatment process and outcome has been discussed by some authors. Studies which have examined the relationship between initial symptom severity and/or initial levels of affect (e.g., anxiety and depression) and therapy outcome (without considering the therapeutic alliance) have indicated that greater initial levels are either unrelated to outcome or are correlated with a positive outcome (see Luborsky, Crits-Cristoph, Mintz and Auerbach [1988] for a 207 review). Malan (1976) found that severity of pathology and duration of complaints were not significant predictors of psychotherapy outcome. However, empirical studies considering the relationship between initial symptomatology and therapeutic alliance have not, for the most part, been supportive of a relationship (e.g., Marmar, Horowitz, Weiss & Marziali, 1986). Investigations which have examined these somewhat counter-intuitive findings indicate that upon closer examination, the critical factor tends not to be symptomatology per se but rather “psychostructural“ variables such as ego-functioning, level of defensiveness, and object relations/interpersonal relations capacity. Gaston, Marmar, Thompson and Gallagher (1988), commenting on the relationship between pre—treatment symptomatology and the development of the therapeutic alliance, state: Symptom severity is conceptualized as a complex of interpersonal, intrapsychic and neuro-biological processes that are, in large part, independent of ego developmental level, of quality of object relationships, or other factors related to the development of the alliance. For example, patients with stable, trusting, and intimate relationships may present with severe depression or anxiety, whereas patients with serious characterological problems may not complain about high levels of symptoms. (p. 484) 208 Truax, Wargo, Frank, Imber, Battle, Hoehn—Saric, Nash and Stone (1966) found that the number of initial complaints, as measured on the Symptom Checklist (SCL-90), were related to improvement. Tollington (1973) had similar findings using the Symptom Inventory. Moras and Strupp (1982) found that pre-therapy ratings on the patient's subjective distress, severity of problems were not as predictive of alliance formation as ratings of pre-therapy interpersonal functioning. Morgan, Luborsky, Crits— Cristoph, Curtis, and Solomon (1982), using the Health— Sickness Rating Scale as assessment of initial pathology in patients, did not find a significant relationship between helping alliance and initial pathology. Marziali (1984b) found that ratings of general symptomatology were not significantly correlated with alliance ratings, but patient pre-therapy social relations were. Initial level of symptomatology was not predictive for two dimensions of patient involvement in the therapeutic alliance in two recent studies from the Langley Porter Institute (Gaston, Marmar, Thompson & Gallagher, 1988; Marmar, Horowitz, Weiss & Marziali, 1986). Piper, deCarufel and Szkrumelak (1985) found that initial severity of disturbance was not a significant predictor of process and outcome variables, instead object choice and defensive style of the patient accounted for most of the variance. Gurman (1977) noted a lack of empirical support in his review of six studies which 209 considered the degree of emotional disturbance in the patient and various relationship measures. The one study with contrary findings (i.e., that was supportive of a relationship between patient pre-therapy symptomatology and alliance formation) did not control for patient pre-therapy interpersonal capacity or ego-functioning (Eaton, 1987). Taken as a whole, psychodynamic theory and empirical research point to the importance of specific patient variables in predicting therapeutic alliance, psychotherapy process, and to a lesser degree, psychotherapy outcome. In particular, variables which address underlying psychostructural factors (which are indicative of the resources the patient brings to therapy) such as quality of object relations, ego-functioning, and motivation or, more general expressions of these factors--such as interpersonal adjustment, have been most predictive. 210 Therapists' Contribution There are a number of factors which the therapist brings to the psychotherapeutic situation; these can be conceptualized and organized in a variety of ways. For example, Luborsky and Auerbach (1985) have distilled two primary sets of therapist factors: (a) therapist empathy and relationship facilitating conditions, and (b) therapist technical factors and competence. In my survey of various articles and, in particular, literature reviews, one especially cogent organization was developed by Beutler, Crago and Arizmendi (1986) in their review article on therapist variables in psychotherapy process and outcome. The authors developed a circumplex model which attempts to "distinguish between those (therapist variables) that exist independently of and coincidentally to the treatment relationship, and those that are specifically designed to have an impact on the treatment process" (p. 258). They proceed to note that therapist characteristics differ in the degree to which they are objectively observable versus inferred internal aspects. Thus there are two interactive therapist dimensions--(a) therapy-specific versus extratherapeutic characteristics, and (b) externally observed versus inferred internal characteristics, resulting in four quadrants. 211 Briefly, these four quadrants are: (1) External, extratherapy characteristics: age, gender, SES, ethnicity, etc. Because these factors are not immediately pertinent to the current study, they will not be discussed here (The reader is referred to Beutler, Crago, & Arizmendi [1986]; Luborsky, Crits-Cristoph, Mintz, & Auerbach [1988]; Orlinsky and Howard [1986]; and Parloff, Waskow and Wolfe [1978] for literature reviews); (II) Inferred internal extratherapy characteristics: Therapist personality style, emotional well-being, personal attitudes and values, etc; (111) External, therapy-specific characteristics: Therapeutic style, competence, experience, therapeutic interventions, theoretical orientation, etc; and (IV) Inferred internal therapy-specific characteristics: relationship attitudes, expectations, and other aspects of the therapist's contribution to the therapeutic alliance. These quadrants will provide the general framework from which to review therapist factors. Therapist Personal Factors This section deals with inferred, internal extratherapy therapist characteristics. While these factors bear upon the process and outcome of psychotherapy, they appear to be less central than the various therapy-specific therapist characteristics. There are a limited number of studies dealing with this variable and, because of their lack of methodological sophistication, the question of the effect of therapist personality patterns on therapy--either directly 212 or in terms of the most efficacious therapist-patient match- -remains unanswered. The following represents a brief overview of the area. Therapist Personality Factors. Studies which have investigated the effects of the therapist's personality on the psychotherapy process and/or outcome have often been limited by design (e.g., simplistic or unidimensional) and seem to indicate that no rudimentary causal relationship exists. One relatively well researched therapist personality style is the A-B dimension. The A-B therapist personality refers to the observation that “A” therapists were seen as humanistic, person-oriented and were found to be more effective in treating schizophrenic patients, in contrast to "B" therapists who were more problem-oriented, authoritarian and ostensibly more helpful with neurotic patients (Razin, 1977). This personality dimension has been fairly extensively researched over the years and has resulted in rather weak and equivocal findings (e.g., Beutler, Crago & Arizmendi, 1986; Cox, 1978; Razin, 1977). Luborsky, Crits- Cristoph, Mintz and Auerbach (1988), in reviewing 12 studies, concluded that Type A therapists are slightly more effective (i.e., 50% positive, 8% negative and 42% nonsignificant in comparing Type A with Type B). 213 One large-scale and methodologically complex study (Berzins, 1977) which examined the effect of matching between therapist and patient personality types found a complementarity in style. They noted that dominant, individualistic therapists were more effective with dependent, submissive patients, and that dependent therapists did better with autonomy-oriented patients. In their review of the literature, Beutler et al. (1986) conclude that there is no clear finding concerning how therapist personality style influence the process and/or outcome of psychotherapy. There are some studies which indicate that therapist-patient dissimilarities on certain variables (e.g., independence-dependence) can facilitate the treatment process, but the findings are uncertain and comparatively minor, directing Beutler et al. (1986) conclude "It is unlikely that any single dimension of personality or personality similarity (of the therapist) is a major facilitator or inhibitor of therapy benefit” (p. 271). Therapist's own Psychotherapy and Emotional Well-Being. Luborsky, Crits—Cristoph, Mintz, and Auerbach (1988) reviewed 5 studies noting that neither the fact that the therapist was in therapy nor length of therapy related to patient outcome. One difficulty in assessing this relationship is that the therapist's relative level of 214 disturbance interacts with the potential for improvement or resistance to the personal therapy they receive. Some studies are supportive of improved self-esteem, ability to establish warm therapeutic relationships, and other process variables, but the relationship to subsequent psychotherapy outcome is only mildly supported (e.g., Buckley, Karasu, & Charles, 1981). Beutler, Crago, and Arizmendi (1986) contend: Overall, a very mixed picture of this relationship has emerged in the outcome literature...Therefore, while the best available literature suggests that the therapist's emotional well-being at least moderately facilitates both effective treatment processes and outcome, the evidence does not consistently support the value of personal therapy. (p. 273) However, there is more conclusive support for the relationship between the therapist's emotional well-being and positive patient outcomes. A dilemma which immediately presents itself is how therapist well-being or psychological disturbance will be defined. Further, as Beutler et a1. (1986) observe, the therapist's level of emotional health or disturbance must be measured independently of the psychotherapy they provide, otherwise such interactions would obfuscate the examination of patient benefit from 215 therapy. Some studies have shown significant differences in the therapy process, finding that troubled therapists were more likely to mishandle the affective climate of the therapy (e.g., Wogan, 1970) as well as other aspects of the therapy relationship and therapist interventions (e.g., Bergin and Soloman, 1970). In terms of the relationship between therapist well—being and therapy outcome, Garfield and Bergin (1971) determined that greater patient improvement on depression and defensiveness was linked to healthier therapists. Reviewers (e.g., Parloff, Waskow, & Wolfe, 1978) have noted other studies with similar findings. Wogan (1970) found that higher level defensive styles as measured by the MMPI were associated with patient's rating their therapeutic progress as accruing more quickly. VandenBos and Karon (1971) found, using independent therapist T.A.T. ratings, that therapist pathogenesis (i.e., ‘where the dominant person does not take into account the needs of the dependent person, Karon and VandenBos, 1981, p. 123) was a significant predictor of therapeutic effectiveness with schizophrenic patients. Luborsky, Crits- (Zritoph, Mintz, and Auerbach (1988) reviewed 7 studies which [examined therapist emotional well-being and concluded that WHealthier therapist personality qualities (were) related to better patient outcomes" (p. 343), finding approximately 60% gxasitively related and 40% nonsignificantly related to pat ient treatment outcome . 216 Therapist Values/Attitudes. Research in the area of values and attitudes is thought to be important because therapist values and belief systems are hypothesized to significantly affect patients who are receptive. It is especially difficult to objectively define such amorphous terms as "attitudes" or “values“ in a way that allows for meaningful questions to be researched within viable psychotherapy research paradigms. Beutler, Crago, and Arizmendi (1986) list several problems and confounds with research in this area, including (a) difficulty in determining causal relationships and directionality; (b) translation of abstract values or attitudes into measurable behaviors; and (c) it is complicated to design experiments in which therapist's attitudes can be controlled for and measured. Two of the most prominent areas of research include religious values and therapist-patient value similarity or dissimilarity. One general finding with regard to the few studies done on religion and morality is that therapist and patient values tend to differ, and that these differences are not often significant to the selection of therapist or the process of psychotherapy (e.g., Beutler, Pollack, & Jobe, 1978). The greatest number of studies has been conducted in the area of initial value dissimilarity. The studies indicate that initial therapist-patient value incongruity are gradually transformed into the patient's 217 values slowly approaching those of the therapist (e.g., Beutler, 1981). Beutler et al. (1986) conclude their literature review in this area by stating: ...the observation that therapist and patients have different belief systems, particularly around issues of religion and morality. Once in treatment, however, there is a decided tendency for successful therapy dyads to be associated with the patients' acquiring therapists' belief systems. (p. 275) The authors are quick to point out that initial value dissimilarity does not, in and of itself, predict positive psychotherapy outcome. Therapist Training, Style, and Interventions This section deals with those observable therapy-specific behaviors whose presence is, by and large, designed to have a specific effect on the psychotherapeutic process. Therapist Trainingvand Experience. There is no clear agreement in the field about what constitutes adequate training or experience or how one measures this. For example, one study's 'inexperienced' therapist may be another study's 'experienced' therapist (see Lambert, Shapiro, & Bergin [1986] for a discussion). In a related 218 vein, one's theoretical orientation and level of experience influence which domain of phenomena is being examined, which, in turn, interact with how psychotherapy process and outcome will be considered; this is especially important when therapist ratings are an integral part of the study (e.g., Marziali, 1984a). A corollary issue is the type of training as well as the affiliation or professional discipline with which the therapist is associated. Before examining some of the research, attention should be directed to some of the general conclusions and beliefs from the clinical field to provide the conceptual framework from which empirical studies have originated. For example, some writers have suggested that with experience comes maturity and life experiences that can be integrated into the psychotherapy process. Several writers (e.g., Auerbach and Johnson, 1977; Parloff, Waskow, & Wolfe, 1978; Rice, 1965) in the area suggest that experienced therapists are more able to: establish a quicker and deeper therapeutic alliance, communicate better, provide more empathy (Mullen and Abeles, 1971), show more positive regard (Barrett- Lennard, 1962), facilitate the client's expression of affect and receive more favorable evaluations from patients. Further, while Auerbach and Johnson (1977) state that no *while steadfast conclusions can be drawn from the equivocal :research.on therapist experience and psychotherapy process and outcome, certain qualified conclusions are suggested by 219 'the material: (a) inexperienced therapists, especially ‘untrained.paraprofessionals, are more directive and confrontational than trained therapists; (b) experienced therapists have been shown to be more active (Grigg, 1961); (c) inexperienced therapists elect to use “safer" interventions which tend to be explicit and brief; (d) experienced therapists tend to talk more, commit themselves through the use of interpretations, and take the initiative with greater frequency than inexperienced therapists (e.g., Ornston, Cicchetti, Levine, & Fierman, 1968; Strupp, 1955b; Strupp, 1958). Other authors have also suggested that beginning therapists make more errors (Matarazzo, Wiens and Saslow, 1966). There have been a number of literature reviews of research pertaining to therapist level of experience and/or type of training and there are significant discrepancies in how the reviewers have interpreted the relationship between therapist experience and outcome. For example, Smith, Glass and Miller (1980), in their meta-analysis of nearly 500 psychotherapy studies concluded that “there was no relationship between years of experience of the therapists in the study and the magnitude of the therapeutic effect produced in that study" (p. 117). On the other hand they did notice a relationship between type of training and outcome, stating that those trained in psychology obtained the largest effects (this conclusion is qualified because of 220 the observation that there was an interaction between type of training and type of problem treated). Other researchers, however, cite correlations from 30 to 60% with regard to the impact of therapist experience. (1986) reviewed 35 studies of therapist level For example, Beutler et al. of experience and found a positive relationship in 42% of the studies‘and a nonsignificant relationship in 58% of the studies. Luborsky, Crits—Cristoph, Mintz and Auerbach (1988) reviewed 16 studies (some of which overlapped with Beutler et al.'s review) and found virtually the same percentages, and concluded that therapist level of experience is positively related to outcome. In an attempt to reconcile these inconsistent findings, Stein and Lambert (1984) reviewed 27 studies which examined therapist experience and treatment outcome and formulated a list of mediating variables. Three variables which appeared to mediate therapist experience and outcome seemed (a) when the especially pertinent to the present discussion: distinction between levels of experience is explicit and (b) when more intensive psychotherapy is considerable; utilized; and (c) when more severely disturbed patients were treated. 1988) which In contrast, a recent review article (Garb, investigated the relationship between clinical judgement, clinical training and professional experience is not supportive of the general value of experience or the 221 validity of clinical judgments and somewhat supportive of the impact of training. One study concerning type of training (and, secondarily, amount of experience) came from the Vanderbilt Psychotherapy Project (Gomes-Schwartz, 1978b; Strupp and Hadley, 1977). Among other factors which were investigated, the researchers compared college professors with trained psychotherapists. Overall, they found a very different process (e.g., trained clinicians were more likely "to investigate the psychodynamic roots of the patient's 1031]) but the outcomes problems" [Gomes-Schwartz, 1978b, p. were not significantly different. There are a limited number of studies which have investigated type of training or affiliation separately. Briefly, professional affiliation has been shown to be associated with a different therapeutic stance and therapy process (e.g., Orlinsky and Howard, 1975); these differences have occurred mostly in the context of patient perceptions of the therapist. There is some support for differential effectiveness between psychologists and psychiatrists (e.g, Smith, Glass, & Miller, 1980) but this may be due in part to different patient demographics (Taube, Burns, & Kessler, 1984) and different diagnostic and/or treatment concerns (e.g., Beutler, Crago, & Arizmendi, 1986). 222 Therapist Style. Therapeutic style refers to those generalized and not directly conscious therapist behaviors which are not usually linked explicitly to a theory of change and are often relegated to the amorphous category of non-specific factors. Parloff, Waskow, and Wolfe (1978) proposed that the type of therapist expression and communicative pattern can be viewed as independent of the content of the therapist's communication. Furthermore, these stylistic factors may not be incidental but can be as important as the intended meaning of what is said. There are several fundamental categories of therapist style--most of which are verbal in nature--which have been researched (this will be discussed in greater detail below). There is a literature on various non-verbal therapist behaviors, and some of the areas which appear to exert at least an initial effect include: physical proximity between therapist and patient, therapist attire and office decor, and therapist posture in relation to the patient (see Beutler et a1. [1986]; Orlinsky & Howard, 1986; Parloff, Waskow & Wolfe and Pope [1977] for a discussion of these [1978]; (1986) point out, variables). However, there have been virtually no studies of the enduring impact as Beutler et al. of such variables on psychotherapy process and/or outcome. Beutler et al. (1986) discuss three general categories of therapist verbal styles which have been researched: (1) Therapist directiveness--patterns that direct 223 patients' attention, action or thought; (2) lexical characteristics that direct therapist language including pauses, interruptions, change of topics, and word counts; and (3) therapist self-disclosure. (p. 288) There are various operationalizations of therapist directiveness, but, in general, high degrees of therapist directiveness are characterized by high initiative, comments outside the immediate context of the patient's statements or lead, and, often, “advice: recommending following a course of action“ (Kiesler, 1973, p. 297). In contrast, low levels of therapist directiveness are usually distinguished by an emphasis on the therapist reflecting the client's comments and/or affective state, and encouragement of the client's further exploration of the material presented. It is apparent that the directive versus non-directive dimension overlaps considerably with other fundamental aspects of the psychotherapy situation. Two closely related dimensions are therapist activity level and therapist level of ambiguity versus specificity. The limited scope of this paper permits only a brief sununary statement of each of these two dimensions. Therapist activity level is a more encompassing, non-specific variable which includes three attributes: initiative, inference and ambiguity (Pope, 1977). Research indicates that one of the 224 promising aspects has been the quantity of the therapist's verbal output and its relationship to the emerging synchronization with the patient's verbal output (Pope, 1977). With regard to the ambiguity/specificity dimension, psychoanalytic writers such as Bordin (1955, 1979) have asserted that ambiguity can encourage the flow of associations which lead to the patient focusing on "his major conflicted feelings no matter how unaware he is of them“ (Bordin, 1955, p. 13). Pope (1977) reviews a number of relatively older studies and concludes that, consistent with an informational exchange model, “low therapist input (high ambiguity) elicits high patient productivity; high therapist input (high specificity) elicits low patient productivity “ (p. 390). As intimated earlier, one of the difficulties with research in this area is attempting to isolate only one of a number of therapist variables in a unidimensional approach. Beutler et a1. (1986) explain that therapist directiveness seems to interact with other variables, and state, "Investigations that have looked specifically at patient variables in conjunction with therapist directiveness have produced more promising results (p. 288)." Some support exists for therapist directiveness being correlated with such factors as (a) an increase in negative ratings of the psychotherapy session (Hoyt, Marmar, 225 Horowitz, & Alvarez, 1981) and (b) a greater number of noncompliant behaviors in aggressive adolescents (Patterson and Forgatch, in press). Studies measuring the relationship between therapist directiveness and outcome have yielded inconclusive or equivocal support for various levels of therapist directiveness (see Beutler, Crago, & Arizmendi [1986] for a review), clearly indicating the need to further examine variables which interact with directiveness. One therapist variable which may moderate directiveness is therapist level of experience. McCarron and Appel (1971) compared similar therapist styles (i.e., directive and confrontative) and found that this generated patient resistance when practiced by inexperienced but not experienced therapists. The most thoroughly researched multidimensional paradigms have focused on the interaction of therapist directiveness and various patient factors. Ashby, Ford, Guerney, and Guerney (1957) found differential patient responsiveness to therapist's directive or reflective psychotherapeutic intervention style. Patients high in [defensiveness and aggressiveness showed increases in these resistant behaviors in response to directive but not to reflective therapist style. Beutler et a1. (1986) cite a number of studies which examined the interaction between therapist directiveness and patient attributional style and conclude that "it has been sufficiently persuasive to 226 suggest to some...that therapeutic interventions should selectively encourage either controllable (internal) or uncontrollable (external) attributions, based upon patients' attributional style" (p. 289). For example, Forsyth and Forsyth (1982) found that patients for whom the locus of control is external may show decrements in functioning with therapeutic approaches that stress personal control and responsibility (i.e., non-directive therapeutic style). While some studies have shown that therapist "lexical 2” characteristics" are correlated with therapeutic process variables, the relationship to outcome is equivocal. There are a limited number of studies which examine therapist expressiveness. For example, Rice (1965) found that therapists' voice quality and expressive stance was positively related with outcome. Some studies are supportive of moderate amounts of therapist verbal activity (Grigg & Goodstein, 1957) while others indicate that low levels of therapist verbal activity with more severely disturbed patients are more conducive to therapeutic gain (e.g., Staples & Sloan, 1975); still others (e.g., Scher, 1975) found a non—significant relationship. It may be the case that therapist lexical characteristics may be too broad and unspecified a variable since it appears to interact with therapeutic content and connotative forms of therapist expression as well as some patient characteristics. Patient factors have not been well-researched yet, but one study 227 indicated that therapist and patient verbal pattern congruency was significant (Tracey & Ray, 1984). Another related variable is therapist self-disclosure. The research literature is replete with analogue studies showing a clear, positive relationship between therapist self-disclosure and patient therapeutic engagement, etc. When the examination is switched to either outcome studies or actual psychotherapeutic contexts, the number of studies is limited and the findings are largely non-supportive of the significance of self-disclosure. Therapist Interventions. Therapists' interventions represent intentional, theoretically-grounded and systematic efforts to help the patient in the therapeutic context. Beutler et al. (1986) distinguish three classes of therapists' intervention-—which will provide the organization of the following section: (a) therapist competency; (b) theoretical orientation of the therapist; and (c) specific technical procedures (i.e., actual interventions). Therapist Competence. Schaffer (1982), who has written extensively on the notion of therapist competence and skillfulness, has proposed that therapist skillfulness may be "rated by the extent to which the qualities of the actual operation correspond to the qualities of a skillful 228 operation according to the theory“ (p. 488). Schaffer (1982, 1983) believes that in evaluating therapist factors such as skillfulness, it is essential to consider the multidimensional nature of therapist behavior and thus differentiate between type--"the category of the tactics and goals prescribed by a given theory", skillfulness--”the extent to which these skills are carried out competently" and interpersonal manner of the therapist-- "the manner in which the therapist relates to the patient“ (1982, p. 670). Furthermore, Schaffer (1982, 1983) believes that most studies of therapist factors are limited by a unidimensional approach which can result in weak or non-significant findings in otherwise well-designed research. He proposes that it is necessary to incorporate multidimensional .measures to assess the above-mentioned therapist dimensions, stating "these problems can be overcome by rating therapist Ibehavior on multiple factors and relating the factors to outcome in a multivariate way" (p.670). Kernberg, Burstein, Coyne, Appelbaum, Horwitz, and Voth (1972) in summarizing the Menninger Project, emphasized the .meortance of therapist skillfulness especially in the case <3f patients with more limited ego-strength. Kernberg (1975) :notes that the "skill of the therapist...is the most crucial :factor in the outcome of the treatment of patients with low .level of psychic functioning, that is, ego weakness“ (p. 147). Approaching the issue of competence inversely, Sachs 229 (1983), using the Vanderbilt Negative Indicators Scale, found that therapist "errors in technique", when judged from outside observers, was negatively related to patient improvement; in fact, errors in technique was the most salient predictor of the several therapist factors considered. Luborsky, Crits-Cristoph, Mintz, and Auerbach (1988) identified a strong positive relationship between therapist skill, expertness and therapeutic outcome (i.e., over 80% of the eight studies reviewed indicated this). Orlinsky and Howard (1986) reviewed five studies and concluded “all in all, these few studies show an impressive consistency in finding therapist skillfulness to be a significant determinant of patient outcome" (pp. 330-331). Beutler, Crago, and Arizmendi (1986), in their review of the literature, remark that competency is evaluated either in terms of (a) some external standard (e.g., criterion levels based on a theoretical model, concurrence of experts, etc.) or (b) based upon improvement rates (e.g., comparing therapists whose patients have high improvement rates versus therapists whose patients have comparatively low rates). They cite some studies which correlate psychotherapy process with competency and find support for the importance of therapist competence. For example, Sloane, Cristol, Pepernik, and Staples (1970) found that patients reported that therapist skillfulness was an important aspect of their improvement in over 70% of 230 successful psychotherapy cases. In a recent study (O'Malley, Foley, Rounsaville, Watkins, Sotsky, Imber & Elkin, 1988) evaluating the effects on treatment outcome of therapist competence in interpersonal psychotherapy of depression, therapist competence was positively associated with patient-rated change and with some depression outcome measures, even after various patient pre-therapy variables were controlled for. Beutler, Crago, and Arizmendi (1986) summarize by stating that, while investigations in this area are still in an incipient stage, the emerging data is suggestive of a significant relationship between therapist skillfulness and therapeutic benefit. Theoretical Orientation. Different theoretical orientations hold different perceptions and expectations of clinical data and the psychotherapeutic process, and are guided, via their theory, toward differing therapeutic interventions. Wolfe (1956) delineated basic dimensions of theoretical orientation, proposing two fundamental types of psychotherapy: one utilizing formulas of interpretation and following a preconceived, systematic theoretical approach, while the other focused primarily on an evolving personal relationship. In a similar vein, Sundland and Barker (1962) have conceptualized two basic dimensions codified by research dealing with theoretical orientation, the first 231 being ideological versus experiential psychotherapy, and the second being therapists' activity level (Sundland, 1977). Other authors have emphasized the basic similarities among different approaches and, more specifically, among effective therapists. Frank (1982) believes that all successful psychotherapies deal with the patient's demoralization and convey that addressing this demoralization can be instrumental in bringing about a modification in symptomatology. He states: "All psychotherapies aim to break the resulting vicious circle and to restore morale by providing experiences with a helping person that offer general encouragement and support" (Frank, 1982; p. 33). Sundland (1977) suggests that it may be critical to assess the extent to which the therapist's actual interventions parallel the professed theoretical orientation. Theoretical orientation can be manifested only through specific therapist behaviors and until it is reified in behavior it remains abstract and essentially unmeasurable; this will be discussed further in the upcoming section. Several major treatment outcome reviews are in agreement that the variations between schools of thought have not been predictive of differential treatment outcome (e.g., Luborsky & Auerbach, 1975; Shapiro & Shapiro, 1982; Smith, Glass, & Miller, 1980). Smith, Glass and Miller (1980) remark that only about 10% of the variance in effect size is attributable to different treatment approaches. And 232 Brown (1987), in his review of six major meta-analyses of treatment outcome, states "Most (studies) reported little or no difference between the various schools of psychotherapy while also reporting at least a slight efficacy in psychotherapy in general“ (p. 19). There is, however, an emerging viewpoint which suggests 5...; 'q _i that meaningful differences among therapies lie in the 1“" rial. i ‘ ‘- t 1' interaction of type of therapy or theoretical orientation with certain patient variables (e.g., patient pre—therapy object relations capacity, general level of psychopathology, etc.). Beutler and colleagues (1979c; 1981; 1984) have explored the relationship between therapeutic approaches and patient dimensions. In one study (Beutler & Mitchell, 1981), a comparison was made between analytic and experiential approaches and a differential effect was found in that patients who tended to externalize conflict responded better to experiential psychotherapy. However, the authors concluded that overall, the differential effect was less a function of a general theoretical dissimilarity and more due to specific technical procedures on the part of the therapist. Other authors have discussed the value of matching theoretical orientation and/or techniques with patient variables (e.g., Beutler, 1983); this issue will be discussed in greater depth in the next section. Luborsky, Crits-Cristoph, McLellen, Woody, Piper, Liberman, Imber and Pilkonis (1986) observed that variations 233 in success rates across four psychotherapy outcome studies were due more to the therapist than to therapeutic approach. Among other factors which mediated this, one included the purity of the therapist's technique (i.e., the extent to which the therapy included qualities of the intended approach). In a study which examined various therapist and patient contributions to effective psychotherapy, Comes-Schwartz (1978a) considered, among other variables, the impact of the level of training and theoretical orientation of therapists and untrained helpers. She found that the theoretical orientation and status (professional versus non- professional) of the therapists influenced the process but not the outcome of therapy. Analytically oriented therapists invested in uncovering the dynamic origins of the conflictual processes whereas untrained and Rogerian therapists offered a more warm and personal relationship. The different approaches did not, however, seem to affect the patient's attitudes towards the therapy or therapists; the patient invested in the treatment process and a therapeutic alliance was developed independent of approach. Commenting on the relative importance of "non-specific“ factors over theoretical orientation, Gomes—Schwartz writes: The variables that best predicted change were not related to therapeutic techniques but to the positiveness of the patient's attitude toward his 234 therapist and his commitment to work at changing. (p.103) Therapist Technical Factors This section will consider those specific therapist procedures generated from theoretical formulations from a particular model of psychotherapy and particular training. This represents an intentional and the most substantive aspect of the therapist's behavior. The relational aspects (i.e., empathy, warmth, therapeutic alliance) of the therapist's technical procedures will be examined separately in the following section. Beutler, Crago, and Arizmendi (1986), among others, view the fundamental distinction in therapy techniques as one of relative emphasis upon affective aspects (emotional expression and abreaction) versus cognitive aspects (interpretive and educational); this has also guided the emphasis in research. For example, in one study Wenegrat (1976) found that therapist verbal response styles which emphasized patient affect versus cognitions was more strongly associated with higher empathy levels. Mitchell and Hall (1971) compared high facilitator and low facilitator therapists and concluded that the former were more effective at confronting and interpreting the patient's affect. Beutler, Crago and Arizmendi (1986), as well as Orlinsky and Howard (1986), review a number of studies which 235 support the relationship between techniques focusing on patient affect and various therapy process measures. The relationship between the affective/cognitive variable and psychotherapy outcome is much more equivocal and complex. For example, Mintz, Luborsky, and Auerbach (1971) found no relationship between outcome and the degree to which therapist's interpretations were affectively or [ cognitively focused. Sachs (1983), cognizant of Gomes- M Schwartz (1978) findings of a null relationship between ‘ interpretations of defense and outcome, examined successful therapists' interventions and noted that the distinguishing characteristic of the therapists' interventions seemed to be the extent of focus and consistency of the interpretation. Thus, it appears to be more critical to consider the specific type (Schaffer, 1982) and the content focus of an interpretation. Hoyt, Marmar, Horowitz, and Alvarez (1981) developed rating scales to assess the variable emphasis of particular therapist and patient actions in the therapeutic context. One advantage of this approach is that the unit of study is more specific, objective and measurable. This is in contrast to many studies which have investigated psychodynamic variables where the unit under investigation remains vague, broad and theoretically-bound. Cluster analysis of the 34 items revealed three primary dimensions: (a) reaction to the therapist, (b) working through the 236 stress event, and (c) termination. Hoyt, Xenakis, Marmar and Horowitz (1983) related therapist's and observer's ratings of good and poor psychotherapy sessions to therapist's actions using the therapist action scale (TAS). They found a relationship between good psychotherapy sessions and therapist actions which emphasized encouraging the patient's expression of thoughts and feelings, and F ~ M- "kn-e J explorations of the patient's reactions; conversely, therapist's rated sessions as poor when the therapist's actions focused on detailed recall of the presenting problem, discussion of therapeutic goals, etc. Windholz, Weiss, and Horowitz (1985) used the TAS to examine therapist's actions in time-limited psychotherapy with bereaved patients. Seven general factors emerged as significant for the TAS, most of which showed significant effects across sessions. These items included transference and termination, clarification, affect expression, meaning of event, reassurance, relationships and errors. They also noted a significant shift in the relative importance of these factors across 12 sessions. For example, transference and termination issues shifted in an ascending manner across sessions indicating a connection between the loss of the original object (i.e., bereavement), transference and the termination of the brief therapy. The authors summarize their findings and recommendations for more effective time- limited psychotherapy: (1) clarifications of the patient's 237 thoughts, feelings or ideas, including how they relate to the initial stress event, (2) care to notice how expression of affect may be warded-off or avoided, and (3) attention to external and within—therapy relationships. Interpretation. One general definition of an interpretation comes from Orlinsky and Howard (1986): "An interpretation is essentially an explanatory statement intended to clarify the meaning of an action or experience“ (p. 324). The specific focus of the content of an interpretation may encompass a broad range of possibilities, but it remains a dominant type of intervention and a central technical variable to be examined. Among the studies which have focused on any one of a number of specific content areas, Orlinsky and Howard (1986) considered three general areas to be primary: affective focus, here and now focus, and transference. Orlinsky and Howard's (1986) review of five studies where the interpretive focus was on the patient's affect indicated 50% of the ten specific measures positively correlated with patient outcome; however, an equal number of measures did not correlate. They summarized “that focusing on affect is sometimes but not always helpful" (pp. 324-325). Examining nine measures across five studies which explored the therapist's focus on here and now interpretations, Orlinsky and Howard (1986) found that over 65% had a null relationship to outcome, and the remaining 238 measures showed a negative correlation with psychotherapy outcome. This seems to indicate that the "here and now focus" is not singularly sufficient to describe or capture the therapist's interpretive activities. A more traditional and psychoanalytic conceptualization of interpretation has a somewhat different and potentially more specified focus. Garduk and Haggard (1972) note that in addition to the idea that “interpretations are explanatory in nature; they bring intelligibility to a patient's productions" (p. 16), that they are also focus on unconscious material. Langs (1982) alludes to the emphases upon unconscious processes when he asserts: An interpretation is an attempt at a propitious moment in therapy to enable the patient to become aware of some content or function of which he or she had been unaware. (p. 643) A classic definition of an interpretation is offered by Bibring (1954): Interpretation...refers exclusively to unconscious material: to the unconscious defensive operations (motives and mechanisms of defense), to the unconscious, warded-off instinctual tendencies, to hidden meanings of the patient's behavioral patterns, to their unconscious interconnections, etc...On the basis of derivatives, the analyst FA... .21—q J . 239 tries to "guess" and to communicate (to explain) to the patient in the form of (hypothetical) constructions and reconstructions those unconscious processes which are assumed to determine his behavior. (p. 757) There are several noteworthy studies which have examined the effects of psychoanalytically-oriented interpretations on psychotherapy process and outcome. The operational definitions vary from general interpretations to basic transference interpretations to transference interpretations which differentially emphasize various objects and relationships. A brief review of some of the major studies follows. A comprehensive study which considered several dimensions of the immediate effects of interpretation on patients was done by Garduk and Haggard in 1972. They conducted an intensive analysis of four cases (two psychoanalytically oriented psychotherapy and two psychoanalysis) by extracting 15 examples of interpretations and non-interpretations from each case (as controls) and then examining the five minute period immediately following each type of intervention. Garduk and Haggard defined interpretations as focusing upon the explanation of unconscious processes and further differentiated interpretations into three levels of comprehensiveness. 240 There were a number of findings, owing to the fact that there were 17 original hypotheses which were considered, but the following results were deemed the most salient and were statistically significant in differentiating interpretations from non-interpretations: in the five minute period immediately following therapist interventions, interpretations (versus non-interpretations) were associated with (a) greater defensive and oppositional associations, (b) more transference related material, (c) increased presence of affect, and (d) greater understanding and insight. The authors suggest what they believe to be the link between interpretation and outcome: More, rather than less, understanding and insight was also associated with interpretations, and is favored by analytic theory...On the assumption that it is through the analysis of transference that therapeutic progress is achieved, interpretations again have the theoretically preferred effect in that they stimulated greater reference to such material. (p. 70) However, one chief limitation is the fact that this study examined the therapeutic process of interpretative versus non-interpretive activity and did not attempt to focus on the relationship of interpretation to therapeutic change or outcome . 241 Malan's empirical investigations into psychodynamic psychotherapy, beginning with his study of brief psychotherapy (1963), and continuing with his 1976 study, examined the effects of interpretation on psychotherapy process and outcome. Some of the findings of the earlier study included: (a) comprehensive interpretations of transference were associated with positive outcome and follow-up; (b) more successful outcomes were associated with the following--(1) early use of transference interpretations, (2) adequate interpretation of negative transference, and (3) careful exploration and patient participation in discussion of termination issues; (c) the most salient transference interpretations focused on the transference/parent link (Malan, 1963). Malan's (1976) subsequent study sought to replicate the original study with more methodological rigor and extended follow-ups. Extensive case notes, dictated from the therapist's memory, supplemented by records of the project's case discussions, served as the primary source of process data on the 39 patients used in this study. The major positive prognostic indicators in this study were: (a) motivation for insight based on a psychodynamic understanding of problems; (b) “focality”, which indicated the therapist's ability to successfully use the therapeutic plan developed at the outset of therapy; and (c) interpretations which emphasized the transference/parent link; this headed the ranking of a 242 number of types of interpretations. One significant methodological shortcoming of the study was the reliance upon therapists' case notes; the following study sought to correct this. Building upon the work of Malan's (1976) study of the effects of interpretations in psychoanalytic therapy, Marziali (1984b) refined and objectified the methodology by using audio tapes rather than therapists' summaries of psychotherapy sessions. Raters listened to tapes and located interpretations, and codified them into one of three categories: reference to a parent or sibling (P), reference to someone in the patient's interpersonal sphere (O), and a reference to the therapist or therapy (T). The findings supported Malan's original data: There was a positive association between more favorable outcome...and the frequency with which therapists' interpretations referred to emotions experienced in the transference relationship that were similar to those experienced in relation to the parents and other important persons. (p. 301) Specifically, interpretations linking the therapist and patient (T-P), as well as those interpretations linking therapist, patient and other (T-P-O), were correlated with two measures of dynamic outcome. Luborsky, Bachrach, Graff, Pulver, and Cristoph (1979) V 3::flmu—w 243 investigated the preconditions and immediate consequences of interpretations by looking at 16 segments (each consisting of 250 patient words before and after interpretation) for each of three patients in psychoanalytic psychotherapy. Interpretations were defined as: ...the analyst making references to unconscious material including defensive operations, motives and warded-off instinctual derivatives, as well as the underlying meanings of the patient's patterns of behavior and how these were connected. (p. 393) The findings indicated a differential responsiveness to interpretation among the three patients (from negative to some positive to very positive) and "a clear parallel between the positivity of the immediate response to interpretations and the outcome of treatment" (p. 391). Further, "there was a parallel in terms of the amount of this immediate positive or negative response to interpretation and the ultimate response to the treatment" (p. 398). Therapist factors (e.g., timing, adequacy of interpretation, etc.) did not account for much of the differential responsiveness observed between patients. Instead patient factors were more salient in accounting for the patients' dissimilar responsiveness to interpretations. Specifically, the following were predictive: (1) less initially healthy ratings on the Health-Sickness Rating 244 Scale and (2) readiness to experience a helping relationship (evident as early as the fifth session). Crits-Cristoph, Cooper, and Luborsky (1988) assessed the relationship of accuracy of interpretation to psychotherapy outcome within a core conflictual relationship theme (CCRT) (Luborsky, 1984) paradigm. While the "J contribution of a positive helping alliance was not VT: wan—f correlated with outcome, the interpretive accuracy was predictive of outcome. Correctly interpreting the patient's stereotyped patterns of needs and wishes, followed by addressing the response of others was found to be helpful, while limiting interpretations to addressing the patient's affective states and usual responses was not associated with outcome. Silberschatz, Fretter and Curtis (1986) examined the degree to which the suitability of interpretations, more than category of interpretation (i.e., transference versus non-transference), would be predictive of patient process and outcome in psychotherapy. With a small sample he developed a plan compatibility schema for each case. This consisted of developing a dynamic formulation based upon the patient's treatment goals, resistances, hypothesized ways that the patient would test the therapist, and insights which would be helpful to the patient. Their findings indicated that the suitability of the therapist's interpretations was more highly correlated than type of 245 interpretation with both process and outcome measures. In an attempt to address some of the questions raised by the findings of the Menninger Psychotherapy Research Project (wherein some data was suggestive of the importance of an interpretive approach [Kernberg, Burstein, Coyne, Appelbaum, Horwitz, & Voth, 1972], and some a supportive approach [Horwitz, 1974]). Gabbard, Horwitz, Frieswyk, Allen, Colson, Newsom and Coyne (1988) did an extensive inquiry into a psychoanalytic single case design with a borderline patient. They examined the shifts in patient collaboration (the patient's contribution to the therapeutic alliance) which occurred in response to the therapist's intervention, primarily in terms of transference interpretations. They found, in a sample of six sessions over the course of therapy, that the shifts in collaboration occurred in about 85% of the cases in relation to interpretations. This suggests that even with severely disturbed patients (i.e., borderlines), their is a connection between therapist's interventions and an increase in collaboration. Confrontation and Reflection. In contrast to interpretation, confrontation attempts to facilitate a directly meaningful experience--whether it is a realization of something that was defended against, or an understanding of the meaning or consequences of a behavioral pattern, 246 etc.--in reaction to a therapist's intervention. Orlinsky and Howard (1986) emphasize that it is the direct experiential quality of a confrontation that is definitive. Although few studies have been done isolating this particular type of intervention, some authors suggest that there is strong initial support for the potential for confrontation (Orlinsky and Howard, 1986). The use of reflection, a technique summarizing the meaning and/or affective quality of a patient's communication, has rarely been researched as a specific variable or technique alone, however, Orlinsky and Howard (1986) review seven studies which examine reflection. The general findings indicate that reflection as a specific technique does not appear to be particularly helpful nor detrimental and is probably best considered as an aspect of more complex interventions. Therapeutic Technique and Diagnostic Severity: Supportive versuspEypressive Psychotherapy. A brief consideration of how therapist technical factors might interact with the severity of the patient's diagnosis--in particular, psychostructural factors--should be made here before concluding this section. Clinical wisdom has repeatedly contended that more disturbed patients require modifications in therapeutic technique; such modifications typically suggest an emphasis on supportive psychotherapy 247 and a relative decrease in the interventions associated with expressive psychotherapy. Given the generality of the distinction, there is no real consensus as to the description of these two approaches, but the following seems to capture some of the essential differences: supportive psychotherapy relies upon the develOpment of a largely unanalyzed positive transference which seeks to maintain or enhance ego-functioning via stabilizing defenses in order to make the repression of the salient conflicts more complete; expressive psychotherapy relies upon the analysis of defenses (resistance and transference) leading to the uncovering of major conflicts and resolution, through interpretation, eventuating in reintegration and structural change. The following studies were based upon an examination of the expressive versus supportive dimension. Horowitz, Marmar, Weiss, Dewitt, and Rosenbaum (1984) examined a number of patient pretreatment variables in a bereaved outpatient population receiving brief psychotherapy. They found an interaction of patient predispositional factors-—motivation for treatment and developmental level of self-concept (a measure of object relations)-—and therapist intervention type. The general interactions observed were: (1) patient's who were higher functioning (i.e., greater motivation and more highly organized) were more responsive to exploratory and interpretive actions of the therapist, while such actions 248 were less suitable for lower functioning patients; and (2) patients who functioned at lower levels received more benefit from supportive psychotherapist actions than was the case for higher functioning patients. One of the concluding comments of the authors addressed the complexity of the relationship between technique and patient factors: The single clearest implication of our results is that future efforts to predict outcome of psychotherapy from process or dispositional variables appear unlikely to succeed unless they grapple with 1mutua1lyinteractingOinfluences on outcome. (p. 447) Jones, Cummings, and Horowitz (1988), in an effort to address some of the issues raised by the non—specific hypothesis of therapeutic change, developed an 100 item Psychotherapy Process Q-Sort which examined specific aspects of the psychotherapeutic process. In brief, the results indicated a relationship between specific factors and psychotherapy outcome. Jones et a1. (1988) found an interaction between level of patient pre-treatment disturbance and these specific factors. Further, the distinction between supportive approaches with more disturbed patients and expressive treatment techniques with less disturbed patients clearly emerged in the investigation of successful psychotherapy process. For example, therapist 249 factors such as "T (therapist) gives explicit advice and guidance", "T is directly reassuring“, "T acts to strengthen defenses", and “T behaves in teacher-like (didactic) manner”, etc., were all associated with more successful outcomes with more severely disturbed patients. In contrast, therapist factors connected with successful outcomes in less initially disturbed patients included such Psychotherapy Process Q—Sort items as "T remarks upon specific features of the interaction“, ”silences occur during the hour", “T emphasizes patient feelings in order to help her experience them more deeply“, "T draws connections between the therapeutic relationship and other relationships", etc. The authors conclude that (a) the conceptualization of the non-specific hypothesis has perhaps served to ”obscure more hypotheses about how patient change may be set in motion" (p. 55), (b) that more complex, interactional models of therapeutic change are needed, and (c) "that specific factors were indeed predictive of outcome, though usually in interaction patient pretreatment disturbance level“ (p. 48). Wallerstein (1989) reviewed follow-up studies on 42 subjects from the original Menninger Psychotherapy Project which compared patients in psychoanalysis and other forms of expressive and supportive psycho-analytically oriented psychotherapy. In the most recent treatment of the data, Wallerstein considers the change mechanisms of the 250 psychotherapy process. Briefly, the analysis of the data indicated that psychoanalysis was often less effective than had been predicted while the other psychotherapies were often more effective. Further, through the examination of the change processes involved in psychotherapy through the comparison of expressive (with its emphasis on interpretation) and supportive (with its emphasis upon the unanalyzed positive transference), the findings demonstrated that "supportive mechanisms infiltrated all therapies, psychoanalysis included, and accounted for more of the achieved outcomes (including structural changes) than anticipated" (p. 195). This contradicts some of the fundamental assertions about these two forms of psychotherapy. With regard to expressive psychotherapy, there was the finding that it included significant supportive procedures; and the finding connected with supportive psychotherapy suggested that patients achieved structural change without directly addressing underlying conflicts. This seems to indicate that the distinctions between expressive and supportive psychotherapy may be more apparent than real, given the overlap of technique, and the actual change mechanisms (i.e., interpretation -> conflict resolution -> structural change) are less specific and more subtle than current theory and practice had suggested. To provide a thorough explanation of the meaning of the apparently contradictory findings regarding the validity of 251 the supportive/expressive distinction as it interacts with patient diagnostic severity is beyond the scope of this overview. However, it is worth noting that another dimension of factors appeared to interact with the above factors. These factors include--intensity and duration of therapy, and clinical expertise of the therapist. Therapist Negative Factors and Patient Resistance. This section will review some of the factors associated with the case of the patient who does not improve or gets worse in psychotherapy. This brief overview will consider the general phenomenon, therapist errors and countertransference, and patient contributions. Bergin (1971) coined the term "deterioration effects“ to account for the observation that a portion of patients get worse during the course of psychotherapy. Hadley and Strupp (1976), in their review of negative effects in psychotherapy, state the problem directly: Change in psychotherapy has often been associated with change for the better...Yet, if a treatment ...produces beneficial results, it must also be capable of producing harmful effects. To deny this possibility is to accept the proposition that psychotherapy is not very potent under any circumstances. (p. 1291) 252 Such a phenomenon is difficult to experimentally produce and there are, most likely, a number of extratherapeutic causes and contributions (e.g., Thoits, 1985) to negative effects in psychotherapy (which is beyond the scope of the present discussion). In any case, there is some empirical evidence, and a strong clinical literature supporting the existence of therapy and therapist-induced negative effects. Meta—analytic studies have provided some general parameters of the occurrence of negative effects. Smith, Glass, and Miller (1980) found 9 percent of the overall effect size of their extensive meta-analytic study was attributable to negative effects. Shapiro and Shapiro (1982a) indicated that about 11 percent of the effect sizes calculated in their meta-analysis were negative (and another 30 percent was null). Orlinsky and Howard (1980) noted a 7 percent rate of negative changes in a psychotherapy outcome study (and that a small subgroup of therapists were over- represented in producing negative effects). There is a substantial clinical literature which supports the existence of, as well as suggesting the etiological agents of, negative effects in psychotherapy. One of the major clinical investigations into this area was done by Hadley and Strupp (1976). In a systematic exploration of some of the major questions associated with negative effects, these researchers sent questionnaires to 253 150 “experts" and received responses from approximately half of those polled. The general questions that they asked were: Is there a problem of negative effects in psychotherapy? What constitutes a negative effect and by what criteria can this be inferred in outcome? and, What are the major factors associated with negative effects? The responses were illuminating, varied and insightful, and the following represents a selective overview of some of the salient findings with an emphasis on therapist contributions to negative effects. Hadley and Strupp (1976) noted four constituent aspects of a negative effect: (a) exacerbation of old, or development of new, symptoms; (b) the patient's misuse or abuse of therapy; (c) undertaking unrealistic goals; and (d) loss of trust in therapy or the therapist. Many of the respondents emphasized inaccurate or deficient assessment as a critical factor which created a foundation susceptible to subsequent therapeutic problems. Patient factors which contributed to negative effects included low motivation, limited ego resources and a masochistic character style. Two broad therapist factors were identified, the first of which was deficiencies in training and the second was therapist personality factors and technical style. There were a number of significant therapist factors which were subsumed under "errors in technique“, by and large the most important area discussed, these included: (a) 254 false assumptions concerning the scope and potency of psychotherapy, including an omniscient or cure-all stance; (b) problems with therapeutic goals--too abstract or unspecified, lack of patient and therapist agreement, exceeding the patient's capabilities; (c) misplaced focus of therapy--too much or too little emphasis upon the patient's inner world, sometimes at the expense of considering environmental factors; (d) mismatch of technique to the patient--in particular with more disturbed or fragile patients; (e) technical rigidity; (f) overly intense psychotherapy--becoming too close, giving advice, pushing the patient; (g) misuse of interpretations--wrong focus, transference interpretations to the exclusion of all else, destructive or ill-timed interpretations; and (h) dependency fostering techniques—-inadequate understanding and interventions with inevitable dependency issues. Hadley and Strupp (1976) also outline some of the problems in the patient-therapist relationship that contribute to negative effects. In addition to problems in rapport, they discussed the concept of an interactional neurosis (Langs, 1982) between therapist and patient, wherein transference and counter-transference issues have become entangled. Luborsky (1984) believes that complications may arise when the therapist fails to recognize and adequately deal with the patient's conflictual relationship themes. Various countertransferential problems 255 may arise, such as the therapist not respecting the patient's autonomy, becoming hostile or disappointed in the patient, etc. Communication problems can occur and can be a function of the therapist's limitations in articulating his or her interventions, or in the patient's difficulties and/or distortions in communication. Hadley and Strupp (1976) conclude their study with: "It is clear that negative effects of psychotherapy are overwhelmingly regarded by experts in the field as a significant problem requiring attention and concern of practitioners and researchers alike (p. 1302)“. Finell (1987) provides an extensive and thoughtful review of the psychoanalytic literature on the negative therapeutic reaction. She notes that the negative therapeutic reaction is better understood as a multidimensional concept. Definitions of negative therapeutic reaction include both a restricted version--a paradoxical worsening in response to an ostensibly effective technique, and a broader definition-—that the patient is, for some reason, resistant to recovery and/or the analysis has become interminable. These definitions tie in with the - two general conceptualizations offered to explain the negative therapeutic reaction. One explanation is that the negative therapeutic reaction is a result of the patient's psychopathology, encompassing such aspects as narcissistic conflicts and envy. The other major conceptualization 256 emphasizes the interpersonal aspects of the negative therapeutic reaction. The critical feature here is the interaction of transference and countertransference between the patient and therapist by means of projective and introjective processes. There is a broad range of intrapsychic factors which I have been cited as contributing to the negative therapeutic g reaction including envy, entrenched negativism, anal defiance, undoing, guilt, persecutory anxiety, and narcissistic greed. Freud (1918), in his analysis of the negative therapeutic reaction observed in the Wolf Man, initially believed this phenomenon to be a matter of anal defiance or need for control. Later, he considered that the patient had begun to associate recovery with some form of danger (conscious or unconscious) and speculated that the patientlhad a need to continue to be ill. This he linked with an underlying guilt which required the patient to maintain a certain level of suffering. As mentioned, there are a number of other intrapsychic explanations, but most of them can be subsumed into envy or narcissistic issues. While there are differing versions, most notions of envy arise from Klein's (1957) elaboration of infantile dynamics involved in envy. This conceptualization emphasizes the infant's aggressive reaction to positive or nurturing supplies coming from the mother, its inability to respond with gratitude and subsequent guilty self-punishment over 257 aggression. Overlapping with the concept of envy, explanations emphasizing narcissistic conflicts see as important the denial of infantile dependency needs and a subsequent flight into omnipotence, as well as an attack against the therapist whose work had increased the patient's awareness of infantile needs (e.g., Rosenfeld, 1975). The various interpersonal explanations emphasize the conflict over wishes for fusion and the need for separateness as it is exhibited in the interactional matrix of transference and countertransference. Freud (1923), in some of his later thinking, offered an interpersonal explanation of the negative therapeutic reaction when he suggested that it was due to a continuing identification with a lost love object, as in the case of melancholia. In a related vein, the term resistance, a concept central to psychoanalytic and psychodynamic theory, is frequently offered as one source of explanation of negative therapeutic process and outcome. A critical tenet in psychoanalytic theory is that the patient's resistance subsides in response to the therapist's interpretive activities and that this process produces insight into the patient's problems and facilitates therapeutic change and progress. Thus negative effects could occur when the patient's resistance was too great or too entrenched, or when the therapists interpretive activities were insufficient or misguided. However, this explanation 258 implies that resistance is a unitary and strictly patient- produced phenomenon. Concerning the later point, Meisels (1988) has argued that "Resistance...is not an event...but it is...a particular attitude toward an event. Resistance is in the eye of the therapist, and is not an event in the patient's repertoire (p. 4). The (classical) idea of resistance fails to include the contribution of the analyst (p. 10)." Meisels goes on to state that the concept of resistance must take into account the therapist's ] contribution, since often it can be a matter of the patient's reaction to an inaccurate interpretation (e.g. Crits-Cristoph, Cooper, & Luborsky, 1988) or to the therapist's countertransference. With this qualification in mind, the therapist's contribution and interactional aspects will be considered shortly; for now, the discussion will be concerned with the patient's contribution. The concept of resistance has developed and expanded from its more limited and subordinate original role (i.e., an occasional and specific impediment to the process) to one which potentially encompasses all patient activity in the therapeutic process. Freud's pithy and somewhat circular definition of resistance is still appropriate-—"whatever interrupts the progress of analytic work is a resistance“ (1900, p. 517). Freud's understanding of this phenomenon eventually shifted to a view of resistance as unavoidable and necessary to the treatment process. Freud (1926) discussed various origins 259 and aspects of resistance, noting such factors as keeping material out of consciousness, or away from the analyst; using the resistance for secondary gain; also resistance could arise from the id (repetition compulsion), or from the superego (need for punishment). Subscribing to the notion that resistance is a complex and multidimensional concept, Blatt and Erlich (1982) have discussed three levels of resistance. The first type of resistance is episodic and is described as a limited, situational and usually transient interruption of the therapeutic process. Transference resistance is characterized as a more pervasive and unconsciously determined mode of relating and, as such, is resistant to more mature and realistic modes of interpersonal interaction. The third level of resistance is referred to as a fundamental resistance to change and can be considered to be, when minimally present, an inevitable undercurrent in all therapies and, in more pronounced levels, a major disruptive force sometimes leading to an overall therapeutic impasse, or negative therapeutic reaction. The centrality of resistance to the psychotherapeutic process is illustrated when Schuller, Grits-Cristoph, and Connolly (1988) state that: The entire process of psychoanalytic therapy can be seen as organized around the resolution of these forces which are responsible for dis- ruptions occurring at various levels. (p. 9) .F 260 Unfortunately, there have been only a few empirical studies which have examined this phenomenon in any detail; those that have note contradictory findings. One source of difficulty is the imprecision in defining resistance. Schuller, Crits-Cristoph and Connolly (1988) review the very limited research literature and conclude that resistance is not a unitary phenomenon. There appear to be various types and amounts of resistance in both conscious and unconscious forms. In the Schuller et al. (1988) study, episodic resistance was investigated by examining the therapies of the ten most and ten least improved cases, in a paradigm based upon Garduk and Haggard's (1972) research. Schuller et al. (1988) developed a 19 item Resistance Scale, developed from their review of the literature. Their findings support the notion of resistance as a multidimensional construct. The inter-judge reliabilities were relatively low (especially with regard to items which required greater clinical inference). Statistical analyses revealed four relatively separate subtypes of resistance: (a) abrupt/shifting, (b) oppositional, (c) flat/halting, and (d) vague/doubting. Another finding was the high degree of variance from segment to segment, as well as significant individual differences, supporting an episodic perspective of resistance. The apparent episodic nature of resistance, along with the relatively independent dimensions of 3"“ 261 resistance, may help to explain some of the findings of session to sessionfluctuations, and individual differences in various process factors, especially in the therapeutic alliance research literature (for example, the four identified dimensions of resistance may correspond to different diagnostic or characterological defensive styles). The development of the Vanderbilt Negative Indicators Scale (VNIS) grew out the general psychotherapy process research at Vanderbilt and was an attempt to investigate the problem of negative effects in psychotherapy. Based upon Hadley and Strupp's (1976) survey of clinician's thoughts on negative effects, Gomes-Schwartz (1978b) constructed the first version of the VNIS. The current version of the VNIS has 42 items grouped into five subscales: patient qualities, therapist personal qualities, errors in technique, patient- therapist interaction, and global factors. Rating of the scale is a two step decision process beginning with the rater's determining if a given characteristic is present or not, then rating the frequency or intensity on a five point scale. The VNIS has had at least three empirical applications with clinical populations (Eaton, 1987; Sachs, 1983; Strupp, Keithly, Moras, Samples, Sandell, & Waterhouse, 1980). The earliest study (Strupp et al., 1980) found overall VNIS ratings, but none of the subscale ratings, to discriminate between high and low outcome groups. The Eaton (1987) study 262 had some difficulties with reliabilities, especially with regard to the errors in technique subscale, but found an overall relationship between negative indicators and the therapeutic alliance. Several findings were unique to this study, including (a) the observation that it took some time for negative indicators to emerge, (b) that therapist's positive contributions had more of an impact than negative indicators, and (c) that the VNIS was only correlated with one of ten patient symptom dimensions at outcome. Suh, Strupp, and O'Malley (1986) note that the VNIS has had difficulties with inter—rater reliabilities and they attribute this in large part to the level of sophistication needed to rate the items. They believe that it requires value judgments based upon clinical expertise and a psychodynamically-informed perspective. Sachs (1983) used audiotaped excerpts from the first three sessions of 18 patients seen in brief psychodynamic or experiential therapy. Twenty-five (of 42) items were deemed reliable and found to be correlated with outcome measures, showing significant correlations with measures from the therapist's and independent clinician's, but not from the patient's, perspective. Correlations between the VNIS subscales and therapeutic outcome indicated that errors in technique demonstrated the strongest correlation, followed by global session ratings, and patient-therapist interactions; patient qualities and therapist personal qualities were 263 nonsignificant. Further analyses showed that these relationships applied only to psychodynamic approaches and not the experiential approaches. Sach's examination of the particular items of the subscales, in particular the errors in technique subscale, allows for an elaboration of detail and infers some possible mechanisms of negative effects. Some of the salient items included: (a) the therapist's passive acceptance of problematic aspects of the client's attitude or behavior (e.g., resistance, oppositionalism, etc.); (b) unaddressed deficiencies in the therapeutic relationship; and (c) presentations of poorly timed or destructive interpretations (p. 562). Thus, empirical support for the primary importance of therapists' specific negative therapeutic actions--whether they are errors of commission or omission--in the occurrence of negative therapeutic outcome was clearly supported. Foreman and Marmar (1985) attempted to identify which therapist behaviors were differentially present in those cases where the initially negative alliance shifted to a more viable working relationship and eventuated in a positive outcome (improved) versus those cases where the alliance remained poor and the outcome was negative 264 (unimproved). In reviewing the psychoanalytic literature concerning negative therapeutic effects, they concluded that the negative alliance must be confronted and various misperceptions, destructive transferences, and pervasive defenses--particularly with regard to major relationships in the clients life--must be dealt with early on in order to avert a misalliance (Langs, 1982) which is based upon avoidance and/or passivity toward these therapeutic obstacles. After a review of cases which represented these two populations (i.e., improved--poor alliance that developed into a more positive alliance and had a positive treatment outcome versus unimproved--poor alliance that did not turn around) the authors suggested a number of therapist actions that would address the negative therapeutic alliance. While a number interventions were outlined, the following is a list of those that were found to be differentially present in the improved versus unimproved group: (a) the therapist addressed the defenses the client used to avoid dealing with feelings in relation to the therapist. This was the most predictive variable of their study. (b) From Malan (1976), the "triangle of punishment" was more directly and completely addressed in improved cases. This deals with the cycle of client anger/responsibility over another's suffering leading to guilt leading to self-punishment; (c) In unimproved cases, an“ 265 the therapist did not address or avoided—-despite knowledge of the presence of--the clients problematic feelings toward the therapist. Specifically, there was an emphasis on manifest problems, external events and past relationships. In contrast, these issues were addressed and important linkages were established in the improved cases. Further studies which consider the difficult patient and therapist's interventions have been conducted by the Langley Porter group (discussed in an earlier section; e.g., Gaston, 1990). In summary, the concept of negative effects--with therapist, patient and interactional contributions—-is a well-documented phenomenon which appears to permeate many central aspects of the psychotherapy process. Preliminary research suggests that early identification and an active confrontation of patient resistances and therapist errors in technique may help to transform an initially negative therapeutic alliance into a more productive one. Therapist Relational Factorg This last section deals with inferred internal therapy- specific characteristics. This includes various attitudinal aspects of the therapist's behavior which convey certain values concerning the therapeutic relationship, social influence attributes and therapist's treatment expectations. These factors have been among the most productively 266 researched. Such inferred therapist characteristics are often subsumed under the heading of non-specific factors (e.g., Frank, 1973). That is, they tend to be present in most successful psychotherapies while not particularly affiliated with a distinctive approach or technique. Social Influence Attributes. A number of variables from the social psychology research literature have been investigated with regard to how therapists engage the patient in the psychotherapy process. If one starts with the premise that a major variable responsible for therapeutic motivation is cognitive dissonance and that the patient and therapist begin psychotherapy from two different viewpoints, a fundamental task for the therapist is to resolve this discrepancy in the direction of the therapist. This can occur only to the degree that the therapist has power. In the therapeutic context, power is associated with therapist source characteristics--expertness, trustworthiness and attractiveness (Strong, 1968). While analogue experiments have been supportive of this line of inquiry, clinical studies deserve greater focus because of their intrinsic applicability toward the understanding of the influence of therapist source variables. Beutler, Johnson, Neville, Elkins, and Jobe (1975) found, in a study of attitudinal similarities, a relationship between therapist credibility and patient therapeutic gain. 267 LaCrosse (1980) found that, in a diverse selection of psychotherapists, therapist source characteristics were significantly correlated with outcome. Specifically, up to 35% of the variance was associated with attractiveness, expertness, and trustworthiness. Overall, the research has been supportive of therapist expertness being the most influential of the three source characteristics (Beutler, Crago, & Arizmendi (1986); and these factors have been associated more with the patient's initial impressions and the early phase of therapy (Horvath and Greenberg, 1986). Therapist Treatment Expectations. Owing to the fact that therapist expectation is complex variable interacting with patient variables, process variables, and pre-therapy variables (e.g., patient selection), the research literature is inconsistent and divided as to the conclusions drawn concerning the relationship between therapist expectations and treatment outcome. A substantive case has been made by reviewers for app against the influence of this variable on treatment outcome. In support of the relationship between therapist positive expectations and treatment outcome, Goldstein and Simonson (1971) concluded that there is a significant facilitative relationship. LaTorre (1977) states that is it the congruence between therapist and patient expectations which is the critical facilitative factor. Correspondingly, Baekeland and Lundwall (1975) 268 assert that dissimilar therapist and patient expectations, as well as the therapist's failure to affirm the patient's expectations, are associated with early and higher patient drop out rates. In contrast, Lambert and Bergin (1983) conclude that there is insufficient support for the position that therapist expectations play a direct role in treatment outcome. Luborsky, Crits-Cristoph, Mintz, Auerbach (1988), in their review, noted that the limited number of studies and absence of a clear trend in the findings did not allow for a definite conclusion at this point. Other reviewers (e.g., Duckro, Beal, & George, 1979) have also reached similar conclusions and challenged the validity of therapist expectations and outcome. These inconsistent findings are attributable, in general, to methodological concerns and in particular to the specific dimension of expectations which is measured (i.e., patient gain, intra-therapy behavior, symptom decrease or length of treatment). Beutler, Crago and Arizmendi (1986) in their review, note that patient role expectations account for a larger portion of the variance than therapist expectations. In particular, they see a relationship between outcome and patient expectations when the expectations are in the direction of generally held psychological values. There is support that therapist's expectations exert a stronger influence when they are based on direct observations of patient behavior. That is, when therapist expectations arise from accurate information, are 269 flexible, and open to reevaluation over the course of treatment, improvement is enhanced (Beutler et al., 1986). Facilitative Conditions. Rogers (1957) first conceptualized the basic facilitative therapist attitudes as the necessary and sufficient conditions of effective psychotherapy. These attitudes include accurate empathy, warmth, congruence, and unconditional positive regard. While these tenets have been almost universally endorsed, there is some controversy as to the sufficiency, as well as the origin and maintenance of these attributes. For example, Gurman (1977) believes that relationship attitudes are more tied in with the patient's, as opposed to the therapist's, behavior. Furthermore, the relationship between these conditions and therapeutic outcome have tended to be misrepresented by rater bias and/or halo effects due to the perspective from which facilitative conditions are measured (i.e., only the therapist's or only the patient's). Overall, there is a lack of consistency in the research concerning the relationship between therapist facilitative conditions (or as it is also designated--therapist-offered relationship) and therapeutic outcome. Specifically, ratings taken from external observers and therapists have been equivocal. Rogers (1957) had noted that one of the critical factors with regard to these variables was the patient's perception of the therapist-offered relationship. 270 Gurman (1977) found that a majority of the 26 studies reviewed showed a relationship between the patient's perception of therapist-offered relationship and therapeutic outcome. Many of the studies have used the Barrett-Lennard Relationship Inventory (Barrett-Lennard, 1962), an instrument which assesses, via questionnaire, the patient's experience of the therapist. General findings are that patient perception of therapist-offered empathy was the most stable and significant variable of the facilitative conditions. The other three variables were more affected by various factors (including susceptibility to different therapeutic approaches; Gurman, 1977). Orlinsky and Howard (1986) reviewed 40 studies which specifically examined therapist empathy and concluded that about half of the studies showed a clear relationship to therapeutic outcome when measured from the patient's perspective. Orlinsky and Howard (1986) reviewed studies examining the relationship of therapist warmth/acceptance to therapeutic outcome and found that, again, about half (of 47 studies), across a variety of outcome measures, were positively correlated with outcome. The largest proportion of positive findings was associated with measures taken from the patient's perspective. Concerning therapist genuineness, Orlinsky and Howard (1986) analyzed 28 studies and found a mild but inconsistent positive relationship to outcome (up to 38%). Genuineness seemed less potent than 271 the other facilitative conditions. Genuineness has been more robust when it is "refined“ into therapist engagement and credibility. About half of the 18 studies examining how invested the therapist is in his or her role, that is, therapist engagement versus detachment indicated a positive relationship between engagement and patient outcome I (Orlinsky and Howard, 1986). While there was a positive relationship between objective observers' ratings and ‘ outcome, the strongest relationship remained between ratings from the patient's perspective and treatment outcome. In a similar vein, examination of the level of therapist confidence and credibility versus unsureness indicated that approximately 60% (of the 8 studies reviewed by Orlinsky and Howard, 1986) showed a relationship between therapist credibility and psychotherapy outcome; this trend was preserved across all three process and outcome perspectives (i.e., therapist, patient and objective external judge). Beutler, Crago, and Arizmendi (1986) conclude that, despite some of the methodological problems associated with the research: Nonetheless, one can tentatively conclude that patients' positive perceptions of therapist facilitative attitudes has a modest tendency to enhance treatment gains. (p. 279) 272 The emerging understanding that there is a fundamental interaction between (a) therapist-offered and patient- perceived facilitative conditions or relationship, and (b) between therapist and patient contributions to the relationship, has been discussed in the preceding review of the literature. In summary, there appears to be a confluence of research and clinical findings, originating from diverse sources, which lead us to increasingly consider the centrality of the context of the therapeutic alliance in our efforts to understand the psychotherapeutic process. APPENDIX B : California Psychotherapy Alliance Scales (CALPAS) CALIFORNIA PSYCHOTHERAPY ALLIANCE SCALES (CALPAS) Each item is rated on a seven point scale as follows: 1 = Not at all 2 = A little bit 3 = Somewhat 4 = Moderately 5 = Quite a bit 6 = Quite a lot 7 = Very much so I. _atient Working Capacity-Positive Contribution (PWC) Patient self-discloses thoughts and feelings Patient self-observes behaviors Patient explores own contribution to problems Patient experiences strong emotions Patient works actively with therapist’s comments Patient deepens exploration of salient themes 9509939!" 11. Patient Hostile Resistance (PHR) 7 Patient conveys an expectation of easy cure without work on his part 8. Patient acts in hostile, attacking and critical manner toward therapist 9 Patient seems mistrustful and suspicious of therapist 10. Patient engages in power struggle, attempting to control the session 11. Patient defies therapist’s efforts to promote self-understanding 12. Patient holds therapist at arm’s length with flood of words 273 III. Mient Commitment (PTC) 13. 14. 15. 16. 17. 18. Patient is confident that efforts will lead to change Patient is willing to make sacrifices, for example time and money Patient views the therapy as important Patient has confidence in therapy and therapist Patient participates in therapy despite painful moments Patient is committed to go through process to completion IV. Working Strategv Consensus (WSQ 19. 20. 21. 22. 23. 24. Therapy proceeds in accord with patient’s ideas of helpful change process Patient and therapist work together in joint struggle Patient and therapist agree about the kind of changes to make Patient and therapist share same sense about how to proceed Patient and therapist agree on salient themes Therapist rigidly applies technique V. Therapist Understanding and Involvement (TUI) 25. 26. 27. 28. 29. 30. Therapist is understanding of patient’s suffering and subjective world Therapist demonstrates non-judgmental acceptance and positive regard Therapist demonstrates commitment to help and confidence in treatment Therapist does not misuse treatment to serve own needs Therapist demonstrates tact and timing of interventions Therapist facilitates work on salient themes 274 lb. "‘5‘. APPENDIX C : Object Relations Scale 10R Scale) OBJECT RELATIONS SCALE (OR Scale) I. Complexity of Representations of People Principle: Scale measures the extent to which the subject clearly differentiates the self from others; sees the self and others as having stable, enduring, multidimensional dispositions; and sees the self and others as psychological beings with complex motives and subjective experiences. Level 1: The person does not see others as clearly differentiated or bounded, and/or does not differentiate his/her own thoughts and feelings from those of others. Level 2: The person sees others as clearly bounded, separate from self and from other, but without a sense of people’s subjective states, motives or enduring characteristics. The focus is on behaviors and momentary actions. People are seen as unidimensional, existing i_n situations rather than across situations. Where people are understood as having enduring qualities, these are global, evaluative traits like "nice” or "mean". The person conveys no sense of psychological awareness. Level 3: The person begins to show some evidence of psychological awareness, and begins to make inferences about subjective states in addition to focusing on behavior. Understanding on one’s own and others’ psychological processes and experience does not, however, delve far beneath the surface. The person has ideas or "theories" about self’s and others’ enduring characteristics, but these intuitive theories are either unidimensional, overly general, stereotypic or lacking in subtlety. There is little sense that people could do things out of character or experience psychological conflicts. Descriptions may be based upon schemas of how people in general are, rather than on understanding of the particular self’s or others’ mental states, personal meanings or enduring attributes. Level 4: The person has a developing appreciation for the complexity of the subjective states of others. Self and others may be described in an elaborated, rather than simple and relatively unidimensional fashion. Level 5: The person has a recognition of the complexity of the subjective states of self and others, and has a relatively differentiated view of personality dispositions. However, the person has difficulty integrating complex representations of subjective states with complex representations of personality. Component parts of personality are not yet understood as aspects of an interacting system, in which enduring dispositions can come into conflict. The person has some awareness of unconscious processes, but these are not yet integrated with enduring attributes and personal history. Level 6: The person sees others in fairly complex ways. The person is able to reflect on his/her own representations of self and others, is able to depict relatively non- complex personality change over time, and may convey a rudimentary sense of the 275 history (familial, experiential) roots of traits, feeling state and psychological processes. Level 7: The person can simultaneously coordinate and integrate complex, generalized representations of self and others with complex representations of specific mental states. He/she has an integrated understanding of complex psychological processes within the context of the self’s and others’ personal histories and attributes. The person can sees and convey how people are particularly and individual within themselves, while simultaneously different form others, and can see how personal meanings of long standing affect current psychological experience. The person is able to make elaborate inferences about self’s and others’ mental states, motivations, points of view, unconscious processes, and conflicting feelings and impulses. II. Affect-Tone of Relationship Paradigms Principle: Scale measures the affective quality of representations of people and relationships. It attempts to assess the extent to which the person expects from the world, and particularly the world of people, profound malevolence or overwhelming pain, or views social interaction as basically benign and enriching. Level 1: Unambiguously malevolent or overwhelmingly painful; grossly negligent caretakers or significant others. Level 2: Predominantly hostile but not overwhelming; empty; disappointment or loneliness, but not profound; negligence and indifference. Level 3: Mixture of representations, with an overall negative or neutral tone; mildly negative tone. Level 4: Mixture of representations with an overall positive tone; positive tone without a clear sense of enjoyment of the other person as opposed to the situation. Level 5: Predominantly positive tone; sense of enjoyment of the relationship rather than primarily the situation rather than the activity; clear expectations that relationships are benign and pleasurable. III. Capacity for Emotional Investment in Relationships and Morals Principle: Scale measures the extent to which others as treated as ends rather than means, events are regarded in other than need gratification, moral standards are developed and considered, and relationships are experienced as meaningful and committed. Level 1: The predominant concern is self-gratification. Others’ perspectives, needs and desires are not considered, and the impact of one’s actions upon others is frequently not taken into account. People are seen as existing only in relation to oneself: they are treated as tools for the achievement of one’s desires; as mirrors or audiences for one’s displays; or as impediments to one’s gratification. People may be seen as useful or comforting at the moment but are not emotionally invested in for 276 their unique characteristics. Rules and authorities are seen as obstacles unless momentarily useful. Level 2: The person recognizes that there exist differences between the needs and desires of self and others, though the primary aim remains the satisfaction of one’s own wishes. Friendships develop but in many respects remain interchangeable. Relationships consist primarily of shared activities. Attachments are important but the needs of the other are typically experienced as secondary to the needs of the self, particularly when these conflict. There is a rudimentary sense of right and wrong characterized by an equation of prudence and morality (i.e., bad actions are bad because they lead to punishment). Moral injunctions that exist are frequently primitive and harsh. Level 3: The person considers the needs and wishes of significant others in making decisions. Pleasing others, being liked and behaving in accordance with the standards of respected authorities are salient which often override self-interest. Friendships and familial relationships are relatively conventional. The person is concerned with being good and experiences guilt when his/her thoughts, or actions conflict with internalized standards. Rules are respected because they are rules; manners and conventions are seen as important and even natural. Moral rules tend to be relatively rigid and concrete, and there may be a pronounced sense of duty, particularly to certain people. Level 4: The person is capable of forming deep, committed relationships in which the other is valued for his/her unique qualities. Commitment to others often overrides personal desires, but actions on behalf of another are undertaken without a sense of duty or a predominant desire to be liked for one’s good deeds. Moral judgments, values and modes of conflict-resolution remain relatively conventional. The person is concerned with doing the right thing, as defined by society or respected authorities, which is frequently expressed in more abstract terms and is often self-abnegatory. Relationships are seen as lasting over time and involving considerable commitment and intimacy. Level Q: The person treats self and others as ends rather than means. The person is interested in the development and happiness of both self and others, and attempts to achieve autonomous selfhood within the context of real involvement with and investment in others. Conflicts between people with conflicting legitimate interests are understood as requiring compromise. Authorities and rules are not taken to be absolute; the person has a sense of the conventional nature of social rules and believes that at times these must be overridden or changed because they conflict with self- generated or carefully considered standards, or when they do significant harm to people in concrete circumstances. 277 95”.“?‘999’392‘ APPENDIX D : INTERPERSONAL CHECKLIST (ICL) Interpersonal Checklist (ICL) Instructions: Below is a list of descriptive words and phrases which you will use to describe yourself. Read the items quickly and fill in the circle with a 1 in it of each item you consider to be generally descriptive of yourself at the present time. Leave the answer blank when an item does not describe you. Your first impression is generally the best so work quickly and don’t be concerned about duplications, contradictions or being exact. If you feel much in doubt whether an item applies, leave it blank. Well thought of Makes a good impression Able to give orders Forceful Self- respect in g Independent Able to take care of self Can be indifferent to others Can be strict if necessary . Firm but just . Can be frank and honest . Critical of others . Can complain if necessary . Often gloomy . Able to doubt others . Frequently disappointed . Able to criticize self . Apologetic . Can be obedient . Usually gives in . Grateful . Admires and imitates others . Appreciative . Very anx. to be approved of . Cooperative . Eager to get along with others . Friendly . Affectionate and understand. . Considerate . Encourages others . Helpful . Big-hearted and unselfish . Often admired . Respected by others . Good leader . Likes responsibility . Self-confident . Self-reliant and assertive . Businesslike . Likes to compete with others . Hard-boiled when necessary . Stern but fair . Irritable . Straightforward and direct 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. Resents being bossed Skeptical Hard to impress Touchy and easily hurt Easily embarrassed Lacks self- confidence Easily led Modest Often helped by others Very respectful to authority Accepts advice readily Trusting and eager to please Always pleasant and agreeable Wants everyone to like them Sociable and neighborly Warm Kind and reassuring Tender and soft-hearted Enjoys taking care of others Gives freely of self Always giving advice Acts important Bossy Dominating Boastful Proud and self-satisfied Thinks only of him/herself Shrewd and calculating Impat. with others‘ mistakes Self-seeking Outspoken Often unfriendly Bitter Complaining Jealous Slow to forgive a wrong Self-punishing Shy Passive and unaggressive Meek Dependent Wants to be led Lets others make decisions Easily fooled 278 89. 90. 91. 92. 93. 94. 95. 96. 97 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. Too easily influ. by friends Will confide in anyone Fond of everyone Likes everybody Forgives anything Oversympathetic Generous to a fault Overprotective to others . Tries to be too successful Expects every. to admire them Manages others Dictatorial Somewhat snobbish Egotistical and snobbish Selfish Cold and unfeeling Sarcastic Cruel and unkind Frequently angry Hard-hearted Resentful Rebels against everything Stubborn Distrusts everybody Timid Always ashamed of self Obeys too willingly Spineless Hardly ever talks back Clinging vine Likes to be taken care of Will believe anyone Wants everyone's love Agrees with everyone Friendly all the time Loves everyone Too lenient with others Tries to comfort everyone Too willing to give to others Spoils pe0ple with kindness APPENDIX E : Symptom Checklist (SCL—90-R) SYMPTOM CH ECKLIST (SCL - 90 - R) INSTRUCTIONS: Below is a list of problems and complaints that people sometimes have. Read each one carefully, and select one of the numbered descriptors that best describes how much discomfort that problem has caused you during the past week, including today. Do not skip any items, and print clearly. If you change your mind, erase your first number completely. How much were distressed by ? Descriptors: 0 = Not at all 1 = A little bit 2 = Moderately 3 = Quite a bit 4 = Extremely Headaches Nervousness or shakiness inside Repeated unpleasant thoughts that won't leave your mind Faintness and dizziness Loss of sexual interest or pleasure Feeling critical of others The idea that someone else can control your thoughts Feeling others are to blame for most of your troubles Trouble remembering things 10. Worried about sloppiness or carelessness 11. Feeling easily annoyed or irritated 12. Pains in heart or chest 13. Feeling afraid in open spaces or on the streets 14. Feeling low in energy or slowed down 15. Thoughts of ending your life 16. Hearing voices that other people do not hear 17. Trembling 18. Feeling that most people cannot be trusted 19. Poor appetite 20. Crying easily 21. Feeling shy or uneasy with the opposite sex 22. Feelings of being trapped or caught 23. Suddenly scared for no reason 24. Temper outbursts that could not control 25. Feeling afraid to go out of your house alone 26. Blaming yourself for things 27. Pains in lower back 28. Feeling blocked in getting things done 29. Feeling lonely 30. Feeling blue P?°>‘.°‘5":"‘S*’S°!‘ 279 31. 32. 33. 35. 36. 37. 38. 39. 40. 41. 42. 43. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. Worrying too much about things Feeling no interest in things Feeling fearful Your feelings being hurt easily Other people being aware of your private thoughts Feeling that others do not understand you or are unsympathetic Feeling that people are unfriendly or dislike you Having to do things very slowly to insure correctness Heart pounding or racing Nausea or upset stomach Feeling inferior to others Soreness in your muscles Feeling that you are watched or talked about by others Trouble falling asleep Having to check and double-check what you are doing Difficulty making decisions Feeling afraid to travel on buses, subways, or trains Trouble getting your breath Hot or cold spells Having to avoid certain things, places, or activities because they frighten you Your mind going blank Numbness or tingling parts of your body A lump in your throat Feeling hopeless about the future Trouble concentrating Feeling weak in parts of your body Feeling tense or keyed up Heavy feelings in your arms or legs Thoughts of death and dying Overeating Feeling uneasy when people are watching or talking about you Having thoughts that are not your own Having urges to beat, injure or harm someone Awakening early morning Having to repeat the same actions such as touching, counting. washing Sleep that is restless or disturbed Having urges to break or smash things Having beliefs that others do not share Feeling very self-conscious with others Feeling uneasy in crowds, such as shopping or at a movie Feeling everything is an effort Spells of terror or panic Feeling uncomfortable about eating or drinking in public Getting into frequent arguments Feeling nervous when you are left alone Others not giving you proper achievement for your achievements Feeling lonely even when you are with people Feeling so restless that you couldn't sit still Feelings of worthlessness A feeling that something bad is going to happen to you Shouting or throwing things Feeling afraid that you will faint in public Feeling that people will let you take advantage of them if you let them Having thoughts about sex that bother you a lot The idea that you should be punished for your sins Thoughts and images of a frightening nature The idea that something serious is wrong with your body Never feeling close to another person Feelings of guilt The idea that something is wrong with your mind 280 APPENDIX E : Post Therapy Client Questionnaire (PQCL) POST-THERAPY CLIENT QUESTIONNAIRE (PQCL) For each item chose the answer which you feel best describes your therapy experience. 1. How much in need of further therapy do you feel now? 1 - No need at all 2 - Slight need 3 - Could use more 4 - Considerable need 5 - Very great need 2. What led to termination of your therapy? 1 - My decision 2 - Therapist’s decision 3 - Mutual agreement 4 - External factors 3. How much have you benefitted from your therapy? 1 - A great deal 2 - A fair amount 3 - To some extent 4 - Very little 5 - Not at all 4. Everything considered, how satisfied are you with the results of your psychotherapy experience? 1 - Extremely dissatisfied 2 - Moderately dissatisfied 3 - Fairly dissatisfied 4 - Fairly satisfied 5 - Moderately satisfied 6 - Highly satisfied 7 - Extremely satisfied 5. What impression did you have of your therapist’s level of eXperience? 1 - Extremely inexperienced 2 - Rather inexperienced 3 - Somewhat inexperienced 4 - Fairly experienced 5 - Highly experienced 6 - Exceptionally experienced 281 10. 11. 10. 11. How well did you feel your were getting along before psychotherapy? 1 - Very well 2 - Fairly well 3 - Neither well nor poorly 4 - Fairly poorly 5 - Very poorly 6 - Extremely poorly How long before entering therapy did you feel in need of professional help? 1 - Less than one year 2 - 1-2 years 3 - 3-4 years 4 — 5-10 years 5 - 11-15 years 6 - 16-20 years How severely disturbed did you consider yourself at the beginning of your therapy? 1 - Extremely disturbed 2 - Very much disturbed 3 - Moderately disturbed 4 - Somewhat disturbed 5 - Very slightly disturbed How much anxiety did you feel at the time you started therapy? 1 - A tremendous amount 2 - A great deal 3 - A fair amount 4 - Very little 5 - None at all How great was the internal "pressure" to do something about these problems when you entered psychotherapy? 1 - Extremely great 2 - Very great 3 - Fairly great 4 - Relatively small 5 - Very small 6 - Extremely small How much do you feel you have changed as a result of psychotherapy? ll - A great deal 2 - A fair amount 3 - Somewhat 4 - Very little 5 - Not at all 282 12. How much do you feel this change has been apparent to others? (a) People closest to you (husband, wife, etc.) 1 - A great deal 2 - A fair amount 3 - Somewhat 4 - Very little 5 - Not at all (b) Close friends 1 - A great deal 2 - A fair amount 3 - Somewhat 4 - Very little 5 - Not at all (c) Co-workers, acquaintances, etc. 1 - A great deal 2 - A fair amount 3 - Somewhat 4 - Very little 5 - Not at all 13. On the whole, how well do you feel you are getting along now? 1 - Extremely well 2 — Very well 3 - Fairly well 4 - Neither well nor poorly 5 - Fairly poorly 6 - Very poorly 7 - Extremely poorly 14. How adequately do you feel you are dealing with any present problems? 1 - Very adequately 2 - Fairly adequately 3 - Neither adequately nor inadequately 4 - Somewhat inadequately 5 - Very inadequately 15. To what extent have your complaints or symptoms that brought you to therapy changed as a result of treatment? 1 - Completely disappeared 2 - Very greatly improved 3 - Considerably improved 4 - Somewhat improved 5 - Not at all improved 6 - Got worse 16. How soon after therapy did you feel any marked change? _ weeks of therapy (approximate) 283 17. How strongly would you recommend psychotherapy to a close friend with emotional problems? 1 - Would strongly recommend it 2 - Would mildly recommend it 3 - Would recommend it, but with some reservations 4 - Would not recommend it 5 - Would advise against it Please indicate to what extent each of the following statements describes your therapy experience. Disregard that at one point or another you may have felt differently. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 1 - Strongly agree 3 - Mildly agree 5 - Undecided 7 - Mildly disagree 9 - Strongly disagree (use this scale for the following questions) My therapy was an intensely emotional experience. My therapy was often a rather painful experience. I remember very little about the details of my therapeutic work. My therapist almost never used technical terms. On the whole, I experienced very little feeling in the course of therapy. There were times when I experienced intense anger toward my therapist. I feel the therapist was rather active most of the time. I am convinced that the therapist respected me as a person. I feel the therapist was genuinely interested in helping me. I often felt like I was "just another patient". The therapist was always keenly attentive to what I had to say. The therapist often used abstract language. He rarely engaged in small talk. The therapist tended to be rather stiff and formal. The therapist’s manner was quite natural and unstudied. I felt that he often didn’t understand my feelings. 284 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. I feel he was extremely passive. His general attitude was rather cold and distant. I often had the feeling that he talked too much. I was never sure whether the therapist thought that I was a worthwhile person. I had a feeling of absolute trust in the therapist’s integrity as a person. I felt that there was usually a great deal of warmth in the way that he talked to me. The tone of his statements tended to be rather cold. The tone of his statements tended to be rather neutral. I was never given any instructions or advice on how to conduct my life. The therapist often talked about psychoanalytic theory in my sessions. A major emphasis in treatment was upon my attitudes and feelings about the therapist. A major emphasis in treatment was upon my relationship with people in my current life. A major emphasis in treatment was upon childhood experience. A major emphasis in treatment was upon gestures, silences, shifts in my tone of voice and bodily movements. 1 was almost never given any reassurances by my therapist. My therapist showed very little interest in my dreams and fantasies. I usually felt I was fully accepted by my therapist. I never had the slightest doubt about my therapist’s interest in helping me. I was often uncertain about my therapist’s real feelings toward me. The therapist’s manner of speaking seemed rather formal to me. I feel the emotional experience of therapy was much more important in producing change than intellectual understanding of my problems. My therapist stressed intellectual understanding as much as emotional experience. 285 APPENDIXgE : Post Therapy Therapist Questionnaire (PQTH) POST-THERAPY THERAPIST QUESTIONNAIRE (PQTH) 1. Hopkins Psychiatric Ratings Ratings: 0 = None 1 == Slight 2 = Mild 3 = Moderate 4 = Marked 5 = Severe 6 = Extreme 1 Somatization 2 Obsessive - compulsive 3 Interpersonal sensitivity 4 Depression 5. Anxiety 6. Hostility 7 Phobic anxiety 8 Paranoid ideation 9. Psychoticism 10. Global pathology index (GPI) 2. Please rate each of the following items, comparing the client with other clients whom you see in psychotherapy, using the following scale: 1 - Very little 3 - Some 5 - Moderate 7 - Fairly great 9 - Very great 11. Defensiveness 12. Anxiety 13. Ego strength 14. Degree of disturbance 15. Capacity for insight 16. Overall adjustment 17. Personal liking for the patient 18. Motivation for therapy 19. Improvement expected 20. Degree to which countertransference was a problem 21. Degree to which you usually enjoy working with this kind of patient in psychotherapy 22. Degree of symptomatic improvement 286 23. 24. 25 . 26. 27. 28. 29. 30. 31. Degree of change in basic personality structure Degree to which you felt warmly toward the patient How much of an "emotional investment" did you have in this patient Degree to which you think the patient felt warmly toward you Overall success of therapy How would you characterize your working relationship with this patient? 1 - Extremely poor 3 - Fairly poor 5 - Neither good nor poor 7 - Fairly good 9 - Extremely good How satisfied do you think the patient was with the results of this therapy? 1 - Extremely dissatisfied 3 - Fairly dissatisfied 5 - Neither satisfied nor dissatisfied 7 - Fairly satisfied 9 - Extremely satisfied How would you characterize the form of psychotherapy you conducted with this patient? 1 - Largely supportive 3 - 5 - 7 - 9 - Intensive analytical Do you recall any strikingly pleasant experiences that you had during the therapy sessions with this patient? If yes, please mark the number that best indicated the degree of pleasantness. Otherwise mark "0" for no. 1 - Mildly pleasant 3 - 5 _ 7 - 9 _ Extremely pleasant 287