. p .17.}: 4.. . ... . ,3. .. . .1... . .. ,. . (7!. . r, .94....21132 ..,,. .. . . 1,(.k.§|:\..\x I. 5.1 ., I .3 . . 1 ..r.. .. a!» \t. I? {33.43. . 5...: 5.... A! . .... .. 1 .> a» r! ; .. ..v . n. fin . ...: kl . t1:- . 1 . ESE; .... to . : 1...v.v. . £1! . , if! , . mulirltll...‘ fith. L3. , . . .. . Xx. ... h? v. a)~ r . . Ugigkfr}; . 3%}: . . ...... 77.5 n 7 ..Is ,. . .! 113E} gitxhxlr? n 1 1 Elk... , . ...... . ... . , .. J .. Lsnrxlvut. K r v . v. 1 . .2. ~ a .1). 9: tr...” . . , A. 3 8... {Li-3E; 1 . .fifiséufiuzezi rt. .t 5;... .. . '33?!» A 3:11.... . . 3110-1: 1.1.3.1»? ....)r z o... , . 1.33.2135). 1 y 3.1!. . . 5‘ «...... .... -... .. . 1....1.bi..~3J¥51&1:h¥ z I. rt .1 nil?» 12151311.. :13... fax 1 l I 1.13... reapiovrnnl, L r ..J. . . A .3. VD- 11/ 7! .Lfiw t‘égft-,¥t. L (I. E 3L$9Ptr1lizkrrlzirrie§ 3.1.5.)? .1 1 7.1.. (1)....» .ginxgrifllrx.?lixl:k $33.3.ckhaax 1.3 . s. e, .i.r.r.\. L1..... 3 (‘3?3333ii371 E r. PKEhQ-Ltr’. 11.: ll; .. nuns, «(1.18 1 . ,. LEE-1.331 this rig: 33.49.39...) . . E . . . . 1- , ‘i 1113‘ 1131‘ 114.. . . . , .3... . .. . .la . . ,. $7533.}... . 3T. , .. Haggailiaéiiigfi . 313133131! .1113. . . , . 1 .. s. I . 3%)}. giftnlpilsgix‘ inlaiglntui 1.011.324! . .. .... . ix .11. r y . "3 . 1). 3315\gttuaxfi‘1fl-Illix . . I v‘ E . I. 4‘ 1‘. 1 .. Avg“... : ... ... . . 13211....fin1nd11 .11. 33.1.51 1 \.1Tu|>\lLi£{v\:ll-(Xflf17§ . . 1 V . i. , ’rnEW‘EFEi s. . .... 1.02% )1 , . 55151153 .. . , 5.3 i $3131.33}... $135.11... fiuunfl _ v. 22“ . ll . fi . . ~14. . Q'wbfl’iéfil -\£.3rié¥v1liix¥£llt.{111, 331.31%}. 11:3 5111311113. 513.313.... . 1 , . 311.11. 1%.... , . V131: . . xi 11311: 3%.. , . . ,u 53%: .111 V 113.633 g1 “3.19.1... {itkhfihflfi ii“; 1?... .iuuiu 7 DY. ‘lhk .§\.L\L\:V~ . Iin§sks Jun . 3J3§h,§.vv§b&|i!fhflrra 3.7. {31156334 .... 3111A \ . 1 . , k3 I .5151! £33.51. . l k .. . L . tifitvg fig 3:. . p U)“. Frukifltn . {1&1}... b.1151- . r 3 . . . .LL~ “.1915... 15.... .... n . fl .1..L...L.,....~...£Sxfl\ . . . . L l. 1: . 31.4131» .4 . 1.8.329 21 ii!“ \Qisdlxfi353311321 , . 11...... ..§¥§s3§h§...§uusms 313131w..3~331.wn..%. in $35.51 «4.5:... . 55.... 3141.13.91 .111... .1133. . 33.13.. 1.1.1.1}. . . 5.111211%... . .. . ‘31»:11:5.W§1~§3§1\.§; 1%.. ; .. . a 1 Egg; ”ENE: . . a 5.217.... \A...l.n.uu :11. 8331-.ux.-\u11~t. 1. ‘1‘ , 1.333. t .... x 1.“: .. v8.32... , . 21} “an. :7 7. 3 If 4 .1. . . 1. v. . Y. .. . V. . £3 .. . ,. 5. . ‘51! .u‘gfig i , .43i..§.inauullllil. .1332 3.3.9 ,. 111W§x1sixfluf ..- 1%.. 1%.. t <9 7. S... , g . 4 EEK . 131333,.‘55 3:31.511 1.10131 , v 8‘ Ana-l. 1.0... v l 1) . 5. up .. I. . indicrfll JR.) 1 VH1)". A: ...r ...;:r..$rrr.vlh£!: 3931,1111! 51.31.1131 111.37 . fin?!“ 3.311%.-. .1331»... .. t 3117.13.11 51.7.13. 5.1.1.1 1111 r ... ... is}... 51111.5...) s5 #111,333? .E‘ 1. . ... 7, .. .. . .. ,. . .. .«v‘1‘1‘gze‘fiqag 9 In . . . . ... .1“... . .L. i... .. y .....3' r. z . 1 . . ..... , , . V . , .. .1, .7. . .. . .......yf..... : .. . .311 L. 1 ......,J.. ... .... . n. J .... . SA-v.. . ..NlJulthIIV'3-I I x L m..-y ...1 m1 LIBRARY Michigan State University This is to certify that the dissertation entitled META-ANALYSIS OF COMMUNICATION APPREHENSION TREATMENT TECHNIQUES presented by Mike Allen has been accepted towards fulfillment of the requirements for Ph. D. degree in Communication 77/1/1/ /h (C Major professor Date October 19, 1987 MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 lI/llllllljflljl MSU LIBRARIES lflllllllllllll 1, 0 6 RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. FEB 0 7 t994l ,>‘_ x / Q} V} m W3 mg (4.9%.; {QM}! 43? W,“ h 8’k&h WNW I‘Alv 7‘0 II; 3033 META-ANALYSIS OF COMMUNICATION APPREHENSION TREATMENT TECHNIQUES By Mike Allen A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 1987 ABSTRACT META—ANALYSIS OF COMMUNICATION APPREHENSION TREATMENT TECHNIQUES By Mike Allen A quantitative literature review was undertaken to determine which, if any of the treatment techniques for communication apprehension are effective. A search of the literature found 181 experiments that provided data on the effectiveness of treatment in reducing public speaking anxiety. The experiments were examined to determine if the effectiveness of treatment varied depending on the type of treatment used: (1) systematic desensitization, (2) cognitive modification, (3) skills training or (4) some combination of these three treatments. The findings suggest that the effectiveness of the treatments are additive. In addition, differences between measurment techniques and treatment settings were assessed. Generally measurement techniques reach similar conclusions. The exception to this is physiological measurement devices which tend to record smaller reductions in anxiety. The implications that the findings have for the theories and practice for the treatment of public speaking apprehension are discussed. ACKNOWLEDGMENTS I wish to thank all those that helped me during my long years at MSU. Especially Bill Donohue, my adviser, who said "why not?" and allowed me to do. To Jack Hunter for his patient and guidance in statistical and other matters and whose timely help made my efforts be on time. And to G.R. Miller and James Stiff, forced to read all the stuff once I had finished. Especially G.R. since one of my conclusions questions one of his beliefs. Thanks to literally hundreds of reseachers, who over the last 40 years have spend thousands of hours gathering and analyzing information to make this meta-analysis possible. Last to my wife, Nancy, she stuck by the idea and let me spend $1000 to buy disserations and xerox articles to do the meta-analysis. 111' TABLE OF CONTENTS CHAPTER I REVIEW OF THE LITERATURE Previous Summaries of the Literature Methods of Public Comunication Apprehension Treatment Systematic Desensitization Cognitive Modification Skills Training Combinatorial Treatments Problems of Assessing the Effectiveness of Treatments Problems of Comparisons within a Study Problems of Between Study Comparisons Where Should Therapy Take Place Meta-Analysis as a Technique for Comparing Treatments Theoretical Implications of the Meta-Analysis CHAPTER II METHODS Literature Search Coding Scheme Type of Treatment Type of Measurment Device Setting of Therapy Statistical Analysis CHAPTER III RESULTS iv 4)me 0‘ 10 11 12 16 17- 18. 20 20 21 21 24 25 26 28 Data Analyzed by Treatment Type Data Analyzed by Measurement Technique Data Analyzed by Type of Device Self Report Scales Observer Rating Techniques Physiological Measurement Techniques Setting of Treatment Testing Therapy Additivity CHAPTER IV DISCUSSION Treatment Effects Assessment of the Theories about Treatment Systematic Desensitization Cognitive Modification Skills Training Alternative Explanations for the Effectiveness of Treatments That Anxiety has Multiple Causes A Distraction Hypothesis Impact of Measurement Type Effect of Treatment Setting Recommendations Directions for Future Research Advice for Treatment of Public Communication Apprehension APPENDIX A CORRECTING FOR REGRESSION TO THE MEAN APPENDIX B INACCESSIBLE MANUSCRIPTS V 28 29 30 30 32 32 33 34 47 47 48 49 51 53 55 55 56 57 58 59 59 6O 62 68 APPENDIX C APPENDIX D TECHNIQUES APPENDIX E LIST OF REFERENCES STUDIES LACKING STATISTICAL INFORMATION TEST-RETEST CORRELATIONS FOR MEASUREMENT COMPLETE SUMMARY OF THE DATA vi 69 7O 71 92 Table Table Table Table Table Table Table Table Table Table ...: DOOM 10 LIST OF TABLES Results of Analysis by Treatment Type Comparing Treatment Group Correlations Comparing Measurement Techniques Results of Self Report Scale Self Report Scales Broken Down by Treatment Type Comparison of Observer Measurement Techniques Comparison of Physiological Measurement Techniques Comparison of Treatment Setting Reanalysis of Data by Treatment Groups Test of Additivity Model vii 37 38 39 ' 4O 41 42 43 44 45 46 CHAPTER I REVIEW OF THE LITERATURE The ability to communicate publicly has been called one of the skills necessary for democracy (Jeffery and Peterson, 1980). This skill also appears to facilitate success in school (Richmond, 1984) and business (Richmond, 1977; Daly and Leth, 1976). Not surprisingly, situations requiring demonstration of this skill have been labeled one of the most feared aspects of modern living (Neer, 1982). While people may normally suffer some anxiety or tenseness prior to any big event, such as the start of a contest or the beginning of a wedding, this state of physiological arousal should not totally destroy a person's ability to function (Daly and Buss, 1984). Communication departments in the United States often assume responsibility for teaching public speaking to thousands of students every year. Most of the public speaking classes and texts spend the bulk of their effort on improving the skills of public speaking (organization, research, stategies, and tactics). However, many insititutions to deal with this problem have estabished a special section for apprehensive speakers, created an independent course, have workshops, labs, or special groups, or individual tutoring (Foss, 1982). These methods use classroom time simply trying to calm speakers and instill confidence in their abilities so they can concentrate on improving performance rather than overcoming fear. Most Communication 1 2 Departments expect the problem of communication apprehension to be handled in the classroom by the instructor (Hoffman and Sprague, 1982) and do not provide special classes or instruction for those with extremely high anxiety. Most experienced public speaking instructors can tell antecdotes about anxious students who have refusted to speak when asked, started crying, or run out of the classroom never to return. More commonly, students experience the behavioral signs of fright including a cracked voice, shaking hands, or the scratching of legs and face (Paul, 1966). To reduce these problems public speaking texts often mention the problem of public communication apprehension and offer various methods of minimizing the impact this anxiety has on student performance (Ehninger, Gronbeck, and Monroe, 1984; McCroskey, 1982; Nelson and Pearson, 1981; Verderber, 1976). My purpose in writing this dissertation is to conduct a meta-analysis of the communication apprehension literature as a means of assessing the effectiveness of the various treatments. This meta-analysis is the first step in a program of research that will test the theoretical assumptions about communication apprehension that differentiate the various therapies. Previous Summaries of the Research Researchers have examined numerous types of treatment techniques to find the best method of helping a person overcome their level of communication apprehension. In their seminal reviews of the interventions used to reduce communication apprehension Foss (1982) and Glaser (1981) found that therapies vary widely in terms of the 3 resources used to reduce communication apprehension and the time needed to implement these programs. For example, some scholars suggest that simply participating in the normal public speaking class improves the student performance (Brooks and Platz, 1968). Others suggest use of biofeedback equipment for treating public communication apprehension (Gatchel, Hatch, Watson, Smith, and Gaas, 1977; Gatchel and Proctor, 1976). Unfortunately, these lists of treatment methods stand only to explicate the various treatment techniques rather than to evaluate their efficacy. Foss (1982) states explicitly that she seeks only to summarize the available resources for instructors; she devotes no effort assessing the effectiveness of these varied techniques for reducing communication apprehension. Most evaluations generally defend the effectiveness of a given treatment technique but offer no guidance about about the relative effectiveness of the method compared to other methods (Friedrich and Goss, 1984; Kelley, 1984; Fremouw, 1984). For example, Fremouw (1984) summarizes the past research on cognitive modification techniques but does not offer advice about which specific technique of cognitive modification offers the best evidence for reducing communication apprehension. He also does not compare cognitive modification to other techniques like systematic desensitization. Reviews ought to perform more comparisons. Critiques of these techniques should offer practitioners criteria for selecting their own methods of reducing communication apprehension. 4 Methods of Public Communication Apprehension Treatment This review will divide the treatment techniques into three types: (1) systematic desensitization, (2) cognitive modification, (3) skills training through education. In addition, the issue of whether treatment types can be combined will be examined. This last issue is not one well established in the literature but a necessary consideration when treatment techniques have been combined to treat the phobia. Each section will examine the assumptions about public communication apprehension that the therapy makes, then a discussion of the general procedure of the therapy will be given. Systematic Desensitization Systematic desensitization assumes the problem of anxiety evolves from some association between a stimulus and response. The person learns to associate public speaking with negative emotional reactions (Paul, 1966; Wolpe, 1958; McCroskey, 1972). For example, a person experiences or witnesses a public presentation negatively received by an audience. The speaker involuntary associates public speaking with negative rewards (public speaking leads to embarassment). Systematic dessensitization changes that association by exposing the person to the phobia and causing the patient to involuntary associate more pleasant responses the speaking situation than fear and avoidance. For example, during the public speaking therapy every speech the subject gives would receive applause from the audience. This substitutes a positive feedback for the negative feedback that may be associated with public speaking. Generally the treatment involves instructing subjects (or 5 patients) on methods of relaxation (Friedrich and Goss, 1982; Paul and Berstein, 1976). After teching general relaxation techniques, the instructors ask patients to relax while thinking about public speaking and avoid reacting with fear and anxiety to any aspect of the public speaking situation. The subject is asked to relax first about some situation that is rated as only mildly anxious. As the subject learns to successfully relax when thinking about this situation, the subject is asked to then relax when thinking about some more frightening or more involved aspect of public speaking. Often these increasingly anxiety producing statements in systematic desensitization are called "hierarchies" (Paul, 1966; McCroskey, 1972; Goss, Thompson, and Olds, 1978) A subject that can successfully go through the hierarchy of statements about public speaking without becoming anxious is "cured." In practice, when the subjects speak, they can concentrate on the relaxation techniques learned in therapy rather than responding with fear and anxiety. Advocates argue that the training improves significantly the individual's performance. Some of the various methods of relaxation are: muscular relaxation (Paul and Shannon, 1966), biofeedback (Gatchel, Hatch, Maynard, Turns, and Taunton-Blackwood, 1979; Gatchel, Hatch, Watson, Smith, and Gaas, 1977), and mental (imagery) associational methods (the person thinks of good outcomes rather than embarassing outcomes of the event) (Gurman, 1973; Kirsch and Henry, 1979; Kirsch, Wolpin, and Knutson, 1975). The method of presentation also varies, from automated methods (McManus, 1975; Lohr and McManus, 1975; Marshall, Stoian, and Andrews, 1977), group methods (Rimm and Masters, 1979), 6 and large classroom techniques (McCroskey, Ralph, and Barrick, 1970). All of these methods share the behavioristic assumption that the phobia is the result of some stimulus-response relationship that the subject has learned. The therapy is intended to substitute more functional responses for other less functional responses when the person is supposed to give a public speech. Cognitive Modification Cognitive modification assumes that the person possesses the skills to speak effectively but not the ability to use these skills. Cognitive modification focuses on the beliefs of the Speaker about the event and tries to modify those beliefs to permit success (Meichenbaum, Gilmore, and Fedoravicius, 1971; Fremouw and litter, 1978). Cognitive modification therapy assumes that the anxiety results from "irrational" beliefs that people possess about public speaking. For example people incorrectly believe that there is failure, humiliation, or some other undesirable outcome associated with their speaking (Fremouw, 1984; Ellis, 1962). This irrational fear makes a person unable to use the speaking skills they possess. For example, a person may believe that "I never speak well because people laugh at me." This belief creates an irrational overgeneralization that the therapy would correct. Cognitive modification intends to substitute a set of rational and truthful beliefs for the irrational beliefs about perceived failure to allow the subject to succeed in public speaking. The therapist COgnitively changes the patients so that the situation can be used to 7 gain some advantage. By overcoming these irrational fears patients can rely on their talents and skills to be successful at public speaking. Patients are often taught to use "coping" cognitions while they are speaking such as "so far so good" (much like a mantra). The therapy sessions begin by encouraging patients to discuss their fears about public speaking. One by one these fears are shown to be the result of some irrational belief held by the patient that needs to be discarded in favor of some more rational belief (a coping statement). For example, a person might believe that people always laugh at me when I speak. The therapist points out that this is false, people only laugh at things that are funny and the audience will take the speaker seriously. By demonstrating the voracity of ths statement (usually through practice speeches), the therapist intends to replace the irrational belief with a rational belief (Glogower, Fremouw, and McCroskey, 1978). Skills Training Unlike the other two therapies, this approach does not assume that the innate skill of the speaker is adequate to perform the task of public speaking. Both systematic desensitization and cognitive restructuring assume that the requisite skill level exists but cannot be put into practice because some cognitive or affective feature is blocking the use of the talents of the individual. Skills training assumes that some people have skill deficiencies that must be corrected before they can speak. A person is justified in being anxious if he or she lack the skills necessary to be perceived as competent. Once the skill is learned properly, there need be no 8 barrier to performance. More than that, the person should be confident in the ability to speak successfully. This technique assumes that communication apprehension is caused by a person lacking public speaking skills. People experience anxiety because they do not have the training to be successful public speakers (Clevenger and Phifer, 1959). The goal of the therapy is to train the patient in the skills of public speaking. This training gives the patient the confidence to perform well. Raising confidence through training will result in less anxiety as the person speaks with greater certainty of success. Peeple once trained should be confident of their ability to speak successfully. A large part of this educational process is the giving of practice speeches and constructive criticism (Brooks and Platz, 1968; Borin, 1949; Ertle, 1969). Practice sessions and constructive criticism allow for the person to develop their skills over time and grow in confidence as they are rewarded for demonstrating the lessons they learn from the instructor and each other. The use of positive feedback is emphasized and the use of the skills reinforced with every Speech. Skills training is usually done in the form of a public speaking class at an educational institution. This training in the skill of public speaking has been found to confidence of the speaker (Ewing, 1944; Garrett, 1954; Hayworth, 1940). Once the confidence of the speaker is raised; the speaker will atribute nervousness to excitement, a normal part of the process of giving speeches. The outcome (success) will not be in doubt because of the adequate training the person has received. 9 Combinatorial Treatments Several experimenters have used combinations of the three techniques (systematic desensitization, cognitive modification, and skills training). Experiments use one therapy during the one part of the experiment and then another therapy later with the same group of subjects. Theoretically, if public speaking anxiety if the result of multiple causes, then perhaps all causes must be treated simultaneously (Meichenbaum, Gilmore, and Fedoravicius, 1971; Norman, 1975). Often, the experimental design unintentionally uses a combinatorial treatment. Many of the studies done by communication scholars draw their subjects from introductory public speaking classes (Ayres and Hopf, 1985; Borin, 1949; Fremouw and Harmatz, 1975). The subjects, in addition to their public speaking class, attend therapy sessions using COgnitive modification or systematic desensitization, or additional skills training. This results in the subject receiving two or three types of treatment at the same time. In the communication experiments the comparison to a control group consists of a control group drawn from the same introductory public speaking class. The experiment is therefore testing whether or not the other therapies offer any improvement over and above just attending the public speaking class (one form of skills training). Selecting subjects from public speaking classes, however is not used in the majority of studies conducted by psychologists. The psychologists generally draw their subjects from introductory psychology classes or by advertisements in the college or local paper 10 (Gatchel, Hatch, Maynard, Turns, Tauton-Blackwood, 1979; Goldfried and Goldfried, 1977; Grande, 1975; Jarmon, 1972). These subjects are often screened to make sure they have 395 had training in public speaking and are ggt_taking such classes at the current time. These experiments are testing whether any improvement is taking place as a result of therapy. These subjects do not receive the benefit of a public speaking class (skills training) and therapy but only the specific therapy alone. This difference in subject selection procedure results in the same labels being applied to different treatments. For example psycholgoical studies investigating the effect of systematic desensitization have subjects only improving as a result of that particular therapy. Communication experiments often take subjects from public speaking classes, the change over time could be the result of either systematic desensitization or skills training or both. In communication experiments however, the label applied to these groups is not skills training and systematic desensitization but is only systematic desensitization. In any literature review the communication and psychology experiment are put into the systematic desensization without recognition of the method of subject selection. Problems Assessing the Effectiveness of Treatment Analysis of the treatment techniques is made difficult for a number of reasons related to the nature of the published research. To provide support for a particular treatment technique, scholars reviewing the literature have typically relied on using single studies showing the comparisons between treatment techniques (Page, 1980; 11 Phillips, 1980, Fremouw, 1984). These experiments take different treatment techniques, usually two or three, and then compare the effects of the treatments within one experiment (Watson and Dodd, 1984; Marshall, Stoian, and Andrews, 1977; Johnson, Tyler, Thompson, and Jones, 1971; Jarmon, 1969; Karst and Trexler, 1970; Goldfried and Goldfried, 1977; Jaremko, Hadfield, and Walker, 1980; Ayres and Hopf, 1985; Sherman, Mulac, and McCann, 1974). Because the sample size used in these experiments is typically ten subjects per treatment technique, a great deal of sampling error exists. There are few experiments making such comparisons. Thus, very little evidence exists regarding the relative effectiveness of the techniques. Comarative conclusions are difficult to draw from these studies. Scholars reviewing that have reviewed these experiments have not been able to set forth firm conclusions about the relative efficasy of the various treatments (Foss, 1982; Glaser, 1981). This is demonstrated in the unwillingness or inability of the reviewers to advocate some treatment techniques over others on the basis of effectiveness. Problems of Comparisons within a Study Even if a larger body of studies had existed to compare the various treatment techniques within a study, the result would probably be as inconclusive as it is currently with only a small number of studies. This confusion would come from the reviewers' reliance on the significance test to determine results. With the small sample sizes, the type II (false negative) error rate is extremely large (Hedges and Olkin, 1985). Sampling error in the communication apprehension literature is large because the typical sample size is between 10 and 12 20; i.e. very small. Thus, the "counting" of significant results could be misleading. The Type 11 error rate could be as high as 95% and the typical reviewer would ignore this in using vote counting methods when deciding what a majority of the studies conclude. Most studies fail to reject the null and conclude that no difference exists between treatment techniques. This conclusion may be correct but the method used to reach the conclusion is suspect. Without some attempt to take into consideration the effects of sampling error, these summaries of within study comparisons may misrepresent the literature. Given small sample sizes the difference between treatment groups would have to be extremely large to be significant at the .05 level in every study. This is especially true when the reviewer is looking for a consistent pattern of differences to be found. The number of studies is small and the techniques used are not likely to find any differences that do exist. Thus, the inability of reviewers to make clear recommendations is not surprising. Problems of Between Study Comparisons Method effectiveness can be compared across studies. Even if there were no direct within study comparisons of treatment techniques, they could be compared between experiments. This comparison is based on the average amount of change caused by each treatment in those studies where that threatment was used. This method would quantitatively distinguish between treatments that obtain positive effects and those that obtain even greater positive outcomes. Currently, the reviewers do not distinguish between large and small effect sizes; they only distinguish between significant and 13 insignificant effect sizes. Previous reviewers have compared significance test results (Friedrich and Goss, 1984; Kelley, 1984; Fremouw, 1984) These reviews first look at whether or not a given treatment technique has been consistently effective in reducing anxiety. They then compare treatments in terms of frequency of significance. The conclusion depends on the use of significance tests to determine whether or not a treatment is more effective than other treatments. The underlying assumption is that effective treatment techniques will consistently obtain significant results in studies and ineffective treatments will only get significant results occasionally (because of Type I error). This assumes that the studies are all of equal quality and the chance of a significant finding is equivalent for all studies. Unfortunately, almost all experiments conclude that a given treatment is effective in reducing anxiety. This means that in terms of frequency of effectiveness all treatments could be classified as equally effective in reducing public communication apprehension. The reviewers make no attempt to quantitatively assess the impact of the treatments in the studies. They do not average or estimate the effects of each treatment. The situation is made complicated because the studies may not be equally likely to obtain significant results. Sample sizes vary from 10 to well over 400. This means that sampling error varies from study to study. The means that there is variance in the probability that a significant finding will be found. Further, some studies use selection procedures that restrict the range in scores. Selction will effect the change scores. Regression to the mean 14 also occurs in the studies that use a measurement device that has some unreliability for both sample selection and the pretest score. Studies do not correct for attenuation of effect sizes due to measurement error. The result is a whole host of problems that make the significance test inaccurate as a measure of the size of effect. Almost all experiments conclude that a given treatment works to reduce public communication apprehension. The probability of significant findings is enhanced because most experiments use a selection procedure that selects the top 50, 34, 10 or 5% of the population that is most anxious. Some method of pretesting is done to select only the most apprehensive subjects for treatment out of some larger population. The selection procedure has the effect of restricting the range of scores in the initial population. The effect of the selection procedure for therapy on the pre versus post test measurements is to enhance the change score. Those that are most apprehensive can have larger change scores than those that are least apprehensive. This increases the observed effect size for the treatment, which is almost always significant even given the relatively small sample used. The mathematical effects and solutions to such problems are shown in Hunter, Schmidt, and Jackson (1983). Those studies using small samples can select using a more stringent standard, which will increase the probability futher of a significant finding because the change will be overestimated. The possibility of significant results is also increased because regression to the mean is caused by unreliability in the measurement devices. One way to avoid the effects of regression to the mean is to 15 use one scale in the selecting of the sample and a different scale for the pretest. Most texts on experimental design mention regression to the mean as a possible internal reliability problem for an experiment (Babbie, 1979; Cook and Campbell, 1963). These statements about the problem however, only talk about the occurence in general qualitative terms and propose no quantitative means to assess the severity of the effect that regression to the mean has on the effect size obtained. An ability to quantitatively assess this problem exists (Hunter, Schmidt, and Jackson, 1983). This problem is a significant one that can alter results dramatically. For example, the effect this might have on an experiment measuring the effect a given public communication apprehension treatment technique could be quite significant. Suppose the public communication apprehension scale has a test-retest reliability of .90. With a sample chosen from the top 10% of a population (that is, the subjects for the experiment tested in the top 10% during a pretest), the effect of regressing to the mean will show an observed correlation for the change that is .09 larger than the real change due to treatment. (See Appendix A for a more detailed mathematical explanation). This increases the probability greatly that any observed effect will be significant at the .05 level (the standard significance test used does not take this into consideration). Some experimenters do not preselect their sample. Several experiments have used whole sections of public speaking classes. This use of intact groups avoids the problems of restriction in range and regression to the mean. However small effect sizes are still more likely to be significant because of relatively large sample size (as 16 large as 840). This large sample size makes these unselected sample experiments appear congruent with the studies that preselect subjects. Both experiments obtain significant results at the .05 level. All the treatments consistently seem to work to reduce public communication apprehension. However, the results of a significance test is do not measure the magnitude of the effect obtained by the treatment. All the treatments appear to work equally as well (because the experiments all obtain significant findings). This poses a problem for the instructor seeking to find a treatment technique to maximize the potential of the students in the public speaking course. If no argument can be made that a given treatment is more effective than any other treatment technique, the choice is arbitrary with respect to outcome. The instructor cannot use the existing reviews of the published research and literature reviews make a justifiable choice among treament techniques that will maximize the benefits for the students. Where Should Therapy Take Place Many of the experiments on communication apprehension treatment have been done in the therapuetic setting conducted on a one to one basis (therapist-patient) rather than as a mass classroom exercise. Miller (1984) has raised the question about whether or not educators should undertake to "treat" communication apprehensives in the context of a classroom. Many professors/graduate assistants are not trained in treating problems of anxiety and could be ineffective in reducing the level of communication apprehension (Barrick, 1971; Miller, 1984). A survey of 0.5. colleges and universities reveals that 93.2% of the 17 respondents do not operate a treatment program, 73.4% said that they thought the problem should be handled in the classroom (Hoffmann and Sprague, 1982). If the results of therapy given by a therapist on an individual basis differ from the results of the same therapy when conducted in the class setting, the mass application of the therapy may not be as beneficial as the instructor hopes. This is a legitimate concern given the inability of an instructor to spend the same amount of time with a student that a therpist can spend with a patient. The reviews of the literature mention the issue but make no attempt to provide a resolution of this issue on the basis of experimental evidence. Advocates of particular types of treatment have suggested that caution, training, and expertise is needed when using public speaking therapies (Barrick, 1971). Only McCroskey (1972) has tested the efficacy of treatment from the lab to the classroom and concluded that for systematic desensitization the mass application remains beneficial. The issues surrounding the therapist has been examined for psychotherapy in a meta-analytic review of the literature (Glass and Smith, 1981). The findings show that therapists do not differ in effectiveness in treating patients based on their level of education, years of experience, or expensiveness. This provides evidence that many of the concerns about the need for a highly trained and expensive therapist may be misplaced. Meta-Analysis as a Technique for Comparing Treatments What is needed is some summary and comparison of treatment techniques for communication apprehension. Meta-analysis provides a 18 solution to this type of problem. Meta-analysis has been used to asses psychotherapy (Glass, McGaw, and Smith, 1981), coronary artery bypass graft surgery (Wortman and Yeaton, 1983), and patient education (Mazzuca, 1982). The results of a meta-analysis allow for a quantitative comparison of various treatment techniques. The techniques provide a basis for both within and across study comparisons that can take into consideration problems like regression to the mean, restriction in range, and measurement error due to attenuated measurment. The conclusion of a meta-analysis has the effect of reducing the impact of sampling error and allows for comparisons between techniques. Rosenthal (1984) points out that the issue in conducting a treatment techniques meta-analysis is not simply whether or not any particular treatment is effective. Certainly that is an issue, but as important is the issue of relative efficacy. If all treatments work to apprehensive speakers; which treatment type should be preferred over another because of its effectiveness? This dissertation will make no attempt to assess the administrative questions of resource allocation (how much additional resources should be expended per student to gain the optimal amount of treatment), but this dissertation will address the question of the relative effectiveness of existing techniques. Theoretical Implications of the Meta-analysis Even if one therapy type should emerge as superior to the other therapy types, this does not prove that the theory underlying the therapy is necessarily correct. However, if one therapy does prove superior, special attention should be given that particular 19 explanation of the phobia. At the current time the competing explanations offer evidence that each is a correct interpretation of the phobia. Until the evidence can be compared quantitatively, cumulatively, and systematically, theoretical development will be stunted. It is only with the accumulation of information and the formation of facts that better explanations can be sought and then tested. This meta-analysis will also compare the self-report measures of a phobia to observer ratings of performance. This is an important question for any therapy that claims to reduce an anxiety that inhibits performance of a task. Does the therapy really improve performance (as measured by observers) or does it only raise the confidence (as measured by self-reports). Meta-analysis allows a partial answering to these kinds of questions that will allow for futher theoretical development. CHAPTER II METHODS This meta-analysis gathers the existing quantitative literature on the treatment of communication apprehension. After gathering the literature, the studies were coded by design features that could be possible moderator variables. Once the manuscripts had been coded, the results of every study were converted to a common metric (in this case the correlation). When the results are in a common form they can be averaged to estimate the effect size for the population. Groups of correlations were compared on the basis of the observed variance in effect sizes to determine if that variance is greater than would be expected due to sampling error. This is useful in determining if moderator variables exist. Literature Search A search of the literature was conducted to gather all relevant materials. A manual search of the relevant subject listings for articles on treatment of communication apprehension was made of the Psychological Abstracts and the Education Index. A manual search of the table of contents of all Communication Education issues was made to obtain articles on the topic. All materials obtained were examined for references to additional materials. The Social Science Citation Index was also searched. To be included in the analysis a manuscript had to meet the 20 21 following criteria: (1) the manuscript had to contain a quantitative analysis of the effect that one or more communication apprehension treatments had on one or more dependent measures of anxiety (case studies were excluded); (2) the manuscript had to have measured the level of apprehension prior to treatment (pretest) and measured the level of apprehension after treatment (posttest); (3) the manuscript had to be accessible to the author (Appendix B contains those articles that were inaccessible); (4) the manuscript had to contain enough information permit conversion of results to the common metric (Appendix C lists those articles with insufficient information). There were 115 manuscripts that met the criteria containing the results of 181 separate experiments. Coding Scheme Initially each study was coded for the following three characteristics: (1) type of treatment(s) used, (2) type_of‘ measurement device(s) used, and (3) setting(s) used to adminster the treatment. Type of Treatment The study was coded as using one of the following treatment types: (1) systematic desensitization, (2) cognitive modification, (3) skills training, or (4) combinatorial treatment. The decision for coding was not based on necessarily the term used by the experimenter but rather by the description of the therapy given in the manuscript. 22 Studies were coded as using systematic desensitization therapy if the person went through a process that exposed the person to the anxeity provoking stimulus with the expectation that over time the person would become less sensitive to the problem. The "classic" approach seats the subject in a comfortable chair and then instructs the person to think about the first sentence in a provided hierarchy of statements relating to public speaking. When the subject learns to relax when thinking about the first statement, they are then told to think about the second statement in the hierarchy which is one that is slightly more threatening (phobic inducing) than the first statement. The subject then relaxes while thinking about that statement. This process continues until the subject exhausts all items in the hierarchy. Other systematic desensitization procedures have the subject speaking while relaxing the muscles in their body and regulating their breathing. Some methods use group discussion, self-pacing, biofeedback techniques, or other techniques to desensitize the person to public speaking. The common feature in this type of therapy is instructing the person to relax in a situation that is normally anxiety producing. The person learns to relax and control their fear by becoming less sensitive to the fear invoking aspects of the situation through repeated exposure to the situation. Studies were coded as using cognitive modification if the therapy was designed to change the cognitive beliefs of the person about public speaking. This therapy changes the beliefs about public speaking that cause fear. This therapy is also called, "cognitive restructuring", or "insight" (by "insight" the authors mean an insight 23 into the unreasonableness of the beliefs). This therapy usually involves either eliciting from subjects the reasons they fear public speaking or providing the "common" reasons people fear public speaking. The therapy sessions then expose the falseness of these beliefs. Substitute beliefs are provided that allow a person to cope with the situation (coping statements) or show the positive effects of speaking (rewarding statements). These beliefs are reinforced by practice speeches incorporating the statements provided by the therapy. Studies were coded as using skills training if the focus of the therapy was on providing information/practice on public speaking as an art. The skills training approach emphasizes doing research on the speech topic, outlining and organizing the speech, selecting appropriate language to use in the speech, and practicing the speech to improve performance. The subject learns the proper techniques of public speaking which will lead to success and build confidence. The theory behind the teaching of skills to diminish apprehension is that people trained in the skills of the technique will automatically acquire confidence because they know they have the skills rquired for success. Studies were coded as combinatorial if they used more than one technique to reduce communication apprehension. The specific combination used was recognized as a separate treatment type. For example, a study using systematic desensitization and skills training was coded differently than a study using rational emotive therapy and systematic desensitization. 24 Type of Measurement Device A separate coding was made for each type of measurement device used to assess the treatment technique. The types of assessment fall into three categories: (1) self-evaluations, (2) observer ratings of behavior, and (3) physiological assessments. Self-evaluations are paper and pencil tests that ask the person to answer questions about their own level of apprehension before and after treatment. These tests take a variety of forms and each form was coded separately. For example, the Personal Report of Confidence in Speaking is a 30 item true/false questionairre about various aspects of the phobia. The Fear Thermometer is a single item asking the individual to rate their level of apprehension on a scale from 1 to 100. Both instruments are self-reports but operate at different levels of specificity. The PRCS focuses in on specific features and combines the answers to form a score that represents the level of anxiety. The Fear Thermometer asks the individual to make one overall assessment. Each form of self report was tested against the other forms of self report measures to determine if particular forms obtain different effect size estimates. Observer ratings of behavior is the technique of using some trained person to make assessments about an individual speaker's performance. The assumption of this measurement technique is that the anixiety felt by the speaker will produce nervous behaviors. The individual is rated both before and after treatment on those behaviors. Theoretically, a person‘s performance should improve as the behaviors associated with good public speaking increase. This means in 25 practice that those behaviors associated with poor public speaking (grooming behaviors, scratching of the arms or legs, appearing I'nervous") should be reduced as a result of the therapy. These rating systems vary from a single overall assessment of the performance to scales that rate the individual on an entire range of behaviors individually, summing the scores to determine the level of anxiety. Each separate assessment tool will be coded and compared to other observer rating methods to determine if the different methods obtain similar or dissimilar effect sizes. Physiological measurement techniques use machines to record some physiological reaction while giving a speech. The comparison is between the reactions while giving a public speech prior to and after therapy. The theory is that certain processes result from nervousness (high pulse, increased palmar sweat, or high skin tension). The therapy should work to change the level of anxiety which will result in a reduced level of these reactions. Each type of different physiological measurement will be coded separately. A comparison will be made among these different techniques to determine with if type of bodily function measured reacts similarly or dissimilarly to other physiological reactions produced by public speaking anxiety. Setting of Therapy The setting of the therapy will be coded as either clinical or in the classroom. In some studies using combinatorial treatment techniques both the clinical and classroom settings were used for therapy. In these cases each therapy was coded for the appr0priate setting type. The classroom setting was defined as having the subject 26 experience the therapy as part of the normal coursework for which the student received academic credit. The clinical setting was defined as having trained individuals administer the therapy individually or to small groups at times outside of registered coursework. Statistical Analysis For each mansucript that contained adequate information, an effecp size was estimated in the form of a correlation. The correlation was ‘. yfl___,_.-——-'-°-"-v-" MIM— corrected for restrictioninf:apge_(Hunter, Schmidt, and Jackson, 1983), regression to_the mean (See Appendix A), and attentuation of measurement (Hunter, Schmidt, and Jackson, 1983). These corrections required an estimate of the reliability of the measurement device used. When such information was not provided within that particular manuscript, the estimates from other studies for that particular device were averaged to provide the estimate that was used to correct for selection artifacts (restriction in range and regression to the mean) and attenuation of measurement. When no reliability estimates were available for the particular device in any of the literature, the reliabilities of other measurement devices most similar to that device were averaged and that estimate was used for correcting for selection artifacts and attenuation of measurment. Estimates of the test-retest correlation for all devices used in this body of literature is contained in Appendix 0. Many studies selected a sample by asking for subjects to volunteer for treatment if the person experienced public speaking anxiety. This selelection method is known to introduce a "volunteer bias" in evaluating behavior therapy research (Cash and Janda, 1977). The 27 research shows that the self-selected sample will have more anxiety than the general population. Examination of the means of these volunteer samples shows that they have the same mean and standard deviation as a group that would be chosen on the basis of a score on a measure greater than the population mean. Studies using volunteer samples were treated as having a restriction in range equivalent to selecting a sample having a score greater than the mean. The correlations were fiIEELEEEEEEL§9.9§§§CEIEE_1:“£99 variation in the population was more or less than that expected due to sampling error (Hunter, Schmidt, and Jackson, 1983). After the total sepmof_ studies was tested, each moderator and possible combination of moderators was tested to determine if any treatment type, treatment setting, or measurement device obtained discrepant results. The results of the meta-analysis should allow for a comparative evaluation of the treatment types. This evaluation should provide a means to evaluate the theories about treatment and the application of those theories. The results should also provide some evidence on measurement issues and the issue of treatment setting. CHAPTER III Results This chapter contains the results for each method of subgrouping the data (by treatment, by measurement technique, by particular type of scale, and by treatment setting). A final section of this chapter demonstrates potential explanation (additivity of the treatment types) for the results in the treatment subgroupings of the data. A complete listing of all the effect sizes (correlations) for all studies and all subgroups is given in Appendix E. Data Analyzed by Treatment Type There were seven possible treatment types (cognitive modification/CM, systematic desensitization/SD, skills/SK, CM+SD, CM+SK, SD+SK, CM+SD+SK) and all were represented in the data. When studies used multiple methods of measurement the effect sizes (correlations) used in this section were averaged across measurement methods. For each treatment type, observed variance of effect sizes was less than the variance expected by random sampling error. See Table 1 for a summary of the results. Significance tests show that a combination of all three primary treatment techniques (SD+CM+/SK) was significantly more effective in reducing public communication apprehension than any other treatment method with the exception of cognitive modification. The trend for cognitive modification 28 29 comparison was in the right direction but the t value failed to reach significance). All six non skill treatment types (CM, SD, CM+SD, CM+SK, SD+SK, CM+SD+SK) had a correlation larger than that of the skills (SK) treatment group. Finally a significant difference was observed between the CM+SD and SD+SK treatment groups. See Table 2 for a summary of the t tests. A caveat exists, however, regarding the low power of the significance tests. For example, a t test with 9 degrees of freedom is based on an N of 11. The probability of a significant findings is based on two factors: (a) the size of the difference and (b) the size of the sample. A difference between groups is less likely to be significant with a smaller sample siZe. This means that real differences between groups may not be reflected in the results of the t test when the sample size is small. Data Analyzed by Measurement Technique There were three types of measurement techniques: (a) self report questionnaires; (b) observer ratings of behavior, and (c) mechanical measurements of physiological reactions. When studies used multiple types of the same measurement technique the effect sizes (correlations) used in this section represent averages. See Table 3 for a summary of the results for this section. For each measurement technique, the observed variance of the effect size was less than that expected due to sampling error alone. A comparision of the self report correlation LE = .261) to the observer correlation (_ = .212) and the physiological correlation Q3 = .172) showed that significant differences existed between the self report measures and observer 30 ratings (3 = 2.58,_g:=290,.p§.05) and self report measures and physiological measurement techniques (t_= 4.05, gfé222,p<.05). The difference between the observer and physiological techniques of measurement was not significant (p = 1.67,_g: = 176,_p>.05). Data Analyzed by Type of Device Self Report Scales The self report scales were compared to each other to determine if differences existed between scales regarding the observed effect sizes. Six scales were used in minimally least ten studies and subsequently were used for this analysis. The six scales included: (a) Personal Report of Confidence in Speaking/PRCS (b) Fear Thermometer/FT, (c) Anxiety Differential/AD, (d) Personal Report of Communication Apprehenion/PRCA, (e) Stimulus-Response Inventory of Anxiousness (Speech)/SRIA (S), and (f) Affect Adjective Checklist/AACL. The PRCS is a scale that has 30 statements about attitudes towards public speaking that the person either marks as true (the statement reflects the subject‘s attitude) or false (the statement does ppt reflect the subject's attitude). A score is computed between 0 and 30 by counting true responses as one and false responses as zero. The FT is a one item scale (rated 1 to 10) on which a subject marks the level of fear cuased by public speaking. The AD is a semantic differential that has public speaking as a stimulus and 31 five point scales anchored by bipolar adjectival pairs. A score is computed by summing the scores of the pairs. The PRCA has a subject rate 20 items using a scale of one to five how much they agree with a particular statement concerning communication anxiety. A score is 31 computed by summing the responses to the 20 statements. The SRIA (S) has the subject rate on 13 one to five scale how likely they are to experience certain reactions (faster heart beat, loose bowels, exhilaration) to giving a speech before a large group. The subject is scored by summing the responses to these 13 items. The AACL is list of adjectives (minimally 100) that the subject is asked to read and mark the descriptors that apply to the subject's attitude regarding public speaking. The subject receives a score based on the number and content of the adjectives marked. For each scale, the observed variance of the effect size was less than that expected by sampling error. This indicates that the observed mean correlations are not based on a heterogenous sample of correlations. See Table 4 for a summary of the data on the correlations for the scales. However, the scales seem to break down into two clusters, one cluster with the PRCS, PRCA, and AD, and another cluster with the FT, SRIA (S), and AACL. This apparent difference between scalse may have been the result of scales being confounded with treatment type. If the PRCS, PRCS, and AD occur most often in skills studies, then the correlation will be smaller ppt because of some aspect of the scale but rather because of the scale's use in studies using treatments with a smaller effect size. This confounding effect was demonstrated by breaking down the scales by type of treatment the subject received. The breakdown by type of treatment indicates shows that the scales with small effect sizes (PRCS, PRCA, AD) have a higher percentage of subjects in the skills treatment than the other scales with larger 32 effect sizes (FT, SRIA (S), AACL). See Table 5. The higher percentage demonstrates that the observed difference between scales can be explained on the basis of a confounding variable (treatment type). Observer Rating Techniques Two types of observer rating techniques had been used at least ten times and were compared to each other. The two techniques of observer ratings were: (a) the Behavioral Checklist/BC, and (b) the Anxiety Scale/AS. The BC is a checklist used by observers to rate the presence or absence on a one to five point scale of certain behaviors associated with nervousness in public speaking (e.g., scratching, rubbing, etc.). The subject is scored for each behavior and a total score is computed by summing the scores for the individual behaviors. The AS is simply a single rating made by the observer on a one to seven point scale of the nervousness of the speaker. For the two scales separately, the observed variance of the effect sizes was less than that predicted by sampling error. See Table 6 for a summary of the results. The BC correlation (1 = .255) and the AS correlation (.262) did not differ from each other significantly (3 = .16, g:?24,'B>.05). These results indicate that the type of observer rating method used did not influence the estimation of the effect size. Physiological Measurement Techniques Two types of physiological measurement techniques (Heart Rate/HR and Palmar Sweat/PS) were sued minimally ten times and were compared to each other. Heart Rate is simply measuring the pulse of the subject both pre and post treatment and observing the difference between the 33 pulse rates. Palmar Sweat measures the change in the amount of sweating in the palms from pretreatment to posttreatment. For both techniques separately, the observed variance of the effect sizes was less than that predicted by sampling error. For a full summary of the data see Table 7. A comparison of the Heart Rate effect size (r = .143) to the Palmar Sweat effect size (r = .218) showed significant differences between the two techniques (t = 1.88, df=61, p>.05). More research is needed examining the connection between physiological and psychological responses. Physiological measures may be ambiguous because they might measure both postive excitement as well as fear. The problem is that two types of people could exhibit high levels of physiological arousal: (a) those people who love excitement and (b) those who are anxious and afraid. One solution would be to get two different groups meeting these criteria (e.g., the thrillseekers and the afraid) and investigate what effect a communication apprehension treatment would have on the level of should show little if any decline in physiological arousal after treatment. The anxious group should decline in their level of arousal (unless they decide public speaking is a form of thrillseeking). Setting of Treatment A comparison was made between the three types of treatment setting: (a) Therapy, (b) Classroom, and (c) Combination of Therapy and Classroom. All studies were coded by the type of setting and effect sizes estimated for each setting. A full summary of the results is found in Table 8. Each setting had an effect size with less 34 observed variance in the individual correlations than that expected by sampling error. The three settings were compared using t tests. The results of the t tests indicate that the therapy/classroom combination have significantly higher (p>.05) correlations than the classroom setting alone. No significant difference was found between therapy and the therapy/classroom combination. Unfortunately, the setting of a particular treatment is almost perfectly confounded with the type of treatment. The classroom setting was almost exclusively used for skills training. The therapy setting was used for systematic desensitization, cognitive modification, and SD+CM. The combination of therapy and classroom setting was used for SK+SD, SK+CM, SK+CM+SD. The important feature to note is that for,no setting was the effect size negative, all settings had positive effect sizes. Testing Therapy Additivity One explanation for the differences in effect sizes among the treatment subgrouping is that the effects of each therapy is additive. That is, the positive benefits of one therapy can be added to the positive benefits of another therapy. To test this explanation the following provisions were made. First, a common measurement technique was needed since differences had been observed among the various measurement techniques. Self report measures were chosen because of their extensive use. Second, the model assumes that the effects of systematic desensitization and cognitive modification will only be 50% as effective in combinatorial treatments than when used singly. The reason for this reduced effectiveness is that experiments using the 35 combinatorial therapies used the same amount of time for a combinatorial therapy as was used for a single therapy. This means that the subject in a combinatorial treatment only received one-half of the time in any given therapy that a person in a single therapy received. For example, in experiments using 8 hours of therapy, the systematic desensitization and the cognitive modification therapies when used alone devoted all 8 hours to the single therapy. The systematic desensitization/cognitive modification combinatorial therapy was 8 hours long with 4 hours devoted to each therapy. The effects of any therapy are probably a nonlinear function of time across a long time interval; after a certain number of hours diminishing returns become apparent. However, given the relatively short time involved in these studies, linearity was assumed. Skill treatment is almost always done is the classroom and when used in combination with the other therapies is used in addition to them. Thus the amount of time spent in skills training is not reduced when used in combination. The effect size for each method was estimated from the data on studies using single therapies. These are the treatment groups with the largest number of studies. The treatment effect correlations were recomputed for studies using only self report measures. Table 9 presents the meta-analysis for each single and combinatorial treatment. For each treatment type separately, the observed variance of the effect size was less than the expected variance due to sampling error. The correlations for the single treatments (CM, SD, SK) were used 36 to predict expected correlations for the combinatorial treatments. The expected correlations for the SD+SK and CM+SK treatments represent simple addition. The expected correlation for the SD+CM treatment represents one-half the SD treatment correlation added to one-half the CM correlation. This division is justified because the studies using this treatment divided time equally between the techniques where other single therapies alloted all the time to one therapy. The CM+SD+SK combinatorial therapy represents addition of the CM+SD therapy to the SK therapy. See Table 10 for information regarding the test of the additivity model. The observed and expected correlations for the combinatorial treatments were compared using t tests. All four t values were insignificant (p>.05). Thus, the additivity model fits the data on treatment combinations. The additivity model assumes linearity of treatment effectiveness for communication apprehension. An equal division of time between two therapies (CM and SD) would result in each therapy being only one-half as effective than if the entire time were devoted to the single therapy. The data confirmed this assumption. The next Chapter will discuss the results and provide directions for future research and practice. Specifically, the theoretical implications for the treatment types will be discussed as well as the need for more specific research. 37 Table 1 Results of Analysis by Treatment Grogpg Treatment Type Correlation N # of Studies Observed Variance Expected Variance Var. due Sampling Error Correlation N # of Studies Observed Variance Expected Variance Var. due Sampling Error .044 100% SD+SK .334 1299 30 .018 .018 100% 100% CM+SK .291 246 .006 .029 100% CM = Cognitive Modification SD = Systematic Desensitization SK = Skills 100% CM+SD+SK .505 20 2 .041 .058 100% 5m .243 142 .011 .051 100% 38 Table 2 Comparing Treatment Group Correlations Treatment Groups _1_ 2 _3 4 5 _6_ 1. CM 2. so .70 (82) 3. SK 6.52* 3,73* (65) (109) 4. CM+SD .82 .56 2.30* (27) (71) (54) 5. CM+SK .04 .54 3.80* 1.12 (27) (71) (54) (15) 5. SD+SK 1.00 1.78 6.52* 2.39* .91 (48) (92) (75) (37) (37) * * 'k * * 7. CM+SD+SK 1.71 2.83 4.63 2.79 2.77 2.07 (20) (64) (47) (9) (9) (30) Number in brackets is the degrees of freedom. * Indicates that t value is significant at p<.05. 39 Table 3 Comparing Measurement Techniques Type of technique Self Report Observer Physiolpgical Correlation .261 .212 .172 N 5857 3251 1197 # of Studies 169 123 55 Observed Variance .025 .027 .008 Expected Variance .024 .035 .037 Var. Due Sampling Error 96% 100% 100% t-tests between techniques Self Report v. Observer t 2.58* (df=290) Self Report v. Physiological t 4.05* (df=222) Observer v. Physiological t 1.67 (df=176) * Indicates that t value is significant at p<.05. 40 Table 4 Results for Self Report Scales Name of Scale PRCS FI_ ‘AQ Correlation .266 .346 .262 N 2810 683 806 # of Studies 104 38 42 Observed Variance .036 .031 .021 Expected Variance .034 .030 .043 Var. due to Sampling Error 95% 97% 100% Name of Scale PRCA SRIA (S) AACL Correlation .277 .364 .350 N 1519 548 539 # of Studies 34 26 3O Observed Variance .020 .032 .043 Expected Variance .021 .044 .046 Var. due to Sampling Error 100% 100% 100% PRCS = Personal Report of Confidence in Speaking FT = Fear Thermometer A0 = Anxiety Differential PRCA = Personal Report of Communication Apprehension SRIA (S) = Stimulus-Response Inventory of Anxiousness (Speech) AACL = Affect Adjective Checklist Self Report Scales Broken 41 Table 5 Down by Treatment Type Treatment Type 1. PRCS 2. PRCA 3. SRIA (S) 4. AD 5. AACL 1172 (42%) 663 (44%) 56 (10%) 198 (25%) 119 (22%) 130 (19%) 900 (32%) 324 (21%) 202 (37%) 317 (39%) 179 (33%) 231 (34%) (31 SK+CM 248 143 (9%) (5%) 48 75 (3%) (5%) 142 0 (26%) (0%) 67 25 (8%) (3%) 105 19 (19%) (4%) 94 113 (14%) (17%) SK+SD 321 (11%) 364 (24%) 116 (21%) 168 (21%) 93 (17%) 81 (12%) SD+CM SD+SK+CM 26 (1%) 31 (2%) 32 (6%) 31 (4%) 18 (3%) 14 (2%) 0 (0%) 14 (0%) 0 (0%) 0 (0%) 6 (1%) 20 (3%) 42 Table 6 Comparison of Observer Measurement Techniques Measurement Type .BC ‘AS Correlation .255 .262 N 2052 516 # of Studies 104 22 Observed Variance .037 .029 Expected Variance .048 .039 Var. due to Sampling Error 100% 100% t test between techniques BC v. AS t = .16 (df=124) BC Behavioral Checklist AS Anxiety Scale 43 Table 7 Comparison of Physiological Measurement Techniques Type of Measurement Correlation N # of Studies Observed Variance Expected Variance Var. due to Sampling Error t test between techniques Heart Rate v. Palmar Sweat Heart Rate Palmar Sweat .143 .218 850 553 39 24 .012 .041 .040 .045 100% 100% t = 1.88 (df=61) 44 Table 8 Comparison of Treatment Settipg Setting of Treatment Therapy Correlation .289 N 2369 # of Studies 113 Observed Variance .022 Expected Variance .040 Var. due to Sampling Error 100% t test between settings Therapy v. Classroom t = Therapy v. Therapy+Classroom t = Classroom v. Therapy+Classroom t = * Indicates that the t value is significant at p<.05. Classroom .163 3763 32 .007 .009 100% 4.50* (df=143) 1.54 (df=147) 5.83* (df= 66) Therapy+Classroom .332 972 36 .020 .029 100% 45 Table 9 Reanalysis of data by Treatment Groppg Treatment Type C_M Correlation .327 N 363 # of Studies 19 Observed Variance .038 Expected Variance .041 Var. due Sampling Error 100% 99:98 Correlation .367 N 561 # of Studies 23 Observed Variance .021 Expected Variance .032 Var. due Sampling Error 100% 100% C_Mfl .412 185 8 .007 .045 100% s_|< my. .142 .332 2858 112 33 7 .010 .015 .012 .050 100% 100% w_+s_9+s_K .505 20 2 .041 .044 100% 46 Table 10 Test of Additivity Model Individual Therapies ‘QM Correlations .327 Combinatorial Therapies CM+SK Expected Correlations .469 Expected Variances .030 Observed Correlations .412 Observed Variances .007 t test value comparing expected and observed correlations .84 df for t (14) SD+SK .443 .029 .367 .021 .81: 99 .142 .301 SD+CM CM+SD+SK .314 .456 .054 .061 .332 .505 .015 .041 .18 23 3:.- CHAPTER IV DISCUSSION The purpose of this synthesis of the literature was to quantitatively summarize the public communication apprehension literature. This chapter discusses the results of each type of subgrouping and points out possible explanations for the findings as well as methodological limitations. A final section presents recommendations for both future research and future practice. Treatment Effects The first conclusion is that all forms of treatment (cognitive modification, systematic desensitization, and skills training) were effective in reducing public communication apprehension. Consequently no discussion of whether or not a given treatment works is unnecessary. ‘All treatments successfully reduced communication apprehension. The second conclusion is that the effect of the treatments is additive. That is, the effect of one primary method of treatment (cognitive modification, systematic desensitization, skills) can be added to the effectiveness of another treatment method. This means that combinations of the treatments are more effective than single treatments. This additive effectiveness assumes that the amount of time devoted to combinatorial therapies is also additive. For example if a systematic desensitization therapy is six hours long and the 47 48 cogntive modification therapy is 6 hours long, then the combinatorial therapy should be twelve hours long to gain the additional reduction in anxiety. This quantitative review of the literature has several limitations. First, not all possible combinations of therapies have sufficient sample size for the drawing of firm conclusions. For example, this is particularly true of the combinatorial treatment combining all three primary treatments (SD+CM+SK). This combinatorial threatment had two studies with a total of 20 subjects. Furthermore, most of the other treatments had less than 1000 subjects represented. This is an important consideration because small sample size makes any test of the additive model one with low power. Deviations from the additive model would be detected only if they are extremely large. Second, the test of the model was only conducted on self report measures. The set of data for observer ratings and physiological data was not large or complete enough for a reanalysis of the data to test the additive model as was the case for the self report data. A construct validation study is needed to assess the difference between measures. Perhaps the assessment between measures could be done using meta-analysis with existing data sets. Another meta-analysis should also be conducted to test the additivity model when data becomes available for the other measurement techniques and from construct validation studies. Assessment of the Theories about Treatment The three primary treatment types (cognitive modification, systematic desensitization, and skills training) all posit 49 explanations about why the therapy works to reduce the phobia. This section will explore whether the theory for any single treatment can explain why all treatments would work to reduce the anxiety. As practiced, the operationalizations of the theories underlying the treatments may overlap or may include features of another treatment. The following sections take the perspective of one theory of treatment while examining the other treatments as practiced. This application assesses whether or not a theory of treatment can explain the results of other treatment types. For example, the first section will take the perspective of the systematic desensitization therapist and examine the practices of the other two treatments (skills and cognitive modification). The practices of the other two treatments are examined to see if the treatment processes involve elements of systematic desensitization. Following sections then examine other treatments from the perspective of cognitive modification and skills training. Systematic Desensitization Sytematic desensitization explains the problem of communication apprehension as resulting from a reinforcement history with an involuntary pairing of a stimulus (public speaking situations) with a reponse (anxiety). The therapy creates a different response to the stimulus by substituting confidence and calmness for anxiety as the affective reponse. This form of treatment was successful in the experiments reviewed. However, the cognitive modification and skills training therapies also worked to reduce public communication apprehension. This section will explore SD and/or CM treatments as 50 practiced to see if elements of systematic desensitization are present. Cognitive modification therapy takes false/irrational beliefs about public speaking and substitutes true/rational beliefs about public speaking. These cognitive changes are what should help the speaker overcome the phobia. An examination of the cognitive statements used in the therapy involve a great deal of emotional content that is repeated over and over. For example, a person might be told that when confronted with a speech situation he or she should think that, "...the event is no big deal. There won't be any serious consequences and therefore, there is no reason for me to get so nervous about it." (Trier, 1974, p107). This type of approach when repeated in session after session using the previously mentioned statement and other similar statements begin to resemble closely the typical systematic desensitization treatment where people think about public speaking and then try to relax. The focus of both treatments is to create a different affective response to a stimulus. One treatment (SD) involves using physiological relaxation while the other treatment (CM) uses psychological relaxation. The cognitive modification technique could be considered a variation of the traditional systematic desensitization treatment. The results of skills training can also be explained in terms of systematic desensitization theory. Almost all skills training has occurred in the classroom setting during a public speaking class. One of the pedagogical devices used in public speaking class to help the communicative apprehensive is the use of small classes that share 51 information and mutual support. For example, students are often required to know each other's names and talk to one another. Knowing the other classmates names is encouraged done to make the setting in the classroom seem more friendly and less threatening. After all speeches there is applause by the class led by the instructor. All criticism and comments are supposed to be made in the spirit of constructive, nonthreatening criticism that will encourage the student to improve. This comparison of skills training to systematic desensitization reveals numerous shared features. The goal of the public speaking class is to instill confidence and good speaking habits by associating success with public speaking rather than failure. The student by experiencing success with speaking rather than failure, becomes desensitized to the event. While the processes in all the experiments in treatment of public communication apprehension can be explained by the assumptions of systematic desensitization, the results of the experiments cannot. If the results could be explained only by the elements of systematic desensitization present in the treatments than the additivity model would not work. For example, if the elements of systematic desensitization were responsible for the reduction of anxiety in skills treatment, then the SK+SD would be no larger than the SD. The additivity model shows this is not the case and casts doubt on the explanation. Alternatively, 50 could explain the results if the effect of treatment is linear and substitutable, then the results would be consistent with all three treatment types and the combinations. 52 Cognitive Modification Cognitive modification explains the phobia as a belief in irrational fears that can be changed by replacing the irrational beliefs about the situation with rational beliefs about the situation. This section examines systematic desensitization and skills training to determine if the explanation for the phobia given by cognitive modification can account for the effectiveness of the other treatments. 0n the surface, systematic desensitization treatment does not appear to contain any significant aspect of cognitive modification. The systematic desensitization treatment has subjects practice relaxation in response to a threatening stimulus. This treatment does not address the beliefs regarding the phobia. Systematic desensitization (SD) does have the therapist mentioning any rational or irrational beliefs. The SD treatment is intended to control emotive reponses (usually physiological responses to the stimulus) and not cognitive beliefs. For example, a subject when experiencing anxiety is told to breathe deeply and relax the arm and leg muscles. Systematic desensitization therapists think of their technique as purely affective and not c09nitive. That is, the therapists interpret positive findings as confirming that theory. For example, however, none of these studies actually assessed the cognitive processes in the treatment. That is, therapists gathered no data on cognitive processes and hence there is no basis for excluding a cognitive modification explanation for the results. Consider the following scenario scenario. Subjects are instructed 53 to relax. In order to relax, they find that they must deal with intrusive irrational beliefs. That is, in order to relax subjects may self administer cognitive modification regarding public speaking to allow themselves to relax. Thus, the systematic desensitization treatment may contain a great deal of self administered cognitive modification. Skills training may be interpretable in terms of CM theory. Many of the public speaking texts discuss the irrational reasons people fear public speaking and suggest alternative beliefs (Ehninger, Gronbeck, and Monroe, 1984; Jefferey and Peterson, 1980, Verderber, 1985). Instructors in the classroom often talk about public communication apprehension and give reasons why those fears should not exist in this setting. This type of information contained in lectures and textbooks suggests that CM may be able to explain the results of skills training because actual classroom practices may work to change the cognitions that persons have about public speaking. Cognitive modification provides an explanation for the effectiveness of skills training and systematic desensitization. However, these explanations would still not account for the additivity model if there were nonlinearity in the data. If the other treatments were effective only because of the CM elements and if any given treatment used up the potential effect of that element, then they would not "add" to CM in a combinatorial treatment (unless the skills therapy is still linear). Alternatively, CM could explain the additivity model if the effect of treatment was linear and substitutable. Then the effect of each treatment is the result of the 54 the elements of CM present and the combinatorial treatments would be the result of adding the effect of each treatment. Skills Training Skill training assumes that the anxiety people feel about public speaking is the result of lack of confidence caused by a lack of communicative ability. The goal of the skills treatment is to train persons in the techniques of successful speaking and thus give them the confidence they need to succeed. This section will examine the other two treatments (SD and CM) to see if they contain important element of skills training. Systematic desensitization involves no skills training whatsoever. The relaxation techniques are unrelated to the level of skills possessed by the subject. The results of SD cannot be interpreted in terms of skills training assumptions. The only possible explanation could be that SD may result in self administered skills training. The person in a relaxed state may choose to spend more time in research and preparation for public speaking. The result of SD therapy may then allow for the subject to self administer skills training. At this time no evidence exists for such a claim. Cogitive modification offers no skills training as part of the treatment. The irrational beliefs affected by the therapy are irrational beliefs regarding the reasons for failure. These reasons however, are unrelated to the level of skill possessed by the people undergoing treatment. The rational beliefs used to replace the irrational beliefs involve no issues of public speaking skills. The cognitions affected by CM are attitudes held towards the action of 55 speaking in public not the cognitions regarding how a proper public speech ought to be given. CM does not involve skills training as part of the treatment. Unless the CM training results in self administered skills training. If CM were to result in a subject self administered skills training then CM may be effective not because of changing beliefs but because the changed beliefs result in some other process taking place. At this time no evidence exists for such a claim. Skills training probably is unable to explain the positive findings of SD or CM treatment. The SD and CM treatments involve no instruction in the "skills" of public communcation. Skills training may contain elements of SD and CM but CM and SD do not contain elements of skills training. Alternative Explanations for the Effectiveness of Treatments That Anxiety has Multiple Causes One alternate explanation to the three perspectives (SD, CM, Skills) is that the three treatments address three different types of motivations that a person can have for anxiety regarding public speaking. Individuals could fear public speaking because: (a) they have an anxiety response to the public speaking situation, (b) they have irrational beliefs about public speaking, or (c) they lack the skills necessary to be a successful public speaker. Support for this position is found in the additive model. The reasoning is that if the entire anxiety can be explained by only one cause than the combinatorial treatments should not observe larger effect sizes. This is because the additional treatment cannot reduce an anxiety that no longer exists. Combinatorial treatments involving 56 skills training should have no additional reduction if there is only one cause for the anxiety. Skills training is typically a 3 or 4 hour semester public speaking course. The addition of a 5 to 8 hour SD or CM treatment should not be able to reduce the anxiety much more. This could be explained alternatively in terms of skills training having little time spent on the task. While the amount of class time is great, the actual time spent on the task of public speaking may be small. If the time spent in the public speaking class is nominal time rather than therapy time, then the skills training may not be less effective. The SD and CM treatments may potentially be more effective because at the current treatment has so few hours devoted to treatment (typically less than 8). The content of the treatments may not overlap entirely and the problem may involve different types of individuals that require separate solutions. More research is needed on identifying these types of people, involving these personality types in treatments targeted at the source of the anxiety and observing what features change over time. Theoretically only those persons with irrational beliefs about public speaking should be helped by cognitive modification. Only those persons that lack skills should be helped by skills training. And finally, only those individuals that involuntarily associate anxiety with public speaking should be aided by systematic desensitization. A Distraction Hypothesis Another explanation for additivity model is the nature of anxiety. For example, suppose that communication apprehension is the result of 57 an internal feedback loop that has the speaker concentrating on the fear. As the speaker becomes more fearful the more the speaker thinks about the fear. That is, individuals may also be afraid to be afraid and this could contribute to the level of anxiety that a person feels about the situation. A solution would be to break this internal feedback loop with some type of distraction. Support for this idea can be found in common folklore regarding tense situations. Athletic teams often have pranksters that help players loosen up before a game by distracting them from thinking about the contest. Soldiers will write letters or pray before battle which comforts and distracts them from the upcoming event. The three treatments (cognitive modification, systematic desensitization, and skills training) all provide forms of distraction to divert the speaker's thoughts away from fear and into some more productive pattern. Skills training has the speaker practicing the speech, going over the outline and, in general, preparing for the presentation rather than thinking about fearing the presentation. Cognitive modification distracts the speaker into thinking about the rational rewards of a good speech and can almost create a mantra to focus concentration. Systematic desensitization has the speaker concentrating on relaxing muscles and breathing deeply rather than thinking about the speech. All three treatments could be described as providing different distractors to break the internal feedback loop. The additivity of treatments could be accounted by the susceptability of individuals to different types of distractors. Some v 58 individuals distract themselves by thinking of rewards, some by using a mantra, some by concentrating on breathing excercises, and some persons will distract themselves by practicing the speech. Impact of Measurenent Type Comparing measurement types, self report measures show the greatest reduction in anxiety. All measurement types (self report, observer, physiological) show reduction in anxiety due to treatnent. Self report measures, however, are more concerned directly with emotional states versus observer ratings which may or may not be connected with emotional states. Self report measures ask how nervous, or afraid the person feels about the situation. Observer ratings of behaviors involve observer counting of behaviors that are thought to indicate anxiety felt by the speaker. The observer may count the number of times the person crosses his or her legs or arms. Such systems of observer ratings depend on how well the coded behaviors match the anxiety felt by the speaker. This match may not always be good. For example, shifty eyes may be an individual phenomenon and not the result of fear. A cracked voice may be the natural voice of a speaker and not the result of anxiety. Physiological measures are proabably just an unreliable indicator of an emotional state. However, even the physiological measures show that treatment will reduce anxiety. The only differences observed at the level of individual scales was observed among the self report measures. This difference, however, was the result of a confounding with treatment type. The scales with the largest effect sizes were used most often in the combinatorial treatments, therefore the effect 59 sizes should be larger. Effect of Treatment Setting All settings were effective in treating communication apprehension. This provides evidence to lay to rest Miller's (1984) concern that classroom settings are ineffective or even counterproductive in treating communication apprehension. The superiority of therapy and theraDY/Classroom can be explained due to the additivity of treatments and the superiority of systematic desensitization/cognitive modification. Skills training is conducted almost exclusively in the classroom and has a smaller effect than systematic desensitization and cognitive modification. Both of those therapies are usually set in the laboratory or clinical setting. The therapy/classroom combination by definition has combinatorial treatments and is therefore more effective because more treatment is offered. Recommendations Directions for Future Research First, this study dictates a need for a construct validity study to assess the relations among the various measures of communication apprehension. This is important because better and more standard measurement would make the assessment of treatent effectiveness more accurate. Second, more studies are needed that involve the use of combinatorial treatments, especially the cognitive modification, systematic desensitization, and skills training combinatorial treatment. At the current time, the estimates regarding those 6O particular types of treatment do not rely on a large number of studies. More data would make testing competing models possible and contribute to more power when analyzing the additivity model proposed here. Third, an examination is needed to uncover what treatment actually changes. Paul (1966) concluded that systematic desensitization works to reduce public speaking anxiety but claimed that explanations about why the treatment works have little data. More then twenty years later there still has been little, if any work providing evidence for an explanation. The conclusion of this synthesis of the literature is that treatment works but there is little data to test why the treatments work. Thus, this report advances the state of knowledge regarding the effectiveness of treatment but unfortunately is not able to address the issue of why the treatments work. Fourth, experiments examining the various lengths of time for treatments and combinations of treatments to test the assumption of linearity should be undertaken. This information is of practical value since minimum lengths of time could be established for treatments. Of particular concern is at what point the diminishing return for additional time is outweighed by the cost in resources of continuing treatment. If the effectiveness diminishes over time, then public speaking courses could more profitably use time in systematic desensitization or cognitive modification than in skills training. Advice for Treatment of Public Communication Apprehension If possible, treatment for communication apprehension should use the widest possible combination of methods. Public speaking classes 61 should probably include an in-class form of systematic desensitization and/or cognitive modification, especially for the highly apprehensive. The issue will depend upon resource availability and other pedagogical concerns of the instructor. This synthesis of the literature does provide strong evidence that the current treatment methods are beneficial and great confidence should be placed in the ability of the practioner to reduce public communication apprehension using any of the methods currently available. This is not to say that all issues have been answered. This report does not recommend the length of treatment necessary or whether some characteristic of the practioner (like training) can improve the results of treatment. This report does not provide evidence why any of the treatments work, only that the seven treatment types all successfully reduce anxiety. The results, therefore, should be taken as a starting place for improvement in both theory and practice. Current practices are successful and should obviously be continued. Future research can uncover the reasons that treatments work and seek to improve treatment. The advice for the practioner is to practice confidently knowing that the efforts to decrease public communication apprehension are not in vain. Future research can only improve what is already a healthy start. APPENDICES I APPENDIX A APPENDIX A CORRECTING FOR REGRESSION TO THE MEAN Many methodologists have commented on the problem of regression to the mean in experimental design. The comments have generally involved qualitative assessments of the phenomena's occurence with no recommendations about how to quantitatively assess or correct for the problem. The term regression to the mean refers to the problem of using a measurement instrument to select a sample and then assess a change at some future point with the same instrument. The sample that is selected based on some minimum score will at a later date have scores (as measured by the selection instrument) that regress to the population mean. This change in scores will occur in groups that have been exposed to no treatment or other typical experimental effects (history, maturation, contamination, etc.,) that would explain such a change. For example, suppose an experimenter wants a group of individuals that are afraid of snakes. The experimenter gives a group of individuals a self-report measure to select the 10% of the sample most afraid of snakes. If the experimenter were to select these individuals and do nothing for three weeks and then give them the self-report questionnaire the group would have appeared to become less afraid of snakes. Assuming there was no exposure to therapy explain the 62 63 difference or any other history, maturation, or event that should change the score, the explanation would be traced to regression to the mean. This is the tendency of a samples initial mean score over time to become less distant from the population mean. This appendix will explain the effect that regression to the mean has in overestimating the effect size in an experiment will be explained and a correction suggested for this effect. Two different conditions will be explored, a situation with no real change in the group mean and a situation with real change in the group mean. A number of assumptions will be made about the data that is being analyzed. First, the data will be assumed to be normally distributed so that x is N(O,l). Second, when a cutoff score for selection (c) is chosen, like choosing the highest 10% of the sample scores, the sample mean for the scores greater or equal to the cutoff score can be expressed as E(x/x3c). The value of the mean of the sample, E(x/x3c) chosen on the basis of a cutoff score (c) is equal to oc/p where p is equal to p(x2c) and o(x) assumes the normal density function. These assumptions are standard for most selection methods and measurement techniques as well as robust to violations of these assumptions. Assume No Real Change in Sample Mean This situation will be examined under two different conditions, perfect test-retest reliability and less than perfect test-retest reliability. The score for the mean of the sample at time one is: X1 = sigmaX (gc/p) +‘X This score is the standard deviation for the p0pulation (sigmax) multiplied by the sample mean (as expressed in standard units) added 64 to the population mean. Suppose the measurement device had a test-retest reliability of 1.0, then the first and second scores would show no regression towards the mean. This is because the score for the mean of the sample at time two is: 72 = r (sigmax) (ta/p) +‘x‘ When the reliability (r) is 1.0 the difference between the score at time two (Xé) and the total population mean (X) is equal to the difference between the score at time one (X1) and the total population mean (7) multiplied by the reliability. When the reliability is 1.0, the difference between the sample mean and the population mean at time two will the same as the difference between the sample mean and the population mean at time one. Any change that is observed in individual scores is random and the sum of the random errors should be zero which means no change will be observed at the group level (the sample mean). Suppose now that the test-retest correlation is less than 1.0. The score for the individual at time one remains the same. But the score at time two, however, will not be the same. The value of the test-retest correlation if not 1.0, will always be less than 1.0. This mandates that the distance between the population mean and the sample mean at time one (Xi) will be smaller than the distance between the population mean and the sample mean at time two. The distance between the sample mean at time two and the population mean if multiplied by a number less than 1.0 will reduce the distance between the sample mean at time two (X2) and the population mean (X). This explanation shows that regression to the mean is a function 65 of two values, the.test-retest correlation and the distance between the sample mean and the population mean. The smaller the test-retest correlation the smaller the portion of distance between the population and sample mean that is retained at time two (Xé). The larger the distance between the sample mean and the population mean (X) at time one (X1) the larger the regression that will occur (assuming the test-retest correlation is not equal to 1.0). Assume Change in Sample Mean The score at time one will be the same as mentioned above: X1 = sigmaX (dc/p) + X. The difference in this section will be that the mean at time two (Xi) will be different than the mean at time one (Xi). Within the context of therapy this means that the sample mean is moving towards the population mean. Such a change indicates that the therapy is reducing the difference that exists between the sample and the population. Unfortunately this reduction can be overestimated if the reduction includes regression to the mean. As demonstrated in the case where no change occurs, the sample mean will regress to the population mean. The standard score at time two is: 72 = r (sigmax) (sic/p) +Y+ d This is true where d is equal to the amount of change. Regression to the mean does not occur when the test-retest correlation is 1.0. The score at time at time two (X2) if the test-retest correlation is equal to 1.0, the score at time two will only show change equal to the value of the change score (d). However, when the value of the test-retest correlation is not 1.0 than the 'w 66 score will change by the value of d plus the value of the regression to the mean. This means that the observed change score, unless corrected for regression to the mean, will be larger than the real change score. The implication is that failure to correct for this problem can lead to conclusions about the magnitude of change that systematically overestimate the effectiveness of treatments. Example of Regression to the Mean Suppose the therapist has gathered a sample that tests in the top 10% of the population (dc/p = 1.76 under these conditions) with regards to fearing snakes. The therapist takes the sample and uses systematic desensitization to reduce the level of fear the patients feel about being near snakes. The scale used has a population mean of 50 and a standard deviation of 10. The scale had a test-retest reliability of .90. Substituting these values produces the following for the mean value of the sample at time one: 71 = 10 (1.76) + 50 = 67.6 At time two the observed value for the sample was 57.6. This shows an observed difference between time one and time two of ten points on the scale or one standard deviation. Substitution for the values at time two produces the following: 57.6 = .90 (10) (1.76) + 50 + d This equation reduces to: -8.24 = d This value shows that the change score was inflated by 21% because of regression to the mean. Had the test-retest reliability been perfect the real change score would have been equal to the observed 67 change score. In most experiments the significance tests are conducted on the observed change score rather than the real change score. This increases the probability of Type I error, concluding a significant change has occurred when no significant change has occurred. Not enough is known about the distribution of the corrected change score to suggest corrections to the significance test. The confidence interval could be corrected by correcting both end points of the interval for regression to the mean. Conclusion Regression to the mean can is a problem whenever a selection score is used to divide or choose a sample and the measurement device has a test-retest correlation less than 1.0. This essay suggests a possible correction for the effects of regression to the mean. The best method for correction however, is prevention. If a different scale is used for sample selection than is used for measurement at time one the whole problem can be avoided. This solution is preferable because it avoids the effect of increasing sampling error inherent in the correction formula. When the effect cannot be prevented than the systematic effect of regression to the mean can be corrected. The correction will provides greater accuracy in estimating the effect size in any experiment. APPENDIX B APPENDIX B INACCESSIBLE MANUSCRIPTS The following manuscripts were not included in the analysis. The manuscripts were not contained in the Michigan State University or the Northwestern University library. The material was not available for interlibrary loan. The University Microforms International did not have the manuscripts available for sale. The universities granting the degrees when contacted could not provide access to the manuscript. Algirdas, F. (1972). Self-instructional and relaxation variable in the systematic desensitiZation treatment of speeCh anxiety. UnpubliShed doctoraT dissertation, Waterloo University, Canada. Garrison, K. (1978). The effect of cognitive modification on communication apprehension in children. Unpublished master's the§is. University of Nebraska, Lincoln, Nebraska. Reid, J. (1978). An investigation of heirarchy properties in systematic desensitizatibn. unpublished d0ctoral dissertation, University of Victoria,TCanada. 68 APPENDIX C APPENDIX C STUDIES LACKING STATISTICAL INFORMATION The following manuscripts were not used in the analysis because the published manuscripts lacked sufficient information to estimate an effect size. Hekmat, H., Lubitz, R., and Deal, R. (1984). Semantic desensitization: A paradigmatic intervention approach to anxiety disorders. Journal of Clinical Psychology, 49, 463-466. Lent, R., Russell, R., and Zomostny, K. (1981). Comparison of cue-controlled desensitization: Rational restructuring, and a credible placebo in the treatment of speech anxiety. Journal of Consulting and Clinical Psychology, 42, 608-610. Zettle, R. and Hayes, S. (1983). Effect of social context on the impact of coping self-statements. Psycholpgical Reports, 52, 391-401. 69 APPENDIX D APPENDIX D TEST-RETEST CORRELATIONS FOR MEASUREMENT TECHNIQUES Technique Test-Retest Correlation Personal Report of Confidence in Speaking .92 Personal Report of Communication Apprehension .81 Stimulus-Response Inventory of Anxiety (Speech) .60 Affective Adjective Checklist .68 Behavioral Checklist .80 Anxiety Scale .72 Anxiety Differential .64 Heart Rate .86 Q-Sort .90 Palmar Sweat Print .63 Overall Anxiety Rating (observer) .79 Subjective Units of Disturbance .93 Speech Anxiety Inventory-Trait .79 Speech Anxiety Inventory-State .76 Speech Attitude Survey .68 Duration of Silence .96 Speech Appraisal Survey .89 Interaction Behavior Measure .84 Propensity for Verbal Behavior .84 Behavioral Assessment of Speech Anxiety .80 Fear Survey Schedule-Speech Anxiety .80 Measure of Elementary Communication Apprehension .85 Fear Thermometer .72 Checklist of Appropriate Speaking Behaviors .80 Time of Speech .78 Self Efficacy Measure .62 Speech Disruption Checklist .81 Personal Report of Public Speaking Aprrehension .84 Word Count of Speech .70 Public Speaking Fear Survey .90 Social Fear Scale .96 Lomas Verbal Report Form .73 Combination Public Speaking Inventory and PRCS .97 Rating Scales .76 Compilation of Behavioral Ratings .81 Number of "ah" statement .85 Speech Performance Survey .84 Public Speaking Anxiety Inventory .77 Speech Composite Index .94 Unwillingness to Communicate .83 Reticence Scale .66 7O APPENDIX E APPENDIX E COMPLETE SUMMARY OF DATA Manuscripts are listed by first author and year of publication. The following abbreviations are used: Treatment Type Systematic Desensitization (SD) Cognitive Modification (CM) Skills training (SK) Setting of Experiment Therapy (T) Classroom (C) Measurement Type Overall Average (0A) Average for Self Reports (ASR Average for Observer Ratings (ABR Average for Physiological Reactions (APR Self Report (SR) Behavioral Rating (BR) Physiological Reaction (PR) 71 72 Particular Measurement Device Personal Report of Confidence in Speaking (PRCS) Behavioral Checklist (BC) Personal Report of Communication Apprehension (PRCA) Stimulus-Response Inventory of Anxiousness-Speech (SRIA-S) Affective Adjective Checklist (AACL) Anxiety Scale (observer) (AS) Lomas Verbal Report Form (LVRF) Anxiety Differential (AD) Heart Rate (HR) Q-Sort (QS) Palmar Sweat (PS) Combination Public Speaking Inventory and PRCS (PSI/PRCS) Speech Attitude Scale (SAS) Silence Duration (SD) Speech Appraisal Survey (SAPS) Rating Scales (RS) Speech Anxiety Inventory-~State Scale (SAI-S) Speech Anxiety Inventory-~Trait Scale (SAI-T) Interaction Behavior Measure (IBM) Propensity for Verbal Behavior (PVB) Behavioral Assessment of Speaking Anxiety (BASA) Fear Thermometer (FT) Fear Survey Schedule—-Speech (FSS-S) Measure of Elementary Communication Apprehension (MECA) Subjective Units of Distrubance Scale (SUDS) Checklist of Appropriate Speaking Behaviors (CASB) Compilation of Behavioral Ratings (CBR) Number of Seconds in Speech (SS) Self Efficasy Measure (SEM) Public Speaking Fear Survey (PSFS) Speech Disruption Checklist (SDC) Speech Experience Inventory (SEI) Personal Report of Public Speaking Apprehension (PRSPA) Count of Words in Speech (WC) Number of "ah" Statements in Speech (NA) Speech Performance Scale (SPS) Public Speaking Anxiety Inventory (PSAI) Speech Composite Index (SCI) Unwillingness to Communicate Scale (UCS) Reticence Scale (RES) Author Akin Altmaier Ayres Benton Biggers Borin Borkovec Brooks Calef Casas Chaplin Connell Cradock Deffenbacher Deffenbacher Year 1974 1982 1985 1974 1987 1949 1979 1968 1970 1975 1981 1987 1977 1980 1977 11 100 42 37 14 14 73 Treatment Measurement Type/Setting Type Device SD/T SR PRCS CM/T BR BC SD/T BR BC SD+CM/T BR BC SK/C SR PRCA SK+SD/C SR PRCA CM/T 0A ASR ABR SR PRCS SR SRIA (S) SR AACL BR BC BR AS SK/C SR PRCA SK/C SR LVRF SK+SD/C+T SR LVRF SD/T OA SR AD PR HR SK/C SR 05 SD/T SR PRCS CM/T OA ABR APR SR PRCS BR BC PR HR PR PS BR AS SD/T OA ABR APR SR PRCS BR BC PR HR PR PS BR AS SD/T SR PRCS CM+SK/C+C SR PRCA SK/C SR PRCA SD/T SR PRCS CM/T SR PRCS SD/T SR PRCS SD/T SR PRCA Effect Size .198 .196 .259 .291 .147 .260 .200 .175 .237 .135 .133 .257 .260 .214 .141 .319 .668 .167 .191 .143 .244 .129 .217 .245 .165 .245 .105 .038 .291 .404 .178 .124 .183 .279 .065 -.047 .412 .183 .363 .207 .247 .629 .873 .188 .268 74 Treatment Measurement Effect Author Year _N Type/Setting Type Device Size Devine 1974 33 CM/T 0A .472 SR PSI/PRCS .477 BR BC .467 Ertle 1969 32 SD+SK/T+C SR PRCA .734 16 SK/C SR PRCA .144 Ewing 1944 200 SK/C SR SAS .223 Fremouw 1975 10 SK/C 0A .076 ASR -.O63 ABR .215 SR PRCS -.012 SR AD -.113 BR BC .237 BR AS .192 31 SD+SK/T+C 0A .551 ASR .492 ABR .610 SR PRCS .529 SR AD .454 BR BC .546 BR AS .674 Fremouw 1978 12 SK/T+C 0A .361 ASR .416 ABR .425 SR PRCS .584 SR PRCA .617 BR BC .424 BR AS .425 PR SD .066 SR AD .048 12 SD+SK/T+C 0A .197 ASR .197 ABR .255 SR PRCS .177 SR PRCA .234 BR BC .258 BR AS .253 PR SD .081 SR AD .179 12 SK/C 0A .085 ASR .016 ABR .166 SR PRCS .012 SR PRCA .011 BR BC .162 BR AS .169 PR SD .131 SR AD .025 Author Garrett Gatchel Gatchel Gatchel Germer Giffin Year 1954 1979 1977 1976 1975 1969 46 27 40 36 177 75 Treatment Measurement Type/Settipg Type Device SK/C ASR SR PRCS SR SAPS SD/T OA SR PRCS BR BC SD/T OA APR SR RS BR BC PR HR PR PS SD/T ABR BR BC BR AS SD/T 0A ASR BR BC PR HR SR SAI-S SR SAI-T SR PRCS SR PRPSA SR PRCA SR AD CM/T OA ASR BR BC PR HR SR SAI-S SR SAI-T SR PRCS SR PRPSA SR PRCA SR AD SD+CM/T+T 0A ASR BR BC PR HR SR SAI-S SR SAI-T SR PRCS SR PRPSA SR PRCA SR AD SK/C SR PRCS Effect Size .128 .062 .193 .184 .282 .086 .256 .311 .140 .254 .391 .231 .200 .140 .261 .177 .239 -.031 .010 .201 .130 .330 .268 .316 .188 .235 .253 .118 .246 .226 .149 .183 .122 .666 .171 .398 .467 .091 .290 .452 .347 .680 .345 .655 .322 .250 Author Glogower Goldfried Goldfried Goss Grande Grayson Gross Year 1978 1977 1974 1978 1975 1978 1982 12 35 28 27 17 16 20 49 26 76 Treatment Measurement Type/Setting Type Device SD+SK/T+C ASR SR PRCA SR IBM CM+SK/T+C ASR SR PRCA SR IBM SD/T OA ASR APR BR BC PR SD PR HR SR AD SR PRCS SR SRIA-S SD/T 0A ASR BR BC PR SD SR AD SR PRCS SR SRIA-S SD+SK/T+C SR PRCA SD+SK/T+C OA ASR SR PVB SR PRCS SR BASA SD/T OA ASR ABR APR SR FT SR FSS-S BR BC BR AS PR HR PR PS SD/T PR HR SD/T OA ABR APR PR HR PR PS BR BC BR AS SR PRCS Effect Size .284 .446 .121 .445 .561 .339 .309 .356 .301 .185 .240 .361 .396 .199 .474 .206 .256 .165 .100 .366 .153 .248 .338 .242 .309 .260 .357 .108 .260 .313 .308 .194 .327 .298 .380 .236 .168 .206 .172 .171 .100 .184 .188 .180 .118 .083 .287 77 Treatment Measurement Effect Author Year _N Type/Setting Type Device Size Gross 1982 23 CM/T 0A .176 ABR .108 APR .068 PR HR .093 PR PS .043 BR BC .121 BR AS .094 SR PRCS .528 Harris 1981/1982 33 SD+CM/T+T SR MECA .271 Hayes 1984 14 CM/T 0A .221 ASR .403 ABR -.051 SR SUDS .380 SR FT .482 SR PRCA .346 BR BC -.151 BR CASB .050 14 SD+CM/T+T 0A .414 ASR .456 ABR .353 SR SUDS .361 SR FT .625 SR PRCA .381 BR BC .327 BR CASB .378 14 CM+SK+SD/T+T+T 0A .689 ASR .616 ABR .798 SR SUDS .540 SR FT .716 SR PRCA .591 BR BC .620 BR CASB .926 14 SK/T 0A .629 ASR .590 ABR .689 SR SUDS .617 SR FT .571 SR PRCA .581 BR BC .606 BR CASB .771 78 Treatment Measurement Effect Author Year _N Type/Settipg Type Device Size Hayes 1984 14 SD/T 0A .425 ASR .420 ABR .434 SR SUDS .394 SR FT .424 SR PRCA .441 BR BC .151 BR CASB .717 14 SK+SD/T+T 0A .381 ASR .380 ABR .384 SR SUDS .328 SR FT .370 SR PRCA .442 BR BC .450 BR CASB .317 14 SK/T 0A .642 ASR .374 ABR .855 SR SUDS .356 SR FT .392 BR BC .890 BR CASB .820 Hayworth 1970 840 SK/C BR CBR .138 Hekmat 1985 10 SD/T 0A .183 ASR .185 SR PRCS .092 SR AACL .143 SR SRIA-S .321 BR BC .176 Hemme 1976 18 SD/T 0A .162 APR .119 SR PRCS .154 BR BC .255 PR FS .141 PR SS .096 Henrikson 1943 205 SK/C SR RS .339 Horne 1974 13 SD+SK/T+C 0A .287 SR PRCS .581 PR HR .020 BR BC .260 5 SK/T+C 0A .315 SR PRCS .700 PR HR .219 BR BC .026 79 Treatment Measurement Effect Author Year _N Type/Setting, Type Device Size Horne 1974 7 SK/C 0A .162 SR PRCS .446 PR HR —.O32 BR BC .071 22 CM+SK 0A .300 SR PRCS .591 PR HR .145 BR BC .163 Jaremko 1980 37 SD+E/T+C 0A .250 ASR .278 SR AACL .343 SR SEM .213 BR BASA .194 25 E/C 0A .083 ASR .004 SR AACL .029 SR SEM -.022 BR BASA .242 Jaremko & 1980 6 SD+SK/T+C 0A .185 Hadfield ASR .110 SR AACL .179 SR FT .041 BR BASA .334 9 CM+SK/T+C 0A .422 ASR .407 SR AACL .547 SR FT .266 BR BASA .453 6 CM+SK+SD/T+C+T 0A .245 ASR .246 SR AACL .393 SR FT .099 BR BASA .242 10 SK/T 0A .161 ASR .039 SR AACL .105 SR FT -.027 BR BASA .405 Jaremko 1973 10 SD/T SR PSFS .283 Jarmon 1972 16 CM/T 0A .101 ASR .138 ABR -.012 SR PRCS .008 SR FT .419 BR AS -.011 BR SDC -.012 Author Jarmon Johnson Kanter Karst Katz Kirsch Kirsch Year 1972 1971 1979 1970 1976 1977 1975 Kleinsasser 1968 E 13 16 19 16 94 33 47 3O 80 Treatment Type/Settipg SD/T SD/T SK/T CM/T SD/T SD+CM/T+T CM/T SK/C SD/T SD+E/T+C SD+SK/T+C Measurement Type Device OA ASR ABR SR PRCS SR FT BR AS BR SDC SR PRCS SR PRCS 0A ASR SR SRIA-S BR BC PR HR SR AD OA ASR SR SRIA-S BR BC PR HR SR AD OA ASR SR SRIA-S BR BC PR HR SR AD OA ASR SR PRCS SR FT BR SDC SR PRCS OA ASR BR BC SR PRCS SR AD 0A SR PRCS BR BC 0A ASR SR AD BR BC SR PRCS SR SRIA-S Effect Size .126 .090 .147 .182 .028 .234 .059 .066 .077 .284 .387 .369 .137 .225 .404 .265 .400 .418 .110 .151 .384 .305 .475 .527 .194 .077 .423 .269 .425 .302 .549 -.044 .056 .239 .190 .338 .106 .274 .379 .315 .442 .284 .363 .344 .046 .382 .364 Author Kleinsasser 1968 Krugman Lamb LeTendre Leyden Lieng Lima Little Littlefield 1987 1975 Lohr Year 1985 1965 1977 1941 1976 1975 1976 H 10 37 24 66 22 205 24 81 Treatment Measurement Type/Settipg Type Device SK/C OA ASR SR AD BR BC SR PRCS SR SRIA-S SD/T OA ASR SR PRCS BR BC SR FT SD+SK/T+C 0A SR PRCS BR BC SD/T BR BC SK/T 0A ASR SR SAS SR SEI BR AS SD/T 0A ASR SR PRCS SR AD SR SRIA-S BR BC SD/T ASR SR FSS-S SR AD SR PRCS SD+SK/T+C 0A ASR SR PRCS SR PRPSA BR BC SK/C OA ASR SR PRCS SR PRPSA BR BC SK/C SR PRSPA SD+SK/T+C SR PRCA Effect Size -.003 .064 .211 -.202 .117 —.137 .219 .247 .237 .163 .256 .173 .229 .117 .057 .089 .119 .163 .074 .030 .415 .424 .370 .490 .413 .385 .499 .901 .035 .562 .321 .364 .388 .340 .238 .273 .240 .286 .194 .340 .086 .281 82 Treatment Measurement Effect Author Year _N Type/Setting Type Device Size Longo 1984 38 SD+SK/T+C 0A .243 ASR .242 APR .287 SR PRCS .427 PR PS .421 BR BC .155 SR AD .058 PR HR .153 19 SK/C 0A -.054 ASR .008 APR -.028 SR PRCS -.010 PR PS —.120 BR BC -.230 SR AD .026 PR HR .064 Lynd 1976 18 SD/T 0A .245 ASR .363 APR .126 SR PRCS .347 SR AD .380 PR HR .209 PR SS .044 Mannion 1984 64 SD/T 0A .305 ASR .410 SR PRCS .440 SR SUDS .380 PR HR .095 Marshall 1982 5 SK/T 0A .414 ASR .486 ABR .367 BR CASB .332 SR SUDS .411 BR BC .384 SR FT .561 7 SD/T 0A .397 ASR .434 ABR .372 BR CASB .132 SR SUDS .304 BR BC .473 SR FT .565 7 SD+SK/T+T 0A .510 ASR .404 ABR .581 BR CASB .397 SR SUDS .307 BR BC .674 SR FT .501 Author Year .fl Marshall 1982 24 Marshall 1976 31 Marshall 1977 6 6 6 McCroskey 1972 435 McCroskey 1970 24 24 McKinney 1982 14 28 McManus 1975-1976 24 McSweeny 1975 27 83 Treatment Measurement Type/Settipg Type Device SK/T OA ASR ABR BR CASB SR SUDS BR BC SR FT SD/T 0A ASR SR FT SR SUDS BR BC SD/T 0A ASR SR FT SR SUDS BR BC SK/T OA ASR SR FT SR SUDS BR BC SK+SD/T+T OA ASR SR FT SR SUDS BR BC SD+SK/C+C SR PRCA SD+SK/C+C SR PRCS SK/C SR PRCS SD/T OA APR PR HR PR PS SR PRCS SD+SK/T+C OA APR PR HR PS PRCS SD+SK/T+C SR PRCA SK/C SR PRCA SD/T OA ASR BR BC SR PRCS SR AACL SR AD Effect Size .268 .261 .273 .299 .310 .260 .212 .238 .186 .275 .097 .343 .816 .817 .863 .772 .813 .567 .388 .448 .329 .925 .883 .924 .985 .863 .801 .285 .626 .240 .370 .212 .233 .190 .688 .318 .222 .177 .267 .511 .129 .000 .174 .218 .044 .145 .298 .210 84 Treatment Measurement Effect Author Year _N Type/Setting Type Device Size Meichenbaum 1971 11 SD/T 0A .374 ASR .534 ABR .377 APR .131 BR BC .647 SR PRCS .603 SR AACL .739 SR AD .261 PR WC .165 PR SD .097 BR NA .107 11 CM/T 0A .395 ASR .565 ABR .371 APR .282 BR BC .656 SR PRCS .724 SR AACL .642 SR AD .328 PR WC .236 PR SD .095 BR NA .086 10 CM+SD/T+T 0A .259 ASR .388 ABR .170 APR .154 BR BC .408 SR PRCS .432 SR AACL .425 SR AD .307 PR WC .207 PR SD .095 BR NA .086 10 SK/T 0A .144 ASR .191 ABR .063 APR .154 BR BC .408 SR PRCS .095 SR AACL .322 SR AD .155 PR WC .083 PR SD .225 BR NA -.282 85 Treatment Measurement Effect Author Year _N Type/Setting Type Device Size Morey 1973 30 CM+SK/T+C 0A .280 ASR .363 ABR .195 SR FT .234 SR PRCS .493 BR BC .109 BR AS .282 Morgan 1970 30 SK/C+T 0A .447 ASR .412 SR PRCS .466 SR AACL .358 BR BC .518 Morley 1974 30 CM+SK/T+C 0A .275 ASR .344 ABR .203 SR FT .345 SR PRCS .344 BR BC .111 BR AS .298 Mulac 1974 108 SK/C SR SPS .081 Mylar 1972 26 SD/T ASR .379 SR PRCS .142 SR SUDS .617 Nichols 1969 38 SD+SK/T+C SR PRCA .523 Nicolleti 1972 20 SD/T SR PSAI .460 Norman 1975 12 SD+CM/T+T 0A .359 ASR .503 APR .213 SR AD .143 PR HR .163 SR PRCS .759 SR SRIA-S .608 BR BC .217 PR PS .263 12 CM/T 0A .302 ASR .403 APR .211 SR AD -.014 PR HR .159 SR PRCS .631 SR SRIA-S .591 BR BC .181 PR PS .264 Author Year Norman 1975 Osberg 1981 Paul 1966 (article) Paul 1966 (book) Paul & Shannon 1966 Paulson 1951 Richter 1974 Robinson 1955 12 45 25 18 3O 15 37 86 Treatment Measurement Type/Setting, Type Device SD/T 0A ASR APR SR AD PR HR SR PRCS SR SRIA-S BR BC PR PS SD/T 0A ASR SR PRCS SR SRIA-S BR BC PR PS SD+SK/T+C SR SCI SK/C SR SCI SD+SK/T+C 0A APR SR AD PR PS BR BC PR HR CM+SK/T+C 0A APR SR AD PR PS BR BC PR HR SK/C OA APR SR AD PR PS BR BC PR HR SD+SK/T+C ASR SR SR SK/C SR PRCS SD/T BR BC SK/C BR BC SK/C SR SAS Effect Size .334 .421 .189 .133 .124 .509 .620 .363 .254 .226 .319 .218 .421 .073 .190 .537 .165 .437 .314 .464 .408 .655 .219 .245 .166 .273 .307 .377 .024 .107 .024 .208 .058 .172 -.010 .738 .791 .686 .082 .357 .357 .095 Author Year Russell 1972 Russell 1976 Saidel 1976 Sanders 1967 Sayner 1972 Schleifer 1978 Schmulowitz 1976 11 22 21 42 30 26 38 87 Treatment Measurement Type/Setting Type Device SD+SK/T+C 0A ASR APR SR AD SR SRIA-S SR PRCS BR BC PR HR PR PS SK/C OA ASR APR SR AD SR SRIA-S SR PRCS BR BC PR HR PR PS SD/T PR PRCS SD/T 0A BR BC PR HR SD/T OA SR PRCS BR AS PR BC SD/T OA SR SRIA-S PR HR BR AS SD/T OA ASR SR AD SR SRIA-S SR PRCS BR BC CM/T 0A ASR SR AD SR SRIA-S SR PRCS BR BC CM/T OA SR PRCS BR 505 PR SD Effect Size .166 .244 .089 .173 .387 .173 .085 .126 .052 .166 .223 .089 -.011 .409 .272 .149 .020 .159 .503 .190 .100 .280 .171 .067 .207 .238 .526 .809 .213 .557 .239 .252 .283 .200 .274 .197 .210 .206 .301 .142 .174 .224 .148 .037 .137 .270 88 Treatment Author Year _N Type/Setting Schmulowitz 1976 36 SD/T Schuler 1982 18 SD+CM/T+T Seiffert 1976 15 SD/T Sherman 1974 9 SD+SK/T+C 10 SK/C Slutsky 1975 21 SD/T Straatmeyer 1974 30 CM+SK/T+C 27 SK/C Measurement Type Device OA SR PRCS BR 505 PR SD OA ASR SR PRCS SR AACL SR PRCA BR BC PR HR 0A ASR SR AACL BR BC SR FT SR PRCS SR SAI-T OA SR PRCS BR BASA OA SR PRCS BR BASA OA ASR APR SR FT SR AD PR HR PR PS BR BC SR PRCS 0A ASR ABR SR FT SR PRCS BR BC BR AS OA ASR ABR SR FT SR PRCS BR BC BR AS Effect Size .131 .052 .152 .190 .150 .175 .124 .307 .094 .179 .045 .459 .501 .408 .291 .746 .418 .433 .411 .367 .455 .235 .204 .267 .188 .213 .175 .180 .275 .076 .274 .140 .183 .338 .363 .315 .166 .556 .127 .504 .238 .235 .241 .180 .289 .065 .418 Author Succerman Thorpe Toy Trexler Trier Trussell Watson Year 1977 1976 1973 1972 1974 1978 1984 .11 28 32 3O 33 4O 3O 18 19 89 Treatment Measurement Type/Setting Type Device SD/T OA ASR SR PRCS SR AACL BR BC CM/T OA ASR SR PRCS SR SRIA-S BR BC PR SS SR AACL SD+SK/T+C ASR SR PRCS SR SRIA-S SR AD CM+SK/T+C OA ASR ABR PR PS BR BC SR FT SR PRCS BR AS CM/T 0A ASR APR SR SRIA-S BR BC PR SD PR HR SR AD SR RS SR RS SR SAI-T SD/T OA SR PRCS BR BASA SK/T ASR SR UCS SR PRCA SR RES SD/T ASR SR UCS SR PRCA SR RES Effect Size .120 .128 .184 .133 .042 .167 .183 .044 .321 .202 .086 .183 .459 .518 .605 .254 .197 .171 .229 .184 .186 .125 .218 .272 .266 .452 .218 .400 .078 .089 .148 .253 .428 .369 .360 .215 .116 .315 .479 .537 .460 .441 .488 .583 .487 .410 Author Year Watson 1984 Weinberger 1976 Weingarten 1973 Weissberg 1977 Weissberg & 1977 Lamb 15 15 24 33 41 22 10 10 90 Treatment Iype/Settipg CM/T SD/T SD+SK/T+C SK/C SD/T CM/T SD/T CM/T SK/T Measurement Type Device ASR SR UCS SR PRCA SR RES 0A ASR SR PRCS SR AACL BR BC SR SAI-S ASR SR AACL SR PRCS SR SRIA-S ASR SR AACL SR PRCS SR SRIA-S OA ASR SR PRCS SR AACL BR BC CA ASR SR PRCS SR AACL BR BC OA ASR BR BC SR PRCS SR AACL 0A ASR BR BC SR PRCS SR AACL 0A ASR BR BC SR PRCS SR AACL Effect Size .540 .597 .510 .513 .332 .255 .207 .268 .562 .291 .155 .208 .137 .120 .055 .081 .010 .073 .399 .450 .321 .579 .298 .524 .591 .440 .753 .378 .399 .321 .526 .394 .278 .509 .386 .753 .522 .251 .525 .463 .648 .650 .276 91 Treatment Measurement Effect Author Year .3 Type/Settipg Type Device Size Worthington 1984 5 SK/T 0A .623 ASR .615 SR PRCS .489 SR FT .665 SR AACL .853 BR BC .654 SR PRCA .453 5 SD/T 0A .601 ASR .582 SR PRCS .753 SR FT .485 SR AACL .465 BR BC .675 SR PRCA .626 11 SK+SD/T+T 0A .700 ASR .671 SR PRCS .742 SR FT .817 SR AACL .684 BR BC .820 SR PRCA .439 Woy 1972 22 SD/T 0A .266 ASR .301 SR PRCS .109 SR SRIA-S .281 SR AD .513 PR HR .213 BR BC .215 Zemore 1975 32 SD/T 0A .523 ASR .588 SR PRCS .537 SR AACL .651 BR BC .327 SR FT .576 Zimmerman 1974 18 SD/T 0A .338 BR BC .324 SR PRCA .352 LIST OF REFERENCES LIST OF REFERENCES Akin, C. and Kunzman, G. (1974). A group desensitization approach to public speaking anxiety. Canadian Counsellor,_§, 106-111. Algirdas, F. (1972). Self-instructional and relaxation variable in the pygtematic desensitization treatment 6f speech anxiety. Unpublished doctoral disseration, Waterloo University, Canada. Altmaier, E., Ross, 5., Leary, M., and Thornbrough, M. (1982). Matching stress inoculation's treatment components to clients' anxiety mode. Journal of Counseling Psychology,_29, 331-334. Ayres, J. and Hopf, T. (1985). Visualization: A means of reducing speech anxiety. Communication Education,_§4, 318-323. Babbie, E. (1979). The Practice of Social Research (2nd edition). Belmont, California: Wadsworth Publishing. Barrick, J. (1971). A cautionary note on the use of systematic desensitization. Speech Teacher, 29, 280-281. Bentson, S. (1974). Symbolic modelipg and the effects of relevant observer-model Similarity in the treatment of public-speaking anxiet . Unpublished doctoral dissertation, Southern Illinois University, Carbondale, Illinois. Biggers, T. (1987). Decreasing communication anxiety through public speech training. Paper presentedTat the International Communication Association Convention, Montreal, Canada. Borin, L. (1949). The construction and evaluation of aygroup procedure designed to raise the confidence levels of beginning students of speech. Unpublished doctoral dissertation, Northwestern University, Evanston, Illinois. Borkovec, T. and Sides, J. (1979). The contribution of relaxation and expectancy to fear reduction via graded, imaginal exposure to feared stimuli. Behavioral Research and Therapy, 11, 529-540. Brooks, W. and Platz, S. (1968). The effects of speech training upon self-concept as a communicator. Speech Teacher, 17, 44-49. Calef, R. and MacLean, G. (1970). A comparison of reciprocal inhibition and reactive inhibition therapies in the treatment of 92 93 inhibition and reactive inhibition therapies in the treatment of speech anxiety. Behavior Therapy,_l, 51-58. Casas, J. (1975). A comparison of two mediational self-control techniques for the treatment of_§peech anxiety. Unpublished doctoral dissertation, Stanford University, Palo Alto, California. Cash, T. and Janda, L. (1977). Evaluation of pertinent sources of volunteer bias in behavior therapy research. Journal of Consulting and Clinical Psychology, 45, 337-338. Chaplin, E. and Levine, B. (1981). The effects of total exposure duration and interrupted versus continous exposure in flooding therapy. Behavior Therapy, 12, 360-368. Clevenger, T. and Phifer, G. (1959). What do beginning college speech texts say about stage fright. Speech Teacher, 8, 1-7. Connell, S. and Borden, G. (1987). Incorporating treatment for communication apprehension into oral communication courses. Communication Education,_§§, 56-61. Cook, T. and Campbell, 0. (1979). Quasi-Experimentation. Boston, Massachusetts: Houghton Mifflin Company. Cradock, C. (1977). Behavioral prophylaxis: A test of Poser's psychological immunization hypothesis on subjects dispTEying sub-clinical pUblic speaking anxiety. Unpublished doctoral dissertation, DePauTTUniversity, Chicago, Illinois. Daly, J. and Buss, A. (1984). The transitory causes of audience anxiety. In J. Daly and J. McCroskey (Editors), Avoidin Communication (pp. 67-80). Beverly Hills, California: Sage Publications Daly, J. and Leth, S. (1976). Communication apprehension and the personnel selection decision. Paper presented at the Internatidnal’CommuniEation Association Convention, Portland, Oregon. Daly, J. and McCroskey, J. (Editors). (1984). Avoiding Communication. Beverly Hills, California: Sage Publications. Deffenbacher, J., Michaels, A., Daley, P., and Michaels, T. (1980). A comparison of homogenous and heterogenous anxiety management training. Journal of Counseling Psychology, 21, 630-634. Deffenbacker, J. and Payne, 0. (1977). Two procedures for relaxation as self-control in the treatment of communication apprehension. Journal of Counseling Psychology, 24, 255-258. V 94 Devine, 0. (1974). Effects of client's committment to change, preference for treatment, and expectation of success on grpup psychotherapuetic outcome. Unpublished doctoral dissertation, University of New Hampshire, Durham, New Hampshire. Ehninger, 0., Gronbeck, 8., Monroe, A., and Moore, L. (1984). Principles of Speech Communication (9th edition). Glenview, Illinois: Scott, Foresman and Company. Ellis, A. (1962). Reason and Emotion is Psychotherapy. New York: Stuart. Ertle, C. (1969). A study of the effect of homogenous grouping on systematic desensitization f0r the reduction of inteppersonai communicative computence. Unpublished doctoral dissertation, Michigan State University, East Lansing, Michigan. Ewing, W. (1944). An evaluation of the individual versus the group speaking methods of teaching the beginning college speech course. Speech Monographs, 1;, 80-87. Finkel, N. (1972). The effects of internal and external feedback on speech anxiety. Unpublished d0ctoral dissertation, University of Rochester, Rochester, New York. Foss, K. (1982). Communication apprehension: Resources for the instructor. Communication Education,_§l, 195-203. Fremouw, W. (1984). Cognitive-behavioral therapies for modification of communication apprehension. In J. Daly and J. McCroskey (Editors), AvoidingCommunication (pp. 209-218). Beverly Hills, California: Sage Publications. Fremouw, W. and Harmatz, M. (1975). A helper model for behavioral treatment of speech anxiety. Journal of Counsulting and Clinical Mchoiogy, 5;, 652-660. Fremouw, W. and litter, (1978). A comparison of skills training and cognitive restructuring-relaxation for the treatment of speech anxiety. Behavior Therapy, 2, 248-259. Friedrich, G. and Goss, B. (1984). Systematic desensitization. In J. Daly and J. McCroskey (Editors), Avoiding Communication (pp. 173-188). Beverly Hills, California: Sage Publications. Garrett, E. (1954). A study of the effect of three classroom orientations upon stage fright in beginningcolleggspeakers. Unpfiblished doctoral dissertation, University of Denver, Denver, Colorado. Garrison, K. (1978). The effect of cggnitive modification on v 95 communication apprehension in children. Unpublished master's thesis, University of Nebraska, Lincoln, Nebraska. Gatchel, R., Hatch, J., Maynard, A., Turns, R., and Taunton-Blackwood, A. (1979). Comparison of heart rate biofeedback, false feedback, and systematic desensitization in reducing speech anxiety: Short- and long-term effectiveness. Journal of Consultingyand Clinical Psychology, 41, 620-622. Gatchel, R., Hatch, J., Watson, P., Smith, D., and Gaas, E. (1977). Comparative effectiveness of voluntary heart rate control and muscular relaxation as active coping skills for reducing speech anxiety. Journal of Consulting and Clinical Psycholugyg 45, 1093-1100. Gatchel, R. and Proctor, J. (1976). Effectiveness of voluntary heart rate control in reducing speech anxiety. Journal of Consulting and Clinical Psychology, 44, 381-389. Germer, W. (1975). Effectiveness of cugnitive modification, desensitization, and rational-emotive therapy in the treatment of speech anxiety, Unpublished doctoral dissertation, University of Texas at Austin, Austin, Texas. Giffin, K. and Bradley, K. (1969). An exploratory study of group counseling for speech anxiety. Journal of Clinical Psyphology, .25, 98-101. Glaser, S. (1981). Oral communication apprehension and avoidance: The current status of treatment research. Communication Education, 39, 321-241. Glass, 0., McGaw, B., and Smith, M. (1981). Meta-Anaysis in Social Research. Beverly Hills, CA: Sage Publications. Glogower, F., Fremouw, W., and McCroskey, J. (1978). A component analysis of cognitive restructuring. Cognitive Therapy and Research,_2, 209-233. Goldfried, M. and Goldfried, A. (1977). Importance of hierarchy content in the self-control of anxiety. Journal of Consulting and Clinical Psychology, 35, 124-134. Goldfried, M. and Trier, C. (1974). Effectiveness of relaxation as an active coping skill. Journal of Abnormal Psycholugy,_§3, 348-355. Goss, 8., Thompson, M., and Olds, S. (1978). Behavioral support for systematic desensitization for communication apprehension. Human Communication Research, 4, 158-163. Grande, J. (1975). The substitution of alpha-feedback for relaxation in the systematic desensitization treatment of publiE speaking 96 anxiety. Unpublished doctoral dissertation, Arizona State University, Tempe, Arizona. Grayson, J. and Borkovec, T. (1978). The effects of expectancy and imagined response to phobic stimuli on fear reduction. Cognitive Therapy and Research, 2, 11-24. Gross, R. and Fremouw, W. (1982). Cognitive restructuring and progressive relaxation for treatment of empirical subtypes of speech-anxious subjects. Cognitive Therapy and Research,_§, 429-436. Gurman, A. (1973). Treatment of a case of public-speaking anxiety by invivo desensitization and cue-controlled relaxation. Journal of Behavior Therapy and Experimental Psychiatry, 4, 51-54. Harris, K. and Brown, R. (1981/1982). Cognitive behavior modification and informed teacher treatments for shy children. Journal of Experimental Education, 59, 137-143. Hayes, 8. and Marshall, W. (1984). Generalization of treatment effects in training public speakers. Behavioral Research Therapy, ‘22, 519-533. Hayworth, D. (1940). A search for facts on the teaching of public speaking, II. Quarterly Journal of Speech, pp, 31-38. Hekmat, H., Deal, R., and Lubitz, R. (1985). Instructional desensitization: A semantic behavior treatment of anxiety disorder. Psychotherapy,_22, 273-280. Hemme, R. and Boor, M. (1976). Role of expectancy set in the systematic desensitization of speech anxiety: An extension of prior research. Journal of Clinical Psychology, 32, 400-404. Henrikson, E. (1943). Some effects of stage fright of a course in speech. Quarterly Journal of Speech, 29, 490-498. Hoffman, J. and Sprague, J. (1982). A survey of reticence and communication apprehension treatment programs at U.S. colleges and universities. Communication Education, 3;, 180-190. Horne, A. and Dougherty, M. (1974). The effects of desensitization, insight, and verbal reinforcement techniques on high and low anxious subjects with specific speech anxiety. Psychological Documents, ( Ms. No. 687). Hunter, J., Schmidt, F., and Jackson, G. (1982). Meta-Analysis. Beverly Hills, California: Sage Publications. Jaremko, M. (1980). The use of stress inoculation training in the reduction of public speaking anxiety. Journal of Clinical 97 Psychology,_3§, 735-738. Jaremko, M., Hadfield, R., and Walker, W. (1980). Contribution of an educational phase to stress inoculation of speech anxiety. Perceptual and Motor Skills, 59, 495-501. Jaremko, M. and Wenrich, W. (1973). A prophylactic usage of systematic desensitization. Journal of Behavioral Therapy and Experimental Psychiatry, 4, 103-105. Jarmon, D. (1972). Differential effectiveness of rational-emotive therapy, bibliotherapy, and attention-placebo in the treatment of speech anxiety. Unpublished doctoral dissertation, Southern Illinois University, Carbondale, Illinois. Jeffery, R. and Peterson, 0. (1980). Speech ( 3rd edition). New York, New York: Harper and Row Publishers. Johnson, T., Tyler, V., Thompson, R., and Jones, E. (1971). Systematic desensitization and assertive training in the treatment of speech anxiety in middle-school students. Psycholqu in the Schools, 8, 263-268. Kanter, N. and Goldfried, M. (1979). Relative effectiveness of rational restructuring and self-control desensitization in the reduction of interpersonal anxiety. Behavior Therapy, 19, 472-490. Karst, T. and Trexler, L. (1970). Initial study using fixed-role and rational-emotive therapy in treating public-speaking anxiety. Journal of Counseling and Clinical Psychology, 34, 360-366. Katz, A. (1976). The rationale, nature,uand evaluation of a group centered approach to the treatment of speech anxiety. Unpublished doctoral dissertation, University of Illinois, Urbana, Illinois. Kelley, L. (1984). Social skills training as a mode of treatment for social communication problems. In J. Daly and J. McCroskey (Editors), Avoiding Communication (pp. 189-208). Beverly Hills, California: Sage Publications,. Kirsch, I. and Henry, D. (1977). Extinction versus credibility inthe desensitization of speech anxiety. Journal of Consulting and Clinical Psychology, fig, 1052-1059. Kirsch, I., Wolpin, M., and Knutson, J. (1975). A comparison of in vivo methods for rapid reduction of "stage-fright" in the college classroom: A field experiment. Behavior Therapy, Q, 165-171. Kleinsasser, L. (1968). The reduction of performance anxiety as a function of desensitizatiOn, pretherapy vicarious learning, and vicarious’iearning alone, Unpublished doctoral dissertation, 98 Pennsylvania State University, University Park, Pennsylvania Kondas, O. (1967). Reduction of examination anxiety and 'stage-fright' by group desensitization and relaxation. Behavioral Research and Therapy, p, 275-281. Krugman, M., Kirsch, 1., Wickless, C., Milling, L., Golicz, H., and Toth, A. (1985). Neuro-linguistic programming treatment for anxiety: Magic of myth? Journal of Consulting and Clinical Psychology, 53, 526-530. Lamb, D. (1965). The effect of group psychotherapy on speech anxiety. Unpublished master's thesis, Ohio University, Columbus, Ohio. Lent, R., Russell, R., and Zomostny, K. (1981). Comparison of cue-controlled desensitization: Rational restructuring, and a credible placebo in the treatment of speech anxiety. Journal of Consulting and Clinical Psychology, 49, 608-610. LeTendre, D. (1977). Anxiety-relief conditioning: An empirical investi ation. Unpublished doctoral dissertation, North Texas State University, Denton, Texas. Leyden, R. (1941). An experimental study of the effects of speech training at the secondary school level. Unpublished doctoral dissertation, University of Minnesota, Minneapolis, Minnesota. Lieng, R. (1976). The effects of self-control rationaleyupon the outcome of systematic desensitization and relaxation training. Unpublished doctoral dissertation, University of Kansas, Lawrence, Kansas. Lima, P. (1975). A comparison of induced anxiety and flooding in the modification of speech anxiety. Unpublished doctoral dissertation, University of Georgia, Athens, Georgia. Little, J. (1976). The relative contribution of thought stopping and covert assertion in the treatment of speech anxiet . Unpublished doctoral dissertation, University of Maryland, Ba timore, Maryland. Littlefield, R. and Sellnow, T. (1987). The use of self-disclosure as a means for reducing stage-fright in beginning speakers. Communication Education, up, 62- . Lohr, J. and McManus, M. (1975). The development of an audio-taped treatment for systematic desensitization of speech anxiety. Central States Speech Journal, gp, 215-220. Longo, D. and vom Saal, W. (1984). Respiratory relief therapy. Behavior Modification, 8, 361-378. 99 Lynd, R. (1976). Anxietyurelief, progressive muscle relaxation, and expectancy relaxation in the treatment of speech phobia. Unpublished doctoral dissertation, North Texas State University, Denton, Texas. Mannion, N. and Levine, B. (1984). Effects of stimulus representation and cue category level on exposure (flooding) therapy. British Journal of Clinical Psychology,_23, 1-7. Marshall, W., Parker, L., and Hayes, B. (1982). Treating public speaking problems. Behavior Modification, p, 147-170. Marshall, W., Presse, L., and Andrews, W. (1976). A self-administered program for public speaking anxiety. Behavioral Research and Therapy, 14, 33-39. Marshall, W., Stoian, M., and Andrews, W. (1977). Skills training and self-administered desensitization in the reduction of public speaking anxiety. Behavioral Research and Therapy, lg, 115-117. Mazzuca, S. (1982). Does patient education in chronic disease have therapeutic value. Journal of Chronic Diseases, 35, 521-529. McCroskey, J. (1972). The implementation of a large scale program of systematic desensitization for communication apprehension. Speech Teacher,_2t, 255-264. McCroskey, J. (1982). Introduction to Rhetorical Communication. New Brunswick, New Jersey: Prentice-Hall, Inc. McCroskey, J., Ralph, D., and Barrick, J. (1970). The effect of systematic desensitization on speech anxiety. Speech Teacher, 19, 32-36. McKinney, M. and Gatchel, R. (1982). The comparative effectiveness of heart rate, biofeedback, speech skills training, and a combination of both in treating public speaking anxiety. Biofeedback and Self-Regulation, Z, 71-87. McManus, M. and Lohr, J. (1975-1976). Automated desensitization for the clinical treatment of speech anxiety. American College Health Association Journal, 23, 218-220. McSweeny, A. (1975). Imagenyuas a variable in the efficacy of systematic desensitization. Unpublished doctoral dissertation, Northern Illinois University, Dekalb, Illinois. Meichenbaum, D., Gilmore, J., and Fedoravicius, A. (1971). Group insight versus group desensitization in treating speech anxiety. Journal of Consulting and Clinical Psychology, pp, 410-421. Miller, G. (1984). Some (moderately) apprehensive thoughts on 100 avoiding communication. In J. Daly and J. McCroskey (Editors), Avoiding Communication (pp. 237-246). Beverly Hills, California: Sage Publications. Morey, E. (1973). Locus of control as a variable in the effectiveness of two rational-emotive therapy styles. Unpublished doctoral dissertation, University of South Dakota, Vermillion, South Dakota. Morgan, J. (1970). Self-modeling versus other-modeling versus practice in the reduction of public speaking anxiety. Unpublished doctoral dissertation, University of Arizona, Tucson, Arizona. Morley, E. and Watkins, J. (1974). Locus of control and effectiveness of two rational-emotive therapy styles. Rational Living, 2, 22-25. Mulac, A. (1974). Effects of three feedback conditions employing videotape and audiotape on aquired speech skill. Speech Monographs, 4;, 205-214. Myers, R. (1974). Validation of systematic desensitization of speech anxiety through galvanic skin response. Speech Monographs, 4;, 233-235. Mylar, J. and Clement, P. (1972). Prediction and comparison of outcome in systematic desensitization and implosion. Behavioral Research and Therapy, 19, 235-246. Neer, M. (1982). Enrolling students in communication apprehension laboratories. Communication Education, 31, 205 5- Nelson, P. and Pearson, J. (1984). Confidence in Public Speaking (2nd edition). Dubuque, Iowa: William C. Brown, Publishers. Nichols, J. (1969). An investigation of the effects of varied rates of training on systemaiic desensitization for interpersonal communication apprehension. Unpublished doctoral dissertation, Michigan State University, East Lansing, Michigan. Nicoletti, J. (1972). Anxiety management training. Unpubished doctoral dissertation, Colorado State University, Fort Collins, Colorado. Norman, W. (1975). The efficacy of self-instructional training in the treatment of speech anxiety. Unpublished doctoral dissertation, Pennsylvania State University, University Park, Pennsylvania Osberg, J. (1981). The effectiveness of applied relaxation in the treatment of speech anxiety. Behavior Therapy, 12, 723-729. Paul, G. (1966). Insight vs. Desensitization in Psychotherapy. 101 Stanford, CA: Stanford University Press. Paul, G. (1967). Insight versus desensitization in psychotherapy two years after termination. Journal of Consulting Psychology, 3;, 333-348. Paul, G. and Bernstein, D. (1976). Anxiety and clinical problems: Systematic desensitization and related techniques. In J. Spence, R. Carson, and J. Thibaut (Editors). Behavioral Approaches to Therapy. Morristown, New Jersey: General Learning Press. Paul, G. and Shannon, D. (1966). Treatment of anxiety through systematic desensitization in therapy groups. Journal of Abnormal Psychology,_71, 124-135. Paulson, S. (1951). Changes in confidence during a period of speech training: Transfer of training and comparison of improved and non-improved groups on the Bell Adjustment Inventory. Speech Monographs, ta, 260-265. Reid, J. (1978). An investigation of hierarchy properties in systematic desensitization. Unpublished doctoraT disseration, Univer51ty of Victoria, Canada. Richmond, V. (1977). Communication apprehension and success in the job applicant screening process. Paper presented at the International Communication Association Convention, Portland, Oregon. Richmond, V. (1984). Implications of quietness: Some facts and speculations. In J. Daly and J. McCroskey, (Editors), Avoiding Communication (pp. 125-144). Beverly Hills, California: Sage Publications. Richter, M. (1974). Systematic desensitization of social anxiety in junior colle e students. Unpublished doctoral dissertation, University 0 Pennsylvania, University Park, Pennsylvania. Robinson, E. (1955). An experimental investigation of certain commonly suggested teach1ng methods for the development of confidence in beginning students ofupublic speaking. Unpublished doctoral dissertation, Indiana University, Bloomington, Indiana. Russell, R. (1972). The use of systematic desensitization and conditioned relaxation in the treatment of public speaking anxiet . Unpublished doctoral diSsertation, University 0 IlI1nios, Urbana, Illinois. Russell, R. and Wise, F. (1976). Treatment of speech anxiety by cue-controlled relaxation and desensitization with professional and paraprofessional counselors. Journal of Counseling Psychology, 23, 563-568. 102 Saidel, M. (1976). Effects of false feedback and relaxation on an experimental analogue of public speaking anxiety. Unpublished d0ctoral dissertation, University of Rochester, Rochester, New York. Sanders, B. (1967). The effectiveness of group administered cue-controlled relaxation in the reduction of public speaking anxiet . Unpublished doctoral dissertation, State University of New York at Albany, Albany, New York. Sayner, R. (1972). An in vivo variation to accelerated massed desensitization. UnpubTiShed doctoral dissertatibn, University of South Dakota, Vermillion, South Dakota. Schleifer, L. (1978). Behavior rehearsal and imaginal desensitization in reducing public speaking anxiety. Unpublished doctoral dissertation, Stanford University, Palo Alto, California. Schmulowitz, J. (1976). Effectiveness of group counseling as a function of state/trait anxiety in reducing problematic speech in children. Unpublished doctoral dissertation, Hofstra University, Hempstead, New York. Schuler, K., Gilner, F., Austrin, H., and Davenport, D. (1982). Contribution of the education phase to stress-inoculation training. Psychological Reports,_§l, 611-617. Seiffert, P. (1976). Symbolic accommodation: An approach for reducing phobic anxiety. Unpublished doctoral dissertatiOn, Georgia State University, Atlanta, Georgia. Sherman, A., Mulac, A., and McCann, M. (1974). Synergistic effect of self-relaxation and rehearsal feedback in the treatment of subjective and behavioral dimensions of speech anxiety. Journal of Counseling and Consulting Psychology, 42, 819-827. Slutsky, J. (1975). The effects of instructional and situational cues on systematic desensitization and a placebo in the treatment of public speakin anxiety. Unpublished d0ctoral dissertation, University of Connecticut, Stoors, Connecticut. Slutsky, J. and Allen, G. (1978). Influence of contextual cues on the efficacy of desensitization and credible placebo in alleviating public speaking anxiety. Journal of Counsulting and Clinical Psychology, pp, 119-125. Straatmeyer, A. and Watkins, J. (1974). Rational-emotive therapy and the reduction of speech anxiety. Rational Living, 2, 33-37. Succerman, K. (1977). The role of imagery and mode of presentation in systematic desensitization. UnpuETished doctoral dissertatibn, 103 Syracuse University, Syracuse, New York. Thorpe, G., Amatu, H., Blakey, R., and Burns, L. (1976). Contributions of overt instructional rehearsal and "specific insight" to the effectiveness of self-instructional training: A preliminary study. Behavior Therapy, 1, 504-511. Toy, T. (1973). Investigation of itemized hierarchy presentation versus progressive time sequence in systematic desensitization. Unpublished doctoral dissertation, North Carolina State University at Raleigh, Raleigh, North Carolina. Trexler, L. (1971). Rational emotive therapy, placebo, and no-treatment effects on public speaking anxiety. Unpublished doctoral dissertation, Temple University, Philadelphia, Pennsylvania. Trexler, L. and Karst, T. (1972). Rational-emotive therapy, placebo, and no-treatment effects on public-speaking anxiety. Journal of Abnormal Psychology, 10, 60-67. Trier, C. (1974). Effectiveness of two versions of rational restruturing in reducing speech anxiety: An experimental analogue. Unpublished doctoral dissertation, State University of New York at Stony Brook, Stony Brook, New York. Trussell, R. (1978). Use of graduated behavior rehearsal, feedback, and systematic desensitization for speech anxiety. Journal of Counseling Psychology, 2S, 14-20. Verderber, R. (1985). The Challenge of Effective Speaking (6th edition). Belmont, California: Wadsworth Publishing. Watson, A. and Dodd, C. (1984). Alleviating communication apprehension through rational emotive therapy: A comparative evaluation. Communication Education, 33, 257-266. Weinberger, A. and Engelhart, R. (1976). Three group treatments for reduction of speech anxiety among students. Perceptual and Motor Skills, 43, 1317-1318. Weingarten, C. (1973). Systematic desensitization vs. accelerated mass desensitization with speech anxious subjects. Unpublished doctoral dissertation, Purdue University, West Lafayette, Indiana. Weissberg, M. (1977). A comparison of direct and vicarious treatments of speech anxiety: Desensitization, desensitization with coping imagery, and cognitive modification. Behavior Therapy, _, 606-620. Weissberg, M. and Lamb, 0. (1977). Comparative effects of cognitive 104 modification, systematic desensitization, and speech preparation in the reduction of speech and general anxiety. Communication Wolpe, J. (1958). Psychotherapy by Reciprocal Inhibition. Stanford, CA: Stanford Publishing. Worthington, E., Tipton, R., Cromley, J., Richards, T., and Janke, R. (1984). Speech and coping skills training and paradox as treatment for college students anxious about public speaking. Perceptual and Motor Skills, S9, 394. Wortman, P. and Yeaton, W. (1983). Synthesis of results in controlled trials of coronary artery bypass graft surgery. In R. Light (Editor), Evaluation Studies Review Annual (Volume 8). Beverly Hills, CalifOrnia: Sage PubTicatiOns. Woy, J. and Efran, J. (1972). Systematic desensitization and expectancy in the treatment of speaking anxiety. Behavioral Research and Therapy, to, 43-49. Zemore, R. (1975). Systematic desensitization as a method of teaching a general anxiety-reducing skill. Journal of Consulting and Clinical Psychology, 33, 157-161. Zettle, R. and Hayes, S. (1983). Effect of social context on the impact of coping self-statments. Psychological Reports, SQ, 391-401. Zimmerman, G. (1974). Augmented systematic desensitization versus systematic desensitiZatiOn f0r redUction of speech anxiety. Unpublished doctoral dissertation, Washington State University, Pullman, Washington. lllslllllllllllllllllllllllllllllllllllll|7|lll6|lll|