+—o_¢ 1f. " 7“ - 41-1.. 9 _, 5, . "T7" .9 I, .21; 11111111111I11 :11I5 ,1, '1' ’1'11"'1 I, ' " "‘ 1‘11 .4: [I3 I] I H I ~ 1 ‘11. 1111 :11? 1:11:15; ' :1, '7 ‘ 1III,L.IT. .j"? ', I ~, . ;,";’-'III;1I1:}I" 53.19111111111111 I‘i " ‘ . IIIIj‘f. III-:I ' - .91, 1,114.1 T; 11111111111 1 111 111111.“.1114' ...... v I I": ..... '''''' . . . A~>. :‘ I 5‘::« . 1.5561,... 11 '1 111111111 :11; Ii .. II. = . 5111,15,; 1015', _ I,II1,I11',‘, . ,. . A , m: , ,II” ' I”. .I 1 '.,.. ”I 31111111 ,.';. III ‘I' ,II 1 '::1:1:11,II,I11I 1191.. '1 1f: ., ’II‘, . 1 '15“ V " .. :11 1111.1 . :. “111 , . 11111 1.111 1, '1,1111 1‘1 1' '1 1'11,1I1 111”“ ,,‘ ‘J 1:'II 1111111111111 1 1 111111Jl111|1fl111111 1 fr 1‘1. . “I .11,‘,1;.1'111"3'I!1 ‘9 --‘1I‘ 15:11 11,1515 "11 III 555113113111. 5,5,1“, .. . 11 II ‘1 11111111111111 ' 1' 1"‘ 1 1 I1.31,,I1.‘151.1I:1;.111.,'-'I~':I,_.”11.11119 :I1 ;: 11,1111I‘1'II1U1III "I1 11, 1111"“ , . 111 '1 11111111111111111‘11111 ,1 1,,1,‘,1,‘1I1,",,., , 11 11 1 ,1111 I1;.1I ‘ 1 11 51115111111111 5, “ 11,1111” 1,» 1 11I ,11,111I111111|,1I , . 1 . 511,315 , 1181.111 11I,11W1IIII1‘I,III11,II IIII.II:.,: III 11:11: ,,,.I,,, III. 1I ,,I,:I 1111511 1111,11‘1I1,1,1,',_91 1111111“, 11'1'II,I,1 “1: [11111111111515 I" ,, . .411111I,I$131,‘l 1ti1IIHI11111I1II111.I1fl,J5” 11 115111111 - lllfllllllflilllflflllll(UH!!!HIWUHIIHIHIIWNW l 10418 2815 This is to certify that the dissertation entitled A SCALE AND CONSTRUCT VALIDITY STUDY OF INFANT TEMPERAMENT presented by Howard Morris Bonem has been accepted towards fulfillment of the requirements for Ph.D. Psychology degree in Major professor Date August, 1982 RETURNING MATERIALS: IV1£31_J Place in book drop to remove this checkout from 4:22;:EEEL_ your record. FINES will be charged if book is returned after the date stamped below. A SCALE AND CONSTRUCT VALIDITY STUDY ' 0F INFANT TEMPERAMENT By Howard Morris Bonem A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1982 (Qt 17(i7/ ABSTRACT A SCALE AND CONSTRUCT VALIDITY STUDY OF INFANT TEMPERAMENT By Howard Morris Bonem The construct of temperament has re-emerged into personality theory, receiving considerable attention since 1963. Defined as the set of underlying formal characteristics which influence behav- ioral style, temperament is typically assumed to have a genetic component and to evidence temporal stability. The literature yields, at best, only moderate support for these assumptions. The con- struct's most central tenet is that response styles are consistent across situations. At present, this has been an assumed, rather than a tested postulate. The present study assessed the validity of the Michigan Infant Temperament Scale (M.I.T.S.) and examined the cross-situa- tional stability of temperamental attributes. A sample of 72 infants, age 3 to 12 months, were visited four times in their homes. During the first 3 visits, the infants were observed for 1% hours during which the temperament attributes of the New York Longitudinal Study (N.Y.L.S.) were rated. The last visit involved the counter-balanced administration of the M.I.T.S. and the N.Y.L.S.'s Parent Interview Protocol (P.I.P.). Howard Morris Bonem The M.I.T.S. was found to be only moderately related to the P.I.P. (rx = .33). Correlation of the M.I.T.S. and P.I.P. with the averaged observational ratings yielded mostly non-significant rela- tionships. Factor analysis of each of the methods showed some con- sistencies in factor structure. It was concluded that the M.I.T.S. was not a valid measure of infant temperament and that parent- report measures, including the P.I.P., are generally poor indicators of infant temperament. These measures may in fact be parent-percep- tion scales, which would alter the interpretation of much of the data obtained by previous investigators. Inter-correlation of the three observational ratings of temperament showed relatively high cross-situational stability (rx = .45). This was interpreted to indicate that the basic assump- tion of temperament was supported, provided that the data was firmly anchored to the behavior of the child. Methodological issues were reviewed including the adequacy of the N.Y.L.S. attribute definitions and scoring criteria. These were viewed as lacking in specificity and clarity. Overall, it was concluded that both the construct of temperament and its measures required substantial empirical review. ACKNOWLEDGEMENTS This doctoral dissertation which is the culmination of seven years of graduate training in psychology, provides me with an oppor- tunity to acknowledge the number of people who have been instrumen- tal in facilitating this accomplishment. However, I would first like to indicate to my mentors and colleagues the value I have come to attach to the research investigative effort. My initial interest in psychology was in the clinical appli- cations of the science; research, methodology, and statistics were means to an end--the Ph.D. However the experiences of my masters and doctoral research and greater sophistication in assessing the extent of knowledge of human behavior has convinced me of the absolute importance of the scientist role for the psychologist. My research has focused on developing and validating a method for assessing temperament in infants. I have learned that such validational efforts provide the "reality testing" for elegant and attractive constructs which so easily become incorporated into psychological theory and applications. I hope that this lesson will remain with me throughout my career and that I can convey it to future students and colleagues. The topic of this research was suggested by my chairman and advisor, Robert Zucker, Ph.D., who has provided me not only with support, but high standards and expectations. As my perspective ii broadens, my appreciation grows for the model of thoroughness and thoughtfulness that Bob has provided. I will also always remember that projects such as this one "end with a whimper and not with a bang." I would also like to thank the other members of my disser- tation committee: Lucy Ferguson, Ph.D., Hiram Fitzgerald, Ph.D., and Helen Benedict, Ph.D. I chose my committee for people who I respected and their quality feedback has sustained my regard. A special note of appreciation to Helen Benedict who arrived at MSU with me. Her clinical guidance and friendship have been quite valuable to me; I believe her move to Texas will be a loss to our department. As with most research efforts, the unsung heroes are the people who did the dirty work. In this case, there were nine who collected data, scored tapes, ran errands, etc. Of these, Renee Rossett deserves a special thanks for being my unofficial lieutenant and team leader. The other assistants were Julie Arbit, Michelle McLaughlin, Larry Pepper, Scott Piper, Debbie Schroeder, David Steinberger, Rick Warren, and Gene Nestervelt. Each of these indi- viduals invested considerable energy and thought in this project. I hope that the hot meals and good conversation helped make up for the long hours and lack of pay. A special thanks is in order to my wife Deborah, who accom- modated to the twice weekly team meetings in our home (while nau- seous and pregnant) and who also served as the third tape rater; this was above and beyond the call of duty. It was her support and iii love that maintained my motivation and confidence. My parents must also be recognized for their support. Now that their son has become a doctor, I officially proclaim them "Jewish parents." A note of thanks to Jim and Shelly Heavenrich for their friendship and the fine accommodations they provided during my many trips back to East Lansing. I only wish we could have taken them with us to Cleveland. My appreciation is also extended to Dozier Thornton, June Jacobson, Art Seagull, and Marsha and Cy Horby each of whom were instrumental in my development as a clinician. Finally, I would like to dedicate this monograph to our daughter, Laura Isabel Bonem. Her emergence into the world has indelibly changed and enriched our lives. iv TABLE OF CONTENTS LIST OF TABLES. LIST OF FIGURES INTRODUCTION THE CONCEPT OF TEMPERAMENT. Inheritance of Temperament. . Temporal Stability of Temperament Commentary. . THEORY AND RESEARCH ON CATEGORIES OF TEMPERAMENT Earlier Work and History . . . . . The Factorial Approach. . The Correlated Attribute Approach A Proposed Integration . . Measurement of Temperament. . . The Michigan Infant Temperament Scale . METHOD Statement of Problem. . . . . . subjeétsIIIIIIIIIIIIZ Instruments. Interviewer and Coder Training and Reliability-~P. I. P. Coder Training and Reliability Evaluation of Behavioral Ratings . . . . . . . . . . Study Procedures . . . . . . . . RESULTS Concurrent Validity of the M. I. T. 5.. Stability of Observations . GENERAL DISCUSSION AND CONCLUSIONS . Adequacy of Behavioral Sampling . Inter- Parent Reliability The N. Y. L. S. Coding System. Conclusions. Page vii ix Page APPENDICES Appendix A. . . . . . . . . . . . . . . . 99 Appendix B. . . . . . . . . . . . . . . . 102 Appendix C. . . . . . . . . . . . . . . . 118 Appendix D. . . . . . . . . . . . . . . . 123 LIST OF REFERENCES . . . . . . . . . . . . . . 136 vi TABLE 10 ll 12 l3 14 LIST OF TABLES Summary of Factor Analyses of the N.Y.L.S. Cate- gories. . . . . . . . . . . . . . The Conceptual Relationship of the Buss and Plomin Factors (EASI-I) and the N.Y.L.S. Attributes . Distribution of Sample on Index of Social Position . Educational Levels of Ingham County Fathers and Fathers in Sample . . . . Parent Interview Inter-Rater Reliabilities (with 3 raters) . . . . . . . Values of Observer Inter-Rater Reliability Studies: Pre-Study and In-Process. . . . . . . Reliability of Temperament Observations--Ratings Averaged Over Three Sessions . . Procedures Employed in the Study . Attribute Inter-Correlations Across Three Methods Measuring Temperament. . . . Attribute Inter-Correlations Across P.I.P., M.I.T.S., and Observations: Sample Divided Into Three Month Age Groupings . . . . . . . . . Comparison of the Michigan Infant Temperament Scale Inter-Attribute Correlations Using Two Samples Intercorrelations of the Parent Interview Protocol Attribute Scales and Deviation From the Michigan Infant Temperament Scale Attribute Intercorrelations Varimax Factor Analysis (with Kaiser Rotation) of the the M.I.T.S., P.I.P., and Observational Data . Intercorrelations of the Attribute Scales--Observa- tional Data . vii Page 34 37 53 55 58 64 67 71 73 74 76 77 78 83 TABLE Page 15 Intercorrelation of Attribute Scales Across Three Observation Sessions. . . . . . . . . . . 86 viii LIST OF FIGURES FIGURE Page 1 Flow Chart of Behavioral Observation Procedures. . . 69 ix INTRODUCTION After many years of dormancy, the concept of temperament has re-emerged into psychological and personality theory. This idea has returned within the context of the nature-nuture controversy--a debate over the relative contributions of genetic-organismic versus environmental factors in the development of human personality. From the time of the ancient Greeks, behavior was, at least partly, attributed to the influence of genetic-organismic variables. Body humors, individual constitution, vital elements, etc. were employed as concepts to help explain behavior and personality. The concept of temperament easily fits among these theoretical constructions; the ideas of inherited, temporally stable character traits and behavioral styles are natural extensions of genetic-somatic-physio- logical explanations (Sheldon, 1942). As the search for causal and explanatory concepts switched to environmental/parental factors, (particularly in the United States) the concept of temperament lost favor among personality theorists. Specific behaviors as well as general personality traits were thought to develop from interactions with significant persons and events (Watson, 1925; Freud, 1935; Erikson, 1950; Sullivan, 1953). Inherited tendencies, if present, were viewed as only weak influences which might easily be shaped or changed by environmental experiences. However, some theorists continued to acknowledge the l basic influence of temperament (Shirley, 1933; Gessell and Ames, 1937). Freud (1935) also, while committing himself to exploring intrapsychic contributions to personality development, clearly accepted the existence of genetic and somatic influences, and made use of "constitutional" concepts in helping to explain human behav- ior. Murphy (1947), Escalona and Heider (1959), and others con- tinued to employ the concepts of temperament, crediting heredity with a role in personality development. However, the review of Diamond (1957), the writings of Allport (1961), and the work of Thomas and his colleagues (Thomas, Chess, Birch, Hertzig, and Korn, 1963) spawned the current focus on temperamental factors. While Thomas et a1.'s work is undoubtedly heavily responsible for the renewal of interest in styles of reactivity, their data collection methods (in-depth interviews with knowledgeable informants) have made it difficult for other researchers to attempt to replicate and/ or expand upon their basic findings. Interview and observational techniques are quite expensive, time consuming, and inefficient ways of collecting data. Such methods limit the range of investiga- tion to an established few who are capable of investing large quan- tities of time and money. Other approaches which are more amenable to mass data collection techniques, such as Q-sorts and self- administered scales, have attracted considerable attention (Buss and Plomin, 1975; Carey, l970)--if for no other reason than the greater ease and lesser expense of the investigative procedures. The above developments led this author to the construction of an infant temperament measuring instrument (Bonem, 1978). This instrument, the Michigan Infant Temperament Scale (M.I.T.S.), has been shown to reliably measure eight temperamental attributes. Using a true-false format, this 164 item scale has the potential for meeting the need for an efficient, self-administered parent- report temperament scale. However, before this potential could be realized, further evaluation of the instrument was necessary. It must be demonstrated that the scales contained in the M.I.T.S. actually tapped the attributes they were designed to measure. The present study was primarily directed towards beginning the validation of the M.I.T.S. scales. In addition, a preliminary effort was made to evaluate the validity of the construct of tempera- ment. A survey of the temperament literature reveals that tempera- mental characteristics have been proposed and utilized without a clear demonstration of the validity of the constructs. At the most elementary level, it must be shown that most of the variance in a measurement technique can be attributed to the construct for which the test was designed (Cronbach and Meehl, 1955). Secondly, con- struct validity (minimally) requires a demonstration of prediction, i.e. the construct has predictive power. To date, most of the proposed temperamental characteristics and measurement techniques have not been subject to validation efforts. In fact, the basic definition of temperament contains testable assumptions that have not received adequate verification. This study is a modest beginning to this process and involves the testing of a postulate that is basic to most definitions of the temperament construct. THE CONCEPT OF TEMPERAMENT Behavior may be viewed as consisting of two aspects: The content of the actual sequence and the how or way in which the instrumental act is carried out (Thomas, Chess, Hertzig, and Korn, 1963). It is this latter "how“ which the term temperament attempts to describe. Bronson (1971) refers to this same characteristic as the "style of life". Such descriptors as the "underlying constant pattern" (Thomas, Chess, Birch, and Hertzig, 1960) and the “formal characteristics" of behavior (Thomas, Chess, and Birch, 1968) have also been used. Thus temperament, in its broadest and most unelab- orated form, refers to a set of underlying formal characteristics which influence or determine behavioral style. These may include attributes such as speed of response, quality of mood, regularity, and sociability. Bronson (1971), who prefers the term "orientation", sets two basic criteria by which the decision to include a personality attri- bute under the rubric of temperament is made: temporal persistence and centrality. Temporal persistence refers to whether a character- istic remains intact and unchanged over a period of time. Tempera- mental variables are expected to be relatively enduring over the span of development. Central variables are those which evidence generality, i.e., apply to a wide range of behaviors in many situa- tions. If a personality attribute fulfills these criteria, it is then said to be a central orientation, one which is central to the organism's functioning and stable over time. Thus, the most fundamental tenet of temperament theory is that response styles are consistent across situations: Behavior is expected to show some basic consistency despite varying stimulus conditions. For example, it would be expected that an infant's response to a new stimulus would be similar at different times in the day and with different stimuli (people, toys, etc.). Although the child might be expected to react somewhat differently to his/her first bath as compared to a new toy, temperament theory predicts that the "withdrawing" type of child will be cautious and tentative in each of these situations. This is not to say that an individual's behavior does not vary across situations or over time. Rather, the construct of temperament assumes some basic consistency in underlying .éEXlE with overt responses varying according to situation, organismic state, etc. It should be noted that there are no studies directly tar- geted towards assessing cross-situational stability of temperamental characteristics. Rather, data collection methods assum§_temperamen- tal characteristics are stable over situations and time. In other words, measurement techniques use the following method: Behaviors occurring in a wide range of situations are rated. These observa- tions are summed and a mean rating is calculated for each tempera- mental attribute or factor (Thomas, Chess, Birch, Hertzig, and Korn, 1963; Carey, 1970; Buss and Plomin, 1975; Bonem, 1978; and others). The effect of this method of measurement and its assumptions is that if cross-situational stability does not exist, the obtained scores are central tendency ratings that falsely indicate temperamental tendencies. The ratings are merely mean scores which can always be obtained on any distribution of scores; such a score may or may not be a good representation of the data. For example, if the dis- tribution is leptokurtic (flat), the mean or mode will be a very poor representative of the scores. Thus, while a temperament score can be obtained, it will be meaningless in terms of representing the pattern of responses on the scale and its description of the infant's behavior. While Buss and Plomin (1975) elaborate upon the criteria of centrality and cross-situational stability, their assumptions are open to debate and may not be essential to a useful concept of temp- erament. They state that a "personality disposition should have an adaptive value and therefore an evolutionary history." This can be interpreted to refer to the Darwinian notion that a trait endures because it allows greater adaptation. In such a context, the assump- tion applies equally well to genetic and environmental explanations of temperament deVelopment. However, while it seems quite reasonable to expect temperamental traits to evolve on the basis of adaptivity, this criterion does not appear to be particularly relevant to the utility of the construct, i.e., it neither adds greater predictive power nor leads to a more focused definition. A second assumption is that of heritability. Buss and Plomin state that temperament is expected to be genetically transmitted. As will be discussed below, this hypothesis is not well supported by empirical evidence nor is it shared by other investigators. However, heritability does add substantial power to the concept of temperament in terms of an understanding of its origin and account- ing for the similarity of behavioral styles within biologically- related individuals. Two further criteria are suggested by these authors: 1) "The personality disposition should be present early in the life and show some stability during childhood..." and 2) the disposition should be present in adults. These assumptions refer to the temporal stability of temperamental traits. They suggest that the human organism displays a continuity in temperament style beginning in infancy and spanning the life course. This suggests that behavioral style does not substantially change over time. It could be assumed that measurement of an infant's temperament would be predictive of his/her future (childhood or adult) behavior characteristics. This also indirectly assumes some degree of resistance by the organism to the shaping influences of the environment. These assumptions are open to empirical validation. If verified, they would increase the theoretical and predictive value of the temperament construct. Thomas and Chess (1977) suggest that the term "temperament" is useful as a phenomenological description with no necessary rela- tionship to past or present behavior, or implication regarding acqui- sition. While the concept of underlying styles of reactivity is useful whether they are acquired genetically or through organismic- environmental interaction, some degree of temporal stability appears essential to the utility of the construct. This author questions the theoretical and clinical usefulness of the temperament construct if it does not contain a predictive implication. There appears to be an adequate methodology and nOmological network for investigation of the organism-environment interaction without the addition of temperament as a "phenomenologic term that has no implications as to etiology or immutability" (Thomas and Chess, 1977). A final criterion proposed by Buss and Plomin states that "If a tendency has sufficient adaptive value to be passed on through the genes, it is likely to be present not only in man but in animals close to man" (p. 11). This last criterion, while of theoretical interest, does not appear to be essential to the definition and classification of temperamental characteristics. While a tendency that has persisted through generations of humans is likely to be present in animals, an absence of this characteristic may be due to differences in animal and human development, genetic characteristics, etc. Attributes influential in human functioning may not have an adaptive value for animals (and vice versa). In addition, a tendency once common to both animal and human, may, through many generations of environmental influence, have changed to such an extent that it does not readily relate back to the original temperamental style. Cross-situational stability is one of the most critical assumptions of the temperament construct. At least a minimum stabi- lity over time, or predictable change over time, seems necessary for clinical and theoretical utility of the construct. Genetic influences in the acquisition of temperamental styles and temporal stability of these patterns are open to direct empirical examination. What follows is a review of the available literature pertaining to these assumptions. Inheritance of Temperament Before reviewing studies that look at inheritance of tempera- mental characteristics, it should be noted that adequacy of the measurement techniques (reliability and scale validity) is a serious concern in each of the studies. As will be shown in later sections of this chapter, investigators of temperament have yet to demonstrate a highly reliable and empirically validated measuring instrument. Thus it has not been verified that the various techniques are in fact measuring the characteristics for which they are intended. The first attempt to look at genetics of temperament was conducted by Rutter, Korn, and Birch (1963). Using the New York Longitudinal Study's interview approach (Thomas, Chess, Birch, Korn, and Hertzig, l963),the authors evaluated, over a three year period, 3 M2 pairs, 5 02 pairs, and 26 sibling pairs on nine temperamental characteristics. They concluded that there is some evidence for a genetic role in temperament: MZ pairs were found to be somewhat more alike than DZ pairs; however, this decreased over time. DZ pairs were generally no more alike than sibling pairs. Large intra- pair differences were also found in the M2 twins. The authors indi- cate that while the data do not strongly support temperamental genetic transmission, they hypothesize that this is due to greater effects of the environment which obscure the presence of initial inherited traits. 10 Buss and Plomin (1975) examined the corresponding scores of 81 Ml and 57 DZ twins on their instrument, the EASI, a forced-choice parent-report scale that measures four temperamental factors; Acti- vity level, Emotionality, Sociability, and Impulsivity. All correla- tions with the exception of Impulsivity in female subjects were found to be significantly higher in MZ twins. Plomin (1976), using an expanded version of the EASI, examined the intercorrelations of 60 M2 twins, 51 same-sex and 26 opposite sex fraternal twins, and parent- child correlations (N = 137). The MZ twins obtained an average correlations of r = .55 while same-sex fraternal twins yielded an r = .07. Average correlations were r = .17 for the mother-child pairing and r = .10 for the father—child group. Canter (1973) in a study of 95 twin pairs confirms a high intra-class correlation for sociability, but found impulsivity to have a very low correlation in both M2 and 02 twins. Torgersen (1977) had similar findings. Using Thomas and Chess' parent inter- view technique, twin pairs (34 MZ and 16 DZ) were evaluated at two and nine months of age. He found significantly greater correlations in the M2 twins for two temperament categories at two months, and on all nine categories at nine months. Plomin and Rowe (1977), in a later work, assessed six temperamental traits with the Colorado Childhood Temperament Inventory (Rowe and Plomin, 1977): sociability, emotionality, activity, attention span-persistence, soothability, and reaction to food. Using 91 twin pairs (36 Ml and 31 DZ) with a mean age of 3.6 years, five of the six categories had significantly higher MZ than 02 correlations (reaction to food showed no genetic ll influences). The authors report that the M2 pair intra-class cor- relations yielded relatively large correlations (median r = .56) as opposed to the 02 twins (median r = .05). Matheny and Dolan (1980) used 23 scales incorporating attri- butes from the work of Thomas and Chess (1963), Buss and Plomin (1975) as well as other personality characteristics. Employing a sample of 105, seven to ten year old same-sex twin pairs (68 DZ and 37 DZ), six temperament "factors" were obtained from a cluster analy- sis. Within-pair correlations for MZ twins were again found to be significantly higher than the DZ correlations (r = .39 to .66 versus r = -.11 to .25, respectively). The above studies all involve parent descriptions of mono- zygotic versus dizygotic twins. With the exception of Rutter et a1. (1963), the data consistently yields significantly higher MZ versus 02 intra-pair correlations. In fact, some of the differences have been much greater than the 2:1 ratio that genetic theory would pre- dict for a genetically determined trait. Bates (1980) suggests that such large ratios may be due to biased parent perception of the degree of similarity in twins. Parents may tend to minimize the differences between MZ twins while maximizing those differences found in 02 pairs. A second procedure for investigating heritability in twin pairs is the use of personality ratings based upon direct observa- tion. However, there are few studies that take this approach. In a sample of six to ten year olds, Scarr (1969) found MZ twins (N = 28) to be more similar than DZ twins (N = 24) on observer ratings 12 of friendliness and social apprehension. Freedman (1974) looked at infants and reported greater MZ similarities for social orientation, stimulus reactivity, and stranger fear. However, these studies are subject to the same critique regarding perceptual bias as mentioned above, i.e., observers may rate MZ twins as more similar and DZ twins as less similar than they actually are. Freedman (1974) examined this problem by having raters separately rate films of each twin. The twins were not presented contiguously and the raters were unaware of which twins were related. This controlled for rater bias due to knowledge of a twin pairs“ zygosity. The differences between identi- cal and fraternal twins decreased, but remained significant, thus providing some reassurance that perceptual-bias is not an important issue in observer studies. Bates (1980) discusses another issue in interpreting MZ-DZ differences. He states that while the correlations might suggest heritability, the similarities may in fact be due to maturation. Thus, for those characteristics that change over time, greater MZ similarity may be a result of a greater matching of rates of physical maturation in identical as opposed to fraternal twin pairs. Finally, one study has attempted to look at genetic trans- mission of temperament characteristics by examining adopted children and their biological relatives. Maurer, Cadoret, and Cain (1980) employed a sample of 162 children who were adopted at birth. They divided the sample into two groups: Group 1 included children who were adopted at birth, had no contact with biological relatives, and had a relative with a diagnosed psychiatric condition. The second 13 group had the same characteristics with the exception of having no relatives with a psychiatric condition. A retrospective question- naire was given to the adoptive parents and scored on nine tempera- mental attributes. The subjects were then placed in a slow-to-warm— up, difficult, or easy group based upon a cluster analysis. Psychiat- ric diagnosis of biologic parents and environmental factors in the adoptive home were used in a multiple regression to assess their ability to predict membership in the temperament clusters. Member- ship in the difficult and slow-to-warm-up groups could not be pre- dicted by the biologic parents' psychiatric diagnosis. On the other hand, having adoptive parents with a psychiatric background did predict membership of the male adoptees in the difficult temperament group. While this study does not directly compare the temperament of biologic parents with the adopted offspring, it is a step in the right direction. The above results suggest that difficult tempera- ment in adoptees may be more related to environmental variables than genetic factors. Further research involving a prospective methodology and a valid adult temperament measure would be an impor- tant next step in evaluating the acquisition of temperament. The above research suggests that there is probably a genetic component in characteristics that have been classified as tempera- ment. The studies to date suggest that the variance accounted for by heritability ranges from 10 to 30 percent. Given the relatively low reliability levels of temperament measures, this is a 14 significant proportion of variance. However, a substantial amount of variance remains unaccounted for by heredity factors. Personality attributes may be best viewed as the products of a dynamic process of genetic and environment interactions (Murphy, 1947). Measurement error set aside, the remaining 70 to 90 percent of variance can be viewed as evidence that the human organism is subject to shaping by the outside world. Even MZ twins, who have identical genetic material, will display divergent development if given differential treatment, e.g., if the twins are perceived and interacted with differently due to such factors as birth order, parent preference, differential skills, etc. To expect a high degree of similarity in children or adults solely on the basis of genetic/ constitutional factors is probably quite unrealistic. Murphy (1974) states that the biological/constitutional dispositions of the organism are what is socialized. In this respect, genetic characteristics and similarities are important influences in individual development. However, the commonalities in twin pairs or family members might be better viewed as products of common genetic predispositions that have been socialized or shaped by a common environmental system (Escalona, 1959). In terms of the concept of temperament, the question then becomes, does an organism develop stable patterns of stimulus reactivity, or not? Temporal Stability_of Temperament Temporal stability implies an underlying stability in behavior that withstands the effects of maturation, skill acquisition, 15 and environmental influences. However, this continuity does not necessarily imply stability in overt behavior. Kagan (1971) describes two types of continuity: l) homotypic continuity--stabi- lity in overt behavior; and 2) heterotypic--stability in underlying process with changes in actual behavior. Tyler (1978) suggests that temporally stable characteristics are more likely to be of the second type--that which have an underlying structure. For example, time spent in fantasy in childhood might not predict adult fantasy time but could relate to choice of employment. The construct of temperament fits this second category and theoretically allows for changes in expression of underlying temperamental attributes or factors. However, given the difficulty in constructing and validat- ing alternate form instruments which sample different behaviors at different age-stages, this definition yields many measurement problems. There is a shortage of relevant continuity studies which attend to the above issue of heterotypicality. The research to date has evaluated temporal stability by using one instrument to examine test-retest relationships over varying periods of time. Katcher (1977) reports low test-retest relationships in a group of very young infants who were initially measured on the N.Y.L.S. tempera- ment attributes at 2 to 3 days and at a three-month follow-up period; observational ratings and a temperament interview with the new-born nursery nurse were compared with an N.Y.L.S. temperament interview with the mother. Birns et a1. (1969) describe the difficulty in 16 making comparable ratings of neonates and three month old infants due to the substantial changes in the infant's behavioral repertoire in the first three month period. In addition, instability in behavior has generally been found during this period and attributed to immature physiological mechanisms. Thus, while Katcher's results suggest poor continuity in the earliest days of life, this does not speak to stability of personality characteristics once the infant has made an initial adjustment to the environment. Ratings obtained at an age where the organism is more stable and has a larger reper- toire of skills and actions (and thus a greater potential sample of behaviors from which ratings may be gathered) would appear to be a better test of the stability issue. Thomas, Chess, Birch, Hertzig, and Korn (1963) reported significant stabilities in temperament ratings over a two year (three months to two years) range; 78 of 80 children maintained a statistically reliable rank pattern in four or more temperament categories. Rutter, Korn, and Birch (l963),also utilizing Thomas et al.'s interview method, report inter-period correlations for a two year period (ages one to three). They found stability in five of seven temperament categories in the one to two year comparison. However, stabilities for years 2-3 and years 1-3 were much lower (two of seven and one of seven categories were found to be stable for these age ranges, respectively). Thus, as the child grows older inter-period consistency appears to rapidly decrease. Using the N.Y.L.S. interview, Cameron (1977) yearly gathered temperament data for five years. He also conducted a focused interview of 17 parenting issues at the time of the child's third birthday. While Cameron does not directly report change scores or temporal stability correlations, he does summarize the relationship of temperament changes to parenting characteristics. Temperament characteristics were found to vary over time and to be related to parental treatment. It was noted that temperamental changes in the third to fifth years of age were found to be highly associated with parental behavior in the first two years. Cameron (1978) further reports an increase in persistence and a decrease in adaptability in those children in the sample who developed behavior problems. Feiring and Lewis (1980) used observational measures of activity changes, motor activity, vigor, and persistence with a sample of 58 children at ages 13, 25, and 44 months. No significant correlations were found in the four characteristics that were assessed. Similar results were obtained by Campbell (1979) and Kronstadt, Obarklaid, Ferb, and Swartz (1979). In each of these studies, the stability of the "difficult" child category was evalu- ated. The difficult child category is derived from a clustering of temperament attributes (typically irregularity in biological func- tions, negative withdrawal to new stimuli, non-adaptability to change, and intense, negative mood). Campbell and Kronstadt et al. report significant shifting of membership in the difficult group over an 8 and 6 month period, respectively. Escalona (1968) conducted a predictive study of personality traits involving a review of observational data of infants in the Infancy Project in Topeka, Kansas. Behavioral predictions were 18 formulated by the author and compared to multiple behavioral and personality measures obtained when the subjects were four to six years of age. Escalona reported moderate success in predicting activity level (62% accuracy) and sensory threshold (61% and 74%) over the four to six year period. Kagan (1971) has obtained differ- ent results using an experimental/observational approach to assess stability of activity, irritability, and other variables. Looking at change from four to thirteen months and 13 to 27 months, Kagan found low to moderate continuity in these variables (correlations generally ranged between r = -.30 to r = .30). A follow-up study on 68 of these subjects at nine to ten years of age found a few moderate relationships between the infant behaviors and childhood characteristics (assimilative smiling and attentiveness, Kagan, Lapidus, and Moore, 1978). Kagan's results provide only mild sup- port for the findings of Escalona. This may be due to the lack of comparability in age groups as well as the use of differing methodo- logies. Escalona utilized an observational approach for data col- lection and predictions were derived from "clinical judgements." However, accurate prediction does not necessarily reflect stability of the trait, instead it may represent the clinican's ability to anticipate the organism/environment interaction and any resultant change in the behavior. Thus, Escalbna's results demonstrate her clinical ability to predict behavior change and does not truly speak to the issue of whether personality attributes remain constant. In comparison, Kagan's results, which avoid the confounding influences 19 of clinical prediction--through the use of pre-post measures--are probably more valid. Two studies of temporal stability do report positive stabi- lities in temperament ratings. Utilizing the Carey questionnaire, McDevitt (1976) assessed 178 infants aged four to eight months and again at between ages three to seven years. Dividing the sample into a three to five year group (N = 97) and a five to seven year group (N = 90), McDevitt found moderate stability between infancy and childhood in the first group in terms of cluster membership, individual stability, and, to a lesser extent, profile stability. The temperamental characteristics of the older group had a low cor-- relation with their infancy ratings; McDevitt suggest that this was due to developmental transformation in how the temperament attri- butes are expressed. The finding of stability in attributes in the younger age groups was replicated by McDevitt and Carey (1978) using the revised Infant Temperament Questionnaire. This scale was administered to infants age four to eight months with follow-up conducted at one to three years with the Toddler Temperament Scale. Correlations for the nine N.Y.L.S. temperament categories ranged from .24 to .58 (median r = .38). These correlations suggest significant stabili- ties given the lowering of reliability that occurs when two scales, each with borderline internal consistency, are employed. However, neither scale has received adequate validation and the inter-corre- lations could also be the results of shared factors other than those related to temperament. 20 Commentary The research to date fails to provide adequate data to ascer- tain whether stability exists in temperamental characteristics. Of the studies which do report significant continuity, the magnitude of the correlations have not been high. If further research does pro- vide evidence of temporal stability, the question will still remain as to whether the obtained relationships are of sufficient size to be of practical importance (Bates, 1980). The current data points to little stability in temperament characteristics. This is in line with the views of Escalona (1968) who suggests that any stability in behavior is due to stability in the organism's actual experience which, in turn, is a product of his/her particular environment and constitution. However, such an approach renders the construct of temperament merely a set of abstracted behavioral descriptors which neither predict nor explain overt behavior. Such a descriptor is probably redundant given the variety of behavior descriptors already available. This author views continuity of personality characteristics as essential to the utility of the temperament construct. If future research does not yield significant temporal stability, then the construct becomes superfluous. In his review of the literature, Bates (1980) calls for multivariate studies of temporal stability. He suggests that the use of a linear combination of measures would lead to more assurance in generalizing from the obtained data. This would involve the use of multiple data sources (observation, parent report, alternate 21 caregiver report, etc.) as well as multiple variables. The efforts of Kagan (1971), Kagan et a1. (1978), Campbell (1979) and Kronstadt (1979) begin to approach this in their use of temperamental factors or clusters. However, there has yet to be a temperament study employing multiple data sources. When observational data are uti- lized, multiple observations are necessary to obtain an adequate and representative sampling of behavior. This is assumed to not be an issue when a knowledgeable informant, such as a parent is used. However, perceptual biases of the informant remains questionable (see above with respect to twin data and below regarding self-report instruments). The current studies use a methodology that implicitly assumes homotypic stability. This is not as important in research that looks at relatively short time spans, i.e., six months to one to two years. However, longer periods of stability need to be examined as was done by Escalona, Kagan, and McDevitt. In these cases, different measures need to be used which attend to the changes in trait expression that might occur as a result of maturation. While most of the studies that have been discussed contained respectable sample sizes, certain methodological issues still remain. As Tyler (1978) suggests, the stabilities reported in the N.Y.L.S. data involved rank order comparison and may be due to consistencies in extreme groups. If extreme rankings are stable, middle range ratings may vary considerably without the rank order correlation being affected. This would yield a spurious estimate of stability. 22 In addition to cautiously viewing rank order statistics, another issue is the use of unvalidated or poorly validated con- structs and measurement techniques. The majority of studies, especially those involving the N.Y.L.S., utilize attributes that are based upon a set of characteristics that evolved from a single content analytic study. Operational definitions are vague and it is not clear as to whether various researchers are actually measuring the same set of constructs. In addition, the measuring instruments that have been used have been poorly validated and are subject to measurement errors due to low reliability, perceptual bias, and excessive homogeniety of test items. Clearly, additional research is necessary to continue the investigation of temporal stability and heritability of temperament. These are questions basic to the current conceptions of the tempera- ment construct. What follows is a review of those studies investigat- ing the existence and nature of infant temperament. Research will be reviewed irrespective of the empirical support for the temporal stability or genetics of the proposed temperamental variables. Examination of the research to date will indicate those variables most commonly viewed as falling under the rubic of temperament and present the techniques and instruments employed in measuring these variables. 23 THEORY AND RESEARCH ON CATEGORIES 0F TEMPERAMENT Earlier Work and History Temperamental characteristics have been a frequent component in personality theories ranging from Freud (1935) to Sheldon (1942) to Murphy (1980). However, until the last twenty years, only a few empirical studies have been attempted. Shirley (1933) conducted a longitudinal study of newborn infants through the 24th month of life. Using parental records, direct observation, and physical measurements, she concluded that there "exists coherent patterns of personality based on factors which are not "ironed-out" by training, are often familial and are formed by age three months. One variable, described as the tendency towards irritability or placidity, appeared relatively stable over the two year period. Gesell and Ames (1937) observed significant individual differences in human infants. Their results suggested the existence of a social responsiveness trait in infants. The Factorial Approach A major portion of current theory and research centers around three or four temperamental factor structures. In a massive survey of over 100 biographies, Heymans (1908) hypothesizes three tempera- mental factors consisting of activity, emotionality, and primary versus secondary function (the latter referring to the tendency to be influenced by impressions of the moment versus a greater influence by residue of past experiences). On a sample of 29 ninth and tenth grade adolescents, a factor analysis of forty personality traits by Levandovskii (1970) yields a similar set of factors: general 24 activity, interpersonal relations, and self-control. Bronson (1971) also correlated a large number of personality variables (34) and found seven bipolar variables to account for most of the variance: reserved-expressive, somber-gay, shy-socially easy, reactive- phlegmatic, explosive-calm, resistive-compliant, and passive- domineering. The first three traits listed above are grouped into a dimension called Expressiveness-Reserve while the second group of three fall in the Placidity-Explosiveness dimension. A third dimen- sion consisting of the variable passive-domineering was also found. However, Bronson reports this to be easily subject to organismic and psychosocial influences. Bronson (1969) indicated that the first two dimensions are relatively independent, unmodifiable, and endur- ing. Examination of their behavioral correlates does reveal one area of commonality: activity level. Birns, Barten and Bridger (1969) examined neonates employing experimental procedures including a visual following task, a cold disk on the abdomen, (at a later age) visual pursuit and following tasks, as well as observational measurements of aspects of soothing and feeding. They found the following infant characteristics as present at birth, and stable over the first four month period: irritability, tension, and sensitivity. Activity level, although showing some variance over time, also appeared to be a consistent trait. The authors do not contend that the four traits are necessarily inde- pendent. 25 Finally, Buss, Plomin, and Willerman (1973) and Buss and Plomin (1975), who attempt an integrated theory of temperament, also arrive at four factors: emotionality, activity, sociability and impulsivity. While these four variables evidence factorial independence (low intercorrelationsL the authors raise questions regarding the heritability of the impulsivity trait. This stems from mixed data where the studies are evenly divided between support and non-support for this hypothesis. Corsini and Doyle (1978) cite additional support for the factorial independence of these four tem- peramental factors. However, they propose the addition of a fifth factor: adapability. Allowing for "poetic license" (the vagaries in labelling factors derived from factor analysis) there appears to be some similarity in the factors yielded by these studies. Activity level consistently emerges as one of the three or four factors found by the above authors. This is in spite of the variations in the operationalizations of this variable and the different methods used to measure it. The fact that this factor consistently emerges and that the term “activity level" is almost universally used as the descriptor provides additional support for the existence of this trait. Described as involving the expenditure of energy, activity level may be evidenced as the amount, vigor, and vitality of a person's behavior. It is very much hgw the individual acts, and may refer to the amplitude of a response as well as the tempo of a person's actions. As such, it tends to be influential in almost all behavioral occurrences. 26 An attempt to find a second common factor involves more problems. Emotionality appears to be the most frequent descriptor used (Heymans, 1908; Buss and Plomin, 1975). This term seems to include Bronson's (1971) Dimension 1 (reserved-expressive, somber- gay, and shy-socially easy) and may be similar to the sensitivity and tension characteristics of Birns et a1. (1969). Unlike activity level, a relatively unitary descriptor, emotionality is used to represent more than one stylistic component of behavior. Buss and Plomin (1975) define this trait in terms of arousal, reactivity, and excitability. They report significant intercorrelations between these three characteristics and treat them as aspects of a single factor. Similarly, Bronson (1971) and Levandovskii (1970) view this category as being composed of a number of interrelated variables. This is in contrast to Thomas, Chess and Birch (1968) who identify three similar attributes (intensity of reaction, quality of mood, and threshold of responsiveness) as separate formal categories of behavior. Although the above characteristics evidence enough similari- ties to justify a grouping into common factors, this becomes a questionable procedure with the remaining traits. Heyman's (1908) primary versus secondary function, Buss and Plomin's (1975) impulsi- vity, Birns et al.‘s (1969) irritability, Levandovskii's (1970) self-control appear to have some commonalities. The attributes of distractibility, attention span and persistence (Thomas et a1., 1963) also fall into this category. While no one term completely describes these characteristics, Impulsivity (Buss and Plomin, 1975) 27 does come the closest. This factor is described as tapping the dimension of behavioral expression-inhibition, and the variables of attention, persistence and response latencies. The Correlated Attribute Approach A second approach to the investigation and conceptualization of temperament involves two major deviations from the factorial approach. These differences, which have both methodological as well as conceptual implications, center around: 1) how the temperamental categories are formed; 2) assumptions regarding the statistical independence of the temperamental traits. Whereas the factor studies started with either a derivation of factors from other personality variables or with an a priori listing of hypothesized factors, the second approach begins with an empirical analysis of infant behavior in an effort to extract temperamental categories. These categories thereafter have been subject to further evaluation for "fit" into the formal definition of temperament traits (see pages 4 to 9). Using this approach,no assumptions are made regarding the independence of the attributes under study. The categories are seen as describing the direct expressions of temperament. They concern the style by which individual behaviors are effected, as opposed to a more abstract, generalized description of behavioral style. The attributes are likely to share significant interrelationships and thus not be factorially independent. While this strategy is theoreti- cally and conceptually more cumbersome, it can be more readily operationalized. Whereas temperamental factors are seen as the 28 "irreducible units" of temperament, this second approach primarily centers around the end product of these "units". As the New York Longitudinal Study (N.Y.L.S., Thomas et al., 1963; Thomas et al., 1968) is the first in-depth empirical investigation of temperamental characteristics using this conceptual framework,examination of the "trait" approach to temperament shall begin with a fairly detailed description of their findings. The authors' interest in temperament evolved from a concern that the organismic contribution to infant development had been ignored and neglected. Personality development is conceptualized as consisting of the "interaction of a baby endowed with character- istics of initial reactivity and an environmental complex including familial and extra-familial factors' (Thomas et al., 1963, p. ix). The authors set out to identify those characteristics of individuality present in the first months of life. The researchers then attempted to examine the degree of persistence of these traits and how they influenced individual psychological development. Thomas et a1. performed a content analysis of twenty-two parent-interview protocols which sampled the behavior of the infor- mants' three month old infants. Nine categories were found that scored continuously and differentiated among individuals within each category. They were: activity level, regularity, approach or withdrawal, adaptability, intensity of reaction, threshold of responsiveness, quality of mood, distractibility, and attention span and persistence. Utilizing a three point nominal scale, eighty infants were rated on the above categories. Assessment was conducted 29 by performing semi-structured open-ended interviews (approximately 1; hours in length) and then scoring the interview protocols on the nine attributes. Stability among the categories over the first two years varied from a low of 27.5% for activity to a high of 92.5% in mood. The authors state that these variations may be partly attributable to the developmental characteristics of the individual temperamental categories. Each attribute may have its own develop- mental sequence, with the process of individuation and differentia- tion occurring at different times and rates in the life of the infant. Despite these variations, the temporal stability of Thomas et al.'s temperamental ratings are quite good. Eight of the nine categories show consistency over a two year period, with only acti- vity level evidencing any significant change. This only occurred in the last six months of the testing period. These results, derived from one fairly homogeneous sample, suggest that the N.Y.L.S. attribute categories can be reliably measured over a moderately long time span. However, this does not address the issue of the con- struct validity of these attributes. Later research (Thomas et al., 1968) extended the investiga- tion of these temperamental attributes. Factor analysis of the nine categories yielded a Factor "A“, loading heavily on mood, intensity, approach/withdrawal, and adaptability. This factor was found to be relatively consistent over a five year time span. Scores on this factor were able to differentiate normal and "clinical" groups of subjects. Two other factorial trends also emerged from the analysis: 30 a factor b was composed primarily of threshold, rhythmicity, inten- sity, and adaptability; and a factor c primarily loaded on activity and intensity. Attempts at replication of the N.Y.L.S. findings have been conducted with mixed success. Graham, Rutter, and George (1973) rated a group of British children on seven categories of behavior: mood, intensity of emotional expression, activity, regularity, malleability, fastidiousness, and approach/withdrawal to new people. While some of these categories differ from those of the N.Y.L.S. (malleability and fastidiousness) and fail to include three other attributes (distractability, attention span and persistence, and threshold of response),their findings provide some support for the N.Y.L.S. categories. The categories could be reliably identified (inter-rater reliability was 80% for all except regularity, 64%). This can be viewed as a partial replication of the N.Y.L.S. study, and giving support to the hypothesized existence of the five attri- butes common to both studies. Lambert and Windmuller (1977) provide additional support for the validity of the N.Y.L.S. attributes. Parents of 327 elementary school children (111 adjusted controls and 216 children with adjustment, achievement, or activity problems) were given a structured interview retrospectively assessing infancy and early childhood behavior. A factor analysis of 17 items yielded six orthogonal factors identified as: attention span, threshold level, activity level, distractibility, adaptability, and rhythmi- city. The authors view these results as an independent replication of the Thomas et a1. temperament traits. However it should be noted 31 that the interview data can be regarded more as a verification of the relative independence of these six categories than an empirical validation of the existence of these constructs. Inconsistencies in the Graham et a1. study as well as in the N.Y.L.S. led Birns et a1. (1969) to the following criticisms of the parent interview non-experimental approach: 1) Thomas et a1. rated their subjects at age three months. Although the nine attri- butes found in these subjects are viewed as being constitutional in nature, the extent to which behavior may have been modified by environmental factors in the first three months of life is unknown. 2) The data used in the N.Y.L.S. came from maternal report. Even given reliable maternal reporting, the ratings may be based upon different stimulus conditions, rater judgements, etc. Birns et a1. avoided these difficulties by rating the infants shortly after birth. Subjects were exposed to a uniform set of stimuli in a neonatal laboratory setting. The stimulus battery included presentation of a moving picture, application of a cold disk to the infant's abdomen, and other such procedures. Behavior was rated by three or four observers who, along with the mother, were out of sight of the infant subject. This solved the second criticism by exposing all children to the same experimental conditions and with their behavior being rated by neutral judges. The temperamental attributes being examined in this study were irritability, soothability, activity level, alertness, vigor of response, sensitivity, tension, and maturity level. At three or four months of age, social responsiveness was added to the 32 behavior ratings. It was found that irritability, sensitivity, tension, and soothability showed consistency from birth to four months. This suggests that certain temperamental attributes are evident at birth and remain consistent through age four months. The similarity of the variables being measured provides some support for the attributes used by the N.Y.L.S. The fact that some of the traits were not found to be consistent may be as much attributable to the rating procedures and the behaviors chosen to be rated as to the actual instability of the behavior characteristics. Factors contributing to the negative findings involved difficulty in the measurement of altertness and activity level, and in rating maturity level. Such difficulties also highlight the problems of using a "controlled" laboratory setting to evaluate infant temperament. Naturalistic observation may not only be more practical but also more sensitive to expressions of temperamental characteristics. A Proposed Integration Discussion of temperament research has been conducted by dividing the studies into two groups. The first group through a variety of means, arrived at a three or four factor classification of temperament. A major common assumption was the relative indepen- dence of the factors from one another. The second group included studies which have attempted to investigate temperament through the identification of attributes present at an early age, and then traced these behaviors for possible consistencies over time. Inter- correlation of the attributes, rather than being avoided, was 33 assumed. Emphasis was upon uncovering the various components of unlearned behavioral styles rather than a superstructure under which all the components might be organized. It is suggested that Thomas et al.‘s (1968) nine temperamental attributes and Buss and Plomin's (1975) four temperamental factors may both be valid ways of cate- gorizing temperament. Temperament may consist of three or four independent dimensions, each consisting of a number of related attri- butes. These attributes are capable of direct translation into operational definitions, and therefore into behavioral description. A dimension such as emotionality may be viewed as a grouping of attributes such as mood, intensity of response, etc. While the dimensional approach might be more useful for research and theoreti- cal purposes, the attributes remains descriptive of an observable behavioral style and may be more suited for clinical usage. Support for this conceptualization is provided by Thomas et a1. (1968) Scholom, Zucker and Stollak (1979), McDevitt and Fox (1976), and Bonem (1978). Each conducted factor analyses of the nine N.Y.L.S. categories. The data for the Thomas et a1. studies utilized a questionnaire format (Carey, 1970) as did Bonem who employed the M.I.T.S. (Bonem, 1978).] A listing of the factors yielded by these four studies is found in Table 1. Examination of the distribution of the attributes among the factors suggest a basic 1It should be noted that Garside, Birch, Scott, Chambers, Kolvin, Tweedle, and Barber (1975), among others have conducted factor analytic studies with the N.Y.L.S. attributes. However, due to the inclusion of additional or modified characteristics, the results are not readily comparable. 34 coo: uposmmggh xpwp_amuwegH spasmmggp xu_P_nwpueeom_o >g_m=mucH mocmpm_mgma »p_meeeeH »p_meepeH >Be__e_pueeam_a »p___e_peeeemwa »g_>wpo< >p_>wao< anw>wau< xuw>wpo< m genome gomw> xwemcm o gouge; mocopmwmgmm upogmmgsh »e_e_5ee»;¢ »e_e_5eoaem seeee_=mem appeaseexem N goaoau xuwuwscpzsm aocmpm_m:ou m gauged ap_~wnepaeu< >u___nepaec< szegugapz —ezegu;upzxnomogaa< Fezegu;p_z\:uoogaa< szmeu:u_3\;oeocaa< eeeeeaa< »u_Feeeeaee< eeeemeeeh seemeeeea eeez suepweeeeeeem_o eeez eeez _ Leaded »B_P_eeweem eeez < eeeeea Emcom xom new uum>wouz .—a pm EoFocum ._m pm mmsogh mmeomoumu .m.4.>.z mg» eo mmmapmc< gouoeu mo zgeEE=m--.P unmwm m=_m= .:o_peposeoo mpa_u—=E a mcwpaaeou use pomnnzm some toe museum Amazoomm :_v mucmpmwmemn cams mcwcwepno an nmumpsu_mu .mucmEmmm mampomue> o>w$ co copes sumo ego: anaemnam cop om u m” .umsgomema mew: mpuonaam m>_w we sumo co mm:_pme so» om n ma .vmsgomgma mew; muommnam op Co comm co mm:_peg m>_m om ".me ummo. u e amen. u a N¢.mm mocwpmwmema xo.Fo xo.¢o xo.mm zpwpwnwuuegpmwo &~.mm N~.mm &~.mm zuwpwnmuamu< em.am am.mm Ne.em Fezeee;e_=\eeeeeaa< Ro.mm fio.mm Rm.Nm >p_m:mu:~ xo.me xo.mw No.¢m too: &O.—o &~.mo xo.em _o>w_ anw>vpu< omaeucmu_> ”co_pmzpm>m nEoo¢>m_m cHum>w4 emamuomuw> mpanpgpu< mmmuoea cw .zuzumtmga mmouogmuca ecu Acaumumga "mm_u=pm >u_p_nm_pmm gmpmauemch Lm>gmmno we mmzpm>--.o momgemeo peeew>egem we “Lego zepw--.F mmszu .Ammewuee PNV meewue>eemee ee>em we meeegm omega aewzeppew eeecwuceemwe mw eewuemppeo open umwoz Acewue>gemee we use epeeweeav eeu eewe> V heueews pv eewue>gemeo Ameeeeem omv meewue>gemeo we meweeem eewgee “we; epecws op ea m Aeawp nemeepe Aeuecws F meueeweeHv eeu meow A eeweem eeee>em we euecwe Pv :ewae>eemeo V ‘/ 70 At the beginning of the first session, the mother was requested to sign an Informed Consent form and completed the Back- ground Information Questionnaire (see Appendix B). At this time the infant was presented with and allowed to keep a new toy, the response to which was rated on Approach/Withdrawal and, later, Adaptability. The fourth session was conducted by one of four trained interviewers. The M.I.T.S. and the N.Y.L.S. interview were admin- istered in a counter-balanced order to control for response bias. Once the data were collected, the mother was allowed to ask questions about the study's procedures. She was also informed that a descrip- tion of the study's group results would be sent to her upon comple- tion of the project. See Table 8 for a listing of the above pro- cedures. 71 TABLE 8.--Procedures Employed in the Study. Home Visit #1 Consent Form Background Information Questionnaire Observation #1 Home Visit #2 Observation #2 Home Visit #3 Observation #3 Home Visit #4 M.I.T.S.a Parent Interview Debriefing a aThe M.I.T.S. and parent interview were presented in a counter- balanced order such that one-half of the mothers first completed the M.I.T.S. while the other half began with the parent interview. RESULTS Concurrent Validity of the M.I.T.S. The primary focus of this study was the examination of the concurrent validity of the M.I.T.S. This can be examined in two ways: 1) The correlation of the M.I.T.S. with the P.I.P. The P.I.P. is a technique that also utilizes the parent as an informant but involves a semi-structured interview rather than a highly structured psychometric instrument; and 2) The M.I.T.S.‘s relation- ship to the behavioral observations. In addition, the concurrent validity of the P.I.P. can be examined in the same manner. 1) Table 9 contains the M.I.T.S.-P.I.P. correlations which ranged from low (r = .08) to high-moderate (r = .60). The median correlation was r = .375 and the mean, r = .33. Table 10 lists the attribute scale correlations when the sample is broken down into three to six, seven to nine, and ten to twelve month age groups; this breakdown allows an assessment of whether there are any signi- ficant changes in the inter-scale correlations as a function of subject age. In general, similar ranges of values are obtained in each sample. Therefore subject age does not appear to be an impor- tant influence in determining the level of M.I.T.S.-P.I.P. inter- relationships. The variations that do occur are likely'a result of the low subsample sizes. When the Table 9 correlations are corrected for unreliability in the P.I.P. method, the level of 72 73 .Ameepv ._e um mesecw Eegw mmewaee eeeepmwmgee ee~wpege=em use eeeepmwmsee we megemeez Peeewpe>gemec use .e.H.e mew .eeeeumwmgee epeeumeo eee eeeeepmwees "eeeepmwmgee we meeemeee e3» eeeweueeu .m.w.H.z use a .meewpe>eemee peeew>esee one cow epeepwe>e ue: me: zuweepemem we mesmeee ”mwoz ._.o. v my .mo. v mas eee. me. Aem.v mm. a gem: ee. me. Ame.v mum. L eeweez _~.- NF. Apo.1V wo.- eeeeumwmgee e_eeemee __.- Ne. Axe.v teem. eeeeem_meee eeeezemeeee m_.- me. Aee.v twee. xuwpweeeeee< me. Nee.- Ace.v teem. .ezeeeeewz\eeeeeee< e e Aee.v tape. »u_eepeeem teem. me.- Ame.v we. aew_we_eeeeem_e .em. ee. Aem.v e_. e_eemeeew m_. ee. “Ne.v 4.5m. »e_meeeee eee. Ne. Am~.V _~. eeez .p. me. flee.e twee. _e>ee xuw>wee< =e_ee>eemee eewee>eemee .e.e.e sew; .m.w._.z epeem eezeweeee new: .¢.H.e saw: .m.h.H.z geesegeesmh meweemeez we muonwwz 00.2.3. mmOLU< .mcomeFmLLOUILwHCH muanwsuu<11.m mqu: 74 .m.w._.x oz“ Lew cum: o-eum ooeeumwwgee o_oeemeo .m.p._.x use Low eon: m_eem gunman—meme veeczum_e== .—o. v «c .mo. v m. a e .m=e_ue>gomec _ege_>e:em u .mao ”neueeege zew>wou=_ usage; a .e._.e mo—eum aeoaegoa=ew ueew=_ cemwgowz u .m.~._.x ”who: e~.- NM. .m.- e_. me.- ee. .mee e..- Ne. eN. .a._.e eeeeeem_xeee __. e~.- .m.- .m.. an. me. .mee e~.- .eme. me.- .e._.e eeeeeeneaeea «N. m_. m~. e..- em. ee. .mee em. a..- me. .a._.e e_eemeeew ..en. ..ee. ..ee. .e._.a »e_ee_=ee¢ em. ee.- ..mm. an.. em. ee. .mee e_. Ne. . me. .e._.e »e___e_eeeeem_e __.- ne.- e..- ~_. me. e_. .mee ._e. ..~e. em. .e._.e »e___eeeeee< _e.- e..- mm. .ee. .e. ~e.- .mee .ezeeeee_= ..~e. ..em. ..~e. .a._.e eeeeeeee ee. ee. ~e.- e~.- em. m_. .mee Nu. auem. Om. .¢._.a sawmcme:_ ~_.- Ne.- eN. .e. me. e_. .mee me. .ee. ee. .e._.a eee: Ne.- e_. me. e_. em.- m~.- .mee ~_. ..~m. em. .e._.e .e._.e .m._._.x .e._.a .m.w._.: .a._.a .m.w._.x »e_>_ee< Aee u zv neeeez N. ea e. ANN . z. Geezer e. ea 5 ANN u zv fleece: e e» m e_eem IlliiL: 1.1.1 4 111 54141511 . Illl I 1.1 1 ,5! (Jilin. . 114 I I III... .1 414.. Ilfllifli 1.4.. ‘1 ll U1 IL :1 i144: amewazeec oo< coco: oecep ea:— eoe_>_a o_;sem "n:e_uo>comno nee ..m.w._.z ..e._.e mmece< m:e_8e_ogceu-ceue_ meeewcuu<--.o_ uame_ as. v a use an geeuwwwemwm m. mes—e) :. woeeeowwVe ogwe .xeeam Am~m~V Suzem egg eacw o-ee—_e>e no: mac—He; eoememwmeuee .Aee_ . zV 2%..m »e__.ee__et acecceo one cow tome—amwu use .m.w._.z one cow meewae_ecceu ouaewguee-ceue_ umwoz 1c.l 11 III 11 I1. 4111‘IJ.J1 14 lii if I‘m" 14 “J- .111 .8. v m... .3. v a. e e e e e e e e eeeeemwmcue we.. «can. eon. o_. ~_.- mc.- -.- .eOm. «.m¢.- opeeemeo e e e e e e e eueeem_meee m_. .c. m_. we. ~_.- «om. vo.- eeae. eeeezeAVeea .._e. Ne. Awe. A..m~. V A..mm.-V ...em. V AN..-V «ewe. No. m—.1 m—. comm.1 «eke. v3.1 hum—wneuauc< .Ne.-V Ame. V ANV. V e...m~. V A..em. V Re..-V me. _~.- me. ..~e.- ..Nm. ee.- eeeetee< .Ne.-V Aae.-V Ae_. V AN_. V Ame. V me. e..- mm. mm. me. xe_ee_eeez A~_.-V ee. Awe. V ...m~. V o~.- am. co.. co. xu—__e_uoeeuw_o A..em.-V Ame. V AMV. V c.mm.- No. me. e_e;maezw .em.-V A..mm. V n..- «ma. ASVmeueeV A_e. V .N.- eeez >V_>_eo< me .e xe___eeeeee< ceeegee< xuwse_:ao¢ xu___n.auegum_a e_e=meegw xewmeeV=V wee: xVV>Vuo< llillldiii 11 .III. 1‘.ll.i me_;5em exp me_m= weapon—ecceu oe:e_guu<-gou=_ o_eem eeeaeeeesew eeewe_ eea_:o_z use we eemVLeeseu--.—— mee_ mo. v e on» we ueequ—eawm m. eewuew>ee mwzwe e~wm tee eeweeoe_e on“ use «mosueogee e. museum "cue: . 5. v N .5 aqvm. Amm.+V .mm.+V Ae~.+V Ae_.+V AVe.+V Aem.-V Ame..V Ace.-V Aeeeeem_meee epueumeoV Am_.+V Ace.+V .-.+V Aem.+V .__.-V ANN.-V A__.-V Ae V .eeeeem_meea umasaumwvczv NOD. no. we... 88mm... co; cine. ND. ewe. mucoumwmsmm A-.+V Ame.-V AeN.-V Lee.-V “Ne.-V Aee.+V Am_.+V ..ee. me. me. .ee. .e~.- ..em. e_.- »e___eeeeee< Ac...V Ame.-V . Amm.-V Aee.-V A__.+V Am..-V ee.- he. ..em. -.- ..Ve. e_. eeeeeeee .Ne.+V Re..-V Ae~.+V A~_.+V .0...V me. me. .e. ee. me.- xHVLe.eete Ase.-V eAne..V A...-V Ame..V __.- Im~.- Ne. .e.- »B._.e_eeeeem_e Ae~.+V Ame.-V .N_.+V m..- Re. -.- e.eeweeew Am..-V ANN.-V ee. :8. 3.5.35 AVm.-V e_. eee: xe.>_ee< eeeeem_meea ADV—Veeeeee< geeeeee< xe_ee_=ee¢ »e___e_eeeeem.e e.eemetw ae_aeeee_ eee: »e_>.ee< ilTl‘Nltv all All 1| 'JIII anal-1.1411 .U I l. .41! 4444.4 ill1. I11111 . Amw u 2V meewee—ogceeeoee_ ouan.guu< e_eem eee=eeeeeew Seawe_ eeeVee_z age seed ee_ee_>ee eee meVeem eeeeweuee .eeeeete ze_>eeee_ Sewage ego we mee_ee.etteeteee_--.~_ meewee< ~&5.mmqm Louuew xuwmeeuefi euceumwmgee gueeeee< xuwwweepeee< nee: Ae_.V »e_ee_=eee NN. xuwpweeueee< NN. geeeeee< mm. eeez A&N.wv v Louuem mm. eeeeumwmgme Ne. »e_meeeee Ne. zew>wpe< Aem.m¢V m Leeeeu ee.- meeeumwmgee ee. zewp_eweeeeemwe Nm.- xewmeeeee «em.m_V N teeeee a..- »e_meeeee Ne. auwpweeeeee< om. soeeLee< mm. e_eemeeew mm. eeez mme_eees flee.m_q V eeeeea meeweeeuflem.mmV _ teeeeu wo.- mm.- mm. mm. mm. museumwmgee epeeemee mocmvm mema eeee=DWVeee xuw>Vee< Aem.eFM m eeeeee zew_we_eeeeemwe zprepzmem ewesmeegw zuwmeeuefi Axm.omv N wouoew xawmeeuew xuwpweeueee< geeegee< eeez meeweeee Aem._mV _ eeeeee meewue>tmmeo .¢.H.m .m.h.~.z .eaeo Fecowge>gmmno vce ..m.~.a ..m.H.~.z use we Acowuepom gmmwex sawzv mwmzpec< wouueu xeewge>nn.m_ mqme___e_ueecumVo «.mm.- .._m.- am~.- e_esmec=w .MN. «.em. xe_meeeeV «mm. see: a~.>_ae< uezvemeeea xV___eeeeee< zeeeeee< xa___e_ueecum_a upesmeczw >u_meee=_ eeez xew>Vue< i 14 14554 I .eueo —ece_ue>cemaoi-mu_eum muse—Lee< one we m=e_ue_egeeegea:_--.v— meagmmno 092:. mmogu< mmpmom muznwgauk .._.o cowuwpmggousmpcmil.mp m4mh' 10. 11. 12. 13. 15. 16. 117 Highest school grade completed by father: (circle) 1 2 3 4 5 6 7 8 9 10 ll 12 1 2 3 4 Some graduate Graduate College training degree What kind of job do you do? (if housewife, please list) job title Job Description Highest school grade completed by mother: (circle) 1 2 3 4 5 6 7 8 9 10 ll 12 1 2 3 4 Some graduate Graduate College training degree On weekdays, on an average day, what is the average amount of time each parent spends with the infant? a. Mother: hours per day b. Father: hours per day On weekends, on an average weekend day, what is the average amount of time each parent spends with the infant? a. Mother: hours per day b. Father: hours per day Does your child cared for by someone other than an immediate family member (father, mother, or sibling) for three (3) or more hours per week? Yes No (circle) If yes, what is the average amount of time per week spent with this (these) other person(s) hours per week. In question 14 above, in what kind of situation and by whom is your child taken care of? For example, is he cared for by a grandparent or babysitter? Comments: APPENDIX C 118 Description of Parent Interview The mother interview which is to be conducted during the fourth home visit is adapted from the temperament interview schedules of the New York Longitudinal Study (Thomas, Chess, Birch, Hertzig, and Korn, 1963) and Graham, Rutter, and George (1973). The entire interview is to be tape recorded, for later coding. The interview is divided into three parts. In the first section, the mother is asked to describe how her child has behaved in a variety of situations during the past 24 hours. The second part of the interview is concerned with the eating, sleeping, and bowel habits of the child as they have occurred over the past two weeks. Finally, the mother is asked to describe how her child has behaved over the past six months in less frequently occurring situations such as meeting strangers, being in a new place, behavior during a doctor's visit, etc. ? The interview is kept as unstructured as possible. The inter- 1 viewer introduces content areas and allows the mother freedom to describe her infant in her own words. Descriptions of actual behav- I ior are encouraged and generalizations or vague descriptions are not 1 accepted by the interviewer. When necessary the interviewer will offer the mother a wide variety of adjectives to help her focus on her infant's actual behavior. At no time will the interviewer make any judgemental or value-laden statements about the infant. All questions are to be open-ended and neutral in tone. 119 The goal of the interview is to cover the following areas of infant functioning: sleep, feeding, soiling and wetting, bath, nailcutting, hairbrushing, washing of face, nose, ears, and hands, dressing and undressing, sensory, neuromuscular, motility, response to people, response to illness, play, learning of limits, verbaliza- tions, crying, and anything else the parent thinks is important. The specific questions asked of the parent are as follows: I. II. III. First, in as much detail as possible, I would like you to describe your infant's activities and behavior over the past 24 hours. Be sure to include his (her) feeding, sleeping, playing, and dressing times. Now let's look at the past two weeks. Please describe your child's sleeping, and waking, feeding, and bowel activities as they have occurred during the last two weeks. Finally, I have some more specific questions for you. Tell me about how your infant has been when given a new toy; when ham 2'. A mung meeting a stranger; when left with a babysitter; when taken to the doctor; when given his first bath; when given first new shoes; regular bed; own room; when ill; when in a car; and when given a new food. 120 NINE CATEGORIES OF TEMPERAMENT Mood Activity level Intensity of response Approach/Withdrawal to new stimuli Adaptability Distractibility Persistence and Attention span Threshold of response Regularity of biological functions General Categories to Ask About During an Interview Feeding Sleeping Bathing Playing Changing/Toileting Dressing Going to the Doctor Response to sound light touch sensory Illness Procedures: cutting nails Response to people New places, situations, and things Feeding What type of eating schedule is the child on? How does the child deal with hunger? :5 How does the child deal with new foods? , How does the child respond to other things while eating? How much does the child eat? What happens when there is a change in the feeding schedule? What is the~child's mood while eating? What is the child's activity level while eating? .. How do you know when the child is full? ' Sleeping What type of sleeping habits does your child have? How does the child respond to stimuli while sleeping? How does the child respond to new sleeping places? First night? Thereafter? Where will the child sleep? Activity during sleep? 121 Bathing What was the child's reaction to his/her first bath? How does the child react to the bath now? What is the childs mood during the bath? Is your child active in the bath? How does the child respond to the temperature of the bath water? Intensity of bath. Playing How does your child react to new play things? Over time? What is your child's mood while playing? How long does your child play with a certain toy? How does your child react if some one takes away the toy that he/she is playing with? What is the child's activity level while playing? What happens if the child is playing and some one walks by? If a toy rolls out of reach will the child go after it? Changing/Toileting What is the child's mood during changing? Does the child give some indication when he/she has soiled? Can you distract the child during changing? regularity activity during changing process intensity of response Dressing What is the child's mood during dressing? Can you distract the child while dressing him/her? How does the child react to new clothes? Over time? What does the child do while you are dressing him/her? Going to the Doctor What is the child's response to the doctor? How does the child respond during the examination? What is the child's response to a shot? Can you distract him/her after the shot? What is the child's mood during the examination? 122 Illness What is the child's mood when he/she is sick? intensity of response Sensory What does the child do when she/he hears a loud sound? Bright light? Does your child react to small/large changes? phone ringing, lights on How does your child react to the phone ringing? Procedures What is your child's mood during these activities? How did your child react the first time these procedures were done? Over time? Are you able to distract the child? Response to People Does your child act differently if there are different people around? How does the child respond to new people? Over time? How does your child respond to familiar people? intensity of reaction New Places; Situations, and Things How does your child react to new things? places? Does this reaction change over time? Are you able to distract him/her when he she is interested in a new a thing? 5 intensity 2 ‘ ‘ ' ".‘.0-.fl:1 APPENDIX D 123 Activity Level This category refers to the degree of motor activity in the infant's behavior. All activities that have a motoric component contribute to the activity level score. Activity may manifest as gross body movement, use of arms and legs, vigor of eating, sucking, etc., and amount and interest in exploratory behavior. The rating of activity level should be made for each scoring interval and is based upon the frequency of activity occurring during the two minute rating interval. High activity refers to the exhibi- tion of fairly continuous (and usually considerable) amounts of motoric activity. Examples of this are: moves a great deal in sleep; spends most of interval wriggling around or off bed or couch; steadily kicks, splashes or plays in tub without much inaction; much squirming during dressing; moves a lot during feeding - grabs spoons, bangs table, etc. High moderate activity refers to behaviors with a motor compo- nent that is less than continuous but in which the breaks in activity are generally short and not very frequent. Example: enjoys bath - or quiet play; in playpen entertains self for significant amounts of time by manipulating objects, moving around, etc. - periods of rest or inactivity are short and widely spaced. Low moderate activity is characterized by intermittent bursts of activity with fairly frequent breaks - rests or periods of 124 inactivity. Examples are: feeding is characterized by some motor activity interspersed with periods of quiet and relative inactivity; play alone involves intermittent activity such as crawling or explor- ing, manipulating toys, etc. combined with moments of quiet, visual inspection or passivity. Low activity level describes little or small amounts of move- ment. This would include lying quietly with little or no kicking in the bath; little movement during sleep; stays in one place when left on floor or courch; little excitement around food; accepts food but eats slowly and intermittently; remains in playpen and quietly plays or watches people and objects. Remember, developmental level must be taken in account when rating activity level. At different ages, the infant has differing ways of being active. Examples: At three months activity will be mostly observed as grasping, head movement, and arm and leg movements. Coordinated motoric activity is infrequent and much of the activity may appear to be random or uncontrolled. Banging, kicking, grasping, and head turning is the extent of the purposeful behavior likely to be observed. At this age Activity level is rated on the amount of motor movements observed. Constant arm movements would be rated as "high" even if the movement is only a non-productive swinging or banging. By age 6 months a greater sophistication in motor movement is likely to be observed. Sitting with only minimal support and possibly crawling will be in the infant's repertoire. Arm and leg coordination should be better and a more sophisticated grasp is 125 available. More of the infant's motor activity is likely to be purposeful and thus activities such as picking up and playing with objects, manipulating a bottle, and rolling over will be observed. With more purposeful actions characterizing the infant's movement, changes in the tempo of activity is more likely. An interval may still be scored as high activity when short lulls or decreases in motoric involvement separate periods of constant motor activity. By age nine months the infant may be trying to stand, is usually able to crawl with great skill, shows good eye-hand coordi- nation, and is usually quite purposeful in his/her actions. The full range of motor behaviors should by now contribute to the scoring of activity level. The intervals will more likely contain episodes of movement from one location to another, experimentation at standing, and coordinated activities with another person. These behaviors should all go into the scoring of activity level. Again frequency and length of the non-activity periods is the critical factor in the L scoring. 1 126 Mood Mood describes the amount of pleasant, joyful, and friendly behavior as contrasted with unpleasant, crying, unfriendly behavior. Thus Mood describes the nature of the feeling being expressed by the infant. Positive Mood should be scored when expressions of happy/friendly feelings predominates during the interval. Examples of positive Mood include: smiles and coos when sees bottle being warmed; likes bath - splashes and plays; smiles during washing or dressing; friendly and smiles at new faces. Moderate Mood describes an absence of observable positive and negative affect. The infant may briefly smile or seem fussy, but overall, he/she remains fairly placid and unemotional. Examples: shows no objection to dressing and undressing; sits quietly in high chair or on lap with neither protest nor excitement; plays easily - verbal and facial expressions are neutral. Negative Mood describes negative or protestive expres- sions such as: cries when given food doesn't like; fusses and cries during dressing; cries and attempts to escape from new people. .Ngtg: A child who is affectless - experiences an absence of feelings, should pgt be scored for Moderate Mood. Score affectlessness as Negative Mood and note this special state with an "a" beside the rating. When more than one affective state occurs during a scoring interval, your rating should be the mood state that had the highest 127 relative frequency. Thus, if the infant expressed positive feelings for 55% of the interval and showed moderate mood for 45% of the time, score the interval as positive. When all three mood states are observed relative frequency should again be examined. Relatively equal amounts of positive and negative expressions should be con- sidered as cancelling each other out. Thus if the infant was 50% negative, 30% moderate and 20% positive, a rating of negative should be given. If a proportion of 40% positive, 25% moderate, and 30% negative occurs, the extreme ratings can be considered as equal and moderate should be scored. If a 40% positive, 40% moderate and 20% negative is observed, score moderate; a 40-30-30 would also be moderate. 128 Intensity of Response Intensity refers to the amount of energy involved in a response irrespective of its quality. This category will be scored by observ- ing the "how much" of an infant's reaction and is unaffected by the type of response. The behavior to be rated may be a vocalization, energy of play, bath behavior, response to a dirty diaper, etc. Any discrete intentional behavior may be rated for intensity. The only time intensity is not assessed is when the infant is not "doing something" such as sleeping, sitting quietly, etc. Intensity may be scored as "intense" or "mild." An intense response is one in which the behavior has considerable energy, drive, loudness, or forcefulness. Examples: cries when sun is in eyes; when full spits food out of mouth and pushes spoon away; when dressing does much wiggling and cries when told 'no'; seems "terrified" with F* loud unfamiliar noises; during diapering smiles, helps, and seems . quite excited; "adores" bath; goes into spasms of laughter when played with roughly; great interest in a new person. Mild intensity describes low energy reSponses; the behavior of the infant is mild, P ggjet, "easy to tolerate", etc. Examples include: squints at bright light but doesn't cry; with loud noise startles but doesn't cry; fusses momentarily when clothing pulled over head; quite cooperative when shoes and socks taken off; seems to enjoy bath but doesn't mind if bath is skipped; plays alone without lots of noise; mild reaction to strangers. 129 Regularity This category refers to the rhythmicity or regularity of repe- titive (biological) functions. These include the sleep/wake cycle, rest and activity periods, eating and appetite behavior and bowel and bladder function. What is to be evaluated is whether the infant has established a pattern of functioning in each of these areas. If there is a clear "schedule" which the infant normally adheres to, then regularity should be scored. If the infant does not engage in the biological function in a consistent or regular pattern, scbre non-regularity. Scoring examples for regularity: goes to bed and wakes up at approximately same time every day; naps at about same time each day; one to two bowel movements a day, regular, usually within same part of day; 3 regular meals a day, no extra bottle; food intake similar from day to day. .13.: :wx. macs. .Ja‘uaw 0 vs . Irregular behaviors include: sleeping does not follow a schedule, sometimes awakens late other times early; hunger at dif- “1:41. ferent times, cannot predict when will need to be fed; day naps irregular; bowel movements irregular either in terms of time or frequency; eats different amounts of food from day to day. 130 Approach/Withdrawal This cateogry is defined as how an infant initially responds to a new stimulus. This includes any type of new experience such as a toy, person, food, a procedure such as nail cutting, or a new auditory (balloon popping) or visual experience. Please note the infant, i.e., there must be a real opportunity for the infant to interact with the stimulus. Thus, if a toy is shown to a child but not easily available for play this category should not be scored. A behavioral sequence may be scored either "approach" or "with- drawal". Approach should be scored for positive, interested, inves- tigatory, curious, quick-adjustment behaviors. Examples of Approach include: smiling at observer when first meet; quickly investigates and plays with new toy; reacts with interest to new animal; tries new foods when offered. Withdrawal responses are behaviors which indicate an avoidance of the new stimulus. This may take the form of active rejection of the stimulus, an avoidance of contact with the stimulus, or a lack of attention to the stimulus. Examples are:. cries when first introduced to Observer; exhibits little interest when first shown a new toy; makes face, gags on, or rejects a new food. Ngtg: that this attribute category only refers to first exposure to a new stimulus. Adjustment over time is scorable on the next category - Adaptability. 131 (Ngtg: There will probably be few opportunities to make ratings on the attribute. Likely instances are your (the observer's) first interactions with the infant, introduction by the mother of a new toy or balloon provided by the O, a new food, or a new non-aversive noise such as a bell. 132 Adaptability This category is concerned with the sequence of response to a new stimulus. Emphasis is on the ease or difficulty with which an initial response is modified in a socially desirable direction. An adaptive response describes the acceptance or acquiescence to a previously rejected or disliked stimulus - that which he/she pre- viously withdrew from. Examples: An infant initially rejecting a new toy but later starts to play with it; eats a food that he started out not liking. A non-adaptive response is a continued non-acceptance of a stimulus to which the infant, upon original presentation, had a withdrawal response. This would take the form of continuing to reject a new toy; lack of adjustment to a new procedure such as nail cutting so that nails must be out while infant is distracted or asleep; makes face and refuses to eat a food even after many expo- § sures. Ngtg that Adaptability includes a temporal dimension and the : observer needs to compare present behavior with that observed earlier. % Adaptability will be scored on those stimuli than were earlier scored E on the Approach/Withdrawal dimension. Thus, likely stimuli are: i the new toy and balloon, response to Observer, and adjustment to a new food. Ngtg: Adaptability should be scored for stimuli that receive either Approach or Withdrawal responses. It is likely that previously obtained Approach responses will receive high Adaptability ratings, but be sure to score this. 133 Threshold of Responsiveness This category refers to the amount of stimulation necessary to evoke (elicit) an observable response. The quality or nature of the eliciting stimulus is not relevant to the rating nor is the ultimate response evoked by the stimulus. What is important is the intensity of the stimulus that is applied to the infant. Thus, this category is concerned with the "sensitivity" of the child. Responses to sensory stimuli, environmental objects, and social contacts are to be rated. Examples of High Threshold are: does not indicate (by crying, fussing) that diaper is wet; pays little atten- tion to new people; no startle reaction to noises or lights; shows no difference between breast and bottle feeding; spits out a disliked variety of food but will eat it if mixed with a little of another food. Moderate Threshold refers to a middle range of threshold sensi- F“ tjyity, Examples are: startles to loud noises but not to sounds of more moderate to low levels of intensity; takes foods similar to that which he/she normally likes but rejects food dissimiliar in texture, taste or smell; cries when new person comes close but toler- ates at a distance. Low threshold describes a low level of stimuli necessary to evoke a response. Examples are: startles if a door is closed; if diaper is full cries and fusses until changed; loves fruit but doesn't eat if put in with a little cereal; cries when sees a new person; notices any little change; strong likes and dis- likes of food. 134 Distractibility This category describes whether the infant's current behavior can be interferred with or altered by the introduction of an external stimulus. Thus the attribute refers to the degree to which the child can be directed away from an on-going activity. High Distractibility describes a readiness to be diverted from an activity_upongpresenta- tion of an external stimulus. Examples of high Distractibility are: child stops crawling to electric plug when presented with toy truck; when picked up, stops crying even though hungry; doesn't like to be washed unless a game is conducted at same time; tantrum can be stopped by initiating a new activity. Low Distractibility refers to a ten- dency to not be diverted from an activity or behavior unless substan- tial stimulation is given. Examples: cries and struggles during dressing and cannot be diverted; when hungry continues to cry until k. fed; if toy is lost or taken away, cries even when given an alterna- tive toy; continues to play with own toy despite waving of another toy nearby - eventually takes new toy after many presentations. ‘W'... ;~ ,3 Note: If an infant is engaged in an activity and a person tries to "seduce" him/her away from that activity, if the child quickly returns to the activity after a brief interaction with the person, score Low Distractibility. 135 Persistence and Attention Span This category refers to the length of time spent on an acti- vity and the ability to maintain an activity despite obstacles. The attribute can be assessed on both self-initiated and planned activi- ties. The category is scored in all situations in which the infant is described to have had an interest in an object or activity and there existed a potential for engagement in the activity or object. What is rated is the length of time an infant remains involved with/pursues an activity or the effort in maintaining an activity despite impediments or obstacles. The obstacle may be external such as mother's refusal to allow the infant to touch an object, or internal for example an infant continuing to stand up although he falls down repetitively. The latter would be scored as high persis- tence. Others scoring examples of high persistence: refuses cereal for breakfast and yells and bangs until given preferred food; in crib, will look at birds 1} hours; entertains self through play in playpen for I to 1 hour. Examples of low persistence include if object is slightly out of reach, does not attempt to get it, just goes to another toy; if frustrated, quickly stops upset without need for distraction; cannot play alone for very long periods of time (less than I hour). - —-II 'l'...‘ Ra‘s-Mm." :IX'S-I m- .M 3:; C; LIST OF REFERENCES LIST OF REFERENCES Allport, G. W. Pattern and growth in personality. New York: Holt, Rinehart and Winston, 1961. Bailey, M. M. An observational study of mothers' commands and their young children's responses. Unpublished doctoral disserta- tion, Michigan State University, 1977. Bates, J. E. The concept of difficult temperament. Merrill-Palmer Quarterly, 1980, 26, 299-309. Bates, J. E., Freeland, C. A. B., Lounsbury, M. L. Measurement of infant difficulties. Child Development, 1979, 59, 794-803. Billman, J. The relationship of temperament to peer interaction at the preschool level. Dissertation Abstracts International, 1979, 32(A), 2131-2132. Billman, J. and McDevitt, S. C. Convergence of parent and observer ratings of temperament with observations of peer interaction in nursery school. Child Development, 1980, 51, 395-400. Birns, B., Barten, 5., and Bridger, W. H. Individual differences in temperamental characteristics of infants. Transactions of the New York Academy of Science, 1969, 31, 1071-1082. p Bonem, H. Measurement of infant temperament. Unpublished masters thesis. Michigan State University, 1978. Bonem, H. Unpublished data. Michigan State University, 1979. Bronson, W. C. Stable patterns of behavior: The significance of enduring orientations for personality development. In i Hill, J. P. (ed.), Minnesota Symposia on Child Psychology " (Vol II). Minneapolis: University of Minnesota Press, 1969. Bronson, W. C. Central orientations: A study of behavior organiza- tion from childhood to adolescence. In Jones, M. C., Bayley, N., MacFarlane, J. W., and Honzik, M. D. (Eds.), The Course of Human Development. Waltham: Xeroz College Publications, 1971. 136 137 Buss, A. H., Plomin, R., and Willerman, L. The inheritance of temp- erament. Journal of Personality, 1978, 31(4), 513-524. Buss, A. H. and Plomin, R. A temperament theory of personality development. New York: John Wiley & Sons, 1975. Cameron, J. R. Parental treatment, children's temperament and the risk of childhood behavioral problems. American Journal of Orthopsychiatry, 1978, 58, 140-147. Campbell, 5. B. G. Mother-infant interaction as a function of maternal ratings of temperament. Child Psychiatry and Human Development, 1979, 10, 67-76. Carey, W. B. A simplified method for measuring infant temperament. Journal of Pediatrics, 1970, 11, 178-184. Carey W. B. Measuring infant temperament. Journal of Pediatrics, 1972, 81, 414. Carey, W. B. Clinical applications of infant temperament measure- ments. Journal of Pediatrics, 1974, 84(5), 756-758. Carey, W. B. Personal communication. December, 1976. Carey, W. B., Lipton, W. L. and Myers, R. A. Temperament in adopted and foster babies. Child Welfare, 1974, 53(6), 352-359. Carey, W. B. and McDevitt, S. C. Revision of the infant temperament questionnaire. Pediatrics, 1978, 61, 735-739. Chronbach, L. J. and Meehl, P. E. Construct validity in psycholo- E“ gical tests. Psychological Bulletin, 1955, 22, 281-302. - - I'fi’f‘e' I... Corsini, D. A. and Doyle, K. Temperament traits of preschool chil- dren: Across setting consistency. Paper presented to the Society for Research in Child Development, San Francisco, 1 March, 1979. i u‘ 4. A». M. Diamond, 5. Personality and temperament. New York: Harper and Brothers, 1957. Erikson, E. H. Childhood and society. New York: W. W. Norton & Company,TT950. Escalona, S. K. The roots of individuality. Chicago: Aldine, 1968. Escalona, S. K. and Heider, G. M. Prediction and outcome. New York: Basic Books, 1959. 138 Feiring, C. and Lewis, M. Temperament: Sex differences and stabi- lity in vigor, activity and persistence in the first three years of life. Journal of Genetic Psychology, 1980, 136, 65-75. Feiring, C. and Taylor, J. The influence of the infant and secon- dary parent on maternal behavior: Toward a social systems view of infant attachment. Unpublished manuscript. Univer- sity of Pittsburgh, 1976. Freedman, D. G. Human infancy: An evolutionary perspective. New York: Erlbaum, 1976. Freud, S. A general introduction to_psycho-ana1ysis. Liverwright Publishing Company, 1935. Frodi, A. M., Lamb, M. E., Leavitt, L. A., Donavan, W. L., Neff, C. and Sherry, D. Fathers' and mothers responses to the faces and cries of normal and premature infants. Developmental Psychology, 1978, 14, 490-498. Garside, R. F., Birch, H., Scott, D. McI., Chambers, 5., Kolvin, I., Tweedle, E. G., and Barber, L. M. Dimensions of temperament in infant school children. Journal of Child Psychology and Psychiatry, 1975, 16, 219-231. Gessell, A. and Ames, L. B. Early evidences of individuality in the human infant. Journal of Genetic Psychology, 1937, 41, 335. Graham, P., Rutter, M., and George, S. Temperamental characteris- tics on predictors of behavior disorders in children. American Journal of Orthopsychiatry, 1973, 43, 328-339. Heymans, G. Ueber einige psychischen korrelationen, zeitschr. f. agrew/ Psychology, 1908, 1, 313-381. Hollingshead, A. B. Two factor index of social position. New Haven: S.N., 1957. Kagan, J. Change and continuity in infancy. New York: John Wiley & Sons, 1971. Kagan, J., Lapidus, D. R. and Moore, M. Infant antecedents of cognitive functioning: A longitudinal study. Child Develop- ment, 1978, 42, 1005-1023. Katcher, A. Unpublished data. Cited in Thomas, A. and Chess, S. Temperament and Development. New York: Brunner/Mazel, 1977. 139 Kelly, R. The relation of infant's temperament and mother's psycho- pathology to interactions in early infancy. In K. F. Riegel and J. A. Meachan (Eds.), The developingyjndividual in a changing world (Vol II). Chicago: Aldine, 1976. Kronstadt, D., Oberklaid, F., Ferb, T. E., and Swartz, J. P. Infant behavior and maternal adaptations in the first six months of life. American Journal of Orthopsychiatry, 1979, 42, 454-464. Lambert, N. M. and Windmiller, M. An exploratory study of tempera- ment traits in a population of children at risk. Journal of Special Education, 1977, 1111), 37-47. Lemly, E. B. and Schwartz, J. C. Temperament and child- -rearing antecedents to two-year olds' reaction to male and female strangers. Paper presented at the Society for Research in Child Development, San Francisco, March, 1979. Levandovskii, N. G. On the structure of personality. Voprosy psikhologii, 1970, 16(3), 51-56. Lounsbury, M. L. Acoustic properties of and maternal reactions to infant cries as a function of infant temperament. Disserta- tion Abstracts International, 1979, 19, 4585-4586. Lyon, M. E. and Plomin, R. The measurement of temperament using parental ratings. Jouranl of Child Psychology & Allied Disciinnes, 1981, 22, 47-53. Matheny Jr., A. P. and Dolan, A. B. A twin study of personality P“ and temperament during middle childhood. Jouranl of Research in Personality, 1980, 14, 224-234. 5 ‘1;th Maurer, R., Cadoret, R. J. and Cain, C. Cluster analysis of chi1d- ; hood temperament data on adoptees. American Journal of Orthopsychiatry, 1980, 59, 522-534. McDevitt, S. C. A longitudinal assessment of continuity and stabi- lity in temperamental characteristics from infancy to early childhood. Unpublished doctoral dissertation, Temple Univer- sity, 1976. McDevitt, S. C. and Carey, W. B. The measurement of temperament in 3-7 year old children. Journal of Child Psychology and Psychiatry, 1978, 21, 245-253. McDevitt, S. C. and Fox, M. Reliability and construct validity of the Carey Infant Temperament Scale: A reanalysis. Unpub- lished manuscript. Temple University, 1976. 140 Murphy, G. Personality: Biosocial approach to origins and struc- ture. New York: Harper & Brothers, 1947. Murphy, L. B. Explorations in child personality. In Rabin, A. 1., Aronoff, J., Barclay, A. M., and Zucker, R. A. (eds.), Further explorations in personality. New York: John Wiley 8 Sons, In press. Murphy, L. B. and Moriarty, A. E. Vulnerability, coping:and1growth: From infancy to adolescence. New Haven: Yale University Press, 1976. Nunnally, J. C. Psychometric theory. New York: McGraw-Hill, 1967. Pederson, F. A., Anderson, B. J., and Cain, Jr., R. L. A methodology for assessing parental perception of infant temperament. Paper presented at the Fourth Biennial Southeastern Conference on Human Development, April, 1976. Peery, J. C. and Toney, M. K. Sociometric status and parent reports of temperament in preschool children. Journal of Psychology, 1979, 192, 97-99. Perrson-Blennow, I. and McNeil, T. F. A questionnaire for measure- ment of temperament in six-month-old infants: Development and standardization. Journal of Child Psychology and Psychiatry, 1979, 29, 1-13. Plomin, R. A twin and family study of personality in young children. Journal of Psychology, 1976, 24, 233-235. Plomin, R. and Rowe, D. C. A twin study of temperament in young children. Journal of Psychology, 1977, 21, 107-113. Rothbart, M. K. and Derryberry, D. Development of individual dif- ferences in temperament. In M. Lamb and A. Brown (eds.), Advances in developmental psychology, Vol. I. Hillsdale, N.J.: Erbaum, 1981. V Rowe, D. C. and Plomin, R. Temperament in early childhood. Journal of Personality Assessment, 1977, 41(2), 150-156. Rutter, M., Korn, S., and Birch, H. B. Genetic and environmental factors in the development of primary reaction patterns. British Journal of Social and Clinical Psychology, 1963. g, 161-173. Scarr, S. Social-introversion as a heritable response. Child Develop: ment, 1969, 4Q, 823-832. 141 Scarr, S. and Salapatek, P. Patterns of fear development during infancy. Merrill-Palmer Quarterly of Behavior and Develop: ment, 1970, 16, 53-90. Scholom, A. H. The relationship of infant and parent temperament to the prediction of child adjustment. Unpublished doctoral dissertation, Michigan State University, 1975. Scholom, A., Zucker, R. A., and Stollak, G. E. Relating early child adjustment to infant and parent temperament. Journal of Abnormal Child Psychology, 1979, 1, 297-308. Sheldon, W. The varieties of temperament: A psychology of consti- tutional differences. New York: Harper, 1942. Shirley, M. The first two years: A study of twenty-five babies. (Vol. III: Personality manifestations). Minneapolis: University of Minnesota Press, 1933. Sullivan, H. S. The interpersonal theory of psychiatry. New York: W. W. Norton & Co., 1953. Thomas, A., Chess, 5., Birth, H. and Hertzig, M. A longitudinal study of primary reaction patterns in children. Comprehen- sive Psychiatry, 1960, 1, 103-112. Thomas, A., Chess, S. Birch, B. G., Hertzig, M. E., and Korn, S. Behavior individuality in early childhood. New York: New York University Press, 1963. Thomas, A., Chess, S., and Birch, H. G. Temperament and behavior w* disorders in children. New York: New York University, 1968. ~ Thomas, A. and Chess, S. Temperament and development, New York: Brunner/Mazel, 1977. Torgersen, A. M. Temperamental differences in infants: Their cause as shown through twin studies. Unpublished doctoral disser- tation, University of Oslo, 1973. i U.S. Bureau of the Census. Census of Population: 1970 Detailed Characteristics, Final Report Pc (1)-024 Michigan. U.S. Government Printing Office, Washington, D.C., 1972. Vaughn, 8., Crichton, L., Teraldson, B. and Egeland, B. The assess- ment of infant temperament: A critique of the Carey Infant Temperament Questionnaire. Unpublished manuscript, Univer- sity of Minnesota, 1980. Watson, J. B. Behaviorism. New York: The People's Institute Publishing Co., 1925. 142 Wilhoit, P. 0. Assessment of temperament during the first months of life. Unpublished doctoral dissertation, Florida State University, 1976. Wilson, C. D. and Lewis, M. Temperament: A developmental study of stability and change during the first four years of life. Unpublished manuscript. Educational Testing Service, 1977. "‘1mmm