SOCIOMETRIC STATUS AND PEER PREFERENCE ‘N THE PRESCHOOL AS RELATED TO ORDINAL POSITION. AGE. SEX, RACE, SOCIOECONOMIC CLASS. AND FAMILY BACKGROUND CHARACTERISTICS Thesis for the Degree of M. A. MICHIGAN STATE UNIVERSITY KAREN ELIZABETH TRACY 1974 ll! 0 i 1.. WUQHIILIIEIMINIWill!!!”(HIIHIHWIIHUIHH 10453 8990 MSU ; LIBRARIES RETURNING MATERIALS: PIace in book drop to ~remove this checkout from your record. FINES wiII be charged if book is returned after the date stamped beIow. ____ ABSTRACT SOCIOMETRIC STATUS AND PEER PREFERENCE IN THE PRESCHOOL AS RELATED TO ORDINAL POSITION, AGE, SEX, RACE, SOCIOECONOMIC CLASS, AND FAMILY BACKGROUND CHARACTERISTICS By Karen Elizabeth Tracy The major purpose of this study was to examine whether certain family background factors, as well as age, sex, and race influenced peer choice and peer status in the preschool class. Another purpose was to compare samples of 3 l/2- to 5-year-olds taken from two different popu- lations; 67 middle class, mostly white children from Michigan State University's Laboratory Preschool (LPS) and ll6 lower class, mainly black children from Head Start classrooms in Pontiac and Detroit. The third purpose was to investigate various aspects of use of the instru- ment, Boger's (l969) Play Situation-~Picture Board Sociometric Technique (PSPBS). The children were tested using the PSPBS, and a sociometric status score was computed for each child. Regression analysis was used to detenmine whether relationships existed between various characteristics of an individual and his sociometric score level. In addition, charac- teristics of each individual were regressed against those of his first peer choice to test whether certain relationships existed between the two. Regressions were also run on negative peer scores for a subsample is... F. I II Karen Elizabeth Tracy of the LPS population and on individuals from the total sample who chose the same peer in all three choice situations. Distributions of sociometric status scores for the LPS sample and for the HS sample were contrasted. Agreement between samples regarding hypothesis testing results was examined. The LPS sample were retested at 2 to 4 week intervals. First, second, third, and negative choices were compared between test and re- test and the percentages of choices reproduced were calculated. Individual sociometric status scores for both tests were correlated. Negative scores and weighted scores were compared with positive, un- weighted scores. Results indicated that sex, race, age, ordinal position, father- absence or -presence, and presence and spacing of just younger and just older siblings were significantly related to peer choice and/or socio- metric status. The most highly significant results, that the children preferred friends of the same sex and of the same race as themselves, pertained to both LPS and HS samples although some significant results were different for each. Distribution of sociometric status scores dif- fered between samples, a greater percentage of tHe HS than the LPS sample's scores falling into the very high and the never chosen cate- gories. Reliability results (for the LPS sample only) for the PSPBS Tech- nique attained a very respectable stability coefficient of .70. Having each child give a single negative choice did not increase the test's sig- nificance. Weighted scoring held no advantage over unweighted scoring. Karen Elizabeth Tracy Differences between various classrooms indicated that social structure was more disparate among individual classrooms than between the lower and upper age halves of the sample. SOCIOMETRIC STATUS AND PEER PREFERENCE IN THE PRESCHOOL AS RELATED TO ORDINAL POSITION, AGE, SEX, RACE, SOCIOECONOMIC CLASS, AND FAMILY BACKGROUND CHARACTERISTICS By Karen Elizabeth Tracy A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Family and Child Sciences 1974 ACKNOWLEDGEMENTS To Paul Muhs I am grateful for skilled and enthusiastic assistance in testing the Laboratory Preschool children. To Dr. Robert Boger, Director of the Institute for Family and Child Study, Michigan State University, I am indebted for permission to use the Play Situation—-Picture Board Sociometric Technique (l969) which he designed. I am also grateful for his loaning of test equipment, his funding of a testing assistant, and his permission to use data from the 1968-69 Head Start Evaluation. To Anne Stevenson, Michigan State University Laboratory Preschool Coordinator, I owe thanks for arranging for testing time and space and for offering warm encouragement throughout this project. To Michael Siwek, Katherine Lange, and Deborah Hubbell, Michigan State University Laboratory Preschool Teaching Interns, I am greatly obliged for cheerful cooperation and assistance during testing of their classes. To Dr. Eileen Earhart, Chairman of the Department of Family and Child Sciences, I am very grateful for a wealth of ideas, suggestions, and astute criticism from the inception of this project to its conclu- sion. To Ronald L. Tracy, my husband, I am particularly grateful. He served as computer programmer and statistical consultant for this pro- ject and also offered continuous encouragement, displaying keen inter- est in all facets of the project. ii TABLE OF CONTENTS List of Tables ....................................................... vi List of Figures ..................................................... vii Chapter 1: Statement of the Problem and Hypotheses ................... l Assumptions ...................................................... 4 Socioeconomic Differences ........................................ 6 Nursery School Socialization ..................................... 7 Definition of Terms .............................................. 7 Hypotheses ....................................................... 9 Overview ........................................................ 11 Chapter 2: Review of the Literature ................................. 13 Ordinal Position Theory ......................................... l3 Empirical Investigations of Birth Order Effects ................. 18 Achievement...’ ................................................ 18 Responsibility ................................................ 19 Creativity .................................................... 19 Dependency .................................................... 20 Motivation .................................................... 21 Empathy ....................................................... 21 Anxiety ....................................................... 22 Sibling Sex Position Effects .................................. 22 Critiques of Ordinal Position Studies ........................... 23 Explanatory Theories ............................................ 25 Physiological Causation Theory ................................ 25 Sibling Effects ............................................... 26 Dethronement .................................................. 27 Socialization Theories ........................................ 29 Family size effects ........................................ 30 Ordinal position and parental behavior ..................... 31 Role theory ................................................ 32 Parental Variables ...................................... ........32 Family Size ................................................... 33 Socioeconomic Influences ........................................ 34 Preschool Social Behavior ....................................... 38 Selfishness ................................................ 38 Succorance and aggression .................................. 39 Socioeconomic Differences and Preschool Behavior ................ 39 Verbal functionning ........................................ 40 Conclusion ................................................. 40 Birth Order and Sociometric Choice .............................. 41 TABLE OF CONTENTS (cont'd) Sociometric Testing and Preschool Children ...................... 42 Scoring ....................................................... 44 Validity ...................................................... 45 Reliability ................................................... 46 Intensity ..................................................... 47 Peer Desirability and Personality ............................. 48 Implications for the Present Study .............................. 49 Chapter III: Method ................................................. 51 Sampling.....................; .................................. 51 Laboratory Preschool Sample ................................... 52 Head Start Sample ............................................. 53 Both Samples .................................................. 55 Testing Procedures .............................................. 55 Testing Procedure for the LPS Sample .......................... 55 Apparatus .................................................. 55 Setting .................................................... 56 Testers .................................................... 56 Procedure .................................................. 57 Retesting ................................................. 58 Testing Procedure for the Head Start Sample ................... 59 Statistical Procedures .......................................... 59 Regression Analysis and Correlation Coefficients .............. 60 Regression Analysis and ANOVA ................................. 61 Reliability Measures .......................................... 64 LPS and HS Scores Compared ................................... .64 Model ........................................................... 64 Peer Desirability ............................................. 68 Peer Preference Relationships ................................ .68 Chose All Three ............................................... 69 Scoring on LPS Negative Subsample ............................. 69 Negative Peer Choice .......................................... 69 Discussion .................................................... 70 Chapter IV: Results ................................................. 72 Peer Desirability ............................................... 73 Hypothesis Testing ............................................ 76 Peer Preference ............................................ .....79 Hypothesis Testing ............................................ 79 Summary ....................................................... 87 Additional Findings ........................................... 87 Sex differences ............................................ 87 Standard age ............................................... 89 Father-absence or presence .............................. ...89 Race ....................................................... 89 Other Findings ..................................... . .......... 90 Negative Peer Relationships ................................... 90 Same Peer Chosen Three Times .................................... 93 Hypothesis Testing ............................................ 93 Comparison of LPS and HS Samples ........................... .....97 Hypothesis Testing ............................................ 97 iv TABLE OF CONTENTS (cont'd) Additional Findings .......................................... 100 Peer Preference ........................................... 100 Chose all three ........................................... 101 Negative Scores ................................................ 101 Hypothesis Testing ........................................... 101 Reliability .................................................... 102 Unweighted Versus Weighted Scoring ............................. 105 Chapter V: Discussion ............................................. .109 Sociometric Status ............................................. 109 Peer Preference Relationships .................................. 113 Non-Hypothesized Findings .................................... 114 Standard age .............................................. 115 Father-absence or -presence ............................... 115 Race ...................................................... 116 Assorted other results .................................... 116 Negative Peer Relationships .................................. 117 Summary ................................................... 118 Chose All Three ................................................ 119 Comparison of HS and LPS Samples ................... . ........... 119 Other Findings ........................... . .............. .....121 Negative Scores ........................................... .....122 Reliability .................................................... 124 Scoring ........................................................ 127 Chapter VI: Conclusion ............................................. 128 Major Results .................................................. 128 Sociometric Status ........................................... 128 Peer Preference .............................................. 128 Head Start and Laboratory Preschool Results Compared ......... 129 Use of the Instrument ........................................ 129 Suggestions for Improvement .............. . ..................... 130 Instrument ............................... . ........... ... ..... 130 Statistical Analysis ......................................... 131 Implications and Recommendations for Further Research .......... 132 ' Appendix A: Background Characteristics Data ........................ A-I Appendix B: Sociometric Test Data ................................ ..B-I Bibliography ........................................................ 136 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table LIST OF TABLES 1 : Sample Characteristics ..................................... 54 2 : Results: Regression Coefficients and Standard Errors of Peer Desirability Models ............................. 74-75 3 : Summary of Sociometric Status Results; Regression Coefficients and Standard Errors ........................... 78 4 : Results: Regression Coefficients and Standard Errors of the Peer Preference Models for the Whole Sample ...... 80-81 5 : Results: Regression Coefficients and Standard Errors of the Peer Preference Models for the LPS Sample ........ 82-83 6 : Results: Regression Coefficients and Standard Errors of the Peer Preference Models for the HS Sample ......... 84-85 7 : Summary of Preference Relationships Results; Regression Coefficients and Standard Errors ................ 88 8 : Results: Regression Coefficients and Standard Errors for Negative Peer Relationships Models .................. 91-92 9 : Summary of Negative Peer Trait Results; Regression Coefficients and Standard Errors ........................... 94 10: Results: Characteristics of S; who Chose Same Peer Three Times; Regression Coefficients and Standard Errors...95 11: Summary of Characteristics of Ss who Chose the Same Peer Three Times; Regression Coefficients and Standard Errors ................... . ........ .... ............ 98 12: Summary of Negative Score Results: Regression Coefficients and Standard Errors ..... ... ...... . ........... 103 13: Percentages of Rank Choices Reproduced .................... 104 14: Reliability of LPS Sociometric Status Scores in Correlation Coefficient Terms .......................... 106 15: Sociometric Scoring Technique Comparisons: Total Regression Statistics ..................................... 108 vi LIST OF FIGURES Figure 1: Variables, Value Assumed, and Symbol .................... 65-67 Figure 2: Distribution of sociometric status scores for the Head Start sample .................................. 99 Figure 3: Distribution of sociometric status scores for the Laboratory Preschool sample ........................ 99 vii CHAPTER I: STATEMENT OF THE PROBLEM AND HYPOTHESES The major purpose of this study was to examine whether certain family background factors, as well as age, sex, and race influence peer choice and peer status in the preschool class. A second purpose was to investigate various aspects of the use of the instrument, Boger's (1969) Play Situation-Picture Board Sociometric Technique. The third purpose was to compare results of samples taken from two different populations; middle class, mostly white children from Michigan State University's Laboratory Preschool (LPS) and lower class, mainly black children from Head Start (HS) classrooms. The primary family background characteristic of interest in this study is ordinal position. Many ordinal position theorists have asserted that first-horns are less socially competent and popular than later- borns. Most of the research carried out on this issue, however, has been done on college students. Since the family comprises nearly all of a small child's social world, it would seem that family and peer influences would be more evident in a preschool child than in a college student. As personality structure and social patterns of an individual are established during the preschool years, tendencies towards social competence and congeniality, or the lack of them, should be evident by the age of three or four. The major issue concerning ordinal position in this study is thus to test the hypothesis that first-barns are less pop- ular with their nursery school peers than are later-borns. Critics of birth order research maintain that many studies con- cerning birth order effects are of doubtful validity because of the researchers' failures to control possible confounding variables. To avoid this pitfall and to increase the predictive value of this study, many variables other than birth order were included. Personal char- acteristics such as age in months, sex, and race of each subject were examined. Other family background characteristics such as family size, socioeconomic status, father absence or presence, sex of all siblings, and spacing between the subject and his closest siblings were additional variables used. Not only was a child's ordinal position examined but also his position among like-sexed siblings. Many of the hypotheses concerning these variables were based on work done by Toman (1969) and by Koch (1955, 1956, 1957, 1966) who theorized that individuals choose friends who bear similar characteristics in sex, age, and ordinal position to those of close siblings. To establish peer desirability ratings, the children in the sample were tested using the Play Situation-Picture Board Sociometric Technique (Boger, 1969). A peer desirability (sociometric status) score was then calculated for each child on the basis of the total number of times he was selected by peers. Each subject's family background and personal statistics were regressed against his own peer desirability score. The total sample included about 30% middle class children from Michigan State University's Laboratory Preschool (LPS) and 70% lower and working class children from Detroit and Pontiac Head Start classes. The LPS testing was done in the spring of 1974 by the experimenter and another trained tester. The Head Start testing had been done in the spring of 1969 by a team from M.S.U.'s Head Start Evaluation and Research Center. The children ranged in age from 3 1/2 to 5 years when tested. It was assumed that the first choice of three made by a child would represent his most preferred peer choice. Each child's first peer choice was then analyzed. Certain background and personal character- istics of the selector were then compared with those of the selectee. Although most children chose a different peer for each of the three play situations presented, a few chose the same peer all three times. The hypothesis tested was that children who made the same choice all three times tended to have a similar cluster of character- istics which set them apart from those who chose three different peers. Several aspects of use of the PSPBS were investigated. Weighted scoring was compared with unweighted scoring. One negative choice was asked of the upper half of the LPS sample in addition to the usual three positive choices. The LPS subjects were retested two to four weeks after initial testing to establish reliability scores on the test. Another focus of this study was on socioeconomic differences in family background characteristics and in sociometric results. The middle class LPS sample was compared to the largely lower class (some working class) Head Start sample. As family size and child-rearing practices differ between middle and lower class families, it was theorized that the hypotheses regarding peer desirability and sociometric choice, based largely on studies and observations of middle class nursery school children, would be less likely to be supported by Head Start than LPS data. Northway (1967) asserted that in nursery classrooms with programs emphasizing social development, the spread of sociometric status scores would be greater than in those classrooms with lesser emphasis in social development. In the former type of classroom, thus, only one or two children would never be chosen and the highest scores would not be very much greater than modal scores. As LPS programs had more emphasis on social development than did the HS programs, which focused more on cognitive development, it was hypothesized that LPS scores would be more widely spread than HS scores with fewer children per class getting scores of zero or extremely high scores. In other words, more LPS than HS children would fall into the middle range of scores and fewer into either the high or low range. As some researchers assert that asking for negative choices adds significance to one's results, a single negative choice was requested of the older half of the LPS sample. Scores of positive choices only were compared with scores of both negative and positive choices to ascertain whether the negative choice increased overall significance. In addition, the hypothesis that individuals who receive several nega- tive choices have certain characteristics in common was tested. Assumptions Several important assumptions were made regarding the instrument used in this study. They are as follows: (1) Preschool children do have social preferences. (2) The Play Situation-Picture Board Sociometric Technique measures social preferences of preschool children. (3) The choices on the PSPBS made by preschool children represent ideal playmate choices which may or may not correspond to actual play partners in the classroom. (4) An individual's sociometric status score on the PSPBS is an index of his popularity with his peers. (5) An individual's first choice on the PSPBS is his most preferred playmate. (6) An individual's first negative choice on the PSPBS is his least preferred playmate. The first assumption, that preschool children do have social pre- ferences, has been recognized as a fact by enough researchers (Northway, 1967; Biehler, 1954; Dunnington, 1957b) that it will not be discussed here. The other five assumptions are dealt with in the ensuing para- graphs. The sociometric questioning technique used in this study is aimed at eliciting a child's choice of ideal playmates. A child's idea of whom he would like to play with may be very different from his statement of whom he does play with. One child may indeed wish to play with the same peers as he actually does play with; another may admire a peer from afar but be too shy to approach him and thus settle for playing with less desired peers. It is questionable, however, if all preschool children understand the difference between the conditional and straight- fbrward nature of such a question. Some children may interpret the question as a direct inquiry as to whom one plays with while others may graSp the ideal sense of the question. As the hypotheses in this study rest on the assumption that the instrument actually measures peer preferences, the results are very much dependent upon the validity of the instrument. Unfortunately, the validity of a child's opinion about whom he would like to play with is impossible to establish. A tester can sometimes make an accurate judgement as to whether a given child is actually making a considered judgement or just a random response on the test. Although the 4-year- olds tested seemed to make their peer choices deliberately and discrim- inatingly, some of the 3-year-olds simply chose the photographs nearest them. Such subjects often displayed a short attention Span and other signs of relative immaturity. At the lower age ranges, therefore, the test may not have 100% validity, but with most children over 48 months of age, the investigator believes it to be a valid measure. Another assumption which has been made is that the child's first peer choice corresponds to his most preferred playmate. Indeed, many subjects tested prefaced their first peer choice statements with the comment, "Here's my best friend." If this assumption is warranted, one would expect this choice to remain more constant over a two- to four week period than less preferred choices. In like manner, if a child's negative choice represents his least preferred playmate, this choice should also remain constant. To test these expectations, the LPS sub- jects were retested after the time periods stated (half after each time period) and reliability indices were established. Socioeconomic Differences Socioeconomic differences in individual characteristics of young children have been widely examined; most recent studies, however, seem to be oriented towards academic and cognitive potential and achievement. Little research has been done concerning socioeconomic effects on pat- terns of peer interaction within groups of homogeneous socioeconomic composition. In a research project which did focus on such a comparison, Berk (1971) found fewer actual interactional differences based on socioeconomic differences than had been expected. Writers who assume that lower class preschool children are as socially inexpert with their peers as they are academically handicapped compared to middle class children may be on unsteady ground. Lower class children, who come from bigger families and usually have more opportunity for neighborhood peer interaction than middle class children, may be more skilled at inter- action with peers from the same subculture at an early age than middle class children. Nursery School Socialization Why preschool children choose the friends they do is not well understood; it is a complex phenomenon. It is hoped that this study might uncover some of the factors involved in nursery school peer pre- ference and peer idealization. It is also hoped that it will add to the body of knowledge on sibling position effects. More knowledge in these areas might help professionals in early childhood education to enhance the social climate in the classroom and to prevent a lack of skill in early social interaction from becoming a permanent handicap to a child. Such knowledge might also be useful for experts in the area of self- concept development and group dynamics. Understanding the framework underlying the social preferences of very young children might provide increased insight into the development of adult prejudices, preferences, sociability, and group behaviors. Definition of Terms PSPBS - Play Situation-Picture Board Sociometric Technique - The test, developed by Boger (1969), used in this study to measure peer preferences. LP§_- Michigan State Laboratory Preschool H§_- Head Start Peer desirability -($ociometric status)score - Score based on the number of times a child is selected as a positive choice by classmates on the PSPBS. This score reflects a child's popularity among his peers; the higher the score, the higher the degree of popularity. A score of zero means a child was never chosen. Negative peer desirability score - Number of times a child was selected as an undesired playmate; measured in negative units. Ordinal position - The rank order held by‘a child in the birth sequence of the children in his family. First-born refers to an only or eldest child. Later-born refers to a child who is not the eldest or only child in his family. The lowest ordinal position rank is first-born; the highest in this particular study is tenth-born. Fatherjpresent - Father lives with the family. Father-absent - Father lives apart from the family. Mother may be widowed, divorced, separated, or unwed. Standard age - The age of a given individual relative to the average age in his classroom; measured in terms of standard deviations. Dominant or major race - The race which the majority of children in a given class belong to. Age gap between subject and sibling - the difference in age between the two, measured in months. Sole member of sex - Child is the only one of his sex in his sibship. Selector - Individual making a choice on the PSPBS. Selectee - Individual chosen by another on the PSPBS. Single-sex sibship - Members of one sex only are represented in the child's sibship, including the child. First-of—sex - Child is the oldest member of his sex in his sibship. Hypotheses I. Concerning peer desirability A. B. C. Peer desirability increases as ordinal position increases. Peer desirability increases as standard age increases. Peer desirability is higher for children from homes with a father present than for those from homes with a father absent. Peer desirability is positively related to membership in the predominant race in the class. As the age gap between a subject and his closest older sibling decreases, the subject's peer desirability will increase. As the age gap between a subject and his closest younger sibling increases, the subject's peer desirability will increase. Subjects with younger brothers will be less popular than those with no younger siblings or younger sister(s). Males who are the sole members of their sex in their families will be less popular than those with brothers. II. Concerning the first peer choice A. B. The selectee will be of the same sex as the selector. If the selector comes from a family in which the father is present, so will the selectee. If the selector is high in peer desirability, he will choose a friend of low peer desirability. III. 10 If the selector is low in peer desirability, he will choose a friend of high peer desirability. The higher the selector's standard age, the more likely the selectee will be of his own sex. The higher the selector's standard age, the more likely the selectee will be of the major race. The selector's standard age will be dissimilar to the selectee's. The selector will be of a different ordinal position from the selectee. If the selector is from a single-sex sibship, the selectee will be of the same sex. If the selector is first—of-sex, the selectee will be of the same SEX . If the selector is first-of-sex, the selectee will be first- of-sex. Regarding the choice of the same peer three times A. B. C. The selector The selector The selector The selector The selector The selector The selector The selector The selector total sample. will will will will will will will will will be of a low standard age. tend to be female. have a younger sibling. have a low peer desirability score. be from a father-absent family. be from a small family (1- to 3-child). be the first of his sex in the family. be of a low ordinal position (first is lowest). be of a younger age than average for the 11 J. The selector will be of a non-predominant race. K. The selector may be influenced by a specific tester. IV. Regarding a comparison of LPS versus Head Start samples A. Results in the predicted directions for hypotheses under sections 1, II, and III above will be stronger for LPS than for Head Start subjects. B. The spread of the sociometric status scores will be wider in the middle range for LPS than for Head Start classes. C. Fewer children per class will receive sociometric scores of zero in LPS than in Head Start classes. 0. Fewer children will receive very high sociometric scores in the LPS than the Head Start sample. V. Regarding subjects who obtain negative scores in the LPS subsample A. The more negative choices a subject receives, the more likely it is that he will be male. 8. The more negative choices a subject receives, the more likely he is to be male and only one of his sex in his family. C. The more negative choices a subject receives, the more likely he is to have a younger sibling. D. The more negative choices a subject receives, the more likely he is to have an opposite-sexed sibling. Overview Five chapters follow this one. Chapter 11 contains a review of the literature and implications of this review for the present study. In Chapter III, the method used in the study is presented including sampling, procedure, statistical methods used, and experimental design. 12 Results are reported in Chapter IV in tabular form supplemented by textual explications. Chapter V is a discussion of the results obtained with emphasis on the relevance of the results to findings and theories reported in the literature. Finally, in Chapter VI, major results are summarized, implications are drawn, and suggestions are made for further research. CHAPTER II: REVIEW OF THE LITERATURE Since this study involves several important areas, several topics are covered in this chapter. The primary topic, ordinal position, is discussed in the first half of this chapter. Major wholistic theories on ordinal position as well as empirical studies and Specific explana- tory theories regarding ordinal position are presented. Directly related factors such as family size, parental satisfaction, and socio- economic influences are also dealt with. Next, social behavior in the preschool is described with a particular emphasis on socioeconomic, sex, and age influences on behavior. Finally, sociometric testing is intro- duced with a discussion of two studies which relate birth order to sociometric status and choice. This section concludes with a discussion of sociometric testing of young children including procedural techniques, scoring methods, validity, and reliability of the test. Ordinal Position Theory Freud wrote, "...a child's position in the sequence of brothers and sisters is of very great significance for the course of his later life.." (Freud, 1933, p. 281). Freud's major interest was in the ways that sibling structure affects family dynamics. He stressed the enormous hostility which inevitably exists between siblings who are in continual rivalry for parental love and in conflict over rights and possessions. A first-born who is 2 1/2- to 4-years old when a sibling is born will often actively state that he doesn't like the baby at all. Eventually, 13 14 closely spaced siblings may develop true affection for each other, but initially, they have considerable antipathy towards each other. If the age gap between siblings is greater than four years, Freud averred,_the situation is different; the baby is apt to be looked upon as an object of curiosity rather than hate. When the gap increases to eight years, particularly if the elder child is a girl, the baby might evoke protective, motherly impulses in the older child. In some cases, children model themselves after siblings rather than their parents; this phenomenon greatly affects total family dynamics. Also, the birth of a second child will sometimes cause the first child to reject his mother and attach himself from that point on to his father instead; this attachment can not only alter parent-child relationships but can also affect marital relationships. Sibling factors therefore have far- reaching effects on family dynamics and on the individual development of various siblings according to Freud's theory. Alfred Adler discussed ordinal position and family constellation factors in even more depth than did Freud (Adler, 1956). He is perhaps best remembered for his concept of "dethronement", a term signifying the radical shock felt by a first-born child when a second child arrives and usurps some of the affection and attention previously solely given the first-born. Adler asserted that the first-born is particularly trau- matically affected by the birth of a new child as he has had no exper- ience in cooperating with siblings. He stated that first-borns are most often prone to psychological disturbances during childhood than other- borns because of harmful effects due to this dethronement on his develop~ ing personality. 15 In contrast to the first-born, Adler stressed, the youngest child is immune to dethronement. Although the youngest can be spoiled by the whole family and thus become a problem, he may also be so stimulated by his older siblings that he eventually outstrips them all. Another child safe from dethronement is the only child. He, however, has no siblings to relate to and can therefore easily become an egocentric, demanding adult, warned Adler. As a middle child has both a peer model to follow and one to set, he may be better suited to getting along with people than any other-borns. In addition to ordinal position, Adler addressed himself to other issues related to sibling dynamics. He averred that to be the only boy in a family of girls or the only girl in a family of boys could lead to the overstressing of either masculinity or feminity accompanied by life- long feelings of isolation and insecurity. Spacing is also important; Adler stated that a child born many years apart from another resembles an only child. Finally, Adler introduced the idea that siblings can be dynamically related to each other; if one greatly outshines another, the other may fail to the same degree that the first succeeds. Many of Freud's and Adler's original insights into the effects of family position can be of help to the practicing family counselor; they are, however, difficult concepts to test empirically. Forer (1969) and Konig (1963) have both written recent psychoanalytically oriented books on sibling position aimed at counselors. They are, however, more exper- iential than empirical and more focused on the abnormal than the normal so that they are not of much use to the empirical investigator of normal human development. 16 Three books have been published within the last ten years in the area of sibling position which are more empirically based than the books just mentioned; Irving Harris' The Promised Seed (1964), Walter Toman's femily Constellation: its Effects on Personality_and Social Behavior (1969), and Sutton-Smith 8 Rosenberg's The Sibling (1970). Harris sur- veyed biographies of hundreds of eminent men, contemporary and histori- cal. From this study, he concluded that first-borns tend to have a more intense relationship with their parents than do later-barns (Harris, 1964, p. 285) and thus have different psychic tendencies than the latter. First-horns, he stated, are "connectedness-prone" while later-borns are "disconnectedness-prone" in their thinking. First-borns therefore tend to excell in wholistic, abstract thinking while later-borns excel in discrete, concrete thought. While first-borns adhere to inner controls (conscience) in their behavior, later-borns obey external dictates (laws). Later-horns are more apt to be powerful mainpulators or iconoclastic thinkers than are first-borns. First-borns, however, can synthesize and fbrmulate wholistic theories better than can later-borns. Although Harris amassed much support for his theory from the bio- graphies he surveyed, his method was rather subjective; unfortunately, his study was also limited to the male sex. It is, however, a very persuasive and thought-provoking work. Toman's book is a remarkably extensive study of specific sibling sex and position factors. Toman divided ordinal position into four categories: oldest, youngest, intermediate, and only child. He then subdivided each of these categories according to the sex of the child and the sex composition of his sibship. For each of these cases he sketched a complete personality profile. He then predicted what 17 sibling positions would characterize optimal friend and spouse choices for each case. In the last part of the book, Toman also discussed how a parent's particular sibling position would affect his treatment of each of his children. Toman's theory is that people interact most harmoniously with others of different sibling positions from their own; i.e., an oldest son often chooses a younger son as a best friend. Toman described an empirical study which he had carried out on a random sample of 108 divorced couples. He found that wife and husband of the same ordinal position were more prone to divorce than those of complementary position, i.e., an oldest son marrying a younger daughter. He hypothesized that couples whose ordinal positions were identical had trouble establishing a balance of power in their marriages as they were both used to assuming similar power positions in their families of orientation. He also found that divorced individuals were more apt to be from a one-sex or only-child family than were married individuals. This finding, he asserted, was related to the fact that such individuals had had insuf- ficient experience in living closely with peers of the opposite sex. 0n the basis of this research, he concluded that there is an ideal sex and birth ranking of children in a family: first a boy, then a girl, then a boy, then another girl. For healthy personality development, Toman affirmed, all-girl and only-child configurations are the least desirable. Sutton-Smith & Rosenberg's The Sibling is by far the most empiri- cally oriented of the three books; it contains an excellent review of the literature on sibling position. The authors stated that much re- search done on sibling position is defective. One glaring deficiency is that most studies have been carried out on middle class subjects only; 18 many are further limited to college students as subjects. Furthermore, most researchers do not control the many variables affecting sibling position such as spacing, age, specific ordinal position, sibling sex configuration, family size, family completion, and parental character- istics. Sutton—Smith 8 Rosenberg asserted that too many ordinal position studies have been undertaken by first-born academics who emphasize the importance of the parent-child relationship on human development while ignoring the vital impact of siblings and peers. They therefore under- took their research to try to probe into the area of sibling relation- ships. They concluded from their research that siblings often form dyads within the family, each sibling pairing up with another, usually the closest in age, for both stimulation and conflict. Dyads of varying sex and ordinal composition function very differently with correspond- ingly different results for each dyadic member. Sex order of the sib- ship, they concluded, had much more effect on individuals than did simple ordinal position. Empirical Investigations of Birth Order Effects Achievement One of the most widely investigated relationships in sibling order is that between birth order and academic achievement. Altus asserted that first-barns are more likely to go to college than later-borns (1966). Russell Adams (1972), in a study of fourth graders of three races, found first-borns highest in achievement across race and social class lines. Boroson (1970) stated that first-borns, although no more intelligent than later-barns are better readers as children, better high school 19 sstudents, and better performers on National Merit and S.A.T. exams. Unfortunately, most of these studies did not control for family size and socioeconomic class. Solomon and Hirsch (1972) did control for family size and social class in their study on fifth grade achievement in a black ghetto. They found a strong Birth Order X Family Size interaction, with children from small (1- to 3-child) families performing best, first-borns from large families (6 or more children) outdoing their sibs, and last-borns from intermediate-sized families (4- or 5-child) outshining their sibs. Interestingly, the father's presence or absence in the home had no sig- nificant effect on achievement. Responsibility_ Harris and Howard (1968) investigated the frequently made claim that first-barns are more responsible than later-horns. A questionnaire was administered to college and graduate students. On it, they were requested to state at what age a child should be given certain respons- ibilities and at what age he should be given certain privileges. First- born and first-of-sex subjects set the age for various responsibilities lower than did later-borns. Later-borns correspondingly set the age for various privileges lower than did first-borns. First-borns thus believed that children should accept responsibilities at a younger age but obtain privileges at an older age than did later-borns. The authors therefore concluded that first-borns were more responsible than were later-borns. Creativity Another area of investigation is creativity. Cicirelli, in a study of middle class sixth graders, found creativity scores highest for 20 first-horns with brothers (1967). Lichtenwalner and Maxwell (1969) compared middle class and lower class preschoolers in creativity. They found both social class and ordinal position effects; middle class children tended to be more creative than lower class children and first- borns were more creative than later-borns. Dependengy_ Stanley Schachter (1959) stimulated a great deal of research regarding sibling order and social—psychological characteristics with his The Psychology of Affiliation. His experimentation demonstrated that first-borns react to an anxiety-producing situation with more fright than do later-borns. First-horns also prefer to be with others when anxious to a greater extent than do later-borns. Schachter thus postulated that first-horns have greater social dependency needs than do later-horns. To support this position further, he cited a study by Haeberle indicating that the most dependent nursery school children come from 2- and 3-child families, the most highly dependent being first- borns, the least dependent being last-borns. Waldrop and Bell (1964), however, in a study of 2- and 3-year-old males, found that males from large, dense families are more dependent towards the teacher than those from small, low density families. Since this study did not include girls, however, the discrepancy between these results and those of the Haeberle study might be attributable to a sex difference in dependency. Related to dependency is the need for approval. Moran (1967) found that females had more need for approval than did males and that first- born females had a greater need for approval than did later-born females. 21 This sex difference in need for approval might explain the difference in results between the Haeberle study and the Waldrop study. In another dependency-related study, Hines (1973) investigated the relationship between birth order and social versus non-social reinforce— ment effectiveness. Although both first-borns and later-borns performed better under the former condition, first-borns performed better than did later-barns under social reinforcement conditions; under non-social reinforcement conditions, both groups performed equally. Motivation Glick, Grosse, and Zigler (1971) used elementary school subjects in an experiment measuring the performance of a monotonous task under con- ditions of adult-support and adult-non-support. They found that first- borns persevered longer than did later—borns in the non-support condi- tion while later-horns and only-borns persisted longer in the support condition. The authors hypothesized that first-horns are spurred on by non-responsiveness on the part of an adult, perhaps as a result of their experience in trying to regain parental affection after "dethronement." Later—horns and only-borns are motivated more naturally and directly by adult continual support. Empathy Another landmark study was done by Stotland and Walsh (1963) on birth order and empathy. Among their undergraduate subjects, they found later-barns to be more empathic than first-barns. They theorized that later-horns have a greater capacity than do first-barns to identify sym- pathetically with a peer due to their childhood experience of identi- fying with siblings as well as parents. 22 Anxiety Weiss (1970), investigating the often—made assumption that first- borns are more anxious than later-horns, found for his adolescent male sample that heart rates, but not respiratory rates, were higher for first-borns than later-borns under all the kinds of stress conditions induced in the study, indicating that the first-horns were indeed more intensely anxious than were later-borns. In previous investigations, Weiss reported, he had found first-born athsmatics more prone to attacks induced by emotional upset than later—borns. Sibling Sex Position Effects Several studies indicate that sex order and composition of sibship are more useful in explaining certain traits in an individual than ordi- nal position alone. Schmuck (1963), in a study of family position and female tendency to conform, chose a sample of female undergraduates from two-child families. He found no ordinal effects but discovered that a female with a brother conformed more than a female with a sister. In another study in which female undergraduates were used as subjects, Landers (1969) found that women who had younger brothers participated more in masculine sports and scored lower on a femininity profile than did females from other family constellations. Sterner (1973) built on Schachter's research on birth order and affiliation by investigating Birth Order X Self-esteem effects. A sam- ple of Navy enlisted men was divided into two groups, one composed of men whose self-esteems were high, the other of men whose self-esteems were low. He found that the high-esteem subjects, regardless of ordinal position, preferred affiliation when under stress. Low-esteem subjects, 23 if first-borns, followed the same pattern, as did later-borns with older sisters. Later-horns who had an older brother fairly close in age to themselves, however, avoided affiliation when under stress. Critiques of Ordinal Position Studies Numerous critiques of ordinal position studies have been written within the last decade. The most comprehensive is Sampson's "The Study of Ordinal Position: Antecedents and Outcomes" (1965). Although Sampson averred that certain generalizations can be made on ordinal position effects on the basis of numerous research studies, he believed that a child's psychological position in the family, not necessarily related to birth order, is the most important influence on his personality develop- ment. Sampson also asserted that the interaction of Ordinal Position X Sex is highly important but too often ignored by researchers. Finally, he warned that many researchers overstress the advantages of being a first-born. In general, he stressed, although first-horns tend to be responsible, highly motivated, and hard-working, they are less friendly, relaxed, cooperative, sociable, and independent than later-borns. Boroson (1970) took a firmer stance than Sampson on ordinal position effects, avowing that they are proven and powerful. He warned that the U.S. is a society geared to first-borns and discriminatory of later- borns to the detriment of its quality of life. Although first-barns are highly motivated and financially successful, they are not warm, empathic, and congenial as are later-horns. Boroson recommended that parents make a conscious effort to give time and attention to their later-borns and that colleges try to recruit later-horns in the same way that they 24 recruit racial minorities. By raising more later-borns to influential positions, he asserted, the U.S. could become less of a compulsive, driving, competitive society. Schooler's (1972) review of ordinal position studies is full of scepticism regarding the scientific validity of most ordinal position studies. She cautioned that many such works are invalid as they contain no provision for matching the sample population with a control group on the bases of age, social class, and family size. Schooler stressed the fact that family size is inversely related to the level of the father's educational level. Since the level of the father's education is direct- ly related to socioeconomic level, many experimenters who don't control for family size or socioeconomic level may actually be measuring social class effects rather than ordinal position effects. Unless family size, subject age, and social class variables are accounted for, concluded Schooler, ordinal position studies are of dubious validity. Bert Adams (1972) divided ordinal position theory into seven group— ings on the basis of explanatory orientation: (1) physiological causa- tion, (2) only-child uniqueness, (3) dethronement effects, (4) variation in state of tension of parents, (5) waning of affectional display towards first-horns over time, (6) sibling influence, (7) economical factors. Of these, Adams seemed to favor the only-child uniqueness theory and the dethronement theory. He stated that the physiological causation theory, the sibling influence theory, and the family economics theory were each too limited in scope and demonstrated effectiveness to be useful as predictive theories. 25 Explanatory Theories Physiological Causation Theory Some theorists believe that physiological causation factors under- lie the differences found between first-borns and later-borns. Boroson (1970) theorized that since the mother is younger when she delivers her first-born than she is with any succeeding child, she may be healthier and apt to exercise more care prenatally with the first child. In addition, he hypothesized that the uterus might be in poorer condition during later than in first pregnancies and later-borns may thus be less well-nourished prenatally than first-borns. Jensen (1969) stated that world-wide being first-born is advanta- geous in regard to intelligence and success. First-borns thus, reasoned Jensen, must have some kind of genetic or prenatal advantage over later- borns as the variety in world cultural and social conditions preclude any social-psychological explanation. Upon these assumptions, Jensen concluded that the middle class is genetically superior to the lower class as it contains a greater percentage of first-horns (family size is generally larger in the lower class). Many researchers have questioned Jensen's evidence and logic in this study. Sutton-Smith & Rosenberg (1970) presented a very different point of view from Boroson's and Jensen's. They argued that later-borns have physiological advantages over first-barns. They cited evidence indicat- ing that first-horns are more susceptible to both paranatal and prematur- ity disorders than are later-borns. Bert Adams (1972) concluded that physiological causation theories are of limited usefulness as available research indicates that there are both advantages and disadvantages 26 physiologically to being first-born or later-born. Adams' viewpoint seems a judicious one. Sibling Effects The major investigator into direct sibling effects on child development was Helen Koch. She found that sex order and spacing bet- ween children very much affected the degree of influence they had on each other (1955, 1956, 1957, 1966). She established that when spacing between sibs was close, they had more direct interaction with each other than when spacing was wide. When spacing was wide, they had much more direct interaction with their parents than with each other (1956). In regard to sibling-sex factors, Koch discovered that opposite-sexed sibs effected not only more stress but also more stimulation on each other than did same-sexed sibs. In all sibling-sex configurations, differ- ences in spacing have more effect on males than on females. In a large sample study of middle-class 6-year-olds from two-child families, Koch asked each child to name his playmates. First-born girls with a brother named the greatest number of choices while second-born boys with a sister named the fewest. Most children reported three to five playmates (1957). When asked to name a best friend, the largest proportion of children chose a child of the same age. Of those children who chose older friends, more were second-borns than first-borns. Most of these second-barns had elder siblings less than two years older than themselves. TWins tended overwhelmingly to play with others of their own age (1966). Koch therefore concluded that children prefer playmates with similar age characteristics as their sibs, particularly when they are close in age to those sibs. 27 Although Koch's subjects generally preferred friends of their own sex, those from mixed sex, closely Spaced sibships chose opposite—sexed friends more than did any other group (1957). Koch therefore came to the overall conclusion that age and sex of a child's sibling in a two- child family did influence his friendship choice at the age of six. Dethronement In the beginning of this chapter, Adler's dethronement theory was described. Several theorists attribute many differences between first- borns and later-borns largely to the dramatic effects of dethronement for the first-born (Sutton-Smith 8 Rosenberg, 1970; Forer, 1969; Harris, 1964; Sampson, 1965; B. Adams, 1972; Boroson, 1970). After an initial period of being of central importance, a first-born child is suddenly forced out of his supreme position by a tiny, helpless, squalling in- fant. According to Harris (1964), this jolting dethronement may stimu- late a drive for achievement in a first—born which motivates him through- out life: "By doing great deeds, by creating great things, one can demonstrate to himself that he is still invulnerable, that the stopping of external and internal admiration was just a bad dream. Thus the first son tends to ask himself, 'What must I be, what must I do to win for myself a perpetual source of admiration, so that I can feel as omni- potent and overestimated as I once felt?'" (Harris, 1964, p. 76). Sampson expressed a similar point of view to Harris', that the first-born spends his life trying to regain the centrality he once had. He further posited that dethronement might lower the first-born child's self-esteem; available evidence demonstrates that first-borns do indeed tend to have lower self-esteem than later-borns (1965). Especially if 28 the first child is relatively non-verbal when the second is born, he doesn't truly understand why he loses his parents' exclusive attention and thus may attribute the loss to his own undesirability. When a second child displaces the first from his central position, the first-born child is apt to have many of his dependency needs frus- trated and may carry these unfulfilled needs into adulthood. AS a young child, the first-born will often try to win parental affection by striv- ing to show his superiority over his younger sib; this striving to com- pete may persist throughout adulthood. A first-born's dependent use of others as anxiety-allaying objects could handicap them in making friends. His tendency to compete rather than cooperate may also make him less popular than later-horns. Also handicapping a first-born in social relationships is his tendency towards more intense anxiety than later- borns (Sampson, 1965). Sutton-Smith 8 Rosenberg (1970) suggested that the high anxiety level of first-borns may be due to dethronement effects as onlyfborns demonstrate remarkably less anxiety than first-borns. That self-concept may suffer as a result of dethronement is also supported by the finding that only children have the highest self-esteem of all ordinal positions while first-barns have relatively low self-esteem. Forer (1969) stated that first-borns are more likely to be leaders of groups in which intimate contact between members is at a minimum, such as a large political group. In attaining leadership of a large group, competitiveness and drive could be advantageous; in maintaining control of a small group, however, these same qualities could be disad- vantageous. If, as is hypothesized, dethronement effects are responsible 29 for the first-born's competitiveness and drive, they could thus also be responsible for his leadership characteristics. Taylor and Kogan (1973) sought to test the dethronement theory by empirical observation. They observed eight first-born 2 1/2- to 3 1/2- year-old childrens' interactions with their mothers shortly before and shortly after the birth of a second baby. They found a consistent decrease in expression of warmth on the part of both mothers and children after the birth of the new sib. Both children and mothers displayed a significant amount of emotional flatness post partum. The mothers yawned frequently and seemed to respond to their first-born children only with effort. It was the researchers' conclusion therefore that the birth of the second child does indeed lessen the quantity and quality of interaction between the mother and her first child thus lending sup- port to the dethronement theory. Socialization Theories Many theorists believe that parents treat children of varying ordi— nal positions differentially. Sears, Maccoby, and Levin investigated this issue in Patterns of Child Rearing (1957). They found that first- borns were more often awaited for joyfully by prospective parents than were later-borns. First-borns were more likely to be breastfed, weaned late, and be emotionally upset over weaning than were later-horns. Their toilet training was begun earlier than their siblings and tended to be non-problematic. Fathers were more likely to participate in the rearing and disciplining of first-borns than of later-borns. First-horns were permitted more aggression towards their sibs than were later-borns but also were asked to do more chores than were the youngest. By age five, 30 first-borns had more strongly developed consciences than later-horns. Sears, et al, found only-borns to be treated similarly as first-borns were except that their movement outside the home was considerably more restricted. Middle-born children were weaned earlier but toilet-trained later, often with emotional upset accompanying the latter. They were asked to do chores more often than were older or younger sibs but were praised less often for good behavior. In families with more than three children, mothers spent less time playing with middle children than with oldest and youngest children. Youngest children were more indulged by their parents than were other-borns and were least often asked to do chores. Fathers spent less time with them than they did with their older children. Mothers there- fore played the greatest role in disciplining the youngest. Family Size Effects Although the Sears, et a1, study did not include large families, it did cover 2- to 4-child families. In 2-child families, the children were treated less differentially on the basis of ordinal position than those from larger families. The younger child was, however, treated more strictly regarding noise and aggression than was the older. In the 3- and 4-child families, oldest children were shown less affection and were physically punished more than were the youngest. In such fam- ilies, the mother treated the youngest child more warmly than the others. Family size therefore did affect parental treatment of their children from each ordinal position. 31 Ordinal Position and Parental Behavior Although Sutton-Smith 8 Rosenberg stated that sibling differences seem to be caused by more than differential parent-child interactions, they cited several differences found among parental treatments of var- ious children according to ordinal position (1970). They cited Lasko's 1954 study which indicated that parents give their first-borns more intellectual and verbal stimulation during their first two years than they do other-borns. The same study showed that parents treat first- borns more like adults and demand more of them than they do of other- bonm. Only-borns, stated Sutton-Smith 8 Rosenberg, are unique. They have no particularly disadvantageous traits and possess the great advantage of being highest of all-borns in self-esteem. Perhaps this is partly due to the fact that their mothers interact with them to a greater ex- tent than mothers of more than one child interact with any one Sibling. Sampson maintained that first—borns are trained more inconsistently than are later-borns and also have more frustration of needs during in- fancy because of their parents' inexperience (1965). In his early life, the first-born tends to be overprotected and over—indulged by his parents only to experience a sharp reversal in these trends when another sib is born. Sampson quoted a study by Orbison which indicates that parental anxiety and neurotic conflicts are most evident in the rearing of their first child. Sampson cited Lasko's research as showing parents to be warmer and more permissive with second—borns than with first—borns, particularly during the preschool years. 32 Role Theory_ Sampson (1965) suggested a role theory model for the explanation of ordinal differences. As the first-born identifies with and competes against his parents, he is likely to be highly achievement-oriented and responsible but awkward at achieving warm, intimate, cooperative rela- tionships with peers. As a later-born has both sibling and parental models to emulate, he can cooperate and empathize with his peers as well as achieve satisfactorily. Interestingly, it has been often re- ported that later-borns are over-represented in the acting professions; they can apparently assume and cast off various roles much more skill- fully than can first-borns. Parental Variables Individual parental characteristics help to determine how each parent will treat a given child. Toman (1969), Forer (1969), Adler (1956), and Sampson (1965) all stressed the fact that a parent's own sex and birth order will affect how he relates to each of his own children. Sears, et al, (1957) found that maternal age, which changes for each child at birth (except for multiple births), affected child-rearing methods, younger mothers being stricter and more impatient than older mothers. Some theorists base much of their ordinal position theory on the assumption that all parents are anxious and inept with their first chil- dren thus creating personality characteristics in the first-horns which are not found in later-horns. According to Dyer's study of middle class parents' reactions to the birth of the first child, however, parents of even similar age and social class react in very different ways. While 33 some viewed the event of the first birth as a serious crisis, a dis- ruption of major proportions, others considered it only very mildly dis- rupting. Those parents who had the latter.reaction tended to have happier marriages before the child was born, had had some sort of marital training course in school, had been married three or more years, and had planned for the baby (Dyer, 1963). In a similar kind of study to Dyer's, Hobbes, (1965) using a random sample of couples from a small city who had recently had their first baby, found that poverty-level and middle class mothers viewed the birth as more of a crisis than did working class mothers. It would therefore seem that more factors than new parenthood are involved in parental level of anxiety regarding their first-horns and that theories resting on the assumption that all first-time parents are anxious and incompet- ent are faulty. Family Size Some research regarding family size has already been alluded to; in this section, the relationship of family size to marital satisfaction and family functioning is the factor of interest. Christensen (1968) studied the relationship of number and spacing of children to marital success and discovered that actual size and spacing seemed to be less important in marital success than the convergence of the actual pattern with the desired one. Aldous (1969), however, found that marital satis- faction declined as the number of children increased beyond two or three, especially for the husband. As many large families are not planned, Aldous' findings might actually be compatible with Christensen's. Fam- ily size seems therefore to be related to marital harmony, which can affect parenting quality. 34 Hurlock (1972) divided family size into four categories: the one- child family, the small family (2-child), the medium family (3- to 5- child), and the large family (6-child or greater). The one-child family, she theorized, is characterized by a democratic style of child-rearing, close parent-child relationships which contribute to good peer rela- tions, and parental pressures for achievement. The small family (which may include three children if Spacing is wide) is similar except that sibling rivalry is intense and parents give their children little free— dom to determine their own roles. The medium-sized family (which in- cludes 3-chi1d families if spacing is close) bears many of the small family's characteristics except that control becomes less democratic as family size increases and pressures to achieve are focused largely on the first-born. Large families, stressed Hurlock, are the most different from other types of family. Usually unplanned, these families often are beset with marital friction and financial difficulties. Control is authoritarian with strict role assignments, parental crushing of sibling rivalry, and little pressure for achievement except for first-borns. Hurlock judged optimal family size to be three or four children. Least desirable, in general, is the large family. Socioeconomic Influences One difficulty in examining socioeconomic differences in child- rearing is that so few empirical studies have been done concerning Spe— cifically lower class child-rearing techniques. Kohn (1965), however, did not believe this problem to be serious. He stated that the differ- entiating factor in determining what general cluster of child-rearing 35 practices a parent will use is his occupational class. According to Kohn, parents in the entire manual worker category will use working class techniques while parents in the whole non-manual category will use middle class techniques. Child-rearing practices which apply to the "working class", asserted Kohn, also apply to the lower class. Rodman pointed out, however, that lower class families have a dif- ferent structuring than do middle class or working class (1965). He cautioned that lower class adults have a "fluidity of marital bond" (1965, p. 227) which doesn't exist in other social classes. A lower class fatherless family is not the same kind of phenomenon among the lower class that it is among other classes; it is an adaptive solution to economic and social problems rather than a deficient, unwished for family form. Rodman also stressed that lower class individuals have more fluid values than do people of other classes. It may be that researchers have difficulty in specifying a definite established pattern of child-rearing characteristic of lower class parents simply because no such universal, sharply defined pattern exists. I believe also that it is not always possible to differentiate between working class and lower class families. The boundary between the two classes seems to be rather indistinct and easily traversed. In a given family's life cycle, the family could cross back and forth across this boundary several times. In this chap- ter, therefore, the emphasis will be primarily in terms of working class/ middle class differences in child-rearing. Socioeconomic theorists argue that first-borns and only-borns are overrepresented in colleges and successful careers because their fam- ilies gave them the resources necessary to pursue their educations. In 36 large families, they assert, the youngest Similarly often excels be- cause he reaps the most benefit of the family's resources. The economic fluctuations over the family life cycle affect children of varying or- dinal positions differently according to their staggered rates of devel- opment. For example, the youngest child is often spoiled because the family has the material and time resources to lavish on him. A middle child, however, may be comparatively neglected because of a paucity of time and money when he is growing up (Sampson, 1965; Boroson, 1970; Sutton-Smith 8 Rosenberg, 1970; Bert Adams, 1972; Schooler, 1972). Schooler (1972) asserted that many improperly controlled ordinal position studies were actually measuring socioeconomic differences rather than ordinal differences. She noted that family size and social class are closely connected. Since the middle class has a greater pro- portion of first-barns than the working or lower classes, it is not sur- prising that first-borns are overrepresented in the middle class- dominated college population. In like manner, many personality charac- teristics attributed to first-horns versus later-horns may be more a function of social class than ordinal differences. If position in the family does affect personality, moreover, the effects might differ for the same ordinal position depending upon age or social class. Franklin and Scott (1970) found a value difference in goals which parents set for their children between blue collar and white collar par- ents. Blue collar parents strongly desired that their children obey external controls while professional parents equally strongly desired that their children establish internal control. While working class parents set deference and popularity as highly desired goals for their children, professional parents established industriousness and 37 independence as supreme goals for their children. These different value orientations affect child-rearing patterns greatly; such basic value dif- ferences, I believe, would affect parent-child interactions to a much greater extent than would sibling position factors. The degree of marital harmony and cooperation in child-rearing may also be partly a function of social class. Kerckhoff (1972) discovered that working class couples had significantly lower marital value consen- sus than did professional couples. According to Bott (1964), working class husbands participate less in child-rearing than do middle class husbands. Sears, et a1, (1957) found that working class women were more critical of their husbands than were middle class women; the former also fought more with their husbands over child-rearing issues. Sears, et a1, (1957) also discovered other significant differences regarding child-rearing between the two classes. Working class mothers were more severe in toilet training, punishment of dependency, and sex- training than were middle class mothers. The former also used physical punishment, deprivation of privileges, and ridicule more often in dis- ciplining their children than did middle class mothers. Although work- ing class mothers were more worried about their children's progress in elementary school than were middle class mothers, working class mothers were less apt to have aSpirations that their children attend college. Kohn (1965) asserted that middle class parents see child-rearing as more problematic than do working class parents. Jacoby (1969) stated further that the working class woman puts more intrinsic value on having and rearing children than does the middle class woman. While the working class woman may find children a relief to care for in an incompatible marriage, a middle class woman may find child-care duties 38 disruptive of conjugal intimacy. Working class women tend to have more child care experience and access to help from kin than have middle class women. Thus, although middle class couples have fewer problems with un- planned children and with financial worries than do working class couples, they may find child-rearing more arduous, frustrating, and anxiety-producing than do working class couples. Preschool Social Behavior Many people attach little importance to a child's social function- ing in the preschool as they believe that a child of preschool age is simply too young to function truly socially. According to Hurlock (1972), however, patterns of social behavior are established at the pre- school age and tend to persist throughout life. Positive early social experiences, she stressed, tend to enhance an individual's self-concept. Chamberlin and Nader (1971) in a study of upper-middle class grade school children, found that all children who had had social problems in nursery school of an aggressive-resistant nature had been referred to school guidance counselors by elementary school teachers. If such children had been helped in establishing more positive social patterns at the preschool level, perhaps they would not have continued to mani- fest behavioral problems in elementary school. Selfishness. One source of social conflict in the preschool is competition over playthings. Selfishness, stated Hurlock (1972), is at a peak in children from small families and wealthy homes while gener- osity is high among middle class children from intermediate or large families. If generosity is positively related to peer desirability, 39 the latter children could therefore have an advantage. As such children would not be likely to be first-barns, later-barns might, in general have the social advantage. Succorance and aggression. Sutton-Smith 8 Rosenberg (1970) cited evidence on the behavior of nursery school children indicating ordinal differences in modes of interaction. They stated that only and youngest children are the most succorant in nursery school, a finding congruent with the theory that these particular children are themselves well- cared for at home. First-barns were shown to be least aggressive while middle-barns were most aggressive. Younger and only children were most popular with their peers while first-borns were least popular. Socioeconomic Differences and Preschool Behavior Laura Berk (1971) measured the number of frustrating situations with nursery school children encountered in the classroom in two lower class Head Start and in two upper middle class private nursery school settings. She found that the number of instances of frustration was about the same for each group; the lower class children, however, responded more with physical attacks than did middle class children while the middle class children were more likely to use explanations than were lower class children. The total percentage of verbal adapta- tions, however, was not an effect of social class. McKee and Leader (1955) compared lower class and middle class nur- sery school children in aggressive and competitive behavior with nursery school peers; children were observed in pairs of like social class. The experimenters found lower class children both more competitive and 40 more aggreSSive than were middle class children. Lower class boys were more physically aggressive while lower class girls were more ver- bally aggressive. Verbal functioning. Much has been written concerning the verbal deficiencies of lower class preschool children. Although such children are lacking many verbal skills necessary for early academic success, are they lacking verbal skills necessary for interacting socially with their peers? Bernstein (1960), in his comparison of middle class and lower class language use, found that although lower class language is limited in utterance length and syntax, it is effective in communicating concrete information and is often of a very high level of eXpressive functioning. Since at the preschool level, language is more often used concretely, expressively, and in short sentences, perhaps verbal peer interaction is as effective or even more effective for lower class than for middle class children, except in cases of very severe speech defi- ciency. Unfortunately, although much has been written about the aca- demic handicaps of disadvantaged preschool children, little has been written about their social interaction skills. Conclusion. If more were known specifically about special social problems or attributes of children from various sex and ordinal posi- tions and from various family size and socioeconomic backgrounds, the nursery school teacher would be alert to possible characteristic pro- blems or strengths in children right from the time she read their admission applications. She could then arrange strategies for grouping the children, designing the program, and working with parents early in 41 the term so that each child would benefit maximally from his early socialization experience. This study was designed to help further these ends. Birth Order and Sociometric Choice The study which stimulated the investigator to undertake this par- ticular research project was the one carried out by Schachter (1964) on a sample of over 600 college fraternity and sorority members. A socio- metric questionnaire was administered to each subject. Each subject was requested to choose three house-mates with whom he would like to room and to report three house-mates with whom he actually associated with most frequently. He found that first-horns as a group assigned their choices to fewer individuals than did later-borns as a group and that they chose more popular people than did later-horns for desired roommates. He discovered, moreover, that later-horns were considerably more popular than were first-barns. Schachter theorized that first- borns, being more dependent than later-barns, chose their ideal friends on the basis of what others thought rather than on their own judgement. As dependent people tend to be less popular than non-dependent, first- borns were, he reasoned, unsurprisingly low in popularity. Alexander (1966), using a standard sociometric questionnaire (name your three best friends) with male high school seniors, found first- borns more papular than later-barns and more apt to be involved in re— ciprocal choices. First-horns also tended to choose lower sociometric status friends than later-horns. In explaining the opposite direction of his findings from Schachter's, Alexander hypothesized that later-barns who achieve college fraternity 42 or sorority memberships are likely to be of a higher socioeconomic class than first-barns in the same situation since otherwise they would not even attend college and normally could not afford sorority or fra- ternity fees. Since socioeconomic status is correlated with peer desir- ability, Schachter's later-born population therefore could have been more popular than the first-horns solely because of their social status. Alexander's study is not truly comparable to Schachter's. Alexander's sample was limited to males from middle class families of two to three children only while Schachter controlled for none of these variables. In addition, Alexander used high school seniors living at home while Schachter used college students living together as subjects; Alexander's subjects were probably on less intimate terms with their peers than were Schachter's. As previously stated, first-barns are not as good at establishing intimate relationships as are later-horns, but perhaps they do well at establishing more superficial bonds. Finally, Alexander's study differed from Schachter's in that the fbrmer dealt with only actual and not ideal sociometric choices. Since Schachter's reported differences were based mainly on the ideal choice part of the questionnaire, Alexander's findings cannot be said to be truly compar- able to Schachter's. Sociometric Testing and Preschool Children Since the late 1950's, numerous research studies on sociometric testing of nursery school children have been published. A variety of testing procedures and methods of analysis have been used. Dunnington (1957b) used a direct, simple approach. She took each child in turn to a testing room and asked him whom he liked to play with at school and 43 then, whom he didn't like to play with. Each child made three positive and three negative choices. Then, each child was asked whether he did or didn't like to play with each of the other children; these choices were called "forced" choices. Northway (1967) added a situation Specific dimension to the test asking each child to name three peers he would like to play with in each of three different play situations at nursery school. She did not use negative choice or forced choice questions. Biehler (1954) used a more complex procedure than those just de- scribed with kindergarten children. He developed three pictures of line drawings portraying children in play groups of varied sizes; missing on these drawings were the children's heads. Each child was asked to com- plete each picture by choosing photographs of heads of themselves and of favorite playmates. McCandless and Marshall (1957a) developed a picture sociometric technique in which individual pictures of all the children in a class were mounted on a board. Each child was asked to point to the picture of some one he liked to play with three times. An elaboration of the pictureboard sociometric test was designed by Horowitz (1961), who placed a button attached to a timer underneath each child's picture. Each child then indicated his selection of a best friend by pressing the appropriate button. Horowitz thus was able to record not only choice but latency of choice for each subject. Moore and Updegraff (1964) used the same questioning technique that Dunnington had used but also employed a pictureboard presentation. Hartup, Glazer, 8 Charlesworth (1967) also used a pictureboard 44 technique, recording three negative as well as three positive choices but not requesting forced choices. The technique used in this study, the Play Situation-Picture Board Sociometric Technique, was developed by Robert Boger in 1967 to 1968 for the 1968-69 Head Start evaluation (Boger, 1969). As previous nursery school sociometric testing had been confined to university lab- oratory nursery schools whose population was largely upper middle class, Boger felt that no previous test design was adequate fbr use with Head Start children from working and lower class backgrounds. The latter have been demonstrated to be less formally verbal than upper middle class children, especially during the preschool years (John 8 Goldstein, 1967). Boger therefore developed a test which relied very little on verbal instructions. In this test, the subject first chose three fav- orites from the five toy picture cards. Then, from a picture board, he selected the peer whom he wanted to play with for each toy. Thus, even a fairly non-verbal child could be tested. Scoring In analyzing the results of sociometric tests, all the researchers mentioned have compiled peer status scores and rank ordered all the children in a given group according to their status scores. Some re- searchers, having asked only for positive responses from the children, used only positive scoring. Northway (1967) is one such researcher; she stated that negative choices take on so many different signifi- cances that they cannot be handled well statistically. Several re- searchers (Dunnington, 1957b; Hartup, et a1, 1969; and Moore 8 Updegraff, 45 1964), all, however, affirmed that negative choices significantly con- tributed to the information they obtained. Most researchers recorded only the spontaneous choices of each child, usually three positive and three negative (when negative choices were requested). Dunnington reported that forced choices added signi- ficantly to her information, but Moore 8 Updegraff (1964) in a later study fbund that they were of no use. Currently, the trend seems to be towards requesting three positive choices only. In scaling the sociometric peer status scores, unweighted scoring seems now to be preferred. In two reviews of sociometric techniques (Northway, 1967; and Witryol, 1953), it was reported that many studies had established that weighting the scores did not alter the results ob- tained by unweighted scoring. Dunnington (1957b), however, found that weighting the scores added significantly to her results. Perhaps the reason for this significance was that by weighting the scores differ- entially, Northway could count spontaneous choices more heavily than urged choices. As most researchers used only spontaneous choices, how- ever, unweighted scores were as significant as weighted ones. Despite the evidence against weighting sociometric choices, many investigators continue to use weighting schemes. Campbell (1960) sug- gested that although weighting choices in a large group situation has no effect, it may be appropriate in certain small group situations. Validity Quite a controversy rages over the question of the sociometric test's validity for preschool children. Central to this issue is the form in which the sociometric choice question is asked of the child. 46 Dunnington (1957a) adopted the position that a sociometric test should be a direct measure of actual preference; she therefore phrased her question straightforwardly, "Whom do you like at school?" Northway (1967), by contrast, believed that a sociometric test should measure ideal choice; the question should therefore be phrased in the condi- tional tense, "Whom would you like to play with?" Boger (1969) used this particular approach. If one accepts Dunnington's method, it therefore becomes legiti- mate to relate sociometric choice with actual classroom interactions or teacher judgements. Many such studies have been carried out by researchers of Dunnington's persuasion. Hartup, et a1,(1967) found that sociometric choice did correspond to observed classroom inter- actions as did Dunnington (1957b). Dunnington (1957b) and Horowitz (1961) found a high correspondence between teacher predictions and sociometric scores, but Marshall 8 McCandless (1957b) found that indi- vidual teachers varied markedly in their predictive ability. If one accepts Northway's approach, however, sociometric results are neither validated or invalidated by such comparisons. As Lorber (1969) stated, asking "Whom would you like to play with?" is asking for an attitudinal response. Who can judge if an individual's attitude is valid or invalid? Thus, using the ideal choice method, one cannot really establish validity. Reliability Reliability of sociometric tests in the nursery school is another area of controversy. Some critics argue that social relationships in the preschool are of a very changeable nature and that one would 47 expect a retest, especially after a month or so, to yield very differ- ent results than the initial test. Others argue, however, that even nursery school children enter into relatively permanent relationships with school peers. Published reliability results differ very widely according to experimenter, method, time elapsed before retest, and specific group. Dunnington found that although individual choices changed over a two—month period, the status hierarchy of the class re- mained relatively constant (1957b). Witryol (1953), in his review of sociometric studies, found that scores obtained by the partial rank order method were indeed fairly stable over time although the range of coefficients of stability among various studies is extremely wide. Northway (1967) found in a longitudinal study of children from pre- school through sixth grade that sociometric scores correlated well as far apart as two to three years. Intensity One caution is voiced by most experimenters using the sociometric test; that is that the test does not measure intensity of affection or of dislike. Some researchers believe that all positive choices are on the same affective level while others believe that there is a differ- ence of intensity between choices. Horowitz (1961) attempted to measure intensity by having teachers rate bonds between children who were friends as strong, moderate, or weak. He then measured the length of time each child studied the sociometric pictureboard before pressing the button for best friend choice. He theorized that a child with strong ties would choose more quickly than one with weak ties. His theory was upheld by the 3-year-olds in his sample but completely 48 reversed by the 4-year-olds. Because of this puzzling discrepancy, he could offer no firm conclusions. The measurement of intensity thus remains a problem in sociometric testing. Peer Desirability and Personality Several researchers have attempted to relate sociometric status positions with specific personal attributes. Moore 8 Updegraff (1964) found that children are prone to make positive choices of the same sex as themselves and negative choices of the opposite sex (this tendency did not exist among the older 4-year-olds but held for all younger groups). They also found that dependence on adults is negatively re- lated to popularity with children close to three, but not with those close to five. Dependence on other children, however, is positively related to popularity. Dunnington (1957b) discovered that high status children initiated verbalizations independently and showed a greater proportion of positive expressions than did low status children. Low status children sought more attention of adults, but also rejected adult stimulation more than did high status children. Hartup, et a1, (1967) similarly found that high positive status children reinforced others positively more often than average. In like manner, low-rated children reinforced others negatively more often than average. These studies give additional sup- port to the theory that sociometric tests do measure to some degree a given child's skill in social interactions. 49 Implications for the Present Study It is obvious that controversy reigns in almost every area dis- cussed in this section. If solid theory is to be built in any of these areas, therefore, more research needs to be done. The present study is directed towards this goal. Many theoreticians state that first-horns are less popular and sociable than later-barns, but empirical verification of this hypo- thesis is lacking. Although Schachter's (1964) study supports this hypothesis, Alexander's (1966) refutes it. Both of these studies, like so many others on ordinal position, used samples composed of late ado- lescent individuals. This study is designed to test the hypothesis of first-born unpopularity using a sample of preschool children who are at the stage in life when their families influence perhaps more highly than at any other stage. If ordinal position does influence an indi- vidual's sociability, this influence should be clearly evident during the preschool years when so much of the child's experience focuses around his home. Another area of controversy is sociometric testing of young chil- dren. It has been questioned whether young children, particularly underprivileged children, have the conceptual and verbal skills neces- sary to give valid responses on the test. Boger's Play Situation- Picture Board Sociometric Technique, which is used in this study, relies almost entirely on pictorial representation and thus may pro- vide a superior sociometric measure for preschoolers when compared to tests relying solely on verbal questions. Several issues regarding use of the test are also analyzed in this study. To ascertain whether weighted or unweighted scoring yields better results, both are used and 50 compared. To discover whether negative choice adds to the effectiveness of the sociometric technique, the addition of a negative choice is sug- gested for half the children in the LPS sample and positive only scoring compared to positive and negative combined scoring. As reliability results in previous studies varied markedly, a retest is proposed (on the LPS sample) to ascertain how reliable the particular instrument used in this study is. Perusal of the literature revealed several factors as likely to produce important results in this study. Age, sex, race, socioeconomic class, and ordinal position of a child seemed vital factors. Family size and father-absence or -presence seemed valuable. Regarding sib- ling influence on a subject, the sibling just older and the sibling just younger appeared to have the greatest impact on the subject; sex and age gap between the subject and each of these two siblings were important factors. In this study, therefore, each of the foregoing factors is investigated. In the literature, studies done on ordinal position, on nursery school interaction, and on sociometric testing, subjects were recruited mainly from two specific populations; from college student bodies and from university laboratory preschool enrollments. These populations are overwhelmingly white, culturally advantaged, and middle class. In the present study, lower class children are compared to middle class children in an attempt to discover whether class differences do exist in the phenomena of interest and in hopes that some findings could be generalized to a larger population than to just white, middle class preschool children. CHAPTER III: METHOD This chapter is composed of five parts. In the first section, sampling is explained. Next, the testing method used with the PSPBS is described. In the third part, statistical procedures employed in the study are explicated. Then, the models used in the analysis of the data are presented. Finally, issues concerning the statistical analysis of the data are discussed. Sampling Samples from two different populations taken at two different seg- ments of time were used in this study. The Head Start (HS) sample was given the PSPBS during April and May of 1969 by a research team from Michigan State University's Head Start Evaluation and Research Center. Head Start children were mostly black and from lower class homes. The Laboratory Preschool (LPS) sample was given the PSPBS during April and May of 1974 by the researcher and another trained tester. The LPS chil- dren were mostly white and from middle class homes. Children in both samples had been in the same classroom four to five half-days a week for six to eight months prior to testing; class size ranged from 11 to 18. Ages ranged from 41 to 65 months at the time of testing in both samples; the average age, however, was younger for the LPS than for the HS sample. 51 52 The total number of children used in the study was 183, 67 from LPS and 116 from HS. For parts of the study, however, only those chil- dren who gave responses on the PSPBS could be used. A total of 164 children took the PSPBS, 63 from LPS and 101 from HS. Absence, refusal of parents to permit testing, and refusal of the child to be tested were the major reasons that 19 of the total sample were not tested. The LPS sample and the HS sample are described separately in the following para- graphs. Laboratory Preschool Sample All the children enrolled in the Michigan State University Labora- tory Preschool 3-year-old and 4-year-old classes were used, 67 children from four classrooms in all. One child was excluded from the sample as his mental age was below the average for his class while his chronolo- gical age was substantially above average for his class. It was decided not to use the two remaining LPS classes, the 2-year-old and the 3- and 4-year—old mixed classes; the former might be too young to respond I meaningfully to peer choice questions and the latter had such a wide age spread that age differences might confound other variables. The age range was 41 to 65 months at the time of initial testing. All but three of the §§ had been together in the same classroom four half-days a week since mid-September, 1973. The majority of §§ were Caucasian, but other racial groups, black, Oriental, Indian Eurasian, Mexican-American and American Indian were represented. All the children were from middle class homes with fathers in business or the professions. In only a small minority of homes was the father absent. Family size ranged from one to six children. 53 Head Start Sample All the children who were enrolled during 1968-69 in the ten Head Start programs which were evaluated by M.S.U. Head Start Evaluation teams were used. After all the data had been recorded, however, one of the ten classes had to be discarded because so much necessary data was missing. In all, 116 §§ from nine Pontiac and Detroit HS classes were used in the study. The age range of the HS Ss, when tested, was the same as that of LPS SS, 41 to 65 months. All §s had been together in the same class- room five half-days a week from the beginning of the school year. .§§ were predominantly black, although other races, Caucasian and Mexican- American, were represented. They were largely from lower class back- grounds although a minority were from working class homes. No distinc- tions were made between these two classes, however, as it was not pos- sible to determine fine class distinction from available data and the few homes which did appear to be working class were only marginally so. The majority were from fatherless homes in which the mother was unem- ployed. When the father was present and working, he had a manual job. Family size ranged from one to twelve children. In cases in which a child lived together with cousins under 12 years of age, these cousins were treated as siblings in the data of this study. See Table l for a detailed breakdown of Specific sample character- istics. 54 Table 1: Sample Characteristics Sample Characteristic LPS HS Total Male 35 6O 95 Sex Female 32 56 88 41-53 months 32 22 54 Age 54-65 months 35 94 129 Black 7 103 110 Race White 51 10 61 Other 9 3 12 Y Father es 63 42 105 Prese"t No 4 74 78 First" 24 31 55 born 2nd or 3rd born 37 16 73 Ordinal 4th Position or 5th born 5 23 23 6th, 7th, 8th, 9th, 1 26 27 10th born 55 Both Samples Both the LPS and HS samples were included in the study so that a comparison could be made between the two groups which represented popu- lations which differed vastly in family structure, socioeconomic class, and cultural background. The same factor, for example, family size, might have radically different effects for each of the groups. Adding the HS §s also increased statistical power in analyzing results as the total sample size was thus increased yielding additional degrees of freedom. The inclusion of the HS SS, moreover, also expanded the range of family size and ordinal position in the study. Testing Procedures Although the testing procedure followed for both the LPS and the HS samples was similar, there were a few differences, mainly in regard to apparatus used. In the following section, testing procedure for the LPS sample is first explained in detail. Testing procedure for the HS sample is then briefly summarized with emphasis on any departures from the procedure used for the LPS sample. Testing Procedure for the LPS Sample Apparatus The picture board was constructed from a 22" x 28" piece of light blue posterboard. Three 28" x 3/4" strips of posterboard were stapled across the posterboard beginning at the bottom at 6 1/2" intervals so that photographs could be easily inserted between board and strip. Black and white photographs, 5" x 4", taken of the face and shoulders of each child in a given class were inserted in the strips creating three rows of six photographs each (fer the classes of less than 18, 56 photographs were suitably spaced so that no one picture was set apart from the others). The photographs used were exact copies of the photo- graphs attached to each child's classroom locker in order that the photographs would be familiar to and thus easily recognized by the Ss. Because the photographs were placed in the open strips, they were easily removed and replaced. The six play situation cards, designed by Boger (1969), are colored drawings of toys printed on 12" x 10" pieces of heavy cardboard. The pictures are: (I-A) Dolls (a girl rag doll and a boy rag doll); (I-B) Trucks (one dump truck and one cattle truck); (II) Sandbox; (III) Horses (one live brown horse in foreground, one spotted horse of visually smaller dimensions in the background);(IV) Dual swing; (V) Teeter totter. In the Head Start testing only five cards were presented to each child; I-A was given only to girls, I-B only to boys. For the LPS testing, all six cards were presented to each child as it was felt that both I-A and 1-8 would appeal to certain individuals of both sexes. Setting A corner of a basement testing room of LPS was partitioned off using large cardboard dividers to create a 6' x 12' area. The walls were devoid of pictures and there was no furniture in the area. The picture board was placed flat on the carpet and child and tester sat on the rug in front of it. Testers Both testers were graduate students in Child Development and had had much experience in dealing with young children. One tester, who tested the two afternoon classes, had had previous experience in 57 administering the PSPBS. He demonstrated then how the testing was done. A unifbrm procedure was followed by both testers. Each tester tested one 3-year-old and one 4-year-old group. Procedure The LPS §§ were tested during their free play period, the first hour of the preschool session, during April and May of 1974. Prior to actual testing, each tester spent one to two free play periods in the classroom to establish rapport and learn the names of the children. Once this had been done, the tester entered the classroom, located a child who did not seem to be too involved with anything at the time, and asked, "Would you like to go downstairs and play a picture game with me?" If the child consented, he accompanied the tester to the testing room. If he did not, the tester moved on to another child or sometimes enlisted the teacher's aid in coaxing the child to go down to the testing room. Once in the testing room, the tester asked the child to be seated on the rug in front of the picture board and then sat down next to the child. It was felt that the child would be more relaxed sitting on the rug than at a table. First, the tester spread out all six play situa- tion cards (in rows of 2 x 3) before the child. To put the child at ease before asking questions requiring choice, he said, "Can you tell me what these toys are?" After the child had named the toys, the tester said, "What toy would you like to play with the best?" He picked up the card indicated by the child and proceeded, "Next best?" and "Is there anything else you'd like to play with?" The tester then recorded the first three choices, stacked the three cards chosen in order, and put them aside. 58 After clearing the remaining cards away, the tester said, "Here are pictures of all the children in your class. Can you find your pic- ture?" When the child had correctly identified his own picture, the tester asked, "Can you tell me who each of these other children is?" The child was then left free to point out and name each of the other children in whatever order he wished. The experimentor then pointed to the pictures he'd missed, if any, so that each child was named. If the child forgot a name, the tester supplied it. The tester then placed the first play situation card in front of the child saying, "Here is the picture of the toy that you told me you liked best. Put your picture on it. Whom would you like to play with you?" The first response of the child was then recorded. If the child did not respond, the tester coaxed him until he did. In these cases, "urged" rather than "voluntary" was marked on the score sheet. The same procedure as has been described for the first card was repeated with the second and third cards also so that three peer preferences were recorded in all. Between SS, photos and cards were put into a new random order. With the §§ from the 4-year-old classes only, the §_was asked, "Is there anyone you would ggt_like to play with?" If the child said "No", no further response was urged. If he gave another response, the first name he mentioned was recorded. Both testers initially tried to ask the negative question of the §§ from the 3-year-old classes but fbund that those children did not comprehend the question so it was dropped from the testing of the two younger classes of the LPS sample. Retesting. Retesting was done after 2-week and 4-week intervals. Approximately half of each age group was retested at each interval, more 59 than half of the 4-year-old group in fact being retested after 4 weeks. The same procedure was followed as had been used for initial testing. Testing Procedure for the Head Start Sample With a few exceptions, the procedure followed‘was identical to that described for the LPS sample. Testing was done in April and May of 1969 (a pretest had been given in the fall which was not used at all in this study). The apparatus used with the HS §§ was slightly differ- ent from that used with LPS gs; the pictures were smaller but full-view Polaroid photographs and the board was slightly smaller and was made from masonite. As mentioned, with the HS sample, only five play situa- tion cards were presented to each S, the dolls being presented only to girls, the trucks only to boys. No retests or negative choices were done on the HS sample. Statistical Procedures Previous studies regarding family position factors have been criti- cized for the failure of the researcher to include or control for rele- vant variables. One method of control would be to limit the sample to include children from only one Specific family size, social class, and so on. In limiting the sample in this way, however, one limits oneself in terms of the population to which results may be generalized. One also is restricted in sample size from a given population. Information is thus lost. It was therefore decided that a multivariate approach would be used in this study. Multiple regression analysis was selected as the most effective statistical procedure to use with the hypothesized model. 6O Regression Analysis and Correlation Coefficients The basic equation for linear regression, from which the equation for multiple regression is directly derived (see Kmenta, 1971, pp. 347- 380 for the exact derivation), is: y = EU?) + E where y is the observed value, E(yf<) is the predicted value, and Eis the disturbance term (error term). This can also be written as: y = 90 + 31x + 6 where 30 is the y intercept, .6] is the regression slope, and x is the independent variable. The equation 9 = $0 + 31x is the estimated equation of the regression line, that is, the equation of the line minimizing the sum of squared errors. It will now be shown how regression is related to the correlation coefficient. Defining s as the sample covariance and 5x2 and s 2 as KY Y the sample variances of x and y, respectively, and rxy as Pearson's Product Moment Correlation, a] can be written as: A S I /22 51-55; and rXY as. rxy'sxy/ 5x sy . X It then logically follows that " _ s 131 ’ r'xy .12.. Sx Since 'xy is the estimator of PXY’ the correlation coefficient, it is evident that the two processes, regression.and correlation, are closely related in the single variate case. Since, however, correla- tion coefficients are not suitably defined for multivariate analysis while regression analysis is, regression analysis is the better tech- nique to use in this study. 61 In addition, regression analysis is better fer significance test- ing. Testing the significance of a correlation coefficient can be done in two ways. One can either test whether the correlation coefficient is significantly different from zero or that it is different from some predetermined correlation (using Fisher's r to 2 conversion). Since, however, there is no absolute agreement as to what correlation level is definitely significant, this test is of limited value. With regres- sion analysis, however, a t test can be run to either prove or disprove the significance of a given relationship at a pre-defined Alevel (.05 in this study). Regression analysis is thus superior to correlation coefficient techniques as a statistical tool for this study. Regression Analysis and ANOVA Both multiple regression and analysis of variance (ANOVA) are based on a linear model. Both methods take error of prediction into account in their models (unlike correlation coefficient methods). In addition, both can also be used to test hypotheses using the F test. ANOVA can be a less flexible method of analysis than regression, however. One of the assumptions which ANOVA is based on is that depen- dent variables come from a normal distribution. Since a normal popula- tion is composed of continuous random variables, ANOVA requires that dependent variables be continuous. Some basic statistics texts also assert that regression analysis requires continuous dependent variables. Jan Kmenta (1971, pp. 425-7), however, stated that multiple regression is flexible enough to be used with binary dependent variables (which are non-continuous) although efficiency will be lost. 62 In two-way ANOVA designs, all individual data must be grouped while in multiple regression, data can be treated individually. When comparing different groups in two-way ANOVA, therefore, the number of §§ per group must be the same. In regression analysis, however, since observations can be handled individually, no such constraint exists (Kmenta, p. 416); groups of differing size can therefore be compared without randomly eliminating §§ as required by ANOVA. One of the often stated advantages of two-way ANOVA is that inter- action effects between variables can be studied. Interaction terms can, however, also be introduced into regression analysis (Kmenta, pp. 418- 19). In this respect, therefore, ANOVA has no advantage over regression analysis. To illustrate the equivalence of the ANOVA and regression analysis models, the two techniques are compared as follows. The two-way ANOVA model is written as: Yijk ‘4 ”‘3' * xk " Eijk where Yijk is the ith observation on the dependent variable y, in the jth column and the kth row,Ais the grand mean, akj is the deviation of the mean of the jth column from the grand mean, Ak is the deviation of the mean of the kth row from the grand mean, and £13k is the error term. The comparable regression model is: yijk = 31 " ”2’82 +°°"’ fljxij I ’zziz +...+ kzik * 513k where xij is 1 if the observation is in the jth column and is otherwise zero; likewise, zik equals 1 if z is in the kth row but otherwise equals zero. .3 refers to the slope of the regression, and Eijk to the dis- turbance term (error). 63 Comparing the means of y for each cell, we get: M 5(Yi11)‘*+ 41* A] E(1112)“+ ‘1’”‘2 E(yijk)=‘“+ 41+ 2k 5(3’1211‘“ “‘2’“ "1 E(.Y1'221=.«+ elf/12 Eiyijkl=a+ ulj+ 11k Using this table, we obtain: REGRESSION EWm): 81 O 591°19‘81”” 32 E L 8 ‘8 x x 4— 8' O L x x x g x s : X Z I I I I I I I I I I 0 1 2 3 4 5 6 7 8 9 10 Score Figure 2: Distribution of Sociometric Status Scores for the Head Start Sample 16- x g x .2 +3 ‘6 Z x 3; 8*- ‘8 x q. 1‘ O 3 x '2 x x x :2 x __l #1 J J #1 _l J 4 L 0 l 2 3 4 5 6 7 8 9 10 Score Figure 3: Distribution of Sociometric Status Scores for the 99 Laboratory Preschool Sample 100 tal 1y of results supporting hypotheses at the .05 level yields five for LPS and four for HS. Counting the instances in which results, LPS, thus, has a slight edge over HS in supporting the hypotheses. E3. The spread of the sociometric status scores is wider in the middle "iange for LPS than for HS classes. Overall, the raw scores extended From 0 to 9 for LPS and from O to 10 for HS. Breaking these scores into three groups (Low = 0; middle = 1 to 5; high = 6 to 10), 82 percent of- tzhe LPS scores and 60 percent of the HS scores fall into the middle cate- sgory. The mean score is slightly under three for each group (had each Si taken the sociometric test, the mean would have been exactly three). It is seen, therefore, that for the middle scores, 1 to 5, the range is indeed wider f0r LPS, 82 percent versus the HS 60 percent. 2 8 3. See Figures C. Fewer children per class will receive sociometric scores of zero in LPS than in HS classes. In the four LPS classes combined, seven percent received zero scores. In the ten HS classes combined, 25 percent re- ceived zero scores. It is thus apparent that a lesser percentage of LPS SS than HS SS received zero scores. 0. Fewer children will receive very high sociometric scores in the LPS than in the HS sample. In the high range, six to ten, were 15 percent of the HS Ss, but only 11 percent of the LPS Ss. This difference, although small, is in support of the hypothesis. Additional Finding; Peerpreference. It was found that a LPS S was more likely than a HS S.to choose a friend of the same sex at the .018 level. A HS S, 101 however, was more apt than a LPS _S_ to choose a friend of the same race a: 1: the .045 level. In addition, the HS S was more likely than a LPS S_ ‘tzca choose a friend of high ordinal position at the .043 level and to Choose a friend from a fatherless home at the .0005 level. Chose all three. Only one result reached significance. A LPS S_ was more apt than a HS S to choose the same peer all three times at 1:he .027 level. These unhypothesized, but significant, findings will be discussed '1 n the next chapter. Some may simply be a result of differences in pop- tnlation composition between the two groups, but others may be truly illuminating. Negative Scores It was hypothesized that SS who were chosen more than one or two times as a negative choice might have distinguishing characteristics. Regression analysis was used with the negative-only peer desir- ability score as the dependent variable and with the same independent variables as had been used for the positive peer desirability model. Complete results are given in Table 2. Except when specifically men- tioned, the model corrected for multicollinearity in an identical fashion to model la has been used. flypothesis Testing_ A. The more negative choices a S receives, the more likely he is to be first-of-sex in his family. Results were not significant, but in the predicted direction. 102 B. The more negative choices a S_receives, the more likely he is to be first-of-sex in his family. Results were not significant, but in the predicted direction; the original model was used in this instance. C. The more negative choices a S receives, the more likely he is to have a younger sibling. Results were in the predicted direction, but were non-significant. D. The more negative choices a S_receives, the more likely he is to have an opposite-sexed sibling. Results were non-significant, but in the opposite direction. No results reached even the .10 level of significance. Three hypotheses yielded findings in the predicted direction. For one hypo- thesis, however, results were in the opposite direction. See Table 12 for a summary. Reliability The entire peer preference section of this study rests on the assumption that the first choice given by a child on the PSPBS is indeed his strongest preference choice. One way to test this assumption is to investigate whether or not first choices are more reproduceable over time than are second and third choices. The percentage of first choices in the test which were reproduced in the retest was compared against the replication percentages for second and third choice for each class and then for each combined half of the total age group. See Table 13 for complete results. First choices were indeed more reliable than second choices, as seen in Table 13. They were also more reliable than third choices except for 103 Table 12: Summary of Negative Score Results; Regression Coefficients and Standard Errors Result it, b 49" A; A S with a high negative score is male. -.24H 81 Hypotheses A S with a high negative score is first of sex. .52H .49 A S with a high negative score If has a younger sib. -.00007 .004 A S with a high negative score I has opposite-sexed sib. 20.16 14.31 { Résfilt is not in predicted direction. {I Result is in predicted direction. a. is the regression coefficient. b. ‘is the standard error. the morning three-year-old group in which they were equally reliable. Second choices were more reliable than third choices for the four-year- old, but not for the three-year-old classes. Combining both three-year- old classes, reliability of first choice was 44 percent, of second choice 24 percent, and of third choice 28 percent. Combining both f0ur- year-old classes, reliability was 58 percent for first choice, 30 per- cent for second choice, and eighteen percent for third choice. Although age differences in reliability thus seem evident, as shown in Table 13, differences among specific classes are also substantial, particularly between the morning and afternoon three-year-old classes. 104 Table 13: Percentages of Rank Choices Reproduced N Percentage reproduced Number of SS Class tested and" First Second Third Negative retested choice choice choice choice Morning 3-year-old 14 36% 29% 36% -- Afternoon 3-year-old ll 54% 18% 18% -- Morning 4-year-old 16 69% 38% 19% 54% Afternoon 4-year-old 17 47% 24% 18% 45% Both 3-year-old 25 44% 24% 28% -- classes Both 4-year-old 33 58% 30% 18% 50% classes When raw unweighted sociometric status scores were matched on test and retest for each individual and a correlation coefficient calculated for the entire sample, it was .676. When the correlation coefficient was calculated for each classroom individually, however, it was found that the correlation coefficient for three of the classrooms was above the .700 level, while for the fourth, the morning four-year-old class, it was only .460. Looking at the negative only scores, a large difference between classes is also evident. The correlation coefficient for the morning four-year-old class on negative scores was .884 while for the afternoon four-year-old group, the correlation coefficient was .397. Combining SS 105 from the two classes, the overall correlation for reliability of negative scores was at the .736 level. Interestingly, only seven of the seventeen SS from the morning four-year-old class ever received a negative choice while 12 of the 18 afternoon four-year-olds did. All results concerning sociometric status score reliability are shown in Table 14. For three of the classes, positive scores were fairly reliable. For one of the two classes in which negative choices were given, negative choice was extremely reliable. In the next chapter, pos- sible reasons for variation between classes will be discussed. Unweighted Versus Weighted Scoring Since for the LPS sample, first peer choice was more often repro- duced on a retest than second or third choices, it would seem to be a more important choice. It was, therefore, decided to try a weighted scoring method in which a first choice yielded 10 points, a second choice seven points, and a third choice five points. This particular scoring scheme was suggested by Campbell (1960). Also tried was an unweighted method; first, second, and third choices were each given a value of one. All regression analyses were done twice; once using a weighted socio- economic status score; once using an unweighted score. In the peer desirability segment of the study, the unweighted socio- metric status (peer desirability) score and the weighted sociometric status score were used in turn as the dependent variable. In the negative peer desirability section, negative weighted and unweighted scoring were compared with weighted and unweighted negative and positive scoring methods. 106 No. oo.~ on. - - - mm mommmpu ope-Lemz-e :uom - - - mm. Pm. me. um momma-u v FF< ea. 04. 04. ee. ms. -K. m- me-o-eeos-e eooeeooe< om.- me._ om. am.- No. we. N_ me-o-eemz-e me-e-oz - - - «0. us. mu. mp mUFo-semx-m coocempw< - - - mm. mm. an. up nope-Lemz-m unease: pamosmch oaoFm pcmwuwwmmou “goo-oucH maopm ucmwuwmmmou mcowpe>cmmno cowpmhm-Lop comeepmssou we mme—u mmgoum m>wpemmz mmsoom w>wuemoa Loasaz mscmh ucmwowmmmoo cowumpmscou cw museum msumum owsposowoom ma; mo auwpwnMVme neF open» 107 As shown in Table 15, there was no important difference between weighted and unweighted scoring methods for any of the four samples tested. For the negative subsample, both weighted and unweighted and both negative and positive combined scoring methods yielded similar R2's (correlations of multiple determination), F-statistics, and significance levels. Although there were thus no differences between weighted and un- weighted scoring techniques in increasing the predictive power of the model, there were differences between samples. The R2 and F-statistics indicate that both weighted and unweighted scoring techniques were more significant for the LPS than the HS sample. Using unweighted scoring, the R2 was .20, the F—statistic 1.25 for LPS while the R2 was only .05 and the F-statistic .59 for HS. This may indicate that validity of the PSPBS is higher for the LPS than the HS sample. In the following chapter, the results given in this chapter will be discussed. 108 Table 15: Sociometric Scoring Technique Comparisons: Total Regression Statistics b 0 I s 1 1 f s . a R2 F- Signi- Error amp e ype ° cor1ng n statistic ficance £3 Weighted positive 183 .06 .94 .50 17.53 Whole ' Sample Unweighted positive 183 .06 .99 .46 2.32 Weighted positive 67 .20 1.25 .28 15.47 LPS Unweighted positive 67 .20 1.25 .28 2.10 Weighted positive 116 .05 .54 .87 18.83 HS Unweighted positive 116 .06 .59 .84 2.46 LPS Negative weighted 35 .23 .63 .79 15.61 “993' Negative unweighted 35 .22 .60 .81 2.11 t1ve Sub- Weighted negative sam 1 and positive com- 35 .24 .67 .75 20.65 p e bined Unweighted nega- tive and positive 35 .24 .66 .76 3.29 combined a. n = number of observations. b. R2 is the coefficient of determination. CHAPTER V: DISCUSSION This chapter, like the last one, is divided into several sections. The first five sections are discussions of the results reported in the previous chapter on the following categories: Positive peer status, peer preference, choice of the same peer three times, comparison of the LPS and HS samples and negative sociometric status scores. The last three sections are discussions on use of the instrument; they are on re— liability, scoring, and testing procedure. For precise figures regarding results, the reader may want to refer back to Tables two through fif- teen and Figures two and three all,in Chapter IV. Sociometric Status The only hypothesis which was accepted at the .05 level was that as the gap between a S and an older sibling decreases, his sociometric status increases. Only for the LPS sample did this finding reach signi- ficance although for the whole sample and the HS sample, results were in the same direction. Stated more simply, this hypothesis states that children with no older siblings or siblings several years older are less popular with preschool peers than those with older sibs close in age to themselves. Role model theory explains this result well; a child who has a sibling just older than he can model his behavior and learn from his example how to get along with peers. To survive in his family, such a child has had to learn to c00perate and interact with a more powerful 109 110 peer; in the preschool, these social skills might serve him well in interacting with others. The reason that this hypothesis was supported at a more significant level for LPS than HS might lie in differences in family size and density between the two samples. As demonstrated by Hurlock (1972) and by Sears (1957), quality and quantity of parental attention per child, particularly for middle children in the family, declines as family size increases. As LPS children were largely from smaller, less dense families than were HS children, the amount of parental attention the average LPS child received was likely to be greater than that received by the typical HS child. A HS child with a closely spaced elder sibling might often be from a large, closely Spaced family in which case the disadvantage of lack of parental attention might offset the advantage of having a just- older sibling. According to Waldrop and Bell (1964), three-year-old males from large, dense families were more dependent on the teacher than others were; this dependency on adults decreases popularity with peers (Dunnington, 1957). The hypothesis that sociometric status is higher for SS from father- present than father-absent homes was supported in the predicted direction for the whole sample and HS sample, but rejected for the LPS sample at the .05 significance. According to Moore (1964), dependency on adults tends to make a preschool child unpopular with his peers while dependency on peers tends to enhance a child's popularity. Since all of the LPS SS from father-absent families had had fathers at home until 12 to 6 months befbre testing, perhaps they were unsure about their relationships with adults and did indeed depend more on their peers than on their teachers. Only 6 percent of the LPS SS came from father-absent homes; these SS 111 suffered no financial deprivation, compared with their classmates, and came from families of one to three children. By contrast, 64 percent of the HS Ss came from father-absent homes; these _S_s tended to suffer more financial deprivation than did those from father-present homes and came from families with as many as eight children. In many HS homes, a father had never been present. Due to these very different implications of f"ather-absence for the LPS group, therefore, different results are not Surprising. Having no father at home was no handicap for children at LPS in peer acceptance, but it may have been a handicap for H5 children. At the .10 significance level, sociometric status increases as stan- da‘r‘d age increases for the whole sample. Results were non-significant, but in the same direction for the HS and LPS samples. Since social Ski 11$ tend to improve with age, it is logical that peer popularity would a1 30 increase with age. During the preschool years, however, different chi 'ldren develop at different paces so that chronological age is not For this reason, a1 Ways an accurate indication of social maturity level. the relationship between increased standard age and sociometric status i 3 not to be perfect. For two hypotheses in this section, results were non-significant, bUt in mixed directions. These results could very well be attributeable to chance alone, but they could also be indicative of differing trends be- thleen the LPS and HS samples. The hypothesis that sociometric status 1‘ ncreases as ordinal position increases yielded directional support from the LPS sample, but opposite directional results from the whole sample and the HS sample. If these indications Show the actual trends, with dif- ferences between the HS and LPS samples, increasing ordinal position may 112 be aidVfilhimgeous up to a certain point after which further increase be— comes disadvantageous. The other hypothesis is that Ss who are the sole members of their sex in the family will have lower sociometric status tha" ‘1lliose who do not; whereas results for the HS sample were in this direction, results for the whole sample and LPS sample were in the oppo- site direction. The direction of the HS results, if accurate, lends cre- dence to Adler's theory that being the only one of one's sex in a sibship can be devastating to one's self-esteem and ability to get along with (3131\eers. Perhaps the opposite direction of the LPS results is due to the fact that family size is smaller than in the other sample; being the only t35319' in a sibship of two would seem not at all devastating whereas being the only boy in a sibship of six might well be difficult. Because of "“‘J'Iticollinearity difficulties regarding this variable, the validity of ‘theese results is questionable, however. In summary, the only hypothesis which was accepted at a .05 level c>1= significance was that sociometric status increases as the age gap be- 't3~een a child and his closest older sibling decreases. This finding, Vvliich attained significance only for the LPS sample, was explained in 1terms of role model theory. At the .10 level of significance for the ‘VVhole sample only, sociometric status increased as standard age increased. lrhe fact that this finding failed to attain significance for either the l.PS or HS samples individually was accounted for by the fact that devel- <3pmenta1 level of a preschool child does not always correspond exactly Vvith chronological age. Some non-significant results were discussed; it is hoped that further research could be done to either refute or verify the directional, but non-significant, results in this study. 113 Peer Preference Relationships Four hypotheses in this section received support at the .05 signifi- cance level. The first is that a child and his first peer choice will be like-sexed; this was supported at the .0005 level by all three groups. This finding was also reported by Moore, et al., (1964). A related, non- hypothesized finding was that LPS children are more likely than HS chil- dren to choose a friend of the same sex (at the .05 significance level). At the preschool age, perhaps the sex role identification-process is more intense for middle class than for lower class children. Significant at the .01 level for the HS sample and the whole sample and in the predicted direction for the LPS sample were the hypotheses that a child with high sociometric status will choose a friend of lower status than his own and that a child with low sociometric status will choose a friend of higher status than his own. This finding is congruent with findings by Schachter (1964) and by Alexander (1966) in studies on college and high school students, respectively. Schachter's explanation for this phenomenon is that individuals of high sociometric status are more self-confident and independent than are those of low status and thus, choose their friends according to their own judgment rather than choosing them on the basis of how they are rated by others. The reason that the result is more significant for the HS than LPS sample may be that more HS SS scored dramatically low or high in sociometric status than did LPS Ss; contrasts between individuals were thus likely to be greater on the aver- age for the HS sample, therefore, raising the significance level. The hypothesis that a child and his chosen friend will be of dis- similar ordinal positions was significant for the LPS sample at the .05 level while for the whole and HS samples, results were non-significant, 114 but in the opposite direction. The LPS results support Toman's (1969) theory of role complementarity, i.e., that individuals choose friends who play or have played complementary, not similar,roles in their families to their own roles. The reason for the discrepancy in results between the LPS and HS samples may again have to do with family size. According to Hurlock (1972), parents of two to three children compare their chil- dren to each other much more than do parents of five or more children; sibling rivalry is also more intense in the former than in the latter case. It may therefore be that specific ordinal differences have a sharper effect in small than in large families and that the drive to choose a friend of dissimilar, but complementary ordinal position is thus stronger in a child from a small family than in one from a large family. Non-Hypothesized Findings Most of the additional non-hypothesized findings which were signifi- cant at the .05 or .10 level had to do with sex, age, father-absence or-presence, or race. First, sex differences are discussed. At the .10 significance level for the whole sample alone, girls were more likely than boys both to choose a friend of the same race and to choose a friend of a high ordinal position. The first result may indicate that girls are more aware of individual differences at an early age than are boys as they usually display an earlier social interest than do boys and may feel a need to ally themselves with others of similar racial characteristics. This feminine social precocity may also include an earlier adoption of societal racial prejudices, thus leading the girl to avoid friendships with others of a different race. 115 Standard age. For both the whole sample and the HS sample, at the .05 level, a child of low standard age tended to choose a friend of high peer desirability more than did those of higher standard age. As higher standard age is correlated with higher sociometric status, this result is congruent with two results already discussed; that a child of high socio- metric status will choose a friend of lower status and that a child will select a friend of dissimilar age to himself. At the .10 level for the whole sample only, children of low standard age tended to choose girls rather than boys as friends. This finding may be related to the fact dis- cussed by Freud (1933) that many girls have a tendency to "mother" young- er children as part of their sex role identification process. Finally, at the .10 level for the whole sample only, a child of high standard age was more apt than one of low standard age to select a friend who was first-of—sex. This result is of questionnable Significance due to high multicollinearity between the first-of-sex variable and another inde- pendent variable which can result in invalid results. Father-absence or -presence. For the LPS sample only, children from father-absent homes chose friends of a higher peer desirability level than did others at a .01 significance level. As has been previously sug- gested, this subsample of children who had only recently lost their fathers from the home may be more peer-dependent than others. Since according to Schachter (l964),peer-dependency is related to the choice of a friend who is highly valued by peers, this result is reasonable. For the whole sample and the HS at the .05 level, children from father- absent homes chose first-of-sex friends less than did others; these results were in the opposite direction for the LPS sample. Although no explanation is offered for the specific finding, the difference between 116 LPS and HS results again supports the theory that the father-absent home has very different implications for children of the two samples. Due to multicollinearity problems involving the first-of-sex variable however, this result may be invalid. 3355, The most striking non-hypothesized result was that a child of the major race tends to choose a friend of the major race and, in like manner, a child of a minor race tends to choose a friend of a minor race at the .0005 significance level for both the whole and the HS samples. LPS results, although non-significant, were in the same direction. This finding indicates that children of preschool age do indeed realize racial differences and associate preferentiably with peers of their own race. Again, for the whole sample at the .10 level and for the HS sample at the .05 level, a child of the major race tended to choose a friend of greater standard age than did a child of the minor race. This finding may be accounted for by the fact that a child of the minor race who tends to wish, as just reported, to play with a friend of a minor race, has a more limited range for selection than does a child of the major race. Assorted other results. For the HS sample at the .10 level, a child of a high ordinal position tended to choose a best friend who was not first-of-sex; for the LPS sample, results were non-significant, but in the opposite direction. The difference between these two results again points to the possibility that ordinal position effects differ between the LPS and HS samples. Again, for the HS sample, only, at the .05 level, children of high sociometric status tended to select friends of higher standard age. Again, multicollinearity problems may have caused this result to be more significant than it is in reality. Finally, fer the LPS sample only, at a .05 level, a child with no younger sib or a wide 117 age gap between himself and a younger sib chose a friend of a lower ordinal position than did others. This result is directly related to the already discussed result that children choose friends of dissimilar ordinal position; as most of the children without younger sibs are them- selves later-borns in this sample, it follows that they would select first-barns as friends. A child who chose a friend of the same sex was also apt to choose a friend of higher than average standard age at the .01 level for LPS. This result may be related to a sex difference; as already discussed, males tended to choose older friends than females. Perhaps LPS males tended to choose friends of the same sex more often than did females. Negative Peer Relationships No specific hypotheses were made in regard to negative peer relation- ships. It was, however, assumed that relationships between the chooser and his negative choice would differ from those between the chooser and his positive choice. This assumption was borne out by the results which are discussed in the following paragraphs. It was discovered at the .10 level that girls are more likely than boys to name a peer of a non-major race as a negative choice. This result is congruent with a previously mentioned finding that girls are more apt than boys to choose a friend of their own race; girls may be more precocious in social awareness, of which prejudice is a part, than boys. While children of the major race tended to make negative choices of opposite sexed peers, children of a non-major race made negative choices of same-sexed peers at the .10 level of significance. Perhaps this is 118 due to the fact that sexual antipathy is felt strongly by many children of this age, but children of a non-major race might have stronger feel- ings of antipathy towards a like-sexed peer who treats him in a discrim- inatory fashion. While a child who was the only one of his sex tended to make a nega- tive choice of a peer of a different race at the .05 level, a child who was first-of-sex made a negative choice of a peer of the same race at the “7“ .05 level. These results are obviously logically inconsistent and are due to a high degree of multicollinearity between the two independent variables; "-10 they are, therefore, invalid results. TWo results were significant at the .10 level regarding the age gap between a child and his younger sib. A child with a younger sib close in age to himself was more likely to make a negative choice of a peer of high- er than average sociometric status and (2) higher than average standard age. Since a child with a closely spaced younger Sib is likely to be less popular himself than other children, as indicated previously, he may thus be jealous of children of higher status and age and make a negative choice of an older, higher status child, who had perhaps rejected him. Summary. Although few hypothesized results were significant in this section on positive and negative peer relationships, it is obvious that results for positive relationship were quite different from those fer negative relationships, thus indicating that the first positive peer choice is made on a different basis than a negative peer choice and that the test, therefore, has validity. Unpredicted, but significant results as well as differences between HS and LPS samples in results merit further investigation. This model would also be improved if the 119 multicollinearity between the first-of-sex and only-one-of—sex variables were corrected for. Those hypothesized results which did attain signi- ficance indicate that family background factors do indeed influence both the best-liked and most disliked peer choices of a child. Chose All Three The hypothesis that a child who chose the same peer for all three ,-2-. play Situations is from a father-absent home was accepted at the .01 level for all three samples. A child from a father-absent family may wish for a more secure, stronger peer bond than a child from a father— present family and may reflect this desire in choosing only one peer all three times . An additional non-hypothesized finding at the .05 level was that for the LPS sample only, children who chose the same peer three times tended to have either no older sibs or older brothers rather than older sisters. It may be that a child from a father-absent home or a family with no older sibs or older male sibs (who might offer more competition and threat to their younger sibs than do sisters) wants the security of having one very special devoted friend at preschool. He may feel unsure enough about his abilities to form friendships to want to hold on tenaciously to one friend. Of course, the results may also be due to an interaction effect of tester and child. More extensive investigation into this issue is needed. Comparison of HS and LPS Samples The hypothesis that LPS results would support the hypotheses stated in the first three parts of this section more than would HS results was 120 twat accepted by a large enough margin to warrant more than a cautious statement of acceptance. Interestingly, on 30 different hypotheses, the direction, sometimes significant and sometimes not, was the same for both groups 21 times, or about 70 percent of the time. Although it is realized that directionality is not always a valid trend indicator, it often is. Despite differences between the two samples in family background factors, socioeconomic status, and age, and race group composition, many of the relationships accepted as significant or indicated by the direction of results in this study were very similar for the two samples. The hypothesis that a greater percentage of sociometric status scores will fall in the middle range for LPS than the HS sample was substantiated as 82 percent of the LPS scores versus 60 percent of the HS scores fell into that range. Akin to this hypothesis are the hypotheses that fewer LPS than HS children will receive (1) scores of zero and (2) very high scores. These hypotheses were also borne out, the first by a large mar- gin, 7 percent of LPS SS versus 25 percent of HS SS and the second by a narrower margin, 11 percent for LPS versus 15 percent for HS. The dis- tribution curve represented in Table 2 of the LPS scores was similar to that reported by Northway (1967) in many sociometric tests done at other laboratory preschools. The reason for the difference in LPS and HS score distribution may be due to differences in programs.While the LPS program was directed to- wards the development of the whole child, socially, physically, emotion- ally and cognitively, the HS program was focused more on the cognitive than on other aspects of development. A major goal of the LPS program was to integrate each child into the group. Effbrts were directed towards helping each child realize his potential; social engineering efforts were 121 'frequently made to try to prevent a given child from becoming an isolate, a scapegoat, or a prima donna. If the goals of this program were fully attained, therefore, there would be no isolates or "stars" in the class- room. Since HS program personnel were compelled to devote more energy than were LPS personnel to academic goals, however, they could not do as much "social engineering" to integrate isolates and discourage idoliza- tion as could LPS staff. Other FindingS Several non-hypothesized, but significant findings regarding LPS and HS differences appeared in regression analysis results for the whole sam- ple. In regard to peer preference, it was found at the .05 level that a LPS child was more likely than a HS child to choose a friend of the same sex, but that a HS child was more apt than a LPS child to select a friend of the same race. It was found that a HS S_was more likely than a LPS S_at the .05 level to choose a friend of higher ordinal posi- tion and at the .0005 level to choose a friend from a father-absent home. As noted before, LPS children may be encouraged at an earlier age than HS children to adopt a strong sex role identity and, thus are more likely to choose a best friend of the same sex to reinforce this sex role identity. The other three results are best explained as results due to basic differences in composition of the two samples. The most probable reason that a HS child was more apt to select a friend of the same race was that 89 percent of the HS SS were of the same race while only 79 per- cent of the LPS SS were of the same race. It is not surprising that a HS child was more apt than a LPS child to choose a friend of a higher ordi- nal position and from a father-absent family as the average ordinal 122 [Josition for the HS sample is 3.39 in contrast to LPS's 2.04 and 64 per- icent of the HS SS are from father—absent homes while only 6 percent of I_PS SS are. One other finding of Significance at the .05 level is that a LPS child was more likely than a HS child to choose the same peer for all three play situations. It may be that LPS children coming from fairly settled, but more peer-isolated neighborhoods in comparison with the HS children,who experience more flux,like to form a more intense, stable relationship with one other preschool peer than do HS children,who may have more opportunity and experience to form ties with peers. Negative Scores No significant results were obtained for this part of the study. The following hypothesis yielded results in the predicted direction: (1) The more negative choices a child receives, the more likely it is that he will be male, (2) The more negative choices a child receives, the more likely he is to be first-of-sex in his family, (3) The more negative choices a child receives, the more likely he is to have a younger sibling. In the opposite direction were results regarding the hypothesis that the more negative choices a child receives, the more likely he is to have an opposite sexed sibling. The reason that a male would be more likely than a female to receive negative choices is that preschool males exhibit more physically aggres- sive behavior, which tends to be unpopular, than do females. McKee and Leader (1955) verified this fact that males are more physically aggres- sive. It was hypothesized that first-of-sex children would receive nega- tive points as they have no like—sexed, older peer to learn appropriate 123 social skills in interacting with like-sexed peers according to role Inodel theory. Because of the sometimes unsettling effects attributed to dethronement on a young child's developing personality and self-concept, it was hypothesized that children with high negative scores are likely to have a younger sibling. All these theoretical statements are somewhat supported by the direction of the results although not statistically sig- nificantly supported. The hypothesis which yielded results in a negative direction was that a child who receives many negative points will be of a mixed-sex sibship. According to Koch (1957), having a Sibling of the opposite sex is not only more stimulating but also more stressful for a child. Perhaps the stress between siblings does not carry over in preschool peer inter- action, however. A child from a mixed-sex sibship may indeed be popular with peers as he has had experience in dealing with an opposite sexed peer and because he may conjure up more stimulating, imaginative play ideas than children from one-sexed sibships. According to Sutton-Smith 8 Rosenberg (1970), children from mixed-sex sibships tend to be more crea- tive than others. Although no significant findings were elicited from this part of the study, the direction of results lead to interesting theoretical specula- tions. One difficulty with this part of the study was that only 24 chil- dren in all gave a negative choice and that not three, but only one choice was requested. With only 24 negative points available, only a few chil- dren accumulated more than one of these points. Such a small sample made statistical inferences difficult to make. Interestingly, when a negative choice was initially asked of the first three-year-olds tested, no meaningful reSponses were obtained. 124 'The question, "whom would you ppp_like to play with?" was apparently not comprehensible to these younger children. They did, however, seem to understand the question, "whom would you like to play with with this toy?" Since the negative question was asked without the use of a play situation, maybe the question was too abstract for these younger children. Reliability Reliability results done on the LPS sample compare favorably with those of other sociometric studies. Reproduceability of individual peer choices at 2 to 4-week intervals (see Table 13) was higher for LPS chil- dren than it was for the Head Start children in Boger's study (1969) at a two-day interval. Since percentages of reproduceability generally de- clines as more time elapses between test and retest, (Witryol and Thompson, 1953) this result is particularly significant. It must be noted, however, that the children from the LPS sample may have formed more stable peer attachments or ideals, particularly with their first choice, than did children in the Head Start program. Reliability of individual positive sociometric status scores between test and retest as measured by correlation coefficients (see Table 14) was above .70 for three of the four LPS classes, .68 overall. This com~ pares favorably with the results, obtained by Moore and Updegraff (1964) for first and second graders, .62 and .52, fer a l-2-week interval. According to Witryol et a1. (1953, pg. 255), results from socio- metric studies across all age groups generally yield coefficients of sta- bility above .60, but cover the range of .16 to .96. This stability co- efficient, they stated, will generally be lower for younger children than 125 for older ones. That LPS results are comparable with results obtained from older groups makes them the more significant. Results for the reliability of negative only scores were calculated for the two four-year—old classes (see Tables 13 and 14). While about half the children in each class gave the same response in both test and retest, individual negative only sociometric scores varied a great deal in reproduceability between the two classes. While the stability coefficient was .90 for the morning four-year- old class, it was only .40 for the afternoon four-year-old class. In- spection of the distribution of negative choices in each class revealed that only 41 percent of the morning children were ever named as a nega- tive choice while 67 percent of the afternoon children were. Since most of the morning children, therefore, received scores of zero both times, the correlation between the two sets of scores was naturally high. In reproduceability of first, second and third choices, the morning three-year-old class stands apart from the others; its first choice re- produceability is lower and its third ch0ice reproduceability higher than for any other class. Inspection of the data indicates that the levels of reproduceability shown are due almost entirely to the effect of a few chil- dren who chose the same friend all three times on both test and retest. This indicates that for most of the children in this particular class, reproduceability of any one of the three choices was close to zero. Another result which was inconsistent from majority results was the .46 morning four-year-old stability coefficient when all other classes had coefficients of .70 or higher. The fact that differences were greater between individual classes than between three-and four-year-old age groups indicates that differing 126 social structures may exist in various classrooms. The morning three- year-olds, for example, were the only group of children tested in which a majority of the children could not name each picture of others in their class. Testing children in this group was more difficult than in any other because of reluctance of many children to leave the classroom, distractability and extremely short attention Spans, and unusual dif- ficulty of several children in understanding many of the questions. Meme bers of the LPS supervisory staff mentioned that children in that parti- cular classroom at the end of the year were involved in more solitary and parallel play than were children in the other three classrooms who en- gaged in extensive cooperative play. Although children in this classroom were two months older, on the average, than those in the afternoon three- year-old class, it seems as though they may, as a group, have been more socially immature. The low positive sociometric status stability coefficient of the morning four-year-old class might be related to its extremely high nega- tive status coefficient. Since a few children were consistently named as disliked, the least popular segment of that class was relatively fixed. As long as a child was not in that "disliked" segment, he may have been considered an acceptable playmate by most of his other non-disliked peers and could go up or down in social status depending upon daily circum- stances. The majority of the children, united in their dislike of a few individuals, may not have formed into smaller "cliques" in the fashion that children in a classroom with no fixed lower segment might. The reliability results, therefore, indicate that social structures may differ among different classrooms of like-aged children. Two factors are likely to be involved in this difference. The first factor is that 127 classrooms vary in population characteristics; age, sex, race, and person- ality differences might lead to differing social structures. The second factor is that teachers vary in their approach to the classroom as a whole and to different individuals; the effect of the teacher could be a crucial one in detennining the social structure of a given class. Also important is the joint effect of teacher-child interactions in different group situations; different children are affected differently by various teachers, who, in turn, are affected differently by various children or combinations of children. Scoring Although no advantage was revealed in using a weighted over an un- weighted scoring technique, as also asserted by Northway (1967), Witryol (1953) and Campbell (1960), it appears that the sociometric status scores may be more valid for the LPS than the HS sample as shown by their com- parative statistics in Table 15. This hypothesis is supported by the lower reliability results between another Head Start sample obtained by Boger (1969) than were obtained for the LPS sample. Perhaps peer rela- tionships are more changeable in Head Start types of classrooms than in laboratory nursery school classrooms. Perhaps also, LPS children are easier to test. Although verbal instructions are minimal in the PSPBS, the fact that a child fully understands the verbalizations of the tester may be an important factor in eliciting the type of response which the test is presumed to elicit. CHAPTER VI: CONCLUSION This chapter is divided into three parts. In the first part, major results of the study are summarized. In the second section, suggestions for improving the instrument and the statistical analysis of the data used in this study are made. In the final section, implications of the study and recommendations for further research are discussed. Major Results Sociometric Status The relationship of greatest interest to the investigator'was that between ordinal position and sociometric status. Results failed to indi- cate any significant relationship between them; first-born preschool children did not seem to be any more or less popular than later-horns. Results did indicate that a middle-class preschool child with an older sibling close in age to himself is more popular than one with no older sibling or an older sibling distant in age from himself. Although only at the .10 significance level, results also indicated that peer popularity increases as a preschool child's comparative age in a pre- school group increases. Peer Preference Both middle-class and lower-class preschool children tended to choose a best friend of the same sex and of the same race as themselves. They also tended to choose a best friend of dissimilar ordinal position and of dissimilar sociometric status to themselves. A child who came 128 129 from a father-absent family was more apt than a child from a father- present family to choose the same peer in all three choice situations on the sociometric test. Head Start and Laboratory Preschool Results Compared The distribution of sociometric status scores was different for the two groups; a greater percentage of LPS than HS scores fell into the middle status range. More HS children were never chosen at all than LPS children. A LPS child was more apt than a HS child to choose the same peer in all three choice situations and to choose a peer of the same sex as a first choice. Sociometric status scores were more significant measures for the LPS than for the HS sample. Use of the Instrument Reliability results for the LPS sample on the Play Situation- Picture Board Sociometric test compared very favorably with reliability figures published for other preschool sociometric studies. First, posi- tive peer choices were reproduced substantially more frequently than were second and third choices. The single negative choices were reproduced almost as often as were first positive choices. The relative status in the classroom of a given individual changed little over a two-week or four-week period. Differences in reliability were greater between indi- vidual classes than between the older and younger half of the LPS sample. A weighted scoring technique was not superior to the less complicated, unweighted method; the negative choice did not contribute significantly to the results obtained. 130 Suggestions for Improvement Instrument In the course of administering the PSPBS over 60 times, the experi- mentor obtained some insight regarding use of the instrument. In general, SS responded well to the test. For the younger half of the SS particu- larly, the concreteness of the play situation seemed to be a help in con- ceptualizing the preference questions. As mentioned, these younger sub- jects did not seem to understand the negative peer question, which was - not used with a concrete play situation,while they appeared quite comb fortable with the positive preference questions which were introduced as concrete play situations. TWo of the six play situation cards were interpreted by many chil- dren in a way not compatible with the aims of the test. The picture of the horses portrayed a large brown horse in the foreground with a white, black-spotted horse, which was drawn much smaller as it was in the dis- tance. Many children, however, noted that one horse was 'big' while the other was 'little'. One child stated, "I'm going to put my picture on the big horse. I'll pick Susie to ride the small horse because she's dumb." TWo black children placed their own pictures on the white horse and chose another child for the brown horse. Obviously, the particular way in which this card was drawn may lead a child to interpret the play situation and peer choice in a negative or only mildly positive rather than straightforward strongly positive fashion. The other card which presents a problem, depicts a female and a male rag doll. Almost every S who chose that card put his own picture on the doll whose sex matched his own and chose a peer of the opposite sex for 131 the other doll. This choice constraint perceived by the child may cause him to choose a peer who is not especially well-liked by him. Because preschool children tend to prefer like-sexed friends, this unwanted effect is especially likely to interfere with spontaneous positive se- lections. Although the tester is directed to indicate whether each peer choice offered by a child is spontaneous or 'urged', he is not encouraged to make any other comments on the attitude of the child being tested. Some- times the present experimentor tested a child who was very distractible, moody, unresponsive, or hostile through the whole first part of the test, but who volunteered all three peer choices spontaneously. In such a case, it is possible that the child just gave three random choices so that he could get through with the test. In like manner, no instruction is given a tester regarding a child who quickly choose all three peer pictures from the row closest to him on the board. If the tester were to note such factors on each test sheet, it might be easier to determine which children gave honest preferences and which Simply gave random responses. Statistical Analysis The multiple regression technique used in the major part of this study yielded adequate results. Results would, however, have been better if certain multicollinearity difficulties had been overcome. A more de- tailed use of regression analysis to test for interactions between vari- ables might yield additional significant findings. To increase the statistical power, a more complex statistical technique which specifically provides for heteroskedasticity could be utilized. Another way of treat- ing the data, factor analysis, could also be tried. 132 Implications and Recommendations for Further Research On many application forms for preschools and for elementary schools, there are questions not only on the sex, age, and race of the child, but also on certain family background factors. After a parent has filled out such an application, the application is usually quickly scanned and per- manently filed. Many vital statistics concerning a child's family back- ground are thus never used. If more were known about the implications of these factors for the child and for the composition of a group of children, perhaps these easily obtained statistics would be utilized for the benefit of both the individual child and the group. Teachers and school administrators often find it necessary to group children. In the beginning of each school year, they decide which chil- dren are to be enrolled in each particular classroom. Throughout the year, they group and regroup children for small group lessons, sports and games teams, research committees, and the like. When trying to organize the children into groups which will function harmoniously and construc- tively, the teacher often considers sex, age, and what she perceives as the personality of each child. Results from this study indicate that a teacher would also do well to consider ordinal position and popularity of each child. As the preschool children in this study preferred friends of different ordinal positions and popularity levels from their own, heterogeneity of ordinal position and peer status seems to be a desirable goal to strive for when grouping young children. In this study, it was found that young children tended to choose friends similar to themselves in certain salient characteristics. At the 133 preschool age, when children are assuming a personal identity, it is not surprising that they prefer friends of the same sex and same race as their own. Why then do they prefer friends of dissimilar ordinal posi- tions and popularity ranks? The answer to this question may be given in terms of role theory. While a child is assuming a unique identity, he is also learning to play a certain role in his family which complements those roles of his parents and siblings. In the preschool, the child may thus seek out friends who will complement the role which they have learned to play in their families rather than one who will play the same role and have the same status in the group. The finding that preschool children tend to prefer friends of dis- similar peer status from themselves has further implications. It indi- cates that the social structure in the preschool provides for much verti- cal mobility rather than being restricted to horizontal mobility only. In other words, highly popular children tend to play with less popular children rather than playing only with other very popular children. This tendency prevents the formation of a "highly popular" elite clique and permits children with less well developed social skills to learn from others with more highly developed skills. There is evidence, however, that if a child is actively disliked by more than 3 or 4 other children in a group, he may be excluded from participation with children who are accepted by the group. He thus often is compelled to associate with another unpopular child and his maladaptive behaviors may thus be rein- forced. Understanding such dynamics within the preschool group could provide a teacher with a framework for facilitating the entry of a new child into the group and for designing strategies for helping individuals overcome social deficiencies. 134 In this study, there was evidence that various preschool classrooms differ a great deal in social structure. In some classrooms, individual relationships between children seemed to remain constant over several weeks time while in others, such relationships fluctuated a great deal. Whereas peer status rank remained stable for a given individual over a period of several weeks in some classes, in others it fluctuated. In some classes, no child seemed to be particularly highly liked or disliked; in others, a couple of children could be highly admired while a couple of others could be intensely disliked. How much these differences are due to differences in children and how much are due to teacher effects is an issue worth investigation. How a teacher could act as a catalyst to facilitate cooperation, acceptance, and learning for preschool groups of different combinations of individuals would be a valuable focus for fur- ther research. Standard age did not appear as a very significant factor in either peer acceptance or peer relationships in this study. Since in most cases, the age span f0r each group was only 12 months, chronological age differ- ences indeed may be minimal in significance. A more pertinent variable, which should be included in further research, is the child's degree of social maturity. The fact that children with older sibs close in age to themselves were more popular than were others in this study has important implica- tions. Perhaps for only and oldest children, a nursery school experience is especially important as they have no Siblings to model themselves after in developing social skills with peers. Also, mixing age groups within a preschool might prove more beneficial than segregating ages because of the fact that older children can serve as valuable models for younger ones. 135 The results of this study also indicated that children from father- absent homes desire a more intense relationship with a favorite peer than do other children. Particularly in cases in which separation of the parents has been relatively recent, children often want a strong rela- tionship with another child. Perhaps because he is somewhat uncertain about the permanency of ties with his mother and father, such a child turns to a peer for permanency and warmth of relationship. It may thus be especially important for a child undergoing a family transition to have an opportunity to interact with a group of peers on a regular, stable basis. When socioeconomic status was used as a variable in this study, no overall significant differences between the middle class and lower class groups were found. Results which attained very high significance levels were usually significant for both groups. It may be, therefbre, that the reasons that young children choose specific friends are very similar across social classes. Unfortunately, the data used in this study was less exact and less familiar to the experimenter than that for the mid- dle class sample. Further research could be carried out comparing a homogeneous lower class preschool group, a homogeneous middle class pre- school group, and a heterogeneous preschool group in which both classes would be represented, all under supervision of the same researcher. Further research should be done using multiple factors to explain preschool social structure and relationships. Interaction effects among family background, age, sex, and race factors should be examined perhaps using factor analysis technique. The reason that research on specific birth order effects has been relatively unfruitful is that birth order effects cannot be studied apart from family structure and dynamics. APPENDICES APPENDIX A BACKGROUND CHARACTERISTICS DATA Key for Background Characteristics Data Column Numbers Number of Columns Range of Valid Cards Item Description 1-3 3 001-840 Child identifi- cation number 4 1 ------- Blank 5 1 1-9 School identifi- cation number 6 l ------- Blank 7-8 2 01-14 Classroom ID code 9 l ------- Blank lO 1 O=a.m. 1=p m Time of class 11 l ------- Blank 12 l O=male l=female Sex of teacher 13 l ------- Blank 14 l O=male l=fema1e Sex of child 15 l ------- Blank 16-17 2 41-65 Age of child in months 18 l ------- Blank 19 1 O=middle class Socioeconomic 1=lower class class of child 20 l ------- Blank 21 l O=yes l=no Father is present in home 22 l ------- Blank 23 l l=white; 2=black; Race of child 3-Indo-Eurasion; 4=0riental; 5-American Indian; 6=Mexican 24 1 ------- American Blank 25-26 2 01-12 Number of chil- dren in family 27 l ------- Blank 28-29 2 01-10 Ordinal position of child 30 1 ------- Blank 31 l 0=male l=female Sex of just younger sibling 32-34 3 - 1 - -64 Age gap in months between child 8 just younger sib APPENDIX A (Cont'd) Key for Background Characteristics Data Column Numbers Number of Columns Range of Valid Cards Item Description 35 1 ------- 36 1 O=ma1e l=female 37-39 3 .l — 240 40 l ------- 41 l 0=ma1e l=female 42-44 3 -55 to 260 45 l ------- The rest of the columns follow a similar pattern. Blank Sex of just older Sibling Age gap in months between child 8 just older Blank Sex of next closest sibling Age gap in months between child 8 next closest sib Blank Isolated 0's and 1's symbolize sex of other sibs. If the O or 1 is followed by 3 other numerals or a blank and 2 other numerals, the numerals represent the age gap between that particular sibling and the child in months. A-2 HP‘FJHHHHHHHHHHHHHHHF‘HHHHHHHHP‘HHP" HHP—‘P—‘HHP‘F—‘P’JI—‘I—‘F‘ HP: P“ HHPJHFO 01 31 31 ul ol I 01 t") l I.) l 01 Ol 01 31 31 01 03 03 03 03 O3 03 0 O3 03 UB 03 C.(‘C)C‘C)CC_.C. ‘C NhJ-‘\.‘i‘di\)l\)l\fb.‘kx‘b (I C (\J N 02 32 02 02 02 02 02 O 04 O4 \ . -JOC'L‘C C (TI-)L. cmecvoczr C‘: L HHP—‘HF—‘HP‘HP‘D—‘H Far—J r—-‘ i,‘ F4! C 3‘: C - *— r-a ,_, 2... Hp—JHHrdp—FJHPa'-h«HPP-HH--HHHHHF_F45...; Pray—4H...- P4 (F‘F‘P-‘C \ i—ar—r—f( PIC-1'4 C p_.a +14 A7; N r. \J U FAdl bFVI-‘I-d haw r—«i—ar—JU'Ni—tr—ar—Ji—Jr‘ I—‘F r- Fdh‘r 1—4 m: FJ+h.-a’-«*_J*4POH'—FJ he i be H F—J HI—JI—‘F—fh-‘r—JI—‘L- 1 ) (L “31 4k r—-1—fi\.r~JC‘l\;1.el\Jr-.; c o I \ \)f ,1 \._" ( -'\ . ( 1"} u: Jute—J1" .~. r“. L— i\.. 't ( -: l'\‘ - 2.. C ro ru (NM 1“ 1‘ * ‘h . D ' ' V LJACKQIKULJ'HL.) .bNr—INI—I \ l‘~‘)-*r—-'1\)r—'C' *- b-b‘r—‘N \. C. r—J 1-J \- - \1 |\-' C “- P—d . k 7‘ ‘ C L _ I\7 Pk- r—’ r—‘ r-4 b. - ,. .‘ T‘J 1‘.) b '3: 1—4 N (. }> 1 b3 1-48 1‘29 1‘14 1‘26 U—(iv I-bo 1-12 0-22 C-l7 C C- (- #0 H r w I-‘ C- I—‘CI—‘C 17 23 16 34 44 35 27 23 3V .1 35 75 go 29 68 39 I7 21 bu 64 I8 DATA 0 68 O 84 1102 1 5b @113 C) CH b} C O 93 1-44 0107 0123 1137 \J 7t. 81 45 Club '0126 I I7 1 .1 I 62. L31) U ()4 I I 1 57 I 04 I I I I 04 7 I 036 I 04 \J .5 .3 C 1106 1134 26 73 L. U C I but bli- -9 l J S f‘ I I I 1 I. q 04 634 I 0;: O;6 C3 4" .O .11. «IL 96 o I o 9 1 04 I I o (’1‘, U_l 50 1,) 36 J 20 I 26 1‘23 .y‘ :2 01 1‘40 (- ur- bl- L1 I \I \- ’ I - J 6) 04 I I 1 04 I I 3 00 1 (81 062 1 \ U “3 a} .L 11. I. I I 14“ I 4 08 d I :UZ. «1; J5 \ -4 DO I L. VII 1) 11 I12» bljé '04 l 36 I-Hb \J I I 42 m J 1-36 1-44 12 I lo 1‘ I 24 I 'J 8-24 & 1—24 5 it 44 U.) J 506 4 08 d I I 56/ 4 08 0 4 08 C I 5L0 4 08 O I I 61 306 4—4. «In. Du. AU 1.... ..- -.3 _ I be 72 .7 I‘ww we 2 08 08 C- U I 514 4 08 0 l vI92 IIZC I .7") 31 v0 <- I. .10 ..J 3 dl L I U o} I I 2 do U 3 ,J. L) v 114% I I2 0 uIJé hi.) \ V .J ") .1.- 5 CO Vt) 0V -"‘.Lz.' -) -.J ".1 03 2 I J )3 I O Q) I I 2 I I {J O O O 13 8 I3 92'8 13 -6O 8 13 561 ~ I L. .4 L :1 \J -I (1 96 0 . 3-44 C 42 02 2 I JD 0 6% Z .2 72 U I“.JI ~J Z4 0 1‘13 0 66 OI74 I‘I2 .15 b 13 '1 J 602 C'IZv 64 bu "v .9 t7.) .1..- 11. LO 7.; . .3 (U. 1 72 I tsz UI32 ...u 60 0-24 1-37 I-I2 c. 4 IL” 3 36 L) '40 72 36 24 O 1108 I U‘HJ J \- C: ‘-+ 06 TM.‘ GI O-JS v; ,0 3 ,U (0. Q 52 30 Au 36 1-36 24 O C J ’-l u+ 2 C4 1 1 I #1 J. 14 1 7 (‘0 J 65) 656 6'. 0 3 l l l l 1 I4. .Ul .v .0 34 1|. w l O . 1.. .J 0 .v a 447 ..J u .é L v-1 1‘1]; LU/l Lu J 1I‘L lii 1&1 1414. ll Can} 70 3,0 {00 I .1 VII—(:5 0 2.91.1 V 14b 3".) F7 AJ . 0 ~39 7|) I I #9 14 9 9 '. 694 I M Q d” 2 1... b") J I I 'Ju I 12 | 7 7 60C) 733 112.. I c... . \H.’ 6 «.2. _ GI ... ,.+ J :2 _ _ I .u 1 -.. .2 a ....J a, z. .2. I .1 I .u a. D D u. u 1 1 .. U Q .2. 2 I 1 7 7 ”.1. 3 ”v ..U 7 7. r 9193 \ (.V A I. O v 4b 2% \J 1+ 1 1; b0 qt. 4' l... K.) J L 1 '7‘ L 4-. 7 12 7 12 7U4 7 12 7..» 7-1/ 7 7b) 5 L' I ‘§ ‘IVJ ...2 ‘3'. 11 ...HO VIC» 1‘12 \.“‘+ 4:) o M; Ad .11 r0 .1 I. “U all "; 2 I ’ '73: 6) IO 1 Q I+24 :0 J -.. I ., 1L 1 ‘L IJ '. .l. E) L 10 i 7 lIUB 48 (2 ‘1-7 1 I A \J \J in- b 720 ‘ . J \,’ II. 12‘-«- 11. V 6;? 1")" .- :)‘ 1 12 1 2: "1' \l J l‘ 1 I 6 ")1 I 7CD 6 11 7 1|... 17 6 7 "f2 2 21+ 24 ‘J(. \J b 11 O 1 7f; :0 115 1’36 V 1’12 .0 1-61 VI 1-12 1; 6 11 u I U; 792 6 7&4 7 1‘24 1‘36 u—I2 11. Go 6 bu a. IA 1.1. Lu (.0 7!. 96 O 36 J (5 11 fdv (,1 \-' 1|; J”) dd 'J.) Ru “- O (O (O 3 3 U ad 2 ,4). L AU 3.— ”4.. \1. ...: .. A. «C II. J I 1 SJ «1.. ... .D ‘1‘ 1 11 AV x... p: .vql. NJ ....» 2 «I. IQ ”D ,U 0 8 8 1‘42 24 1-44 0 .1 ‘ - 05 O 809 2 0—46 7,— 811 LIVE I C 36 d 64 1 l 2 v6 L6 , I 3 \ K1 ()5 05 812 2 815 814 2 J?) 813 829 a} ‘44 v) ‘44 uQ 2 21 84 1 I ‘4 (.4 1 bl} I J3 I-IJ I 9; db 2 626 821 830 48 x J 1 2 Q7 36 1‘32 0 18 1) u.‘ wJ. I V {4+ 5'! U7 1 2 .11 852 \D I >7 I 07 1 C136 3 . an... -4 6 -9 .J 1.... l ...l. w l J u I I I. «U 1 6 u .41. l l .0 ...? J d . _ “w l 7.17 v--2d u a V ,U ad 2 . v . I A 11‘ .I. l a .. OJ .1 5 J J .4 J u I I I 1.. I 1. qrv-7 ,U ..U U 2 ...L A/_ I) .0 I ....J .3 .3 n0 ,0 ..u 46 11 2 a7 646 APPENDIX B SOCIOMETRIC TEST DATA Key for Sociometric Test Data Column Numbers Number of Columns Range of Valid Cards Item Description 34 35-37 38 39-41 3 l l _a ..a 09—: 99—; mg w_a_a_a —J 001-840 l=urged 001-840 B—l Child I.D. No. Blank Child gave same choice all 3 times Blank Tester I.D. No. Blank Child I.D. No. of lst choice Blank Child I.D. No. of 2nd choice Blank Child I.D. No. of 3rd choice Blank Child I.D. No. of negative choice Blank Attitude of child Blank Child I.D. No. of lst choice on retest Blank Child I.D. No. of 2nd choice on retest Blank Child I.D. No. of 3rd choice on retest Blank Child I.D. No. of negative choice on retest 001 002 003 004 006 007 306 00) 010 011 012 314 015 016 017 018 01) 020 022 023 024 025 026 027 328 029 030 031 032 033 0 4 h. VJJ 336 C37 333 339 240 041 C42 C45 C44 046 C47 0 "rd 1) 4 9 350 091 C-2 055 Ord'DOOOOr-‘OOO'DOHOOOO()OHOOHOOCD'JHOOOOHHOOO OcokaOv~C)Of)~dO(D .‘CQCHHHHP‘HHHP-‘HHHHHHH '— A \a ‘ COL C'JOOOOCJ -3CH"H*“HHHHHO—dr—‘HHHHH r\ \-’ R5”: 017 .15 30' 364 011 )13 017 J17 015 015 C17 “17 016 uI? ‘05 411 016 1413 013 JIL) 007 617 516 ‘v1 311 II 0C4 .12 CO4 04 002 02 052 J26 022 022 019 .3 022 «26 (J25 ~32; 024 .23 033 £25 7733 oz; 027 J27 028 C26 019 ‘27 025 023 031 025 023 026 035 335 334 4} 349 ;45 040 \48 042 035 J46 35 037 :37 043 qu 038 045 047 336 C48 $45 943 C46 C43 05J {.5'J J/l3 J35 £52 )43 ,4? J54 ';4 362 “67 067 662 530 ODD LCD 000 000 00; JUL LOU uLJ 000 000 \1 dc. bv‘J 000 M00 300 000 003 UCU 006 000 who ‘;)\)U \1II ’ \J v‘uU 0L0 :) (I I: 000 000 000 (J00 -47 61A) 041 034 054 045 036 035 C00 037 637 ’3 VJ JUL 000 34 066 u‘ U 1.? 00() buCIUHLTRIL Crv—a HCJ (3 C: C ()H (I -r \‘ ;)hnH(3CZC(DC;O(DFJO(JC)L'5 C) -'.‘L') Or—o ,- \l C) H(-JC)‘)OQ O C: C (_ DATA .—l .L‘ C. (J C; 5. 1,. ...; F4 r-—' .... NCP£~W~0H \ 02‘ 015 ?1( 015 03? 004 017 632 233 033 027 C28 019 323 LL JN 0‘0‘ 1'1 ‘ -, \Jl \fl ’.“ 0’ N Lt? ’2' (-3 d. (_'. .. '£>£mb\obkpr a C}C min. CDUJN 54C L45 J36 J43 ~kf‘. '11.] .35 .943 seat. ..53 cos 015 £11 016 005 001 004 017 C11 312 017 013 CIU 010 004 015 3J6 007 001 QUO 004 017 019 025 023 025 026 027 027 025 925 025 023 039 034 042 050 033 O44 045 C44 038 333 046 043 330 046 235 C41 J64 C61 367 DC“ 000 063 000 03? 000 000 000 003 000 000 000 000 000 000 036 034 041 034 035 034 034 034 035 037 f‘.“ v. 934 036 003 L34 U58 003 061 055 054 0:6 057 C58 059 060 061 O62 063 064 066 067 065 500 501 5J2 503 504 505 506 508 510 511 513 514 516 517 550 553 554 555 556 557 5329 560 561 562 563 564 (J'v'l 602 603 604 635 606 607 606 607 610 612 614 01) 616 651 OOC)OOOOOOOOP—'OC)0530()OOHHHHOOPJOOOOOHDOHOHOOHOOOHOOOOOOOOOO \LUDUWWWWU1uwwwWWkdwngWIOQC\Ou)\C\3\‘JUYWU‘WWLQNWWNWNWWOOOC)OCOOOOCQOO 06” 064 053 063 O59 068 054 063 061 057 054 35;) 053 065 510 508 514 505 502 510 514 514 500 515 511 508 5(16 505 560 I 5') 560 DJ \ I r- . \. N ‘J1 \L} U" _ b) b O 'I‘ U‘ U7 ‘3 U3 U‘U 553 562 56; 021 605 607 605 607 :75 wd/ 604 607 921 605 613 611 607 03 667 054 '76)" 062 055 665 L62 064 053 055 053 154 251 \‘. \J 5‘ Jo r\ ’1 :-' .1 4 55v ‘7‘) U 559 97v 6J7 6 v if.) 921 E) “J 5 ‘7 Z 1 6&1 613 062 068 052 061 061 065 051 057 052 059 054 054 062 060 514 514 510 51 510 510 60 \.' \J‘I U1 \, (.1 k h) h.) l-‘ '. \ L" Lx. 910 560 557 556 990 561 560 5J 41 610 921 610 C: '5 6J7 612 613 645 615 6v1 641 616 721 611 a 0 7 L" \T' 300 056 1464 Vb? .4151) 661 000 056 367 061 058 C52 U66 U63 0511 000 000 000 (JJU gnu 000' ix”) 000 000 600 1,14,) 000 'UUU VVV 000 JLI'U 000 060 000 000 000 1...... ‘C. C) (:0 (.3 O GHQ a", \' (.1 O L; J ‘CC.C.r—4C(- L; x. -\ ‘4 f) C ,-. \ C‘C‘L.‘ ( J68 064 J62 )63 V66 060 054 053 J61 062 054 LA" 053 060 060 116:, 667 061 H68 054 C68 055 L53 653 £154 051 063 065 062 C55 061 053 065 068 066 066 056 062 354 054 053 055 000 000 U59 U67 056 061 062 052 056 061 US3 003 068 056 OJ (C C. 40 P-‘ ~J‘J\J\J\|\l-J \U‘ i \f (.C ('3 C." h C. u; u; (I) C 605 Lu* 31);» 857 608 8d? 810 811 :31: C)OC)C)()OOOOOOOOHOOHOHOOLJOOQOOUOOOOHO )OOOOOOOOC)OOQ ’ \ C.) mmwx‘uy; \‘J \L) x.) l‘rJNl‘JhJfi.‘NR;NPQNNNNNQN‘JNNNNNNO‘O‘O‘O‘O‘O‘O‘O‘O‘OU‘Q-CX‘Qmw03ClTOJO‘DCLCil ia)‘ J—I 763 757 753 750 991 750 754 7:3 753 753 754 787 751 78:) 781 775 751 77:: 7:»: 731 7a") 751 777 81'; 7'92. 75;; 11“ 0;...) x“) '\ 'I\ u U 1;. (J); C), 05; DU: 9;) (1.», b0, :1; 1' uai 71% 71% .:N N. C: 1‘ ' p—a \. k"(‘( C.“ (A C H!” 'NJUJNUJ\ i‘ukob.l‘.‘ , w- ” U1 C‘ \J1 CT.‘ \" ( ...... t "CNN‘JNVN-JNNN~J~J \. k. y... 73& 7: 7S0 7:25 7 3 775 75v 781 778 7&3 777 732 77? 78¢ 73v 772 7b? 7)? on Sgl $ng bl) Clu O32 313 Jv4 pl? ;o; a)? god uc7 (.1117 009 9J1 0.;2' o‘ ' c 'J 5) ‘J Jig Jud buU VUU v'vv \)JO ‘JUU u‘Qu uuU “LU JQJ add ggu UOJ VLC JQO uJO uQU V00 U00 U00 900 000 UOJ J-L‘C UVQ uCO LOO 000 \JV‘u' M,Q bJU CV0 COO UOC DOC JUQ v00 JQU UJO Wu; JVQ «JO QQU QRT'V DUO cm) (“x ‘J \l. \1 00b UJU VUU uQO vu'u A-o V\/ 1(JH (C f‘ \ 'LIOCJOC) C' (A C‘(. C C (7 ( C' C' C. f... (. 814 815 32‘: 85b 52‘) and 831 I 0.14:. 855 .33 636 \ \' CI) 1 O) U) \ (X. 83 8 J 4;; _\ ... (7? b M m 00 0? C17 \1 h H UI \ + HPVI—‘FJ ..T 030307C: \.\WI\\~\V VC‘C“ «L‘J-‘b-L‘ Px OC)OOC)OOOL)r—'OO WKALJN HFJP‘U‘ O7CL‘OL 03C)03CL b) F) I“? 41‘ DD 1541‘ U‘.\,‘\ 00:) C‘! QLQ DLU 90's.; V'LO 'u'v'U 01:0 000 JUL) Lvd JV" ~.) ‘03 L) 3;) C CO ‘u C: UGO U; ’. C‘; C C- C L. ( . C. I‘ C) O g? Q ( BIBLIOGRAPHY BIBLIOGRAPHY Adams, B. Birth order: A critical review. Sociometry, 1972, 35(3), 411-439. Adams, R. L. Motivational and achievement differences among children of various ordinal birth positions. Child Development, 1972, 43, 155-164. Adler, A. The origin of the neurotic disposition. In H. L. Ansbacher & Q R. R. Ansbacher (Ed.), The individual psychology of Alfred Adler. New York: Basic Books, 1956. Aldous, J. Occupational characteristics and males' role performance in the family. Journal of Marriage and the Family, 1969, 31, 707-712. Alexander, C. N. Ordinal position and sociometric status. Sociometry, 1966, 29(1), 41-51. Allen, K., Hart, 8., Buell, J., Harris, F., & Wolf, M. Effects of social reinforcement on isolate behavior of a nursery school child. Child Development, 1964, 35, 511-518. i Altus, N. Birth order and its sequelae. Science, 1966, 151, 44-49. Berk, L. E. Effects of variations in the nursery school setting on environmental constraints and children's modes of adaptation. Child Development, 1971, 42, 839-869. Bernstein, B. Language and social class. British Journal of Sociology, 1960, 271-276. Bernstein, B., and Henderson, D. Social class differences in the relevance of language to socialization. In P. H. Mussen, J. J. Conger, & J. Kagan (Ed.), Readings in child development and personality. (2nd ed.) New York: Harper and"Row, 1970. Biehler, R. Companion choice behavior in the kindergarten. Child Development, 1954, 25(1), 45-50. Boger, R. P., & Knight, 5. The Play Situation-Picture Board Sociometric Technique. (Research and evaluation papers of the Head Start Evaluation and ResearCh center). East Lansing, Wfichigan: *MiChigan State University, 1969. Boroson, w. First-born--fortune's favorite? New York Times Magazine, Dec. 13, 1970, 6, 87—99. 136 137 Bott, E. Conjugal roles and social networks. In R. L. Coser (Ed.), Ih§_ Family: Its structure and functions. New York: St. Martin's Press, 1964. Bragg, B., Dstrowski, M., & Finley, G. The effects of birth order and age of target on use of persuasive techniques. Child Development, 1973, 44, 351-354. Campbell, D. T. A rationale for weighing first, second, and third sociometric choices. In J. Moreno, The Sociometric Reader. Glencoe, Illinois: The Free Press, 1960. r-n Chamberlin, R. w., & Nader, P. The relationship between nursery school E‘ ' behavior patterns and later school functionning. American Journal of i - Orthopsychiatry, 1971, 41(4), 597-601. 1 Christensen, H. T. Children in the family: Relationship of number and spacing to marital success. Journal of Marriage and the Family, 1968, 30, 283-289. Cicirelli, V. Sibling constellation, creativity, I.D., and academic achievement. Child Development, 1967, 38, 481-490. Cutright, P. Timing the first birth: Does it matter? Journal of Marriage and the Family, 1973, 35(4), 585-596. Deutsch, M. Minority groups and class status as related to social and personality factors in scholastic achievement. In M. Deutsch (Ed.), The disadvantaged child. New York: Basic Books, 1967. Donaldson, T. Robustness of the F-Test to errors of both kinds and the correlation between the numerator and denominator of the F-Ratio. Journal of the American Statistical Association, 1968, 660-676. Doren, M. P. Evaluation of studies on birth order and sibling position. Doctoral dissertation. (Dissertation Abstracts International, 1973, 33(lD-A), 5548-9). Dunnington, M. Behavioral differences of sociometric status groups in a nursery school. Child Development, 1957, 28(1), 103-111. Dunnington, M. Investigation of areas of disagreement in sociometric measurement of preschool children. Child Development, 1957, 28(1), 93-102. Dyer, E. 0. Parenthood as crisis: A re-study. Marriage and Family Living, 1963, 19, l96-201. Forer, L. K. Birth order and life roles. Springfield, Illinois: Charles G. Thomas, 1969. 138 Franklin, J., & Scott, J. Parental values: An inquiry into occupational setting. Journal of Marriage and the Family, 1970, 32, 406-409. Freud, S. Introductory lectures on psychoanalysis. London: George Allen and Unwin, 1933. Gabel, Yvonne. Birth order and achievement in Anglo, Mexican-American, and Black Americans. Doctoral dissertation, 1972. (Dissertation Abstracts International, Jul. 1972, 33(1-A) 190.) Glass, G., & Stanley, J. Statistical methods in education and psychology. Englewood Cliffs, New Jersey: Prentice-Hall, 1970. F“1L Glick, S., Grosse, M., & Zigler, E. Birth order, number of siblings and social reinforcer effectiveness in children. Child Development, 1971, 42(2), 1153-1163. Harris, I. 0. The promised seed. London: Collier-Macmillan, 1964. -«Il Harris, I. D., & Howard, K. I. Birth order and responsibility. Journal of Marriage and the Family, 1968, 30, 427-432. Hartup, w. Peers as agents of social reinforcement. Young Children, 1965, 20, 176-184. Hartup, M., Glazer, J., & Charlesworth, R. Peer reinforcement and sociometric status. Child Development, 1967, 38, 1017-1024. Hendershot, G. E. Familial satisfaction, birth order, and fertility values. Journal of Marriage and the Famflly, 1969, 31, 27-33. Hines, G. Birth order and relative effectiveness of social and nonsocial reinforcers. Perceptual and Motor Skills, 1973, 36, 35-38. Hobbs, D. Parenthood as crisis: A third study. Journal of Marriage and the Family, 1965, 27, 267-272. Horowitz, F. D. Latency of sociometric choice among preschool children. Child Development, 1961, 32, 235-242. Hurlock, E. Child develppment, (5th ed.). New York: McGraw-Hill, 1972, 226-282. Jacoby, A. P. Transition to parenthood: A reassessment. Journal of Marriage and the Family, 1969, 31, 720-727. Jensen, A. How much can we boost 1.0. and scholastic achievement? Harvard Educational Review, 1969, 39(1), l-123. Jersild, A. T. Child psychology, (4th ed.). Englewood Cliffs, New Jersey: Prentice-Hall, I954, I9l-298. 139 John, V. P., & Goldstein, L. S. The social context of language acquisition. In M. Deutsch (Ed.), The disadvantaged child. New York: Basic Books, 1967, 163-175. Kerckhoff, A. C. Status-related value patterns among married couples. Journal of Marriage and the Family, 1972, 34, 105-110. Kircher, M., & Furby, L. Racial preferences in young children. Child Develqpment, 1971, 42, 2076-2078. Kmenta, J. Elements of Econometrics. New York: Macmillan, 1971. Koch, H. The relation of certain family constellation characteristics and the attitudes of children toward adults. Child Development, 1955, 26(1), 13-40. Koch, H. Some emotional attitudes of the young child in relation to characteristics of his sibling. Child Development, 1956, 27(4), 393- 426. Koch, H. The relation in young children between characteristics of their playmates and certain attributes of their siblings. Child Development, 1957, 28, 175-202. Koch, H. Twins and twin relations. Chicago: University of Chicago Press, 1966. Kohn, M. Social class and parent-child relationships: An interpretation. In H. Rodman (Ed.). Marriage, family, and society, New York: Random House, 1965, 230-244. Konig, K. Brothers and sisters. Blauvelt, New York: St. George Books, 1963. Landers, D. Sibling-sex-status and ordinal position effects on females' sport participation and interests. Springfield, Illinois: State Department of Mental Health, 1969, (ERIC 042 212). Landreth, C. (Ed.) Early childhood, behavior and learning. New York: Alfred A. Knopf, 1967, 306-320. Lichtenwalner, J., & Maxwell, J. The relationship of birth order and socioeconomic status to the creativity of preschool children. Child Develppment, 1969, 40(4), 1241-1247. Lorber, N. The reliability and validity of sociometric measures. The American Sociologist, 1969, 4(1), 243-4. Marshall, H. R. An evaluation of sociometric-social behavior research with preschool children. Child Develppment, 1957, 28(2), 131-137. 140 Marshall, H., & McCandless, 8. Relationships between dependence on adults and social acceptance by peers. Child Development, 1957, 28, 413-419. (a) Marshall, H., & McCandless, B. A study in prediction of social behavior of preschool children. Child Development, 1957, 28(2), 149-159. (b) McCandless, B., 8 Marshall, H. A picture sociometric technique for preschool children and its relation to teacher judgements of friendship. Child Development, 1957, 28(2), 139-147. (a) participation of preschool children. Child Development, 1957, 28(4), . 421-430. (b) : McCandless, B., & Marshall, H. Sex differences in social acceptance and -run1 McKee, J., & Leader, F. The relationship of socio-economic status and aggression to the competitive behavior of preschool children. Child Development, 1955, 26(2), 135-142. McSweeney, M., & Olejnik, S. Notes, exercises, and diagnostic measures ; for education 969B. Unpublished text. Michigan State University, ’ Department ofiEducational Psychology, 1974. Moore, S., & Updegraff, R. Sociometric status of preschool children related to age, sex, nurturance-giving, and dependency. Child Development, 1964, 35, 519-524. Moran, G. Ordinal position and approval motivation. Journal of Consulting Psychology, 1967, 31(3), 319-320. Northway, M. Apprimer of sociometpy. Toronto: University of Toronto Press, 1967, 2-53. Parsons, T., & Bales, R. Family, socialization, and interaction process. Glencoe, Illinois: The Free Press, 1955. Read, K. H. The nursery school. Philadelphia: w. B. Saunders, 1971. Ripstra, C., & Schmidt, H. Syllabus for education 869. Unpublished text. Michigan State UniverSTty, Department of Educational Psychology, 1973. Rodman, H. Middle class misconceptions about lower-class families. In H. Rodman (Ed.) Marriage, family, and society, New York: Random House, 1965, 219-229. Ryder, R. 6. Longitudinal data relating marriage satisfaction and having a child. Journal of Marriage and the Family, 1973, 604-606. Sampson, E. The study of ordinal position: Antecedents and outcomes. In B. A. Maher (Ed.), Progress in experimental personality research. Vol. II, New York: Academic Press,719651 141 Schachter, 5. Birth order and sociometric choice. Journal of Abnormal and Social Psychology, 1964, 68(4), 453-456. Schachter, S. The psychology_of affiliation. Stanford, California: Stanford UniverSTty Press, 1959, 62-89. Schmuck, R. Sex of sibling, birth order position, and female dispositions to conform in two-child families. Child Development, 1963, 34, 913-918. Schooler, C. Birth order effects: Not here, not now! anchological Bulletin, 1972, 78(3), 161-175. Sears, R., Maccoby, E., & Levin, H. Patterns of child rearing. New York: Harper and Row, 1957, 394-419. Solomon, 0., Hirsch, J., Scheinfeld, R., & Jackson, J. C. Family characteristics and elementary school achievement in an urban ghetto. Journal of Consulting and Clinical Psychology, 1972, 39(3), 462-466. Sterner, G. Birth order and self-esteem as determinants of affiliative behavior under ego threatening conditions. Doctoral dissertation, 1973. (Dissertation Abstracts International, 1973, 33(ll-B), 5501). Stotland, E., & Walsh, J. Birth order and an experimental study of empathy. Journal of Abnormal and Social Psychology, 1963, 66(6), 610-614. Sutton-Smith, B., & Rosenberg, B. G. The sibling. New York: Holt, Rinehart, & Winston, 1970. Taylor, M., A Kogan, K. Effects of birth of a sibling on mother-child interactions. Child Psychiatry and Human Development, 1973, 4(1), 53-58. Toman, W. Family constellation: Its effects on personality and social behavior. New York: Springer, 1969. Waldrop, M., & Bell, R. Relation of preschool dependency behavior to family size and density. Child Development, 1964, 35, 1187-1195. Weiss, J. Birth order and physiological stress response. Child Development, 1970, 41, 461-470. Witryol, S., & Thompson, G. A critical review of the stability of social acceptability scores obtained with the partial-rank order and the paired-comparison scales. Genetic Psychology Monographs, 1953, 48, 221-260. HICHIGQN STRTE UNIV. LIBRQRIES 111 1111 Ill 11111 1111 |||1111111111111Ill 31293104538990