University of Southern California Dissertations and Theses

Children's judgments of the similarity of television series and the similarity of gratifications associated with television series

(USC Thesis Other)

Children's judgments of the similarity of television series and the similarity of gratifications associated with television series

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content CHILDREN'S JUDGMENTS OF THE SIMILARITY OF TELEVISION SERIES AND THE SIMILARITY OF GRATIFICATIONS ASSOCIATED WITH TELEVISION SERIES by Peter Michael Kovaric A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillm ent of the Requirements fo r the Degree DOCTOR OF PHILOSOPHY (Communication Theory and Research) December 1985 UNIVERSITY OF SOUTHERN CAUFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK This dissertation, written by Peter Michael Kovaric under the direction of his. Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillment of re quirements for the degree of LOS ANGELES, CAUFORNIA 90089 DO CTO R OF PH ILO SO PH Y Graduate Studies December 20, 1985 DISSERTATION COMMITTEE Chairperson ACKNOWLEDGEMENTS As w ith any dissertation, many people have contributed to the long and d if fic u lt process of bringing this one to its present form . I would like to thank all of the follow ing for their contributions; w ith o u t them, this piece of w o rk w ould look entirely different, if it existed at all. W ithout data, m ost of the ideas discussed herein would remain idle speculation. For providing m eaningful data, I relied on the cooperation of the children, and their parents, teachers, and school adm inistrators of the follow ing schools: Am erican M artyrs School, Eastwood Elementary School, Franklin Elementary School, La Pluma Elementary School, Loyola High School, Marina Light and Life School, McKinley Elementary School, Redondo High School, St. John Fisher School, South Bay Union High School, and West Los Angeles Baptist School. W orking in these schools was always both pleasant and challenging. A number of people served capably in gathering, coding, and checking data. Kathleen Brown, Danny H ilton-C halfen, Steve Guberman, Walt Kimmelman, and Elizabeth Lewis all conducted expert interview s w ith the children w ho participated in the study. For data entry and checking, I relied heavily on the patience and painstaking care of Danny Hilton-Chalfen, W alt Kimmelman, Elizabeth Lewis, and David Lohr. Elizabeth Lewis carried out her half of the content analysis w ith cheer, and, even more im portantly, w ith high reliability. Phil W atson w rote the com puter program s to organize and com bine the individual children's data matrices. ii In addition to help in w orking through some of the intricacies of ALSCAL, Cathy Doubleday has been a prim ary source o f the intellectual environm ent in which I have worked now fo r many years. She deserves more acknowledgem ent than a paragraph allow s fo r being a w illin g audience and critic, and em bodying collegiality in it's true st sense. To get from printouts of analyses and drafts of the text to this final docu ment, all of which were created on num erous different com puter systems, re quired the help of a grand m aster of sm all and large system com puting. Skip Eastman guided me through that maze, som etim es pulling and som etim es push ing me along. My dissertation com m ittee, Aimee Dorr, Bill Dutton, and Dick Clark, gave me a long leash and trusted me not to hang m yself w ith it. Fred W illiams, com m ittee m ember in spirit if no longer in fact, provided me w ith much needed guidance during the years he was at Annenberg. Finally, I would like to especially acknowledge Aimee Dorr fo r the many years she has been teacher, mentor, occasional adversary, always friend, and col league. Debts like the one I owe to Aimee can never be repaid directly, of course. I only hope I have the opportunity to provide fo r other students what she has provided fo r me. iii TABLE OF CONTENTS ACKNOWLEDGEMENTS i i ABSTRACT x i i 1. In tro d u c tio n 1 1.1 The Content of Television Program m ing as a Variable 2 1.2 The Function of Television Program ming as a Variable 10 1.3 Identifying Program ming Dimensions 14 Age differences in dimensions 18 1.4 Hypotheses 20 2. M eth ods 22 2.1 Pilot Study 24 2.2 Participants 26 2.3 Experimenters 29 2.4 Conditions 30 2.5 Measures 31 2.6 Procedures 36 3. R esults 39 3.1 S im ilarity o f Series' A ttributes and G ratifications 39 3.2 Age Differences in A ttribute Dimensions 51 Content analysis results 51 Correlation analysis results 79 Canonical analysis results 81 Summary 89 3.3 Age Differences in G ratification Dimensions 90 Content analysis results 90 i v Correlation analysis results 94 Canonical analysis results 97 Summary 103 3.4 Condition Differences in Dim ensions w ithin Age 104 Second grade results 104 Sixth grade results 106 Tenth grade results 111 3.5 Predictive Validity of the Dim ensional Solutions 112 4. Discussion 137 4.1 Age Differences in A ttribute Dim ensions 138 4.2 Age Differences in G ratification Dimensions 142 4.3 Differences Between A ttribute and G ratification Dimensions W ithin 145 Age 4.4 Content Analysis and M ultidim ensional Scaling as Means of 146 Representing Children's Judgm ents of Television Series' S im ilarity 4.5 Predicting Beliefs about Social Reality 147 Second graders 148 Sixth graders 150 Tenth graders 153 Summary 154 4.6 Lim itations of the Study 156 4.7 Directions fo r Future Research 158 Categorization of program m ing: Dimensions and clusters 158 Predicting beliefs about social reality 159 REFERENCES 1 6 3 V A. Examples of Cards Used in Sorting Task 171 B. Attribute Condition Scripts 173 B.1 Instructions — 2nd Grade A ttribute Condition 174 B.2 Instructions — 6th/10th Grade A ttribute Condition 177 C. Gratification Condition Scripts 180 C.1 Instructions — 2nd Grade G ratification Condition 181 C.2 Instructions — 6th/10th Grade G ratification Condition 184 D. Viewing Frequency Checklist — Short Version 187 E. Viewing Frequency Checklist — Long Version 191 F. Content Analysis Coding Scheme 196 F.1 INTRODUCTION 197 F.2 ATTRIBUTE CONDITION 198 F.3 GRATIFICATION CONDITION 202 F.4 REVISIONS TO ATTRIBUTE CONDITION CODING 205 G. Social Reality Questionnaire 207 H. Envelopes Used for Sorting Television Series 215 vi LIST OF FIGURES Figure 3-1: RSQ and STRESS fo r A ttribute Condition 47 Figure 3-2: RSQ and STRESS fo r G ratification Condition 48 Figure 3-3: Second Graders' Series A ttributes in Two Dimensions: Dimension 1 by Dimension 2 66 Figure 3-4: Sixth Graders' Series A ttributes in Two Dimensions: Dimension 1 by Dimension 2 67 Figure 3-5: Tenth Graders' Series A ttributes in Three Dimensions: Dimension 1 by Dimension 2 68 Figure 3-6: Tenth Graders' Series A ttributes in Three Dimensions: Dimension 1 by Dimension 3 69 Figure 3-7: Second Graders' Series G ratifications in Four Dimensions: Dimension 1 by Dimension 2 70 Figure 3-8: Second Graders' Series G ratifications in Four Dimensions: Dimension 3 by Dimension 4 71 Figure 3-9: Sixth Graders' Series G ratifications in Four Dimensions: Dimension 1 by Dimension 2 72 Figure 3-10: Sixth Graders' Series G ratifications in Four Dimensions: Dimension 3 by Dimension 4 73 Figure 3-11: Tenth Graders' Series G ratifications in Three Dimensions: Dimension 1 by Dimension 2 74 Figure 3-12: Tenth Graders' Series G ratifications in Three Dimensions: Dimension 1 by Dimension 3 75 Figure 3-13: Tree Diagram o f Cluster Analysis Results fo r Grade 2 A t tribute Condition Dimensions 115 Figure 3-14: Tree Diagram of Cluster Analysis Results fo r Grade 6 A t tribute Condition Dimensions 116 Figure 3-15: Tree Diagram o f Cluster Analysis Results fo r Grade 10 A ttribute Condition Dimensions 117 Figure 3-16: Tree Diagram o f Cluster Analysis Results fo r Grade 2 G ratification Condition Dimensions 118 Figure 3-17: Tree Diagram o f Cluster Analysis Results fo r Grade 6 119 v i i G ratification Condition Dimensions Figure 3-18: Tree Diagram of Cluster Analysis Results for Grade 10 120 Gratification Condition Dimensions viii LIST OF TABLES Table 2-1: Mean Ages, Standard Deviations, and Numbers o f Par ticipants 28 Table 3-2: Number of Second Grade Children in Each Condition Who Sorted Each Series 41 Table 3-3: Number o f Sixth Grade Children in Each Condition Who Sorted Each Series 42 Table 3-4: Number o f Tenth Grade Children in Each Condition Who Sorted Each Series 43 Table 3-5: Television Series in Final Sample by Nielsen Category 45 Table 3-6: Means, Standard Deviations, and Ranges of C ontent Analysis A ttributes fo r Grade 2 52 Table 3-7: Means, Standard Deviations, and Ranges of Content Analysis A ttributes fo r Grade 6 53 Table 3-8: Means, Standard Deviations, and Ranges of Content Analysis A ttributes fo r Grade 10 54 Table 3-9: Means, Standard Deviations, and Ranges of Content Analysis G ratifications fo r Grade 2 55 Table 3-10: Means, Standard Deviations, and Ranges of Content Analysis G ratifications fo r Grade 6 56 Table 3-11: Means, Standard Deviations, and Ranges of Content Analysis G ratifications fo r Grade 10 57 Table 3-12: Correlations between Content Analysis A ttrib u te s and A t tribute Condition Dimensions fo r Grade 2 58 Table 3-13: Correlations between Content Analysis A ttributes and A t tribute Condition Dimensions fo r Grade 6 59 Table 3-14: Correlations between Content Analysis A ttributes and A t tribute Condition Dimensions fo r Grade 10 60 Table 3-15: Correlations between Content Analysis G ratifications and G ratification Condition Dimensions fo r Grade 2 61 Table 3-16: Correlations between Content Analysis G ratifications and G ratification Condition Dimensions fo r Grade 6 62 ix Table 3-17: Correlations between Content Analysis Gratifications and 63 Gratification Condition Dimensions for Grade 10 Table 3-18: A ttribute Condition Dimensions 64 Table 3-19: G ratification Condition Dimensions 65 Table 3-20: Correlations between A ttribute C ondition Dimensions 80 Table 3-21: Canonical Analysis of A ttribute C ondition Dimensions 85 Table 3-22: Correlations between G ratification Condition Dim ensions 95 Table 3-23: Canonical Analysis of G ratification Condition Dimensions 98 Table 3-24: Correlations between A ttribute Condition Dimensions and 107 G ratification Condition Dimensions Table 3-25: Canonical Analysis o f A ttribute C ondition Dimensions and 108 G ratification Condition Dimensions Table 3-26: R-Squared and Cubic Clustering C riterion fo r First Ten 121 Clusters fo r Each Condition Table 3-27: Clusters of Series fo r Grade 2 A ttribute Dimensions and 122 Outcom e Measures They Predict Table 3-28: Clusters of Series fo r Grade 6 A ttribute Dimensions and 123 Outcom e Measures They Predict Table 3-29: Clusters o f Series fo r Grade 10 A ttrib u te Dimensions and 124 Outcom e Measures They Predict Table 3-30: Clusters o f Series fo r Grade 2 G ratification Dimensions 125 and O utcom e Measures They Predict Table 3-31: Clusters of Series fo r Grade 6 G ratification Dimensions 126 and Outcom e Measures They Predict Table 3-32: Clusters o f Series fo r Grade 10 G ratification Dimensions 127 and Outcom e Measures They Predict Table 3-33: Social Reality Outcomes Predicted by Average Viewing 129 Frequency of Tenth Graders Table 3-34: Social Reality Outcomes Predicted by Viewing Frequency 131 o f Nielsen Classified Series Table 3-35: Social Reality Outcomes Predicted by Viewing Frequency 132 of A ttribute Condition Series Clusters x Table 3-36: Social Reality Outcom es Predicted by Viewing Frequency 134 of G ratification Condition Series Clusters X I ABSTRACT Previous research relating children's social reality beliefs to exposure to television has m ainly used overall exposure or exposure to Nielsen genres as predictors. General principles of cognitive developm ent suggested that children m ight not use the same criteria as adults to classify television series, and that classification criteria w ould probably be com prised of content attrib u te dim en sions and function (gratification) dimensions. It was expected th a t older children w ould em ploy m ore and m ore com plex dim ensions, th a t content dim ensions would be m ore distinct from gratification dim ensions w ith age, and that exposure to series grouped by attribute or gratification would predict social reality beliefs better than exposure to Nielsen classified series, w hich w ould predict better than overall exposure. Participants were 47 second graders, 84 sixth graders, and 78 tenth graders who were white, m iddle class, and about evenly divided by sex. Children in dicated how often they watched each of 40 highly viewed entertainm ent series, then sorted the 40 series into groups they considered sim ilar, based either on the content of the series (attribute condition) or the functions o f the series (gratification condition). Next they indicated w hat made each group of series sim ilar or what the group of series made them feel or think about. Finally, a num ber of their social reality beliefs were measured. Judgm ents of series' sim ilarity were analyzed using ALSCAL which represented the series in tw o, three, and four dim ensional spaces. Descriptions xii of the groupings were content analyzed, w ith the results used to label the ALS- CAL dim ensional solutions. Content analysis, correlational analysis, and canonical correlation were used to compare the dim ensional solutions w ithin conditions and across age groups, and w ithin age groups and across conditions. Cluster analysis was used to identify groups of series w ith sim ilar dim ensional structure fo r each age and condition. Stepwise regression analysis, using social reality beliefs as outcom es, indicated that fo r second graders, series clustered by at tributes predicted best, fo r sixth graders, series clustered by gratifications predicted best, and fo r tenth graders, Nielsen genres predicted best. These results are discussed in the fram ew ork of m easuring exposure to television to predict social reality beliefs. xiii CHAPTER 1 INTRODUCTION Much research concerning the ways in which television affects children utilizes one of tw o traditional social science approaches: experim ents conducted prim arily in laboratories which dem onstrate that television can affect its audience in particular ways, and field surveys which attem pt to dem onstrate that it does affect its audience in those ways. Typically, o f course, one finds th a t in experi ments effects are of a much larger m agnitude than in surveys (Comstock, Chaffee, Katzman, McCombs, & Roberts, 1978). That m ost survey-based research has resulted in findings indicating statistically rather sm all effects o f television — typical R2s range from .00 to.15 — has been a source of frustration to resear chers in this area fo r years and has lead to repeated calls fo r m ore sophisticated m ultivariate research on the way(s) television is used by and affects its audience (cf., Anderson, 1981; Blumler, 1979; Pearl, Bouthilet, 8t Lazar, 1982). These calls, however, have been heeded less often than one would like. As w ell as the ob vious econom ic inhibitions to the conduct of such research, another more fun damental constraint exists: W hat are the appropriate variables, on both theoretical and practical bases, to include in it? More precisely, given the concern over the range of possible influences of television, w hat are the televisio n -re la te d predic to r variables one ought to include? 1 1.1 The Content of Television Programming as a Variable Television program m ing has been treated as a variable in survey research on effects in tw o prim ary ways. It has been considered an integrated elem ent of experience, to be measured by variations in overall exposure. Alternatively, it has been treated as com posed of conceptually distinct genres, to be measured by ex posure to each genre. The follow ing paragraphs summarize examples of each of these approaches, highlight th e ir lim itations, and conclude w ith an argum ent fo r an alternative to them. Probably the m ost com m only em ployed television variable in surveys of children (and, fo r that m atter, adults) is som e measure of exposure, usually operationalized as am ount of viewing per unit tim e. Two m ajor studies of te le v i sion influence on school achievem ent used this type of approach. Fetler (1984) reported on the California Assessment Program in which exposure to television was measured by asking sixth graders how much tim e they viewed television on a typical weekday, in half hour increm ents. An additional indicator was an ag gregate of how regularly they watched each of 27 popular series. Roberts, Bachen, Hornby, and Hernandez-Ramos (1984) used a sim ilar measure w ith second, third and sixth graders: the reported am ount of viewing tim e before school, after school, and in the evenings (in half hour increments). Both of these studies, then, presumed that television view ing per se may be related to academic achievem ent, a relationship borne out in their findings. Precisely how such exposure to television program m ing influenced achievem ent was not en tirely clear, however. It m ight be presumed in both studies th a t m ore exposure 2 to television in general resulted in less of whatever was necessary fo r higher academic achievem ent. Data Fetler (1984) reported make this interpretation ten tative, however, since heavier viewers reported w atching different series than lighter viewers. There are also more purely cognitive outcom es w hich have been employed in surveys w hich have used overall exposure as the measure of television. A long term em ployer of this approach to measuring exposure to television pro gram m ing is the Cultural Indicators Project (e.g., Gerbner 8 1 Gross, 1980; Gerbner, Gross, Morgan, 8 1 Signorielli, 1980; Gerbner, Gross, Signorielli, Morgan, 8i Jackson-Beeck, 1979), w hich has been concerned w ith how television influences children's and adults' conceptions of reality. In these studies, exposure to televi sion program m ing has m ost often been measured as the num ber of hours respondents report view ing on the average day. W hile typ ica lly reported in ways that do not use the standard i?2s (e.g., Gerbner et al., 1979; Gerbner et al., 1980), reanalysis of some Cultural Indicators data has suggested th a t exposure to television has typically resulted in Rzs of .05 to .10 fo r measures of fear, anomie, alienation, and the like, after controlling fo r age, sex, education, and race (Hirsch, 1980). Use of this measure o f exposure explicitly reflects the assum ption either that all television program m ing is essentially identical or that w hat little variability there is gets elim inated through exposure to m utually neutralizing con tent. Associated w ith this project, but focusing on another outcom e, Morgan (1982) surveyed sixth through tenth graders to assess the extent to w hich ex 3 posure to television influenced their attitudes tow ard sex-roles. He also found an effect o f television exposure, though lim ited to particular groups of children. Though the findings of the Cultural Indicators Project have been criticized on a num ber of m ethodological grounds (Hirsch, 1980, 1981a, 1981b), the exposure measures suffer from a conceptual lim itation as well. As content analyses have dem onstrated, television program m ing can and does vary on a num ber of variables across "types" of programs. For instance, the num ber of male and female characters, black and w hite characters, old and young characters, sexual acts, and orders given, received, and follow ed by male and fem ale characters have all been found to vary according to the type of program in which they occurred (Greenberg, 1980). If one is to argue, as do Gerbner et al. (1979, 1980), that overall exposure to television program m ing per se w ill con tribute to one's conception of the world, then it is appropriate to dem onstrate that content differences across program m ing, w hich could be argued to be relevant to the outcom e variable and w hich have been dem onstrated to exist, have a com paratively trivial im pact when com pared to overall exposure. This has not been done. An alternative to conceiving of television program m ing as an undifferen tiated w hole fo r which a single measure of overall exposure is an adequate in dicator is to think of exposure to particular genres or types of program m ing as likely to have greater or lesser relationship w ith the outcom e variable of interest. A t one extreme, this approach is represented by studies which examine effects of exposure to a single type of program m ing. For instance, Carlson (1983) found 4 that exposure to 16 different crim e series was negatively related to children's support fo r civil liberties. Buerkel-Rothfuss and Mayes (1981) found th a t college students' exposure to soap operas was positively related to perceptions of the num ber of professional men and w om en in the real world, the num ber of divorced people, the num ber of people w ho have been in jail, and the like. Singer, Singer, and Rapaczynski (1984) found that exposure to realistic action ad venture program m ing was positively related to young children's beliefs that the w orld is a scary place. W hile it can be argued that the particular program type em ployed is related to the outcom e measures in each of these studies on the basis of content, this conceptual connection cannot be used as evidence th a t the outcom e measures are differentially related to exposure to various program types. This kind of test represents the other extreme of this general class of studies — those which measure exposure to a num ber of different types of program m ing, expecting that not all types w ill be related to the dependent measure(s). M cllw raith and Schal- low (1983), fo r instance, em ployed this approach w ith adults and found that ex posure to sports and lack of exposure to n o n -vio le n t drama were related to a masculine sex-role orientation, w hile exposure to 6 other program types was unrelated to the outcom e. Focusing on children, Hawkins and Pingree (1981; Pingree 8i Hawkins, 1981) partitioned the view ing schedule o f a sample of Australian children into ten different types of program m ing: news, situation com edies, crim e adventure, drama, game shows, cartoons, docum entaries/public af fairs, m usic/variety, children's shows, and sports. They found th a t exposure to 5 different types of program m ing resulted in different effects on tw o measures derived form the Cultural Indicators Project: interpersonal m istrust and judgm ents of the am ount o f violence in the real world. These relationships ranged from partial correlations of .00 to .16. A lthough there is a clear advantage to this ap proach over that em ployed by Gerbner et al. in that it allowed id entification of those kinds of program m ing w hich did and did not account fo r variance in the dependent variables, the division of the program m ing reflected conventional in dustry labels, which, although perhaps applicable from an adult standpoint to the series they represented, have never been dem onstrated to be m eaningful to children. This, of course, can be considered a drawback to nearly all research con cerning children w hich em ploys such labels or categories. Until it has been dem onstrated that these categories are generally employed by children it is pos sible that they may be concealing part of the relationship between exposure and effects. There appear, then, to be lim itations to both of the above approaches to representing television program m ing in a m odel of effects. Measuring overall ex posure has resulted in finding very small effects and can be criticized fo r ignor ing dem onstrable differences in the relationship between program types or genres and outcom es. U tilizing measures o f exposure to program types or genres has relied on program categories w hich have not been dem onstrated to be actually em ployed by the children whose attitudes or behavior were being predicted. Although it could be argued that identifying television influences ac 6 cording to conventional adult categories is appropriate because policy makers and child caretakers are adults and would be best able to use inform ation fram ed in standard adult term s, the issue here is how theoretically to represent television program m ing influences in a way w hich w ill best predict outcom es o f interest. Thus adult categories may not be the m ost appropriate. The approach to be argued fo r here is th a t the categories of program m ing em ployed by the audience ought to be represented in a m odel o f television in fluence. As was indicated above, it has been dem onstrated that program type or genre makes a difference in the influence of exposure on outcom es, but w hether this difference is due to differences in theoretically im portant messages across types or to cognitive m echanism s related to how the program m ing is categorized has not been established. That exposure to cartoons, game shows, and crim e adventure program s should all be highly and about equally related to perceptions of violence in society, w hile sports, news, and drama are equally unrelated to those perceptions (Hawkins & Pingree, 1981), suggests that considering alter native categories fo r these program s, at least fo r children, has some face validity. A theoretical rationale fo r this assertion can be drawn from social learning theory (Bandura, 1977). In general, social learning theory as applied to television and children research has led researchers to look at the incidence of modeled behaviors (or attitudes) and the acquisition and/or perform ance of those be haviors, taking into account the age of the child audience, the appeal of the model, social norm s fo r the modeled act, the cla rity of the antecedents and con sequences of the modeled act, and the like. A great deal of experim ental 7 research in this field is either explicitly or im p licitly guided by the theory, as is much non-experim ental research. Essential to the occurrence of learning which is not directly reinforced is the perception of covariance between the modeled behavior and its consequences (Bandura, 1977). There are num erous examples in the social learning and autonom ic conditioning literature of behavior learned or extinguished because of m isperceived or erroneously anticipated consequences (e.g., Bandura & Barab, 1971; Chattergee & Eriksen, 1962; Fuhrer & Baer, 1965; Grings, 1973). It is thus im plicit in this explanation of learning that both the be havior to be learned and its consequences m ust be in som e way recognized, at tended to, and categorized as m eaningful by an observer. If they either are not recognized at all or are som ehow classified as invariant, then learning either does not occur or perform ance is eventually extinguished. In autonom ic conditioning this is referred to as the necessary-gate hypothesis (Dawson & Furedy, 1976). In the television view ing situation, it may be that the cognitive category into w hich a child places the program she or he is view ing is a "necessary gate" through which it m ust pass in order fo r particular kinds o f effects to take place. For instance, McLeod, Atkin, and Chaffee (1972a) found that fo r one sample of children the partial correlation of view ing w estern program s w ith aggressive be havior was -.01 (n.s.), w hile fo r adventure drama program s it was .11 (p<.Q5). Since the rated violence fo r westerns was greater than that fo r adventure dramas, it could be argued that in som e way the incidents of violence in the w esterns were unrecognized, filtered out, discounted, or restructured by the child viewers. If w esterns were classified by this sample of children as in som e sense 8 inconsequential entertainm ent, then one m ight not expect m odeling effects to oc cur. McLeod, Atkin, and Chaffee (1982b) took a sim ilar approach, believing that westerns m ight be judged as less realistic than adventure dramas. They found that children did rate the form er as less realistic than the latter. Another pos sibility exists, however, w hich the data reported do not allow testing. Though the overall means of realism ratings fo r w esterns were low er than fo r adventure dramas, it m ight be that this classification system was not w hat the children were using. If such were the case, it m ight be th a t som e w esterns could be classified as m ore realistic than som e adventure dramas, and that using another way of classifying program m ing m ight have resulted in a greater am ount of ex plained variance in aggressiveness fo r those series classified as highly realistic. In either event, a plausible case can be made fo r considering the categories into which children classify television program s when attem pting to predict ef fects of exposure to them . There is, however, a fu rth e r concern, w hich is w hether there may be age differences in the num ber and kind of such categories. Although even very young children can physically sort objects into groups w ith out being verbally able to name the attributes they use to distinguish among the objects (Nelson, 1973; Ricciuti, 1965), it is generally accepted that the ability to consider more than one attribute at a tim e when categorizing objects increases w ith age (Flavell, 1977; Mussen, 1979). This is considered to be so regardless of w hether that ability is though t to be contingent on language developm ent. If one is concerned w ith how hypothesized cognitive categories m ight m oderate the ef fects of exposure to television program m ing on children, it is clear th a t attention 9 ought to be paid to possible age differences in those categories. 1.2 The Function of Television Programming as a Variable The television program m ing categories w hich adults have tra d ition ally em ployed are prim arily "structural." That is, the divisions are based on the presence or absence of attributes upon w hich m ost adults would agree. For ex ample, situation com edies have laugh tracks, are one half hour in length, have regular characters, and have sim ple plots w hich revolve around a problem w hich is resolved at the end, while m ystery/suspense series are usually an hour long, may have m ultiple and/or com plex plots, and often em ploy music, sophisticated editing techniques, and pacing to m aintain excitem ent and suspense. One m ight even argue that these are w hat Locke (Copleston, 1964) labeled "prim ary qualities," in that they would exist even in the absence of human observers. It is equally possible, of course, to think of program m ing in term s of func tions rather than structure. That is, program s could be categorized based on w hat they are used for rather than on w hat th e ir "intrinsic" qualities are. Indeed, many objects can be conceived of as a com bination of "perceptual attributes" (which w ould probably be considered prim ary qualities) and "functional attributes." (There are also relational attributes, such as em ploying the dim ension of fam ilial offspring to define a child, or a m other o r father, but these attributes are m ostly unim portant fo r the current discussion.) In a series of studies Rosch and her colleagues (Rosch & Mervis, 1975; Rosch, Mervis, Gray, Johnson, & Boyes-Braem , 1976) have consistently found that both children and adults em ploy both percep tual and functional attributes when categorizing sim ple objects. 10 Directly analogous to the findings that objects are categorized according to their function, a large body of television research has focused on people's reported uses of, or functions ascribed to, the medium. This "uses and gratifications" approach (Blum ler & Katz, 1974) is based on several assum ptions, the m ost im portant of w hich are 1) individuals are goal directed, 2) they actively choose som ething from th e ir environm ent to satisfy a fe lt need, and 3) they are aware of and able to relate th e ir reasons fo r making th e ir choices (Katz, Blumler, & Gurevitch, 1974). Technically, it can be argued that there are tw o form s of the functionalist approach. In the "strong" form , any given object, in this case som e unit of television program ming, may be used by any given individual fo r any given purpose(s) on any given occasion. If this were the case, it would be necessary to measure reported function of the program m ing in each instance of use, as one could not know in advance w hat th a t function would be. In its "weak" form , pro gram m ing is characteristically associated w ith one or m ore functions by m ost people m ost of the time. Nearly all uses and gratifications research has taken this latter position, though it has predom inately focused on the gratifications people report associating w ith television as a m edium, either in isolation (Frank & Greenberg, 1980; Greenberg, 1974) or in the context of other media (Elliot & Quattlebaum, 1979; Katz, Gurevitch, & Haas, 1973). Rarely has the focus been on individual series or genres. The results of the studies of the television medium have varied considerably, w ith the num ber of gratifications found ranging from one or tw o upwards to about fifteen. In addition to m ethodological reasons fo r 11 the wide range of findings, it may be that the focus on m edium is m isplaced. The concern w ith gratifications associated w ith the m edium has tended to obscure the possible im portance of the content typically contained in it. As Gantz (1980) has noted, there is a range of levels at which one can investigate gratifications, from com parisons among media to com parisons am ong episodes of particular series. Only one study (Bantz, 1982) has been conducted in an atte m pt to deter mine the appropriate level fo r gratifications analysis. While the results of this study indicated that the facto r structure of gratifications associated w ith favorite types of program m ing did not diffe r significantly from that of television as a medium, the study has been criticized on conceptual grounds fo r not providing an adequate test of the im portant question (Kovaric, Dorr, & Nicol, 1983). That is, favorite type o f program m ing may represent only a fraction of overall viewing, thus other types of program m ing may have different facto r structures from television as a whole. There are tw o other studies w hich have examined differences across kinds or types of program m ing. Rubin (1979) found differences in gratifications as sociated w ith trad ition ally defined genres, and Kovaric et al. (1983) found d if ferences in gratifications associated w ith different clusters of individual series. The Rubin study em ployed gratification statem ents derived from Greenberg's (1974) facto r analytic study of gratifications, and Kovaric et al. (1983) em ployed a com bination o f gratifications reported in the literature and those derived from focus group interviews. Both studies could, therefore, be faulted fo r not develop 12 ing gratification statem ents m ore directly from respondents and m ore clearly at the series or genre level. However, it should be noted that both studies did find differences in gratifications related to program m ing types, in one case using a priori specified types and in the other developing them from the data. Thus it may be argued th a t there is som e reason to believe in the weak form of the functional approach applied to individual series or genres: that particular func tions can be characteristically associated w ith particular series or genres. It remains to be seen w hether these associations are consistent across ages. Rubin's (1979) study included children of different ages, but the way in w hich the data are reported does not allow conclusions regarding age d if ferences. Two studies w hich have examined gratifications associated w ith te le v i sion as a medium have looked at age differences. Greenberg (1974) found that fo r the seven gratifications identified, older children (12 and 15 years) consis tently reported using television less than younger children fo r each gratification. Brown (1976) found the same trend, though fo r slig h tly different functions, begin ning w ith 9 -year-o lds, but he also found th a t 7 -ye a r-o ld s reported associating television w ith each function less than 9-ye a r-o ld s. When Greenberg used these gratifications as covariates o f attitudinal outcom es, fo r 9 -ye a r-o ld s learning, relaxation, arousal, and fo rg e ttin g were im portant; fo r 12-y e a r-o ld s those same five gratifications plus habit, passing tim e, and com panionship were im portant; and fo r 15-year-olds all but com panionship and passing tim e were im portant. It may be expected, then, that som e differences in functional categories fo r children of different ages may exist, as may differences in how these categories m oderate 13 effects. 1.3 Identifying Programming Dim ensions The above discussion suggests the need fo r research to describe the dim ensions — both in term s o f program m ing attributes and program m ing func tions — children use in categorizing or distinguishing among television program ming. One aspect of program m ing w hich has som etim es been studied and which can be thought of as a dim ension is how real the program m ing is perceived to be. When asked directly about the reality of program m ing, older children (about eight or older) usually can distinguish between program m ing that is made up and program m ing that depicts som ething as it is happening (Hawkins, 1977; Kelly, 1981; Klapper, 1981). Beyond this "m agic w indow " assessment, it is not entirely clear how judgm ents of the reality o f program m ing change w ith age. Some studies have indicated that children th in k o f television program m ing as less real the older they become (Greenberg Reeves, 1976; Lyle 8t Hoffman, 1972). Others have found no age differences or mixed results. Hawkins (1977) found that as a fictiona l presentation of w hat life is like and how useful it is fo r drawing conclusions about their own lives, preschoolers and sixth graders were m ore skeptical than firs t and third graders. D orr (1983) found a general trend fo r children to em ploy concepts such as probable, representative, and sim ilar to ac tual experience increasingly w ith age. From this literature it is possible to conclude that if children actually do spontaneously em ploy a "reality dim ension" to distinguish am ong television pro gram m ing one m ight expect to find age differences in both how prom inent such 14 a dim ension is in relation to other dim ensions and how program s rate on that dimension. How, though, can one determ ine w hether such a dim ension is generally used by children, w hat other dim ensions are used, and how they relate to one another? There are tw o general approaches one can take in attem pting to answer these questions. The first, and m ost com m only em ployed in analogous situa tions, is to develop a list of possible dim ensions and the attributes w hich m ight describe them , based on intuition, previous research, focus group interview s, and so on, and then ask subjects to rate the program m ing on them . There are tw o studies, both w ith adults, w hich have used this approach to try to identify te le vi sion program m ing dimensions. Frost (1969) employed 58 sem antic d iffe re n tia l-typ e scales to measure audience assessment of 61 series. Nine factors (labeled general evaluation, in fo r m ation, romance, violence, conventionality, scale of production, noise/activity, ac ceptability, and hum or) were identified through fa cto r analysis, and a cluster analysis based on the series loadings on the factors suggested six "types" of program m ing. These were labeled "Family entertainm ent," "Pop," "Adult crim e/violence," "W omen's rom ance," "Intellectual," and "Sports/new s." In another study carried out a few years later, H im m elw eit, Swift, and Jaeger (1980) em ployed a m ultidim ensional scaling approach to identify the dim ensions adults used in categorizing 20 television series. They used 17 at tributes derived from review ing the w ork of professional drama critics and 15 presented their subjects w ith the 20 series names and a subsam ple of the 17 at tributes. The attributes were classified by the authors as being either "im pact" attributes (i.e., inform ative, m oving, exciting, disturbing, trivial, annoying, convinc ing, absorbing) or "stylistic" attributes (i.e., violent, brutal, realistic, funny, lighthearted, glam orous, sexy, unpredictable, complex). Each of the attributes was fu rth e r labeled as either "em otional" or "intellectual." The analysis resulted in a three dim ensional solution, w ith degree of realism, level or kind o f arousal, and level or kind o f involvem ent as the labels o f the dim ensions. It was fu rth e r found that the attributes were related to reported liking fo r the series. It is evident in both of these studies that the m ost serious flaw in id e n tify ing the dim ensions is that the dim ensions found were tied to the researchers' ideas of w hat they would be. W hile the results may accurately represent the ways in w hich respondents categorize program m ing using the concepts (attributes) presented to them , it may not be that they w ould naturally use (at least) some of these concepts if not forced to do so. It is also evident th a t there are som e com m onalities in the tw o studies. One could even make a case that the attributes investigated in the second study, and the factors derived in the first, can be divided into functional and structural attributes. An alternative approach to identifying program m ing dim ensions is to avoid specifying them in advance. This can be done by em ploying m ultidim ensional scaling as did Him m elw eit et al., but, rather than asking respondents to rate series on prespecified attributes, asking them to assess how sim ilar series are to each other. This m ethod assumes that each object (series) is made up of a num 16 ber of attributes, properties, or features. To the extent that these attributes are sim ilar in kind (though not necessarily in degree) across series, each series can be represented in an n-dim ensional space, w ith the n dim ensions representing the n properties shared by the series. The advantage of this approach is, of course, th a t the respondents use th e ir own set of attributes, not the researcher's, to rate program m ing sim ilarity. A lthough relying on people's application of th e ir own subjective criteria fo r judging sim ilarity among objects has a great deal of appeal to researchers be cause of the apparent validity of the results, m ultidim ensional scaling as a means of representing objects has been criticized by Tversky on several grounds (1977; Beals, Krantz, & Tversky, 1968; Krantz & Tversky, 1975; Tversky & Krantz, 1970). Generally he has argued that psychological sim ilarity should be treated as an hierarchical structuring of qualitative features rather than a geom etric represen tation of quantitative distances. Specifically, the m ultidim ensional scaling ap proach has been faulted on three grounds: 1) fo r not considering the possible asym m etry of object sim ilarity (object 'a' may be judged as m ore like object b' than 'b' is like 'a'; 2) for violating the principle of m inim ality (identical objects should be judged as "same" rather than "different," but in practice are som etim es not); and 3) fo r accepting the triangle inequality principle when in practice it may not hold {'a' may be sim ilar to 'b,' and 'b' to 'c,' but 'a' may be quite dissim ilar to 'c'). Responding to Tversky, Krumhansl (1978) and Cooper (1981) have presented argum ents fo r the viability o f the geom etric model. If it is assumed th at the 17 density of any particular region of space in a geom etric m odel may lead to fine distinctions of sim ilarity, then all of Tversky's objections can be accom m odated. Technically, then, sim ilarity or dissim ilarity measures do not have to behave precisely as m etric distances in order to be represented by a geom etric model. It is further argued that the parsim onious and (usually) interpretable dim ensions provided by a Euclidian m odel o ffe r greater heuristic value in m any settings than the alternative approaches. In practice, m ultidim ensional scaling techniques have been successfully u til ized in numerous research areas, including m arketing (Cooper, 1973), public ser vices campaigns (Levy & Kilburn, 1979), attitudes and attitude change (Woelfel, Cody, Gillham, & Holmes, 1980), interpersonal relations (Wish, 1976), and the per ception of television characters (Alexander, 1980; Reeves 8i Greenberg, 1977; Reeves 8i Lometti, 1979). The use of m ultidim ensional scaling as a means of identifying dimensions children use in distinguishing among different kinds of television program m ing is thus ju stifie d on both theoretical and practical bases. Age differences in dimensions Having presented a rationale fo r the im portance o f id entifying the dim en sions children may use in distinguishing among or categorizing television pro gram m ing, it is appropriate to pull together the thread of expected age d if ferences which has been woven thro u g h o u t the argum ent. In general, of course, children tend to think about things in a m ore co m pli cated manner as they grow older. W hether this general tendency applies to 18 television is problem atic. Reeves and his associates (Reeves & Greenberg, 1977; Reeves & Lom etti, 1979) found little difference in the dim ensions third, fifth, and seventh, and sixth and eighth grade children used to distinguish am ong television characters. They speculated that these results m ight indicate th a t children learn very early th a t television characters are stim uli w hich require only sim ple categorization to make them m eaningful. Sim ilarly, it may be that there are only sim ple distinctions made am ong kinds of program m ing in general, th a t these are made at a very early age, and that they are consistent across ages. There are, however, tw o reasons w hy these hypotheses m ight be incorrect. First, the literature in cognitive developm ent suggests that it is not likely to be so. It is generally asserted th a t concrete operational children are not able to decenter easily. That is, they cannot think about problem s involving tw o or more classes, dim ensions or relations at the same tim e (Flavell, 1970). W ork in percep tual developm ent (e.g., Bruner, Olver, & Greenfield, 1966; Gibson, 1969; Shepp, 1978) has indicated that younger children w ill collapse separate perceptual dim ensions into integrated wholes, thus sim plifying discrim ination tasks. It would then be expected th a t as children age they w ould break some categories into conceptually d istin ct parts. For instance, younger children m ight classify pro gram m ing along a funny-seriou s dim ension, w hile older children m ight em ploy a second dim ension contained w ith in th is one: trivia l-im p o rta n t. Second, as children become older they becom e less and less egocentric in th e ir thinking and m ore and more able to separate th e ir own point o f view, in ter ests, awareness, and so on from the rest o f th e ir environm ent, including other 19 people (Flavell, 1977). For young children thinking egocentrically, then, one m ight expect that there w ould be little distinction between som e program m ing's reported function (e.g., "I w atch it because it makes me laugh") and its attributes (e.g., "It is a com edy"). Older children, however, w ould be m ore likely to make such a distinction. Despite the findings of Reeves, then, general principles o f cognitive and perceptual developm ent indicate th a t older children are likely to use m ore and m ore com plex dim ensions in distinguishing among television program m ing and that older children w ill distinguish between attribute and fu n ctio n dimensions m ore clearly than younger children. 1.4 Hypotheses The prim ary th ru st of the current research is descriptive, intending to iden tify those dimensions children use in distinguishing am ong television programs. There are, however, a num ber of hypotheses derived from the above discussion of the ways in w hich television program m ing has been characterized, and the possibility that children of different ages w ill be likely to categorize identical program s differently. They are: H1a: H1b: H2a: H2b: 20 Older children w ill utilize m ore content attrib ute dim ensions than younger children. Older children w ill utilize m ore com plex content attribute dim ensions than younger children. Older children w ill utilize m ore functional dim ensions than younger children. Older children w ill utilize m ore com plex functiona l dim ensions than younger children. H3: Functional dim ensions w ill resem ble content attribute dim en sions less w ith increasing age. H4: For all ages, view ing frequency of sim ilarly categorized series, when based on dim ension ratings, w ill predict social reality beliefs better than view ing frequency of series categorized using tradition al industry labels, w hich in turn w ill predict the same beliefs better than overall view ing frequency. 21 CHAPTER 2 METHODS The problem presented in the previous chapter was to identify w hat content attributes and gratifications children o f different ages associated w ith television program m ing, how much they viewed the program m ing, and how the social reality beliefs they held m ight be related to view ing of program m ing w hich was categorized in various ways. The general approach to this problem , partly sug gested in the previous chapter, involved tw o m ajor com ponents. The firs t was to present children w ith stim uli representative o f television program m ing and ask them to make judgm ents on the sim ila rity o f various units of that program m ing, either based on som ething contained in the program m ing itself, or on the func tion or functions it served fo r the child. M ultidim ensional scaling analyses o f the data derived from this com ponent w ould provide means of describing the "attribute" and "gratification" dim ensions children used to distinguish am ong the various units of program m ing. The second com ponent was to measure children's exposure to the program m ing and som e of th e ir social reality beliefs in order to assess w hether the dim ensional structures derived in the firs t com ponent would help predict these beliefs. The particular m ethods em ployed in the study are described in the fo llo w ing pages. It is appropriate to m ention here, however, one im portant decision w hich delim its all the follow ing. This decision was in response to the issue of how to represent television program m ing in the study. As m entioned in the pre 22 vious chapter, television program m ing can be thought of as a unit in and of itself, as a num ber of distin ct genres, as individual series, or as individual episodes. The choice here was to em ploy a set of regularly aired entertainm ent series which were dem onstrably popular w ith children of a wide age range. The deci sion to em ploy program m ing o f this kind was made fo r tw o reasons. First, everyday experience suggests th a t individual series is the level at w hich m ost people, children and adults, thin k of and talk about program m ing. Second, only by making judgm ents at the series level could dim ensions be found which m ight allow categorizing the series into traditional genres. Employing a lim ited set of series was necessary because the universe o f series is to o large fo r any human to consider in a reasonable am ount o f tim e. Using regularly broadcast popular series was necessary to ensure th a t m ost children w ould be at least reasonably fam iliar w ith m ost o f the series, thus making th e ir data roughly com parable. A significant portion of com m only broadcast program m ing was autom ati cally excluded from consideration as a result of these restrictions. Specials, sports, and m ovies were not considered because of the irregularity w ith w hich they are broadcast, and news, game shows, and inform ational or educational pro gram m ing were excluded because they are not heavily viewed by children. The series selected were all broadcast between 4 p.m. and 10 p.m. on weekdays and between 8 a.m. and 10 p.m. on weekends. Both networks and independent sta tions broadcast the series, som e of w hich were still in production at the tim e of the study and some o f w hich were syndicated. All had been on the air fo r at least tw o seasons. Thus the study was restricted to a particular portion of avail 23 able program m ing and should be construed to represent children's judgm ents only of that portion. 2.1 Pilot Study Before conducting the form al study several prelim inary questions needed answering. These were: 1. W hether the names o f television series alone w ere sufficient stim uli to elicit ideas w hich children could use to discrim inate among series, 2. W hether the sorting task to be used was manageable by the younger children tested individually and by the older children tested in small groups, 3. W hether the instructions fo r the sorting task were sufficien tly clear that children o f all ages w ould understand them and com plete the task w ith out confusion, 4. W hether the series chosen fo r the study were com m only watched by children o f the ages to be studied, 5. W hether the form ats fo r the questions to be used to test the predictive validity of the obtained dim ensions were suitable fo r the ages of the children to be tested, and 6. W hether the questions them selves were m eaningful to the children. To answ er these questions and to provide guidance fo r making revisions where appropriate in the procedures and measures, the author carried o u t a pilot study at a local private elem entary and ju nio r high school and at a local private high school in December 1983, and January 1984. Thirteen second graders, nine sixth graders, and three tenth graders were tested and then questioned about the 24 testing process. The procedures fo llo w e d in the pilot study w ere essentially the same as those to be described below in the Procedures and C onditions sections below. The results of the pilot study as they influenced the final m easures and procedures w ill be discussed in those sections. The one issue w hich w ill be dis cussed here is the firs t one m entioned above: w hether the nam es o f television series alone were sufficient to e licit enough ideas about them th a t one could ex pect m eaningful distinctions to be made am ong them . To address this issue, each child was firs t asked to fill out a checklist indicating how m uch he or she w atched each of 41 television series w hich were being broadcast in Los Angeles in Septem ber of 1983. The series were selected because they had large child audiences of all ages as determ ined by the Nielsen ratings (Nielsen, 1983) fo r the Los Angeles area in August of 1983, and because they represented a variety of categories or genres Nielsen em ploys to classify program m ing. Later each child was asked to describe half of the series, and the descriptions w ere w ritte n down by the author. The child then viewed a videotape edited fo r the pilot study w hich was com prised o f the weekly opening segm ents of each series. The child was then asked if s/he could tell anything else about the series described earlier. Only a few children, and then only fo r a few series, were able to provide more inform ation about any series after having seen the tape. As a result, it was con cluded that providing children w ith m ore than the names of the series was not essential to th e ir com pleting the study tasks m eaningfully. 25 2.2 Participants Three age groups of children were used in the study: seven-, eleven-, and fifte e n -ye a r-o ld s. These ages were chosen, first, because there were likely to be developm ental differences among them (e.g., Flavell, 1977). If there were age dif ferences in the dim ensions children used to distinguish am ong television series they w ould be likely to show up in these age groups. Second, these were ages often studied in other television effects research, so any findings of age differences in the present study could be com pared w ith results of studies of related issues w ith sim ilar age groups. This consideration is directly reflected in the broader research context in w hich the present study was conducted. Concurrent w ith this study, tw o others were under way. One was a fo u r year project to study the ways in which television fam ilies and real life fam ilies influence children's learning of social norm s fo r expression o f em otion and th e ir acquisition of useful ideas about such expression (cf.. Dorr, Kovaric, Doubleday, Sims, & Beizer Seidner, 1985), fo r w hich data were gathered at sites different from those reported here. The other study was designed to identify the dim ensions children em ploy to distinguish am ong television fam ilies and their ow n fam ilies or th e ir conception of an ideal fam ily (Doubleday, 1985). Some data fo r the latter study were gathered at some of the same sites as the present study but using different participants. Both o f these studies were designed to in vestigate age differences w ithin samples o f w hite, m iddle-class, fluen t English speaking children. In order that the results o f the present study m ight help in form each of these other studies, the participants in the present study were 26 chosen so as to be com parable w ith regard to age, ethnicity, SES, and English language fluency. A final reason fo r restricting the sample to participants w ith the charac teristics described above was that some previous research (Gerbner et al., 1979; G erbner et al., 1980) has been criticized (Hirsch, 1980, 1981a, 1981b) fo r not con tro llin g fo r variables such as SES or ethnicity. By studying a fairly hom ogeneous group of children, who also constitute a m ajor portion of the child audience, som e of the obvious confounding influences on the outcom es could be elim inated a priori. Children were recruited through four public and tw o private Los Angeles area elem entary schools and three public and one private Los Angeles area high schools. Inform ed consent was obtained from the parent or guardian o f each child participating in the study and from the child h im - or herself. To screen out children who watched very little television, those w ho indicated they viewed half or less of the series on the Viewing Frequency Checklist (see below) were ex cused from further testing when the testing situation perm itted. The children w ho participated in and supplied usable data fo r the study were 47 second graders, 84 sixth graders, and 78 tenth graders w ho were w hite (as judged by visual inspection and name) and from m iddle class neighborhoods (as judged by the location of the school or com m ents made by school adm inistrators). They were tested between April 1984 and December 1984. The children's mean ages, by grade, sex, and condition, are presented in Table 1. 27 Table 2-1: Mean Ages, Standard Deviations, and Numbers of Participants Grade Condition 2 6 10 Attribute Girls X 7.40 11.17 15.20 SD .52 .39 .62 N 10 23 20 Boys X 7-31 11-15 15.10 SD .48 .59 .46 N 13 20 19 Gratification Girls X 7.50 11.45 15.18 SD .52 .61 .40 N 12 20 22 Boys X 7.08 11.38 15.24 SD .67 .50 .44 N 12 21 17 28 Eighteen children of other ethnicities were tested also, but they are not in cluded in the present report since there were too few to analyze fo r group d if ferences. An additional 28 children w ho provided com plete data w ere excluded from the analyses because they did not report having viewed at least half the series on the Viewing Frequency Checklist in the last year, or because they put half or more of the series in the "don't know" envelopes in the card sort task (see below). The num ber of second graders is low er than those fo r the other tw o grades because the proportion o f parents returning perm ission slips fo r them was low er than fo r the other tw o ages, even though approxim ately the same num ber of second graders as older children was recruited. The decision was made not to pursue recruitm ent beyond December 1984, because the television program m ing schedule had begun to depart from that used to choose the series to which the children responded. 2.3 Experimenters Children were tested by the author and seven others who were naive to the specific hypotheses of the study but aware th a t its general purpose was to determ ine what dim ensions children use in discrim inating am ong television series. There were three fem ale and five male experim enters, o f w hom six were w hite and tw o were black. Their ages ranged from m id -tw e n tie s to m id -fo rtie s, and all but tw o were graduate students. All experim enters received tw o hours of training in the procedures, and m ost had extensive previous experience w orking w ith children. Which experim enters conducted testing varied from school to school; no m ore than four were present at one tim e. Experim enters w ere ran 29 dom ly assigned to conditions and thus to a child or group of children (see below). 2.4 Conditions As noted previously, objects are usually regarded as being com prised of both perceptual and functional attributes (cf., Rosch & Mervis, 1975). Simple judgm ents o f the sim ilarity between series could then be expected to be com prised of both. Since the literature on gratificatio ns associated w ith television is extensive, and since such gratifications are often assum ed to intervene between exposure and effects, it was decided th a t separating a priori the functional aspects o f the series from th e ir perceptual attributes w ould be desirable. Doing so w ould allow clear identification of differences in series' attributes and gratifications where they existed and provide an o p p o rtu n ity to test w hether gratification dim ensions served a predictive, besides a descriptive, purpose. Children were random ly assigned w ith in age and sex to one of tw o con ditions. In the A ttribute Condition, children were asked to sort 40 three by five yellow index cards w ith television series' names on them into piles of series w hich they though t were sim ilar to each other, based on one or m ore content at tributes of the series. (See Appendix A fo r representations of the index cards, and Appendix B fo r the scripts interview ers follow ed in the A ttrib u te Condition.) In the G ratification Condition, children were asked to so rt the same 40 series into piles based on w hat the series made them feel or th in k about when they watched them. (See Appendix C fo r the scripts interview ers fo llo w e d in the G ratification Condition.) In each condition, each pile o f series was then placed in an envelope. 30 2.5 Measures A num ber of different measures were em ployed in the study. All children responded to five of them ; the rem aining m easures were either unique to con dition or, in one case, to age. Measures of view ing frequency were obtained from all children by having them fill out the Viewing Frequency Checklist (VFC). It was com prised of the names of 40 television series, w ith spaces next to each name fo r children to in dicate how often they had viewed each series since school had started in Sep tem ber. There were six possible responses fo r series aired weekly, ranging from "never w atched" to "just a few tim es" to "about once a week." For series aired m ore than once a week there was an additional response category: "a couple of tim es a week (or more)." This form at had been used successfully before w ith children of the m iddle age group (Kovaric et al., 1983). Inspection o f the Pilot Study data indicated that all the series o riginally selected fo r inclusion on the list had been viewed by m ost o f the children at least occasionally except fo r THE WALTONS, so it was dropped. In addition, when results from the Pilot Study card sort task were com pared to the Pilot Study VFC it was found that older children were often able to sort and make judgm ents about series w hich they indicated they did not watch. Discussion w ith them revealed the need fo r another category on the Viewing Frequency Checklist — "have watched, but not during the last year" — w hich was included in the final version. W ithin the checklist the series w ere divided into tw o groups, those w hich were broadcast weekly and those w hich were broadcast m ore than once a week. W ithin groups the series were 31 random ly ordered. Because the Doubleday (1985) study was conducted at some of the same schools as the present study and required fa m ilia rity w ith som e of the same series, tw o versions of the VFC were used fo r efficiency o f data gather ing: one w ith the series being em ployed in this study (see Appendix D), and one w ith all the series being used in either study (see Appendix E). Because of a clerical error, THE WALTONS was included in the sm aller VFC. Measures of A ttribute S im ilarity were obtained from 104 children. The rem aining 105 children provided data on measures of G ratification S im ilarity. A t tribute S im ilarity and G ratification S im ilarity were measures o f childrens' ju dg m ents of w hether any given pair o f series was considered sim ilar or was viewed fo r sim ilar reasons. For each child, each possible pair of series received a score of " 1" if they had been placed in an envelope together, a "0" if they had been placed in different envelopes, and a m issing value code if the child was not fam iliar w ith one or both of the series' pair. Measures of frequency of use of each of 17 attributes and 15 gratifications were obtained from the children in the corresponding cond ition s by content analyzing the envelopes on w hich were w ritte n descriptions o f w hat made the series in the envelope sim ilar (attributes) or w hat the com m on reasons for w atching them were (gratifications). Coders recorded on the envelopes each oc currence of an attribute or gratification. A ttributes or g ra tifica tio n s used more than once, either on the same envelope or on different envelopes, w ere recorded each tim e they were used, although m ultiple occurrences on the same envelope were rare. The coding from each child's envelopes was then transferred to a 32 single score sheet, w hich was a m atrix o f series by attributes or gratifications cells. Three sources were em ployed in the developm ent of the content analysis scheme. Some of the literature cited in Chapter I provided lists of gratifications (Brown, 1976; Kovaric et al., 1983; Rubin, 1978;) and attributes (Frost, 1969; H im - m elw eit et al., 1980). Referring to the list of series and the ages of the children to be used in the study, these attributes and gratificatio ns were reworded as necessary to make them conceptually appropriate fo r children. Finally, three ran dom ly chosen envelopes from each age group and condition w ere inspected to identify any concepts the children actually used w hich had not been previously identified. From these sources a prelim inary content analysis schem e was devised. Using this prelim inary version of the final content analysis scheme, the au th o r and one coder each coded one envelope selected random ly from each age group and each condition. Disagreem ents and coding d ifficu ltie s were discussed until agreem ent was reached on each problem code or the need fo r revising or creating coding categories was determ ined. Then five envelopes selected ran dom ly from each age group and condition were coded. Intercoder agreem ent across the 30 envelopes was .90. The agreem ent value was determ ined by using the num ber of attributes and gratificatio ns coded by the author as a base, count ing the num ber of codes the other coder em ployed w hich were in agreem ent w ith the author, and dividing them by the base. Each coder was then random ly assigned one half of the envelopes in each age group and condition to code. 33 This coding yielded data fo r each attrib ute or gratificatio n on each series fo r each child. The requirem ents of later analyses revealed that tw o attrib u te category codes. Aesthetics and Realism, precluded clear interpretation o f the data because both concepts “good" and "bad" were included in one code and both "real" and "unreal" were included in the other. All envelopes w ith these tw o codes were recoded prior to further analyses to correct fo r this difficulty. (See Appendix F fo r the final content analysis coding scheme.) Since it was assumed th a t w hatever a ttrib ute or g ra tifica tio n dim ensions children used to distinguish am ong television series m ight be part of a general m odel o f the effects process, outcom e measures unique to this study were deemed undesirable. It was decided to utilize a num ber o f m easures others have used in surveys of children, dividing these m easures between those w hich were related to perceptions of violence and those w hich were not. It was expected that all of the outcom es m ight show differences in the influence of the dim en sions. Four measures of perceptions of social reality were adm inistered to all children, and a fifth only to sixth and tenth graders, in a single instrum ent referred to as the Social Reality Q uestionnaire (SRQ). Tw o of these were scales em ployed by Buerkel-Rothfuss, Greenberg, Atkin, and Neuendorf (1982) to assess fam ily support and helping (SUPPORT) and fam ily direction giving and com pliance (COMPLY). These were eleven-point three and fo u r item scales, respectively, w hich asked such questions as "How many parents help th e ir kids a lot?" (SUPPORT; see the second page of Appendix G) and "How many kids do w hat 34 th e ir parents tell them to do?" (COMPLY; see the third page of Appendix G); they were given to all children. These measures seemed quite likely to show dif ferences in dim ensional influence, as they represented beliefs w hich were m ost likely to be influenced by exposure to program m ing featuring fam ilies. The au th o r m odified the response fo rm a t o riginally employed, since it required children to make judgm ents about how many people out of every ten do som ething, a task w hich was deemed likely to confuse second graders. Instead, sim ple per centages of all people were presented, along w ith graphical representations of these percentages. Pilot te stin g indicated that children had no d iffic u lty under standing either the questions or the response form ats. Two other SRQ measures were derived from the Cultural Indicators Project and were used by Hawkins and Pingree (1981) to measure beliefs about violence in society (GERBNERV) and how mean or scary a place the w orld is (MEANWLD). The GERBIMERV scale was com prised of fo u r item s which required choosing be tw een a "tv" and a "n o n -tv " answer, such as "What percent o f all Am erican men w ho have jobs are police officers or detectives?" (see the last page of Appendix G). The MEANWLD scale em ployed three items, such as "If they got a chance, m ost people w ould try to cheat me," to w hich children responded on a five point, agree/disagree scale w hich was accom panied by sm iling and fro w n in g faces to assist the second graders (see the fo u rth page of Appendix G). Pilot testing in dicated that children had no d iffic u lty w ith either the questions them selves or the response form ats. The MEANWLD scale was given to all children, but the GERBNERV scale was not adm inistered to second graders since it required 35 responses to questions about w hich children of that age were unlikely to have thought, and w hich m ight have provoked anxiety in them . The last SRQ measure, given to all children, was derived from one used by Singer et al. (1984), w hich was another indicator of how much violence children judged to be in th e ir w orld (SINGERV). It was com prised of nine item s, such as "Is it safe to w alk around your ow n neighborhood?", each of w hich had a m ore or less violent response (see the fifth and sixth pages o f Appendix G). 2.6 Procedures The procedures fo r all children were identical except w here noted below. Children firs t filled out the View ing Frequency Checklist (VFC). This was usually done in groups o f ten o r more, depending on w hat arrangem ents w ere m ost ac ceptable to the schools involved. If a child had viewed m ore than half of the series on the list since school had started in Septem ber (or, fo r tenth graders, if s/he reported having at least seen m ore than half the series at least som etim e in the past), then s/he continued on to the sorting task. In som e settings, available tim e made it im possible to determ ine w hether a given child m et the view ing fre quency criterion before continuing on in the card sort task. In these cases, the child com pleted all the rem aining procedures, but his or her data w ere not in cluded in the analysis if he/she did not m eet the criteria. For all the procedures w hich follow ed adm inistration o f the VFC, second graders w ere seen individually and sixth and tenth graders w ere seen in groups o f up to five. The experim enter explained the card sort procedure and led the 36 child or children in tw o practice sorts. The practice sorts had been deemed necessary since several children in the p ilo t study had d iffic u lty separating the concept "reasons fo r w atching series" from the concept "sim ila rity am ong series" w hile actually doing the sorting task, so that rather than basing th e ir so rt on the gratifications they associated w ith w atching the series, they based them on the series' attribute sim ilarity. O bservation o f and discussion w ith these children during and after the testing suggested th a t the d iffic u lty was not th a t they could not grasp the distinction between the concepts, but th a t they could not m aintain it th ro ugho ut the task. Thus it was decided to rephrase the in struction s to em phasize "how the program s make you feel, or w hat they make you th in k about" as reasons fo r viewing, to utilize tw o training sorts w hich em phasized the charac te ristics of each condition, and to provide envelopes bearing printed rem inders of w hat concept the sim ilarity judgm ents were to be made about, into w hich the sixth and tenth graders w ould place th e ir groups of series (see Appendix H). Rather than rating the sim ilarity of all possible pairs o f series, a m ore com m on approach in m ultidim ensional scaling studies, the sorting approach was selected because it was considered to be m ore manageable by children given that the large num ber of series to be sorted (40) w ould result in 780 pairw ise com parisons, an insurm ountable task. The sorting procedure has been used before in MDS studies and is recom m ended when dealing w ith around 20 or m ore objects (Kruskal & W ish, 1978). The child (or children) was then presented w ith 40 cards eight at a tim e, each w ith a series' name on it, and to ld to arrange them , depending on condition, 37 into piles of series w hich were sim ilar to each other or o f series w hich made them feel or think about sim ilar things. Children were told there could be as m any or as few series in a pile as they wanted. They were fu rth e r instructed to place any series w ith w hich they w ere unfam iliar in a blue envelope w ith a ques tion m ark on it, called the "don't know " envelope. All children w ere presented the cards in the same order, w hich had been randomized. It was not judged neces sary to randomize the order of presentation across children since there is no evidence th a t order affects the outcom e o f MDS analysis (Jain & Pinson, 1976). Follow ing com pletion o f the sorting task, the experim enter asked the sixth and tenth graders to place each pile of cards in a separate envelope and to w rite on each envelope w hat made the series in the group sim ilar o r w hat kinds o f feelings or thoughts the child had w hile w atching the series in a group. Ex perim enters asked the same questions of second graders, but they w ro te their answers down fo r them. Finally, the experim enter gave the children the Social Reality Q uestionnaire, directing the sixth and tenth graders to fill it out them selves. It was read to the second graders w hile they circled or checked o ff th e ir answers. To com plete all the procedures took about an hour, w hether they were com pleted in a group or individually. 38 CHAPTER 3 RESULTS The results of this study are presented in tw o main parts. The first, and larger, section is devoted to the analysis of children's judgm ents of the sim ilarities among and between series based on th e ir attributes or gratifications. These analyses relate to the firs t three hypotheses and provide in form ation w hich is prim arily descriptive. In the second section the results described in the firs t section are used to test the predictive va lid ity o f the dim ensional solutions, w hich concerns the fourth hypothesis. 3.1 Sim ilarity of Series' A ttributes and Gratifications The data derived from each child from the sorting task consisted of a below diagonal m atrix of zeroes (series not considered sim ilar or not watched for sim ilar reasons), ones (series considered sim ilar or w atched fo r sim ilar reasons), and m issing data (series not sorted). M atrices fo r children in the same grade and condition were then com bined to create one m atrix fo r each grade in each con dition. Since the m ultidim ensional scaling algorithm used is m ore accurate w ith dissim ilarity data than w ith sim ila rity data, the data were transform ed before creating the group m atrices to conform w ith this "preference." A dditionally, since the num ber of children in each grade in each condition varied, and since children excluded varying num bers of series from th e ir sim ilarity judgm ents, the p ropo r tio n of, rather than the sum of, children in a grade and cond ition judging the series to be dissim ilar was used fo r the value fo r each cell in the m atrices. The final value fo r each cell was the num ber o f children w ith in grade and condition 39 w ho thought the pair of series were not sim ilar divided by the num ber of children w ithin grade and condition w ho thought the pair was either sim ilar or dissim ilar. This approach helped m inim ize the differences in the num ber o f children judging each series. Before proceeding w ith the m ultidim ensional scaling analysis, it was neces sary to ensure that the num ber of children fam iliar w ith each series was high enough to provide a fairly stable estim ate of its (dis)sim ilarity to all other series. The num ber of children in each age group and condition sorting each series is presented in Tables 2, 3, and 4. Based on the apparent lack of fa m ilia rity of THE WHITE SHADOW and 9 TO 5 fo r second and sixth graders, it was decided to drop these tw o series from fu rth e r analysis. Additionally, the rather sm all num ber of second graders w h o were fam iliar w ith DYNASTY, REMINGTON STEELE, SIMON AND SIMON, EIGHT IS ENOUGH, and VIDEO ONE, suggested that these series m ight be dropped also. Rather than "throw out data" im m ediately, however, it was decided that both the set o f 38 series (dropping THE WHITE SHADOW and 9 TO 5) and the set o f 33 series (additionally dropping DYNASTY, REMINGTON STEELE, SIMON AND SIMON, EIGHT IS ENOUGH, and VIDEO ONE) w ould be scaled, and the resulting dim ensional solutions compared. Visual inspection of the tw o, three, and fo u r dim ensional solutions fo r each age group and condition suggested th a t the dropping of the five additional series made little difference in the dim ensions fo r second and sixth graders, but did fo r tenth graders. Since the tenth graders as a w hole w ere quite fam iliar w ith these series, retaining them seemed justified. The resultin g sam ple 40 Table 3-2: Number of Second Grade Children in Each Condition Who Sorted Each Series Condition Attribute Gratification Series N % N % PACMAN 23 100 23 96 POPEYE 23 100 24 100 RICHIE RICH 22 96 22 92 SMURFS 23 100 24 100 SPIDERMAN 20 87 21 88 DYNASTY 7 30 7 29 EIGHT IS ENOUGH 11 48 8 33 FAME 17 74 19 79 FANTASY ISLAND 16 70 14 58 LITTLE HOUSE 19 83 22 92 THE LOVE BOAT 19 83 23 96 THE A-TEAM 21 91 22 92 CHIPS 23 100 24 100 THE DUKES OF HAZZARD 23 100 23 96 THE FALL GUY 17 74 17 71 KNIGHT RIDER 23 100 22 92 MAGNUM, P.I. 18 78 15 62 REMINGTON STEELE 8 35 8 33 SIMON AND SIMON 12 52 8 33 BENSON 17 74 12 50 THE BRADY BUNCH 19 83 21 88 DIFF'RENT STROKES 22 96 21 88 THE FACTS OF LIFE 17 74 13 54 FAMILY TIES 15 65 14 58 GILLIGAN'S ISLAND 22 96 23 96 GIMME A BREAK 21 91 18 75 HAPPY DAYS 23 100 24 100 LAVERNE AND SHIRLEY 21 91 23 96 M*A*S*H 20 87 20 83 THE MUNSTERS 21 91 15 62 ONE DAY AT A TIME 17 74 13 54 SILVER SPOONS 22 96 22 92 THE THREE STOOGES 22 96 22 92 THREE’S COMPANY 21 91 19 79 REAL PEOPLE 15 65 15 62 RIPLEY'S 19 83 18 75 THAT'S INCREDIBLE 18 78 20 83 VIDEO ONE 12 52 7 29 41 Table 3-3: Number of Sixth Grade Children in Each Condition Who Sorted Each Series Condition Attribute Gratification Series N % N % PACMAN 41 95 40 98 POPEYE 43 100 38 93 RICHIE RICH 41 95 39 95 SMURFS 43 100 39 95 SPIDERMAN 39 91 36 88 DYNASTY 31 72 27 66 EIGHT IS ENOUGH 37 86 35 85 FAME 33 77 36 88 FANTASY ISLAND 41 95 38 93 LITTLE HOUSE 37 86 37 90 THE LOVE BOAT 41 95 41 100 THE A-TEAM 40 93 39 95 CHIPS 43 100 39 95 THE DUKES OF HAZZARD 40 93 40 98 THE FALL GUY 39 91 33 80 KNIGHT RIDER 43 100 39 95 MAGNUM, P.I. 38 88 36 88 REMINGTON STEELE 35 81 31 , 76 SIMON AND SIMON 33 77 27 66 BENSON 40 93 36 88 THE BRADY BUNCH 39 91 39 95 DIFF'RENT STROKES 43 100 41 100 FACTS OF LIFE 43 100 41 98 FAMILY TIES 39 91 39 95 GILLIGAN'S ISLAND 42 98 40 98 GIMME A BREAK 42 98 40 98 HAPPY DAYS 43 100 41 100 LAVERNE AND SHIRLEY 43 100 38 93 M*A*S*H 34 79 32 78 THE MUNSTERS 36 84 36 88 ONE DAY AT A TIME 40 93 38 93 SILVER SPOONS 43 100 41 100 THE THREE STOOGES 41 95 37 90 THREE'S COMPANY 43 100 41 100 REAL PEOPLE 37 86 36 88 RIPLEY'S 38 88 34 83 THAT'S INCREDIBLE 41 95 38 93 VIDEO ONE 36 84 32 78 Table 3-4: Number of Tenth Grade Children in Each Condition Who Sorted Each Series Condition Attribute Gratification Series N % N % PACMAN 39 100 27 69 POPEYE 38 97 33 85 RICHIE RICH 38 97 29 74 SMURFS 38 97 35 90 SPIDERMAN 38 97 29 74 DYNASTY 33 85 32 82 EIGHT IS ENOUGH 39 100 38 97 FAME 36 95 34 87 FANTASY ISLAND 38 97 34 87 LITTLE HOUSE 38 97 38 97 THE LOVE BOAT 39 100 38 97 THE A-TEAM 35 90 30 77 CHIPS 39 100 37 95 THE DUKES OF HAZZARD 37 95 36 92 THE FALL GUY 38 97 32 82 KNIGHT RIDER 38 97 36 92 MAGNUM, P.I. 38 97 37 95 REMINGTON STEELE 36 92 30 77 SIMON AND SIMON 35 90 27 69 BENSON 33 85 38 97 THE BRADY BUNCH 39 100 38 97 DIFF'RENT STROKES 39 100 39 100 THE FACTS OF LIFE 39 100 39 100 FAMILY TIES 38 97 33 85 GILLIGAN'S ISLAND 38 97 38 97 GIMME A BREAK 39 100 38 97 HAPPY DAYS 39 100 38 97 LAVERNE AND SHIRLEY 39 100 38 97 M*A*S»H 37 95 33 85 THE MUNSTERS 39 100 34 87 ONE DAY AT A TIME 39 100 38 97 SILVER SPOONS 38 97 38 97 THE THREE STOOGES 36 92 36 92 THREE'S COMPANY 39 100 39 100 REAL PEOPLE 39 100 37 95 RIPLEY'S 38 97 34 87 THAT'S INCREDIBLE 39 100 35 90 VIDEO ONE 37 95 31 79 43 of 38 series, listed in Table 5 by their Nielsen categories, were analyzed as fol lows. The subgroup m atrices were subjected to analysis by ALSCAL (Takane, Young 8 1 deLeeuw, 1977) provided by SAS (Reinhardt, 1980), tre a tin g the data as ordinal. A lthough ALSCAL allow s fo r w hat is called th re e -w a y solutions (in this case, series x series x age-grou p or condition), each subgroup solution w ould then be based on the overall group solution and thus w ould be less likely to ex hibit differences than w ould solutions calculated separately. That is, using the th re e -w a y solutions option, the dim ensions fo r each subgroup w ould be the same, w ith only th e ir overall im portance and/or the ratings of the series on them varying between groups. Given th is treatm ent, one could not find m ore dim en sions fo r one group than another, except in the sense that fo r one group all the ratings on a dim ension m ight be near zero. Therefore, it was decided to analyze each subgroup separately, in order not to obscure any dim ensional differences between age groups and conditions. In analyses like these the m axim um num ber of dim ensions w hich can theoretically be extracted from any set of data is N-1, or in the present case, 37 (the num ber of series [38]— 1). In practice, however, large num bers of dim ensions are rarely needed to explain m ost data adequately, and the SAS version o f ALS CAL lim its the user to a m axim um of six. To provide guidance fo r choosing the appropriate num ber of dim ensions, ALSCAL com putes tw o indices o f g o o dness- o f-fit, STRESS and RSQ. STRESS is Kruskal's (1964) stress form ula 1 and is the square root o f the proportion o f the to ta l sums of squares o f the o p tim a lly scaled 44 Table 3-5: Television Series in Final Sample by Nielsen Category Children's Programming: PACMAN POPEYE RICHIE RICH SMURFS SPIDERMAN General Drama: DYNASTY EIGHT IS ENOUGH FAME FANTASY ISLAND LITTLE HOUSE ON THE PRAIRIE THE LOVE BOAT Mystery/Suspense: THE A-TEAM CHIPS THE DUKES OF HAZZARD THE FALL GUY KNIGHT RIDER MAGNUM, P.I. REMINGTON STEELE SIMON AND SIMON Situation Comedy: BENSON THE BRADY BUNCH DIFF'RENT STROKES THE FACTS OF LIFE FAMILY TIES GILLIGAN'S ISLAND GIMME A BREAK HAPPY DAYS LAVERNE AND SHIRLEY M*A*S*H THE MUNSTERS ONE DAY AT A TIME SILVER SPOONS THE THREE STOOGES THREE'S COMPANY Variety: REAL PEOPLE RIPLEY'S BELIEVE IT OR NOT THAT'S INCREDIBLE VIDEO ONE 45 data w hich is NOT accounted fo r by the model. RSQ is the propo rtion o f tota l variance o f the optim ally scaled data w hich IS accounted fo r by the m odel. Since RSQ has a m ore straightforw ard interpretation, only it w ill be referred to in the text; how ever, inspection of STRESS results provided essentially identical in fo r m ation. Figures 1 and 2 present the RSQ and STRESS values fo r the age groups in each condition. As can be readily seen, six dim ensions accounted fo r nearly all the variance fo r every age group. The am ount of variance explained by the ad dition o f each new dim ension varied considerably, though. Looking at Figures 1 and 2 it was possible to see "elbow s" in the RSQ and STRESS values, such that the inclusion o f an additional dim ension added little to the variance explained. W here these "elbow s" occur is usually though t to be the appropriate num ber of dim ensions to consider initially as an adequate solution. If upon inspection of solutions utilizing a larger num ber o f dim ensions conceptual sense can be made o f them , it is usually considered acceptable to include them . This was the ap proach taken by Reeves and Greenberg (1977), and it is the approach fo llow e d here. In Figure 1 it can be seen th a t there were rather sharp "elbow s" in the RSQ and STRESS curves fo r the dim ensional solutions of the attrib u te data fo r each grade level. For the second graders, tw o dim ensions accounted fo r about 83% of the variance, w ith each additional dim ension adding only about 4% or less to the overall RSQ value. The sixth grade tw o dim ensional solution accounted fo r about 91% of the variance, w ith additional dim ensions, like those fo r the second 46 0 J ----------- H — --------------------- , -------------H 1 2 3 4 5 6 N u r n b e r of D Irriensions LEGEND RSQ — 2nd Grade STRESS - 2nd Grade RSQ - 6th Grade _ __ _ STRESS - 6th Grade RSQ — 10th Grade STRESS - 10th Grade Figure 3-1: R S Q an d STRESS for Attribute Condition 1 .o - .9- o U~> .8 + ct; .7-- .6 - .5- .4- • U~ > cn LjJ O C - h — .2- L D .1 - - H ------------------------ 1 ------------------ 1 ---------- 1 2 3 4 5 N u m b e r o f D im e n s io n s 6 00 LEGEND RSQ — 2nd Grade STRESS — 2nd Grade RSQ — 6th Grade STRESS — 6th Grade RSQ — 10th Grade STRESS - 10th Grade Figure 3-2: R S Q and STRESS for Gratification Condition graders, each adding only about 4% or less to the variance explained. For the tenth graders, a three dim ensional solution accounted fo r about 94% o f the variance, w ith additional dim ensions each adding only about 3% or less to the overall RSQ value. The overall am ounts of variance explained (as represented by RSQ) were fairly high, especially fo r the sixth and tenth graders, indicating that the data could be represented quite accurately w ith a very sm all num ber of dimensions. The RSQ and STRESS curves fo r the dim ensional solutions of the gra tifica tio n data presented in Figure 2 were slig h tly m ore d iffic u lt to interpret. Both the second and sixth graders' data could be argued to have tw o "elbow s," at tw o and fo u r dim ensions. Because the overall RSQ value fo r the tw o dim ensional solution fo r the second graders only accounted fo r about 66% o f the variance and in creasing the num ber of dim ensions in the solution to fo u r increased the am ount of variance explained to about 81% — roughly the am ount o f variance explained in the attribute condition data — it was decided to use the fo u r dim ensional solution fo r the second graders. Sim ilar logic guided the selection o f the four dim ensional solution fo r the sixth graders' data. Tw o dim ensions explained about 73% of the variance, but fo u r explained about 91%, again equivalent to the am ount explained by the tw o dim ensional attrib ute solution. The tenth graders' data resulted in a much clearer bend in the RSQ and STRESS curves w ith a three dim ensional solution accounting fo r about 89% of the variance, and each ad ditional dim ension accounting fo r an increase o f only 4% or less. Again, these solutions resulted in fairly high overall am ounts o f variance explained, indicating 49 that the data could be quite accurately represented by a ju st a few dim ensions. To help interpret these dim ensions the w ritte n descriptions o f w h at made the series sim ilar, or o f w hat though ts and feelings were associated w ith the series, were content analyzed as described in the M ethods chapter above. The content analysis yielded the num ber o f tim es each child em ployed each a ttribute or gratification category fo r each series. Since the focus of the study was on how children as a group distinguished am ong series, these data w ere used to calculate the mean num ber o f tim es each a ttrib ute or g ra tifica tio n was used fo r each series w ith in each age group. All subsequent analyses using the content analysis data em ployed these means as values fo r the attributes or g ra tifica tio ns and em ployed the series as cases. The approach to in terpretation of the dim ensions involved id entifying w hich attributes or gratifications the children in each age group used a lo t and used in a consistent manner. The overall mean use of each attrib u te and gratification and the correlations of each attrib ute and g ra tifica tio n w ith the dim ensions chosen were used to select w hich attributes and gratifica tio n s should be retained fo r interpretation. The means fo r the attrib ute s are reported in Tables 6 -8 ; the means fo r the g ratificatio ns are reported in Tables 9-11. The corre latio ns of the attributes and gratificatio ns w ith the dim ensions selected are presented in Tables 12-17. It was decided that to retain an attrib ute or gratificatio n, it should have a mean of .05 or greater and a high correlation (p<.01) w ith one o r m ore dim en sions. A dditionally, the last three attributes and g ra tifica tio n s fro m Tables 6-11 were dropped fro m consideration because they were not relevant to the con 50 ditions. Finally, it was judged th a t the series' attrib ute s or gratifications w ith higher means should be considered m ore im portant or salient fo r the dim ension w ith w hich they were associated. For those w ith identical or nearly identical means, the variable w ith the stronger correlation was considered m ore im portant to the dim ension. Tables 18 and 19 indicate w hich attrib ute s and gratifications, in adjudged order o f im portance, com prised the dim ensions suggested by the RSQ criterion. These tables, along w ith Figures 3-12, w hich graphically sum marize the findings of the MDS analysis, were used to label or name the dim en sions. The labels used in the results th a t fo llo w do not always represent the conceptual com plexity of the dim ensions, since to do so w ould have som etim es resulted in rather unw ieldy phrases. It should be kept in m ind th a t the dim ension labels are used as a kind of shorthand fo r the concepts discussed in the text. 3.2 Age Differences in A ttribute Dimensions C ontent analysis results Inspection of the lists of attributes associated w ith the provisionally selected set of dim ensions presented in Table 18 and of the locations of the series on the dim ensions presented in Figures 3 -6 suggested that there were som e conceptual and structural sim ilarities am ong the dim ensions fo r the three grade levels. The firs t dim ension fo r the second and tenth graders tended to separate w hat adults th in k of as com edies from the crim e dramas, as did the second dim ension fo r the sixth graders. For all ages, the a ttrib ute s of A ction and Crim e tended to be related to the series at the negative ends o f the dim ensions, and Hum or and Family tended to be related to the series at the p o sitive ends of 51 Table 3-6: Means, Standard Deviations, and Ranges of Content Analysis Attributes for Grade 2 Attribute X SD Range Humor • 34 .22 0 -.77 Action .25 .22 .04 -.71 Crime .12 .16 0 -.63 Unreal .12 .16 0 -.60 Good .10 .07 0 -.25 Character Attributes .08 .05 0 -.18 Family .06 .08 0 -.26 Music .05 .12 0 -.58 Novelty .04 .09 0 -.33 Problem Solving • 03 .05 0 -.19 Age Appropriate .02 .03 0 -.10 Occupation .01 .03 0 -.12 Era .01 • 03 0 -.11 Importance .00 .02 0 -.11 Bad .00 .01 0 -.06 Real .00 .01 0 -.06 Gratification .06 .05 0 -.17 Non-Content Attribute .10 .08 0 -.33 Other .11 .10 0 -.37 NOTE: N = 38 Series 52 Table 3-7: Means, Standard Deviations, and Ranges of Content Analysis Attributes for Grade 6 Attribute X SD Range Humor .30 .24 0 -.67 Action .17 .22 0 -.68 Crime .10 .20 0 -.61 Unreal .20 • 30 0 -.98 Good .02 .02 0 -.06 Character Attributes .04 .02 0 -.09 Family .16 .18 0 -.59 Music .03 .13 0 -.75 Novelty .03 .07 o -.32 Problem Solving .09 .08 0 -.24 Age Appropriate .02 .03 0 -.10 Occupation .02 -03 0 -.16 Era .03 .04 0 -.15 Importance .04 .06 0 -.24 Bad .02 .02 0 -.08 Real .04 .04 0 -.10 Gratification .07 .05 0 -.14 Non-Content Attribute .07 .05 0 -.15 Other .10 .12 0 -.43 NOTE: N = 38 Series 53 Table 3-8: Means, Standard Deviations, and Ranges of Content Analysis Attributes for Grade 10 Attribute X SD Range Humor 00 eg • .20 0 -.69 Action .11 .16 0 -.49 Crime .10 .20 0 -.58 Unreal .13 .28 0 -.84 Good .02 .02 0 -.05 Character Attributes .05 .04 0 -.14 Family .13 .16 0 -.49 Music .02 .10 0 -.59 Novelty .08 .11 0 -.49 Problem Solving .12 .09 0 -.28 Age Appropriate .06 .09 0 -.29 Occupation .00 .01 0 -.05 Era .05 .05 0 -.18 Importance .02 .04 0 -.13 Bad .07 .04 .03 -.21 Real .02 .05 0 -.18 Gratification .10 .05 0 -.22 Non-Content Attribute .26 .11 .08 -.44 Other .11 .13 0 -.72 NOTE: N = 38 Series 54 Table 3-9: Means, Standard Deviations, and Ranges of Content Analysis Gratifications for Grade 2 Gratification X SD Range General Pleasure .44 .13 .17 -.71 Amusement ♦ 33 .19 0 -.68 General Displeasure .08 .10 0 -.50 Learning .09 .09 0 -.36 Involvement .12 .09 0 -.57 Excitement .18 .13 0 -.62 Passing Time .03 .04 0 -.17 Social Guidance .04 .04 0 -.17 Cur ios i ty/Novelty .04 .05 0 -.21 Relaxation .00 .00 0 Aesthetics .03 .04 0 -.14 Companionship .00 .02 0 -.08 Social Interaction .00 .01 0 -.07 Escape .00 .00 0 Other Gratification .00 .00 0 Content Attribute .05 .04 0 -.14 Other .00 .02 0 -.09 NOTE: N = 38 Series Table 3-10: Means, Standard Deviations, and Ranges of Content Analysis Gratifications for Grade 6 Gratification X SD Range General Pleasure • 37 .16 •13 -.63 Amusement .22 .15 0 -.51 General Displeasure .18 .11 0 -.44 Learning .15 .10 0 -.40 Involvement .15 .06 .03 -.28 Excitement .14 .12 .02 -.45 Passing Time .10 .06 .02 -.22 Social Guidance .09 .06 0 -.26 Curiosity/Novelty .05 .04 0 -.17 Relaxation .01 .02 0 -.06 Aesthetics .01 .01 0 -.04 Companionship .01 .01 0 -.04 Social Interaction .00 .01 0 -.04 Escape .00 .00 0 Other Gratification .01 .01 0 -.04 Content Attribute .09 .04 .03 -.21 Other .01 .02 0 -.05 NOTE: N = 38 Series 56 Table 3-11: Means, Standard Deviations, and Ranges of Content Analysis Gratifications for Grade 10 Gratification X SD Range General Pleasure • 35 .12 .06 -.54 Amusement .24 .15 .03 -.54 General Displeasure -31 .13 0 -.58 Learning .14 .10 0 -.41 Involvement .14 .06 0 -.26 Excitement .10 .09 0 -.32 Passing Time .06 .04 0 -.15 Social Guidance .08 .06 0 -.19 Curiosity/Novelty .04 .04 0 -.12 Relaxation .01 .02 0 -.05 Aesthetics .06 .04 0 -.13 Companionship .00 .00 0 -.03 Social Interaction .00 .01 0 -.03 Escape .02 .02 0 -.05 Other Gratification .01 .01 0 -.04 Content Attribute .17 .06 .08 -.31 Other .02 .04 0 -.15 NOTE: N = 38 Series 57 Table 3-12: Correlations between Content Analysis Attributes and Attribute Condition Dimensions for Grade 2 Dimension Attribute 1 2 Humor .81* -.19 Action -.69* • 33 Crime -.81* .08 Unreal .04 -.24 Good • 23 -.10 Character Attributes .01 -.45* Family .67* .17 Music • 03 .35 Novelty .21 .24 Problem Solving -.05 .18 Age Appropriate -.12 -.80* Occupation .04 -.10 Era .11 .08 Importance .06 .39 Bad -.05 .06 Real .01 -.29 Gratification .05 .58* Non-Content Attribute .66* -.38 Other .47* • 37 NOTE: N = 38 Series * p < .01 58 Table 3-13: Correlations between Content Analysis Attributes and Attribute Condition Dimensions for Grade 6 Dimension Attribute 1 2 Humor .00 .85* Action .58* -.60* Crime .61* -.55* Unreal -.81* -.06 Good • 31 -.24 Character Attributes .08 -.04 Family .18 .78* Music .04 -.20 Novelty -.15 -.41 Problem Solving .36 .72* Age Appropriate -.73* -.11 Occupation .25 .25 Era -.07 • 35 Importance -.20 -.33 Bad -.07 .10 Real .06 .44* Gratification .18 .34 Non-Content Attribute -.50* .65* Other .07 -.31 NOTE: N = 38 Series * p < .01 59 Table 3-14: Correlations between Content Analysis Attributes and Attribute Condition Dimensions for Grade 10 Dimension Attribute 1 2 3 Humor .84* -03 -.27 Action -.71* .43* -.16 Crime -.69* .44* -.17 Unreal -.26 -.89* -.31 Good -.04 .39 .11 Character Attributes -.42* .26 -.46* Family .72* .19 -.17 Music .02 .05 • 33 Novelty .04 .16 .87* Problem Solving .61* .53* -.27 Age Appropriate -.17 -.90* -.31 Occupation .15 .04 -.04 Era .47* .02 -.28 Importance • 03 .06 .87* Bad .08 -.04 -.47* Real .00 .05 .80* Gratification .84* -.02 .18 Non-Content Attribute .47* .63* -.30 Other -.08 .34 • 35 NOTE: N = 38 Series * p < .01 60 Table 3-15: Correlations between Content Analysis Gratifications and Gratification Condition Dimensions for Grade 2 Dimension Gratification 1 2 3 4 General Pleasure .22 .08 .48* .35 Amusement .69* .52* .26 -.02 General Displeasure .14 -.33 -.28 -.55* Learning -.22 -.26 -.38 .51* Involvement .05 .04 .12 .18 Excitement -.80* -.14 .21 -.01 Passing Time -.01 -.24 -.42* -.25 Social Guidance .06 -.59* -.11 .01 Curiosity/Novelty .12 ■ 03 -.13 .33 Relaxation .00 .00 .00 .00 Aesthetics -.46* -.03 .02 .10 Companionship .24 .05 .06 -.06 Social Interaction -.26 .11 .00 -.04 Escape .00 o o • .00 .00 Other Gratification .00 .00 .00 .00 Content Attribute -.34 .64* -.11 .20 Other .15 .14 -.07 -.31 NOTE: N = 38 Series * p < .01 61 Table 3-16: Correlations between Content Analysis Gratifications and Gratification Condition Dimensions for Grade 6 Dimension Gratification 1 2 3 4 General Pleasure .91* -.01 .25 -.01 Amusement .78* -.46* .30 -.08 General Displeasure -.65* -.45* -30 • 30 Learning .00 .66* -.67* -.03 Involvement -.32 .60* -.08 • 38 Excitement -.74* .41 .46* -.16 Passing Time -.19 -.54* -.49* -.06 Social Guidance .56* .10 -.03 .56* Curiosity/Novelty -.24 .00 -.49* -.48* Relaxation .48* -.20 -.09 -.22 Aesthetics .19 .14 .11 .01 Companionship .39 .18 .25 • 29 Social Interaction -.12 .14 -.65* -.02 Escape .00 .00 .00 .00 Other Gratification -.66* .41 .46* -.02 Content Attribute -.04 -.15 -.52* .04 Other .11 .18 .07 • 37 NOTE: N = 38 Series * p < .01 62 Table 3-17: Correlations between Content Analysis Gratifications and Gratification Condition Dimensions for Grade 10 Dimension Gratification 1 2 3 General Pleasure .69* \14 .17 Amusement .72* -.08 .27 General Displeasure -.19 -.54* -.53* Learning -.25 .51* .42* Involvement -.03 .74* -.05 Excitement -.70* .49* .00 Passing Time -.05 . 18 .12 Social Guidance .66* .57* -.15 Curiosity/Novelty -.45* .17 .51* Relaxation .47* .08 -.15 Aesthetics .56* .22 -.07 Companionship -.26 .12 -.10 Social Interaction • 33 • 32 -.40 Escape .25 .38 • 37 Other Gratification -.06 -.87* -.07 ‘ Content Attribute -.60* -.52* -.30 Other -.17 -.88* .01 NOTE: N = 38 Series * p < .01 63 Table 3-18: Attribute Condition Dimensions Grade Dimension 10 Humor + Action - Crime Family + Unreal Action Crime Character - Humor Attributes Family + + + + Action - Crime Problem + Solving Humor + Family + Action - Problem + Solving Unreal Problem + Solving Action + Crime + Age Appropriate Novelty + Bad NOTE: Attributes are listed in decreasing order of means. Signs following attributes indicate end of dimension attribute is associated with. 64 Table 3-19: Gratification Condition Dimensions Grade Dimension 10 Amusement Excitement General + Pleasure Amusement + General Displeasure Excitement- Social + Guidance General + Pleasure Amusement + Excitement - Social + Guidance Amusement Amusement - General Displeasure Learning + Involvement+ Excitement + Passing Time General Displeasure Involvement+ Learning + Excitement + Social + Guidance General h Pleasure Learning Excitement + Passing Time General Displeasure Learning + Learning + Social + General - Guidance Displeasure MOTE: Gratifications are listed in decreasing order of means. Signs following gratifications represent end of dimension gratification is associated with. 65 Figure 3-3: Second G raders' Series A ttributes in Tw o Dim ensions: Dim ension 1 by Dimension 2 2 . 5 ♦ 1 1 1 1 2 . 0 * 1 CLUSTER 1 1 1 V I ? ^ 1 1 . 5 + 1 / @ 1 1 1 / ) 1 1 . 0 + 0 1 1 1 1 • 0 0 . 5 + t 1 2 £ r m i >KR % i P 1 E E A N j ZC S F I | T 0 | H N ! B G 1 L 2 - 0 . 5 * 0 1 1 l X I i - 1 .0 ♦ 1 3 1 1 1 M “ 1 . 5 ♦ 1 CLUSTER 5 t 1 1 s 8 5 I 1 - 2 . 0 + 1 ( 7 / 1 1 I ( ^ 9 1 - 2 . 5 ♦ -2.5 -2.0 - 1 . 5 - 1 . 0 “ 0 . 5 0 . 0 0 5 1.0 1.5 2.0 2.5 D IM E N S IO N 1 L e g e n d 6 s PACMAN B = BENSON 7 = POPEVE 2 = THE 6RADV BUNCH 8 = R I C H I E R IC H 0 = D I F F ' RENT STROKES 5 * SMURFS i a THE FACTS OF L I F E 9 * SPIDERM AN F a F A M IL Y T IE S X a GX L L IG A N * S IS L A N D $ = DVNASTV G a GIMME A BREAK E » E IG H T I S ENOUGH H a HAPPY DAYS * 2 FAME L a LAVERNE AND S H IR L E Y 3 = FANTASY IS L A N D ♦ a M * A * S * H & = L I T T L E HOUSE M a THE MUNSTERS 0 s THE LOVE BOAT 1 a ONE DAY AT A T IM E s a S IL V E R SPOONS A a s THE A -T E A M 3 = THE THREE STOOGES C a C H IP S T = T H R E E 'S COMPANY z = t h e d u k e s o f h a z z a r d > s t h e f a l l g u y P * REAL PEOPLE K = K N IG H T R ID E R ? a R I P L E Y 'S % s MAGNUM. P .I. I = T H A T 'S IN C R E D IB L E R = REMINGTON STEE LE V = V ID E O ONE & = SIMON AND SIMON 66 in z m 3 Figure 3-4: Sixth Graders' Series Attributes in Two Dimensions: Dimension 1 by Dimension 2 F SG[ D IM E N S IO N J Legend 6 = PACMAN 7 = POPEYE 8 - R I C H I E R IC H 5 = SMURFS 9 - SPIDERM AN $ * DYNASTY E = E IG H T I S ENOUGH * = FAME ) = FANTASY IS L A N D © ~ L I T T L E HOUSE 0 * THE LOVE BOAT A = THE A -T E A M C * C H IP S 2 a THE DUKES OF HA2ZARD > a THE f a l l g u v K = K N IG H T R ID E R * a MAGNUM. P . I - R = REMINGTON STEE LE 6 a SIMON AND SIMON 6 a BENSON 2 = THE BRADY BUNCH D * D IF F E R E N T STROKES I a THE FACTS OF L I F E F a F A M IL Y T I E S X a G I L L X G A N ' S IS L A N D G a GIMME A BREAK H a HAPPY DAYS L * LAVERNE AND S H IR L E Y + = m* a * s * h M = THE MUNSTERS 1 a ONE DAY AT A T IM E S * S IL V E R SPOONS 3 a THE THREE STOOGES T * T H R E E 'S COMPANY P a REAL PEOPLE ? a R I P i . £ Y 'S I = T H A T 'S IN C R E D IB L E V = V ID E O ONE 67 Figure 3-5: Tenth Graders' Series Attributes in Three Dim ensions: Dimension 1 by Dimension 2 2 . 5 * 1 1 1 1 2 . 0 + 1 1 i 1 1 . 5 + 1 GENERAL DRAMA 1 1 s. 1 r - 1 A 1 . 0 + 1 I % > i__ C / --------------- —^ 1 j K } © I R ~~~--------- J 0 D 0 . 5 + ___ * E / F I J ^ [ 8 M [ z + G H 1 E 1 n i V S T S 0 . 0 ♦ ------------- * I 1 P I ? o I N 1 2 - 0 . 5 ♦ t i i 2 X 1 i - 1 . 0 + 1 1 3 M 1 1 - 1 . 5 + 1 1 1 1 1 9 - 2 . 0 + 1 7 I 6 I 1 5 8 i - 2 . 5 •* -2.5 -2.0 -1.5 -1.0 -0.5 0 0 0.5 1.0 1.5 2.0 2.5 D IM E N S IO N 1 Legend 6 = PACMAN 8 * BENSON 7 * POPEYE 2 s THE BRAOV BUNCH 8 * R I C H I E R IC H D = D I F F ' RENT STROKES 5 » SMURFS I = THE FACTS OF L I F E 9 * SPIDERMAN F » F A M IL Y T I E S X * G I L L l G A N 'S IS L A N D $ = DYNASTY G * GIMME A BREAK E * E IG H T IS ENOUGH H = HAPPY OAYS • - FAME L « LAVERNE AND S H IR L E Y } * FANTASY IS L A N D ♦ = M * A * S * H @ a L I T T L E MOUSE M * THE MUNSTERS 0 * THE LOVE BOAT 1 * ONE DAY AT A T IM E s » S IL V E R SPOONS A = THE A -T E A M 3 * THE THREE STOOGES c = C H IP S T a T H R E E 'S COMPANY 2 = THE DUKES OF HAZZARD > s THE F A L L GUV P a R E A L PEOPLE K = K N IG H T RID E R ? = R I P L E Y ' S r. r MAGNUM. P . I . I * T H A T 'S IN C R E D IB L E R = REMINGTON STEE LE V s V ID E O ONE & = SIMON AND SIMON 68 Figure 3-6: Tenth Graders' Series Attributes in Three Dimensions: Dimension 1 by Dimension 3 GENERAL DRAMA 0 . 5 0 - 2 .0 - 2 . 5 - 2 . 5 - 2 . 0 - 1 . 5 - 1 . 0 - 0 . 5 0 . 0 0 . 5 1 . 0 1 . 5 2 . 0 2 - 5 D IM E N S IO N 1 L e g e n d 6 a PACMAN B s BENSON 7 s POPEYE 2 a THE BRADY BUNCH 8 s R I C H I E R IC H D s D IF F E R E N T STROKES 5 S SMURFS £ = t h e f a c t s o f l i f e 9 - s p i d e r m a n F = F A M IL Y T I E S X s G I L L l G A N ' S IS L A N D $ - DYNASTY G S GIMME A BREAK E = E IG H T I S ENOUGH h = HAPPY DAYS * - FAME L s LAVERNE AND S H IR L E Y } s FANTASY IS L A N D = M » A * S * H ® a s L I T T L E HOUSE M = THE MUNSTERS 0 - THE LOVE BOAT 1 = ONE DAY AT A T IM E S = S IL V E R SPOONS A - THE A -T E A M 3 s THE THREE STOOGES c = CHIPS T = T H R E E 'S COMPANY Z S THE DUKES OF HAZZARO > - THE F A L L GUV P * REAL PEOPLE K - K N IG H T r i d e r ? = R I P L E Y ' S % - MAGNUM. P . I . I s T H A T 'S IN C R E D IB L E R - REMINGTON STEE LE V = V ID E O ONE & = SIMON AND SIMON 69 Figure 3-7: S e c o n d G r a d e r s ' S e r i e s G r a t i f i c a t i o n s i n F o u r D i m e n s i o n s : D i m e n s i o n 1 b y D i m e n s i o n 2 2 . 5 + 1 1 I i j 1 1 1 I 8 I 2 . 0 ♦ ! I I 1 1 I 6 I X 1 5 i 1 . 5 ♦ i 3 1 7 j I M 1 1 | I 9 I j 1 l . 0 « ■ I 1 i i i 1 H I I [ G 1 D 0 . 5 ♦ i 1 L I Z 1 M | > 1 2 E | K A t N 1 C v i S 0 . 0 — - — i i 0 j i D [ N j i 0 1 TE 2 - 0 . 5 * i i % i i i i i © i - 1 . 0 . 1 i i j 1 F 1 1 & z ! 1 \ • 1 r i - 1 . 5 + I 1 1 ? 1 i J a 1 1 - 2 . 0 + 1 ! 1 i 1 1 1 1 1 - 2 . 5 . I 1 i s - 2 5 - 2 . 0 - 1 . 5 - 1 . 0 - 0 . 5 0 . 0 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 D IM E N S IO N L e g e n d 6 = PACMAN B = BENSON 7 = POPEYE 2 * THE BRADY BUNCH 8 * R I C H I E R IC H D = D I F F ' RENT STROKES 5 * SMURFS t * THE FACTS OF L I F E 9 - s p i d e r m a n F * F A M IL Y T IE S X * G I L L l G A N 'S IS L A N D $ = DYNASTY G - GIMME A BREAK E = E IG H T I S ENOUGH H = HAPPY DAYS * s FAME L = LAVERNE AND S H IR L E Y ) * FANTASY IS L A N D ♦ = M * A » 5 * H © = l i t t l e HOUSE M * THE MUNSTERS 0 - THE LOVE BOAT 1 * ONE DAY AT A T IM E S * S IL V E R SPOONS A = THE A -T E A M 3 * THE THREE STOOGES C = C H IP S T = T H R E E 'S COMPANY z = THE DUKES OF HAZZARD > s s THE F A L L GUY P * RE A L PEOPLE K - K N IG H T R ID E R ? = R I P L E Y 'S % = MAGNUM, P . I . I = T H A T 'S IN C R E D IB L E R = REMINGTON STEE LE V = V ID E O ONE & = SIM ON AND SIMON 70 Figure 3-8: Second Graders' Series Gratifications in Four Dimensions: Dimension 3 by Dimension 4 2 . 5 ♦ I 2.0 D IM E N S IO N 3 L e g e n d 6 - PACMAN B s BENSON 7 s POPEVE 2 s THE BRADY BUNCH 8 = R I C H I E R IC H D s O I F F 'R E N T s t r o k e s 5 = SMURFS [ = THE FACTS OF L I F E 9 3 s p i d e r m a n F S F A M IL Y T I E S X S G I L L l G A N ' S IS L A N O S = DYNASTY G = GIMME A BREAK E - E IG H T I S ENOUGH H HAPPY DAYS * 3 FAME L = LAVERNE ANO S H IR L E Y } 3 FANTASY IS L A N D ■ + ■ = M * A * S *H S L I T T L E HOUSE M = THE MUNSTERS 0 3 1 THE LOVE BOAT 1 ONE DAY AT A T IM E s = S IL V E R SPOONS A = THE A -T E A M 3 s THE THREE STOOGES C = C H IP S T = T H R E E 'S COMPANY z 3 THE DUKES OF HAZZARD > 3 THE F A L L GUY P - REAL PEOPLE K - K N IG H T RID E R ? = R I P L E V ' S z - MAGNUM. P . I . I = T H A T 'S IN C R E D IB L E R ~ REMINGTON STEELE V = V ID E O ONE & = SIMON AND SIMON 71 Figure 3-9: Sixth Graders' Series Gratifications in Four Dimensions: Dimension 1 by Dimension 2 CLUSTER 4 CLUSTER 2 CLUSTER 7 CLUSTER 6 SG CLUSTER 8 CLUSTER 1 CLUSTER 3 CLUSTER 5 - 2 . 5 - 2 . 0 - 1 . 5 - 1 . 0 - 0 . 5 0 - 0 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 D IM E N S IO N 1 L e g e n d 6 _ PACMAN B = BENSON 7 _ POPEYE 2 = THE BRADY BUNCH e _ R I C H I E R IC H 0 a s D I F F 'R E N T STROKES 5 _ SMURFS I = THE FACTS OF L I F E 9 * SPIDERM AN F X s F A M IL Y T I E S G I L L l G A N 'S IS L A N D $ _ DYNASTY G a GIMME A BREAK E _ E IG H T i s e n o u g h H = HAPPY DAYS * - FAME L = LAVERNE AND S H IR L E Y } & _ FANTASY IS L A N D = M* A * S *H _ l i t t l e h o u s e M s THE MUNSTERS 0 * THE LOVE BOAT 1 = ONE DAY AT A T IM E S s S IL V E R SPOONS A _ THE A -T E A M 3 = THE THREE STOOGES C - C H IP S T T H R E E 'S COMPANY 2 _ THE DUKES OF HA2ZARD > - THE F A L L GUY P = REAL PEOPLE K K N IG H T RID E R ? = R I P L E Y ' S % s MAGNUM. P . I . i s T H A T 'S IN C R E D IB L E R £ REMINGTON STEELE V - V ID E O ONE & = SIM ON AND SIMON 72 Figure 3-10: Sixth Graders' Series Gratifications in Four Dimensions: Dimension 3 by Dimension 4 ♦ X . CLUSTER 1 CLUSTER 7 CLUSTER 2 \ CLUSTER 4 0 . 5 0.0 CLUSTER 5 - 0 . 5 CLUSTER 8 CLUSTER C L U S T E R - 2 . 5 ♦ 1 . 5 0 . 5 1 .0 0.0 - 2 . 5 D IM E N S IO N 3 Legend 6 _ PA C M A N B s B E N S O N 7 - P O P E V E 2 = T H E B R A D V B U N C H 8 - R I C H I E R I C H 0 = D I F F ' R E N T S T R O K E S 5 5 S S M U R F S I = T H E F A C T S OF L I F E 9 - S P I D E R M A N F = F A M I L Y T I E S X as G I L L l G A N ' S I S L A N D $ = D V N A S T V G = G IM M E A B R E A K E — E I G H T I S ENOUGH H = H A P P Y D A Y S * s FA M E L = L A V E R N E A N D S H I R L E Y } - F A N T A S Y I S L A N D = M « A * S * H & S l i t t l e H O U S E M = t h e m u n s t e r s 0 - T H E L O V E B O A T 1 - O N E D A Y A T A T I M E S = S I L V E R SPO O NS A s T H E A - T E A M 3 = T H E T H R E E S T O O G E S c - C H I P S T = T H R E E ' S C O M P A N Y z - T H E D U K E S OF H A Z Z A R D R E A L P E O P L E > - T H E F A L L GUY P = K - K N I G H T R I D E R ? = R I P L E Y ' S * - M A G N U M . P . I . I a T H A T ’ S I N C R E D I B L E R s: R E M IN G T O N S T E E L E V = V I D E O ONE & - S IM O N A N D S IM O N 73 Figure 3-11: Tenth Graders' Series Gratifications in Three Dimensions: Dimension 1 by Dimension 2 2 . 5 0 5 GENERAL DRAMA 0 5 .0 - 0 . 5 0 5 0 5 0 . 5 0 1 M E N S IO N 1 Legend B - B E N S O N 6 * P A C M A N 2 _ T H E B R A D Y B U N C H 7 = P 0 P 6 V E D _ D I F F ' R E N T S T R O K E S 8 ~ R I C H I E R I C H I - T H E F A C T S O F L I F E 5 * S M U R F S F _ F A M I L Y T I E S 9 * S P ID E R M A N X - G I L L l G A N ' S I S L A N D G _ G IM M E A B R E A K $ = d y n a s t y H _ H A P P Y D A Y S E = E I G H T I S ENO UG H L - L A V E R N E A N D S H I R L E Y * s FA M E M * A * S *M ) * f a n t a s y i s l a n d M T H E M U N S T E R S < § • = l i t t l e H O U S E = O N E D A Y A T A T I M E 0 * t h e l o v e b o a t s - S I L V E R SPOONS T H E A - T E A M 3 - T H E T H R E E STO O G E S A * T _ T H R E E ' S CO M PANY C = C H I P S Z = T H E D U K E S OF H A Z Z A R O p R E A L P E O P L E > ~ T H E F A L L GUY ? _ R I P L E Y ' S K * K N I G H T r i d e r I _ T H A T ' S I N C R E D I B L E % s m a g n u m , p . I . v _ V I D E O ONE R = R E M IN G T O N S T E E L E & = S IM O N AND S IM O N 74 Figure 3-12: Tenth Graders' Series Gratifications in Three Dimensions: Dimension 1 by Dimension 3 2 . S * 2.0 0 . 5 0.0 • 0 . 5 GENERAL ORAMA - 2 .0 - 2 . 5 - 2 . 5 - 2 . 0 - 1 . 5 - 1 . 0 - 0 . 5 0 . 0 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 O X M E N S X O N 1 6 s P A C M A N B a B E N S O N 7 = P O P E Y E 2 = T H E B R A D Y B U N C H 8 a R I C H I E R I C H D = D I F F ' R E N T S T R O K E S 5 - S M U R F S I = T H E F A C T S OF L I F E 9 = S P I D E R M A N F = F A M I L Y T I E S X = G I L L l G A N ' S I S L A N D S a d y n a s t y G a G IM M E A B R E A K E s E I G H T I S ENO UG H H = H A P P Y D A Y S * = FA M E L = L A V E R N E A N D S H I R L E Y 3 = F A N T A S Y I S L A N D * s= M * A * S * H = L I T T L E H O U S E M = T H E M U N S T E R S 0 s T H E L O V E B O A T 1 = ONE D A Y A T A T I M E S = S I L V E R SP O O N S A a T H E A - T E A M 3 a T H E T H R E E S T O O G E S c s C H I P S T a T H R E E ' S C O M P A N Y z S T H E D U K E S OF H A Z 2 A R O > s t h e f a l l g u y P = R E A L P E O P L E K a K N I G H T R I D E R 7 a R I P L E Y ' S % s M A G N U M . P . I . I a T H A T ' S I N C R E D I B L E R = R E M IN G T O N S T E E L E V = V I D E O ONE & = S IM O N ANO S IM O N 75 the dim ensions. The prim ary differences between the age groups w ere the in creased absolute im portance o f Family fo r both the sixth and tenth graders, as reflected in the means, and the addition by the older children o f Problem Solving as associated w ith the series at the positive end of the dim ension. Despite these age differences in the com positions of the dim ension, the sim ilarities were strong enough to w arrant labeling this the H um or/A ction dim ension fo r all three ages, based on the strong associations of these attributes w ith the series at the poles o f the dimension. The rem aining dim ensions were less sim ilar across age. W hile there was som e sim ilarity among the second graders' second, the sixth graders' first, and the tenth graders' second dim ensions, it was not as consistent as am ong the H um or/A ction dim ension at each age. A lthough cartoon series w ere at the ex trem e negative end of the dim ension at each age, this was the only com m onality across ali three ages. The second graders em ployed only one attrib ute consis tently fo r this dimension. Character A ttributes, and an inspection of th e ir responses revealed no uniform concept underlying this code. Responses ranged from sex o f the main characters ('they are all men") to the clothing o f the main characters ("they all wear hats"). Neither sixth nor tenth graders used Character A ttributes much at all or in any consistent way. Further distinguishing the second graders from the older children was th e ir placem ent o f the magazine fo r m at series and the general dramas at the other extreme o f the dim ension, w hile the older children tended to place the crim e dramas in that position. In addition to the m oderate sim ilarity between series' placem ents on the 76 sixth graders' firs t and tenth graders' second dim ensions, there was som e sim ilarity between the attributes these age groups used w hich w ere associated w ith the dim ension. Both had Unreal stro n g ly associated w ith the (cartoon) series at the negative end and A ction and Crim e associated w ith the series at the positive end. The tenth graders, however, also had Problem Solving associated w ith the series at the positive end, and Age A ppropriateness associated w ith the series at the negative end, suggesting th a t they perceived cartoon program m ing as appropriate fo r younger children. Since these dim ensions had som e sim ila rity to each other, and since they appeared in part to represent d istin ctio n s in how realistic or unrealistic the series were, it was decided to name th is the Realism dim ension fo r all three ages. Finally, of course, tenth graders em ployed a th ird dim ension. As can be seen from Figure 6, this dim ension prim arily distinguished m agazine fo rm a t pro gram m ing from narrative form at program m ing. As is clear fro m the figure, s tru c tura lly this was a sim ple dim ension, w ith little va ria b ility in series scores on it apart from the form at differences. The co n te n t analysis suggested th a t it was conceptually fairly sim ple, also. The a ttrib u te s associated w ith the dim ension were N ovelty and Bad, w ith the positive end com prised o f series w hich were judged to be unusual, w hile series at the negative end tended to be regarded as low er in quality, although no series received a very high negative loading on the dim ension. It was decided to rely on the placem ent o f the series on the dim en sion to label it the Form at dim ension. From the above description it is evident th a t a sim ple in terpretation o f the 77 A ttrib u te C ondition dim ensions was not justified, and fro m th e perspective of w hat se lf-re p o rte d attributes w ere associated w ith the dim ensions, H ypothesis 1a was at least partially supported: there was a slig h t increase in age in the num ber o f dim ensions used, based on the second and sixth graders' use o f tw o, and the tenth graders' use o f three. If one regards increasing dim ensional com ple xity as a possible precursor to an increase in the num ber of dim ensions used, then sixth graders' use o f m ore com plex dim ensions than second graders m ay indicate an interm ediate phase between a sm all num ber o f sim ple dim ensions and a larger num ber of m ore com plex dim ensions. Because there is no te st fo r significance fo r num ber o f dim ensions, it m ight be asserted that the dim ensional solutions chosen, although explaining a high am ount of the subjects' variance (as evidenced in th e ir respective RSQs), were not the optim al solutions, at least based on interpretability. In o rder to explore th is possibility, the next highest dim ensional solution fo r each age group was in spected, em ploying the same criteria o f correlations and means as described above. For the second and sixth graders, no a ttrib ute m et the m ean (X>.05) and correlation (p<.01) criteria fo r in terp re tin g a third dim ension. The tenth graders' fo u rth dim ension did have one attrib ute associated w ith it w h ich m et these criteria, Humor. Since this dim ension im proved the overall so lu tio n only 2.6%, and since Humor was already the prim ary attrib ute on the firs t dim ension, and since the inclusion of an additional dim ension w ould not a lte r the conclusions regarding the support o f Hypothesis la , it was decided th a t an adding another dim ension to the tenth graders' solu tio n w ould not be appropriate. 78 Correlation analysis results Based on the content analysis data and inspection o f the plots of the series on the dim ensions, one can conclude that, at least descriptively, the dim ensional solutions o f the different age groups bore som e sim ila rity to each other, though not enough to suggest that they should be considered the same. It was also possible to com pare the dim ensions in ways w hich did not rely on the content analysis. By looking at the correlations between dim ensions, em ploying the series as cases, it was possible to obtain som e sense o f the degree to w hich the series were "located" by the different ages in the same dim ensional spaces rela tive to one another. These correlations are presented in Table 20, and it is clear that fo r the second and sixth graders, the H um or/A ction dim ensions w ere fa irly strongly correlated and that the Realism dim ensions were som ew hat less strongly correlated. The correlations between the second and tenth grade dim ensions were som ew hat higher. The H um or/A ction dim ensions are quite stron gly corre lated, and the relationship between the Realism dim ensions was m oderately strong. The Form at dim ension fo r the tenth graders was som ew hat related to the second graders' Realism dim ension. The sixth graders' H um or/A ction dim ension was strongly related to the tenth graders', though th is relationship was slig h tly less strong than that between the second and te n th graders. The Realism dim ensions, however, were even m ore stron gly related fo r the sixth and tenth graders. The tenth graders' Form at dim ension was som ew hat related to the sixth graders' H um or/A ction dim ension, but was unrelated to th e ir Realism dim ension. 79 Table 3-20: Correlations between A ttribute Condition Dim ensions Grade 2 Grade 6 Dimension Humor/Action Realism Humor/Action Realism Grade 6 Humor/ Action .67* -.17 Realism -.23 .59* Grade 10 Humor/ Action .88* .03 .82* .08 Realism -.08 .73* .01 .91* Format .28 .50* -.44* -.04 NOTE: N = 38 Series * p < .01 80 Canonical analysis results Extending the correlational analysis of the dim ensions, it was possible to com pare the dim ensional solutions across age groups in another way, again using the series as cases, but this tim e sim ultaneously com paring the series "loadings" on each dim ension between age groups. The dim ensions, then, were the variables, and the series' loadings on them w ere the values. This analytic ap proach is canonical correlation, w hich is often used in MDS analysis, and par ticularly in research w hich has used television stim uli (Alexander, 1980; Lom etti, Reeves, & Bybee, 1977; Reeves & Greenberg, 1977; Reeves & Lom etti, 1979). In canonical analysis the overall relationship between tw o sets of variables is ex amined by taking a linear com bination of the variables in one set to p redict a linear com bination o f the variables in the other set. These linear com binations are term ed the canonical variates o r canonical variables. This process is repeated, em ploying new variates uncorrelated w ith all previous variates o f the same set, until there are as many pairs o f variates as there are variables in the sm aller set. The role of each set of variables can then be reversed, predicting in the other direction. This kind o f analysis yields several bits of in form ation w hich can inform the in terpretation o f the dim ensions. First are the canonical co rre la tion coefficients, w hich are m erely the Pearson corre latio ns between each pair of canonical variates. These are often quite high, som etim es approaching 1. However, they are not to be understood as an index o f the relationship between the sets of variables, a use to w hich they have som etim es been put in the past (cf.. Reeves & Greenberg, 1977). They m erely represent the strength o f the relationship between the variates. 81 The second im portant outp u t o f canonical analysis is the canonical co e f ficie n t m atrix, w hich is a m atrix of the w e igh ts of the variables on th e ir variates. These w eights are beta w eights and allow one to evaluate the extent to w hich each variable is influential on the variate. This inform ation provides guidance in determ ining the extent to w hich any given variable (dim ension) in a set is related to any given variable in the othe r set through the variate. Canonical analysis also provides a structure m atrix, not reported on here, w hich is a m atrix of the co r relations between the original variables and the variates. A lthough the structure m atrix is often the "m atrix o f choice" to aid in in terpreting the relatio nship be tw een variables w ith in and between sets, this preference is p rim arily in response to the frequent problem of m u ltico llin e a rity of predicto r variables. Since ALSCAL scales sim ilarity data into essentially uncorreiated dim ensions, and since these dim ensions are the variables in the canonical analysis, m u ltico llin e a rity is not a problem here. A dditionally, because the series' scores on the dim ensions are standardized, the structure m atrix values differed only a little fro m the canonical coefficient m atrix values. See Levine (1977) and Tucker and Chase (1980) fo r fu r ther discussion o f the differences between the m atrix of w eights and the m atrix of correlations. Finally, canonical analysis can provide an index of "redundancy," w h ich is the m ost appropriate means of representing how much one set o f variables ex plains the other set (Cooley & Lohnes, 1971; S tew art & Love, 1968). It should be noted here th a t the am ount o f variance explained in one set by the o th e r is an asym m etrical measure. That is, the am ount of variance each set explains in the 82 other is typ ica lly not equal when the num ber o f variables in the tw o sets is not equal. The larger set w ill alm ost always explain a larger percentage o f the sm aller set than the sm aller set w ill o f the larger set. It should also be noted th a t the literature on canonical redundancy analysis uses a variety of te rm s to describe the results of the analysis. Som etim es th is index is said to reflect the am ount of variance one set o f variables explains in the other set of variables, som etim es it is said to reflect the am ount o f variance one set of canonical variates explains in the other set of variables, and som etim es it is said to represent the am ount o f shared variance or overlap between the sets o f vari ables. Because there are a large num ber of redundancy analyses reported below, these variations in phrasing w ill be em ployed to avoid tedious repetition. Again using the series as cases, the dim ensions as variables, and the series loading on the dim ensions as values, canonical corre latio n analyses w ere con ducted to explore fu rth e r the relationship between the dim ensions as represented by the series' locations on them . Beginning w ith the second and sixth graders, it can be seen in Table 21 that both the canonical correlations were significant. In spection of the canonical coefficie nt m atrix in the table suggests w hat accounted fo r these correlations. The w eights o f the second graders' tw o dim ensions on the firs t variate w ere both substantial, though the signs indicate th a t they operated in opposite directions. The sixth graders' w e igh ts on th e ir firs t variate were sim ilar. This is consistent w ith the content analysis results — th a t the sixth graders' second dim ension corresponded roughly to the second graders' first. For both grades, the relative m agnitude of the dim ension w e igh ts on the second 83 variates were approxim ately the reverse o f th e ir w eights on the firs t variates, w ith the exception that th e y all are positive. These results are partly consistent w ith the content analysis results in th a t the sixth graders' firs t and second dim ensions both had attrib ute s of A ction and Crim e associated w ith them , as did the second graders' firs t dim ension. That these dim ensions should all be related in the canonical analysis, then, is not surprising. That the second graders' Realism dim ension was related to the others {even though less strongly) is surprising. A lthough this conceptual overlap between dim ensions is apparent, the redundancy analysis suggests th a t it is not extensive. The am ount of variance in the sixth grade tw o dim ensional solution explained by the tw o canonical variates o f the second grade solution was about 48%. If looked at the other way, from sixth grade to second, it was 44%. In either case, slig h tly less than half the to ta l variance was explained, w hich, although substantial in m ost regression procedures, should probably be interpreted conservatively in th is case as m ean ing th a t the tw o are not very sim ilar. The sixth and tenth graders, how ever, looked much m ore alike. Both canonical correlations w ere high and highly significant, and the co e fficie n t m atrix was m ore easily interpreted. When com pared to the second dim ension, the w e ig h t fo r the sixth graders' firs t dim ension on th e ir firs t variate was quite high. On the tenth graders' firs t variate th e ir firs t dim ension w e ig h t was high, w ith the second and th ird dim ensions substantially lower. On the second variates, the sixth graders' Realism dim ension had a high w e ig h t and the H u m or/A ction 84 Table 3-21: Canonical Analysis of Attribute Condition Dimensions Canonical Variate I II Grade 2 with Grade 6 Rc .87 .42 F 21.40 7.38 df 4,68 1,35 p .00 .01 Grade 2 Coefficients Humor/Action .80 .60 Realism -.68 .74 Grade 6 Coefficients Humor/Action .74 .67 Realism -.66 .75 Redundancy (Variance Explained) Grade 2 .34 .10 Grade 6 .39 .09 85 Table 3-21 (cont'd) Canonical Variate I II Grade 6 with Grade 10 Rc .94 .93 F 71.71 103-30 df 6,66 2,34 p .00 .00 Grade 6 Coefficients Humor/Action .89 -.45 Realism .46 .89 Grade 10 Coefficients Humor/Action .74 -.49 Realism .51 .87 Format -.51 .08 Redundancy (Variance Explained) Grade 6 .43 .43 Grade 10 .28 .29 86 Table 3-21 (cont’d) Canonical Variate I II Grade 2 with Grade 10 Rc .94 .85 F 48.20 42.83 df 6,66 2,34 p .00 .00 Grade 2 Coefficients Humor/Action 1.01 .00 Realism -.10 1.00 Grade 10 Coefficients Humor/Action .95 .01 Realism -.21 .81 Format .25 .50 Redundancy (Variance Explained) Grade 2 .43 .36 Grade 10 .29 .26 87 dim ension a low (and negative) w eight. The tenth graders' firs t tw o dim ension w eights w ere nearly identical to the sixth graders' on th e ir variate, and th e ir third dim ension had a virtu a lly zero w eight. These results generally confirm ed the findings of the content analysis, in th a t the sixth and tenth graders' H u m or/A ction dim ensions corresponded to each o th e r fa irly closely, as did Realism, but since the firs t tw o dim ensions fo r both ages had a com ponent o f A ctio n and Crim e to them , these dim ensions all had noticeable w e igh ts on the firs t tw o variates. The tenth graders' third dim ension had a m oderately strong negative w e ig h t on th e ir firs t variate, but was not associated at all w ith the second variate. The redundancy analysis indicated th a t there was a som ew hat stro n ger relationship between the ways the sixth and tenth graders discrim inated am ong series than between the second and sixth graders. The sixth grade canonical variates explained 57% of the variance in the tenth graders' solution, and the tenth grade canonical variates explained 86% o f the sixth graders' solution. The sixth and tenth grade solutions, then, shared a little over 71% of th e ir to ta l variance, a large proportion by any standards. The second and tenth graders, oddly, w ere alm ost as m uch alike as the sixth and tenth graders, again w ith tw o high and highly sig nifica nt canonical c o r relations. The canonical coefficie n t m atrix indicated an even stro n g e r relationship between the corresponding dim ensions fo r these ages than betw een the sixth and tenth graders. The H um or/A ction dim ension w e igh ts w ere very high on the firs t variates fo r each age, and the rem aining dim ensions were quite low. The w eights fo r the Realism dim ensions on the second pair o f variates w ere also 88 high, w ith only the tenth graders' th ird dim ension show ing additional m oderate association w ith the variate. These results do not correspond to the results o f the co n te n t analysis, in th a t the second graders' dim ensions as derived from the content analysis w ere th o u g h t to be m uch sim pler than the tenth graders'. However, it m ust be repeated here th a t these canonical results indicate the relative sim ila rity between age groups o f the dim ension structures based on the loading o f the series on the dim ensions. It is not im possible th a t the dim ensions m eant d ifferent th in g s to these tw o groups of children. The redundancy results w ere very sim ilar to those fo r the sixth and tenth graders, w ith the second graders' canonical variates explaining 55% of the variance of the tenth graders' solution, and the te n th graders' canonical variates explaining 79% of the second graders' solution, fo r an overall average o f ju st about 67%. This finding suggests that, fo r the dim ensions used to distinguish among series (at least as represented by the series' loadings on them ), second graders and sixth graders w ere both m ore like tenth graders than they were like each other. Summary The com bined results o f the content analysis, the MDS analysis, and the canonical correlation analysis suggest m oderate support fo r Hypotheses 1a and 1b. Com bining the content analysis and the MDS analysis resulted in tw o sim ple dim ensions fo r the second graders, tw o som ew hat m ore com plex dim ensions fo r 89 the sixth graders, and three dim ensions fo r the tenth graders w hich w ere at least as com plex as those fo r the sixth graders. A lthough none o f the differences w ere extrem e, they unam biguously supported the hypotheses th a t children w ould use m ore and m ore com plex dim ensions w ith age. The results o f the canonical analysis o f the MDS solutions w ere not quite so straightfo rw ard, how ever. In th e ir placem ents of the series on the dim ensions the second graders and sixth graders both resem bled the tenth graders m ore than they did each other, though the resem blance was slig h tly greater between the tenth and sixth graders than between the tenth and second graders, as w o uld be expected. If the results were to parallel those fo r the conte n t analysis/M DS analysis, the second and tenth graders should have exhibited the least sim ilarity. It should be kept in mind, how ever, th a t the canonical analysis results represent sim ila rity o f the placem ent o f the series on the dim ensions, not s im ila rity of the content of the dim ensions. 3.3 Age Differences in G ratification Dim ensions C ontent analysis results The gratificatio ns associated w ith the dim ensional solutions suggested by the RSQ criterion are presented in Table 19. Figures 7 thro ugh 12 display the series' locations on the various dim ensions. From Table 19 it is im m ediately ap parent that, as w ith the a ttrib u te dim ensions, the second graders used m uch sim pler dim ensions than the sixth o r tenth graders. Unlike the A ttrib u te Con dition solutions, however, second and sixth graders used m ore dim ensions than tenth graders. Because o f the conceptual com plexity o f the o ld e r children's dim ensions, the sim ple labels applied to the dim ensions are necessarily som e 90 w hat m isleading. A lthough the gratifica tio n s associated w ith each grade level's firs t dim ension (as reflected in Table 19) varied, there w ere som e consistencies across ages. For all three ages this dim ension was p o sitive ly associated w ith Am usem ent and negatively associated w ith Excitement. Inspection o f where the series fell on the dim ensions in Figures 7, 9, and 11 indicated th a t com edies were typ ically at the positive end of the dim ension and crim e dram as were at the negative end. Because of these com m onalities, the firs t dim ension was named A m usem ent/E xcitem ent fo r all three grades. However, the tw o older groups had additional gratifications associated w ith th e ir dim ensions w hich suggested that the dim ension functioned differe n tly fo r them than fo r the second graders. Social Guidance was positively associated w ith the dim ension, suggesting th a t the series at the positive end of the dim ension (m ainly com edies) w ere view ed in part to learn about social interaction. A dditionally, both older groups associated General Pleasure w ith the series at the positive ends of th e ir dim ensions, and the sixth graders associated General Displeasure w ith the series at the negative end, suggesting that Am usem ent was generally associated w ith enjoym ent, and that fo r sixth graders only, Excitem ent was not associated w ith enjoym ent. Each grade's second dim ension resem bled the o th e rs' less than the firs t dim ension did. The second graders only associated a single g ra tifica tio n . A m use m ent, w ith it, w hile the sixth and tenth graders each associated five gratifications. Inspection of the plots in Figures 7, 9, and 11 indicated th a t the only clearly con sistent sim ilarity across all three grade levels was th a t the cartoon series always fe ll at one extrem e of the dim ensions. Thus they w ere stro n g ly associated w ith 91 Am usem ent fo r the second graders, associated w ith Am usem ent but also General Displeasure and Passing Time fo r the sixth graders, and only associated w ith General Displeasure fo r the tenth graders. The series at the opposite ends of these dim ensions from the cartoons had no strong associations fo r the second graders, but they were associated w ith Learning and Involvem ent fo r both the sixth and tenth graders and w ith Excitem ent fo r only the tenth graders. Using the poles of these dim ensions, it was decided to call them the Inform ation/D iversion dim ensions fo r the sixth and tenth graders, and the Sim ple Am usem ent dim ension fo r the second graders. The third dim ensions were sim pler and seemed less related than any pre vious dim ensions. Second graders only associated General Pleasure w ith th e ir third dim ension, a gratification no other grade level em ployed. A lthough tenth graders associated General Displeasure, along w ith Learning, w ith th e ir third dim ension, there did not seem to be a relationship between th e ir th ird and the second graders' th ird dim ension based on the plots in Figures 8 and 12. The sixth and tenth graders both associated Learning w ith their third dim ensions, but there was no apparent relationship between th e ir series' plots either (Figures 10 and 12). In addition to Learning, sixth graders also associated Excitem ent and Passing Tim e w ith this dim ension; oddly. Passing Time and Learning w ere both negatively associated w ith the dim ension. Because of the lack o f conceptual overlap or apparent sim ilarity am ong series placem ent, different labels w ere selected fo r each grade level's third dim ension. For second graders, it was called Sim ple Liking, fo r sixth graders it was called Learning/Excitem ent, and fo r tenth 92 graders, Learning/Displeasure. Only the second and sixth graders had fo u rth dim ensions. The second graders' was sim ilar to the tenth graders' th ird dim ension, being associated w ith Learning and General Displeasure. Inspection o f the plots o f the series' locations on the dim ensions (Figures 8 and 12) revealed th a t the magazine fo rm a t series w ere located tow ard the positive end of the dim ensions, w hich were associated w ith Learning. The series at the negative ends w ere m ore varied. Like the tenth graders' th ird dim ension, the second graders' fo u rth was also labeled Learning/Displeasure. The sixth graders' fo u rth dim ension was only associated w ith Social Guidance. Inspection o f the plots of the series' locations did not suggest any relationship w ith the second graders' fou rth dim ension, although the series at the dim ension poles bore som e resem blance to the series at the poles of the tenth graders' Learning/Displeasure dim ension. Since there was no con ceptual com m onality fo r these dim ensions, the sixth graders' fo u rth dim ension was labeled w ith the only gratificatio n associated w ith it. Social Guidance. To explore w hether an additional dim ension fo r each age group was in te r pretable, the next highest dim ensional solutions were inspected in the same m anner as they were fo r the A ttrib u te C ondition solutions described in the pre vious section. As was the case fo r the A ttrib u te C ondition, no gratificatio n m et the tw o criteria fo r association w ith an additional dim ension fo r the second or sixth graders. For the tenth graders, one gratificatio n. Am usem ent, was as sociated w ith a fourth dim ension. W ith the im provem ent in RSQ fo r a fourth dim ension only 4.1%, and Am usem ent having been very prom inently associated 93 w ith the firs t dim ension, it was decided th a t an additional dim ension w o uld not add appreciably to the understanding o f the tenth graders' g ra tifica tio n categories. Correlation analysis results Overall, the content analysis results suggested a m oderate conceptual relationship am ong the firs t tw o g ra tifica tio n dim ensions of the d iffe re n t age groups, and a weak relationship am ong the rem aining dim ensions. Looking at the correlations am ong the age groups' dim ensions tended to confirm th a t re la tio n ship, at least in term s of how the series loaded on them (see Table 22). C om par ing the second graders' and sixth graders' gratifica tio n dim ensions, it was clear th a t only the firs t dim ensions, A m usem ent/E xcitem ent, were very stro n g ly related to each other and that they were unrelated to any other dim ensions. The second pair of dim ensions, Simple A m usem ent and Learning/Diversion, w h ile fa irly stron gly related to each other, were also m oderately related to the th ird dim en sions o f the opposite age group, w hich in turn were only m oderately related to each other. The fourth dim ensions, Learning/Displeasure and Social Guidance, w ere unrelated to each other. The sixth and tenth grade solutions exhibited a sim pler relationship. Each age group's firs t and second dim ensions were strongly related to the other group's firs t and second dim ensions and unrelated to any other dim ensions. The content analysis results suggested th is relationship fo r the firs t dim ension, but only partially fo r the second. The tenth graders' th ird dim ension, Learning/Displeasure, was m oderately stro n g ly related to the sixth graders' fo u rth 94 Table 3-22: Correlations between Gratification Condition Dimensions Grade 2 Amusement/ Simple Simple Learning/ Dimension Excitement Amusement Liking Diversion Grade 6 Amusement/ Excitement .90* -.01 .17 -.10 Learning/ Diversion -.13 -.72* .50* -.35 Learning/ Excitement -.18 .43* .55* -.04 Social Guidance .20 -.02 -.05 -.26 Grade 10 Amusement/ Excitement .89* .10 .01 -.09 Learning/ Diversion -.04 -.70* .45* -.41 Learning/ Displeasure -.03 -.05 .13 .48* 95 Table 3-22 (Cont'd) Grade 6 Amusement/ Learning/ Learning/ Social Dimension Excitement Diversion Excitement Guidance Grade 10 Amusement/ Excitement .9^* Learning/ Diversion .07 Learning/ Displeasure .11 MOTE: N = 38 Series * p < .01 -.17 -.0 2 .12 .92* .03 .08 .05 .09 -.69* 96 dim ension, Social Guidance, but the sixth graders' th ird dim ension, Learning/Excitem ent, was not related to any tenth grader dim ension. As w ith the sixth graders, the second graders' A m usem ent/E xcitem ent dim ension was stron gly related to the tenth graders’ corresponding dim ension, and neither o f these was related to any o th e r dim ension. The second dim ensions fo r each o f these tw o age groups exhibited a fa irly strong relationship, but the second graders' th ird dim ension was also m oderately related to the tenth graders' second dim ension. The th ird dim ensions fo r these tw o age groups w ere unre lated to each other, but the fo u rth dim ension fo r the second graders, Learning/Displeasure, had a m oderately strong relationship w ith the tenth graders' Learning/Displeasure dim ension, as w ould be expected fro m the content analysis results. Canonical analysis results Pursuing these relationships, canonical analyses w ere perform ed on the G ratification C ondition dim ension data; the results are presented in Table 23. For the second and sixth graders, the firs t three variates had sig n ifica n t canonical correlations; the fourth did not. Looking at the m atrix o f coefficients, it is evident th a t there was a fa irly sim ple relationship between the dim ensions and the variates, w ith each dim ension fo r each age group having a high beta w e ig h t on its corresponding variate and all other w eights being lo w to negligible. Since the canonical correlation fo r the fo u rth pair o f variates was not significant, these variates should not be interpreted as containing any useful in form ation, even though the pattern o f relationships is consistent w ith the previous variates. 97 Table 3-23: Canonical Analysis of Gratification Condition Dimensions Canonical Variate I II III IV Grade 2 with Grade 6 Rc .97 • 93 .74 .24 F 30.22 19.16 8.42 1.98 df 16,92 9,76 4,64 1,33 P .00 .00 .00 .17 Grade 2 Coefficients Amusement/ Excitement .98 .17 -.01 .11 Simple Amusement .20 -.80 • 49 -.28 Simple Liking .08 • 34 O ' 00 • .30 Learning/ Displeasure .01 -.36 -.18 .92 Grade 6 Coefficients Amusement/ Excitement .90 .28 .21 .26 Learning/ Diversion • 32 • 92 .22 .08 Learning/ Excitement .14 -.19 .97 -.09 Social Guidance .21 .14 -.02 -.96 Redundancy (Variance Explained) Grade 2 .23 .24 .13 .01 Grade 6 .24 .22 • 13 .01 98 Table 3-23 (eonfc'd) Canonical Variate I II III IV Grade 6 with Grade 10 Rc .97 .95 .70 F 51.78 35.14 15.78 - df 12,82 6,64 2,33 - P .00 .00 .00 - Grade 6 Coefficients Amusement/ Excitement .83 .55 -.05 Learning/ Diversion -.52 .86 .07 - Learning/ Excitement -.07 .11 -.11 - Social Guidance .07 -.05 .99 - Grade 10 Coefficients Amusement/ Excitement .90 .42 .08 Learning/ Diversion -.40 .91 .14 - Learning/ Displeasure .01 .24 i • OO - Redundancy (Variance Explained) Grade 6 .24 .21 .12 Grade 10 .32 .28 .17 _ 99 Table 3-23 (cont') Canonical Variate I II III IV Grade 2 with Grade 10 Rc .94 .86 .48 - F 21.49 13-65 5.06 - df 12,82 6,64 2,37 - P .00 .00 .01 - Grade 2 Coefficients Amusement/ Excitement .76 .62 .17 - Simple Amusement .56 -.47 .30 - Simple Liking -.27 • 36 • 39 - Leaaning/ Displeasure .17 -.44 .87 - Grade 10 Coefficients Amusement/ Excitement .75 .66 .13 - Learning/ Diversion -.64 .76 .13 - Learning/ Displeasure -.04 -.11 1.00 - Redundancy (Variance Explained) Grade 2 .22 .20 .06 - Grade 10 • 30 .25 .08 * 100 This relationship was confirm ed in the redundancy analysis. As is evident fro m the table, the firs t tw o dim ensions fo r each grade to g e th e r accounted fo r alm ost half the variance in the solutions fo r the other grade, w ith the third dim ensions adding another 13%. The fo u rth dim ension added v irtu a lly nothing, as one w ould expect from the n o n -s ig n ific a n t canonical corre latio n associated w ith it. in all, the second and sixth grade variates each explained about 61% of the variance in the other solution. W hile still leaving a substantial am ount of variance unaccounted for, this was about 15% m ore variance explained fo r these tw o groups than in the A ttrib u te C ondition. All three variates had highly sig n ifica n t canonical corre latio ns fo r the sixth and tenth grade com parisons. The dim ension w eights on the variates indicated th a t the dim ensions overlapped som ew hat. The A m usem ent/E xcitem ent dim en sion fo r each age group had high positive w e igh ts on the firs t variate, though the Learning/D iversion dim ensions each had m oderate negative w e ig h ts associates w ith the variates. This is not surprising given the results of the co n te n t analysis, w hich showed some overlap in the gratifica tio n s associated w ith the firs t and second dim ensions, but reversed signs fo r shared concepts. The second variates were also dom inated by single dim ensions, Learning/D iversion, though again there was some representation of the firs t dim ension fo r each age group, but this tim e the signs were both positive. The th ird variate fo r each grade was the "purest." The sixth graders' fo u rth dim ension, Social Guidance, was w eighted highly on th e ir variate, and the tenth graders' th ird dim ension, Learning/Displeasure, was highly negatively w eighted on th e ir variate. 101 The results of the redundancy analysis of the sixth and te n th graders' g ratificatio n dim ensions indicated a m oderately large am ount o f shared variance between the solutions. The sixth graders' canonical variates explained about 77% of the tenth graders' variance; the tenth graders' canonical variates explained about 57% of the sixth graders' variance, fo r an overall average of about 67%. The second and tenth graders' g ra tifica tio n dim ensions showed a som ew hat weaker relationship. The firs t tw o canonical correlations were highly significant, but the third ju s t barely achieved significance. As w ith the sixth and tenth graders, on alm ost all variates m ore than one dim ension was at least m oderately w eighted. On the second graders' firs t tw o variates th e ir A m usem ent/E xcitem ent dim ension was m ost heavily w eighted, w ith Sim ple A m usem ent m oderately w eighted also. This pattern was nearly identical fo r the tenth graders, except th a t A m usem ent/E xcitem ent and Learning/D iversion reversed th e ir relative im p o r tance on the second variate. The third variates were much sim pler, w ith each grade's Learning/Displeasure dim ension having a high w e igh t and all other dim ensions having low w eights. The second graders' third dim ension, Simple Liking, was not highly w eighted on any variate. Unlike the results in the A ttrib u te C ondition, the second graders shared less variance w ith the tenth graders than they did w ith the sixth graders, though the differences w ere not very large. The second graders' canonical variates explained about 63% o f the variance o f the tenth graders' solution, w hile the tenth graders' canonical variates explained about 48% of the second graders' solution, fo r an overall average o f about 56% of the variance in com m on. 102 Overall, then, it can be argued th a t the g ra tifica tio n s associated w ith the television series in th is sam ple differed m oderately across the three ages. The m ost succinct results w ere fo r the redundancy analyses reported above, w hich suggested that the sixth graders had m ore in com m on w ith the tenth graders than they did w ith the second graders, even though they used fe w e r dim ensions than the tenth graders and the same num ber as the second graders. Sum m ary Com bining the results o f the co n te n t analysis and the MDS analysis sug gests no support fo r Hypothesis 2a, in th a t the num ber of g ra tifica tio n dim en sions remained constant or decreased w ith age, rather than increased. The results suggest th a t the second and sixth graders both em ployed fo u r g ra tifica tio n dim ensions, and the tenth graders three. Hypothesis 2b received som e sup port, in that the com plexity of the dim ensions was som ew hat greater fo r the sixth graders than fo r the second graders, though it was no greater fo r the tenth graders than fo r the sixth and could be argued to be less. The results o f the canonical analysis of the MDS solutions were m ore in the expected direction, w ith the placem ents of the series on the second graders' dim ensions resem bling those o f the sixth graders m ore than those o f the tenth graders. Again it should be stated that sim ilar placem ent o f the series on the dim ensions does not mean that the dim ensions represented identical gratifications. 103 3.4 Condition Differences in Dim ensions w ithin Age As the preceding section makes clear, there was m oderately strong evidence fo r age differences in the num ber and com position o f dim ensions children used in distinguishing am ong televisio n series based on the series' at tributes and based on the gratificatio ns associated w ith the series, although not all the age differences were in the expected direction. It is th e o re tica lly possible that these dim ensions, though referring to d ifferent conceptual dom ains (i.e., be tween series' attrib ute differences and between series' g ra tifica tio n differences), w ere in fact different ways of representing a single phenom enon. That is, series m ight be sim ilarly positioned on one or m ore dim ensions relative to one another in the different experim ental conditions, thus suggesting th a t series' attributes corresponded m ore or less perfectly w ith gratificatio ns. For exam ple, the m ag nitude of a series loading on the H um or/A ction dim ension m ight alw ays be re lated to the same series' loading on the A m usem ent/E xcitem ent dim ension fo r one or m ore age groups, and so on. It was hypothesized that th is w o u ld tend to be m ore so fo r younger children than fo r older children. Second grade results The m ost im m ediately apparent conclusion fo r the second graders is that the A ttrib u te C ondition solution and G ratification C ondition solution are not the same, if only because they have different num bers of dim ensions. However, it is m athem atically possible th a t a single dim ension in one condition could be related to m ore than one dim ension in the other cond ition . Examining the series' at tributes and gratifications associated w ith d iffe re n t dim ensions it was clear th a t 104 the firs t dim ensions from each condition could be argued to be conceptually re lated, in that Hum or can be th o u g h t o f as the cause of Am usem ent, and A ction and Crime m ight be considered to be causes o f Excitem ent, all o f w hich were stron gly associated w ith the firs t or second dim ensions. A nother possible con ceptual relationship am ong the dim ensions was betw een the attrib u te dim ension Realism and the gratificatio n dim ensions w hich had a Learning com ponent as sociated w ith them . It seemed' plausible that those series w hich w ere judged to be m ore realistic (or, m ore accurately, less unrealistic) could be associated w ith greater learning of factual m aterial. A dditionally, the sixth and tenth grade H um or/A ction attribute dim ensions had com ponents of the Family and Problem Solving attributes associated w ith them , w hich it was th o u g h t m ight be related to those gratification dim ensions w hich had Social Guidance associated w ith them . The correlations between the second grade A ttrib u te C ondition and G ratification Condition dim ensions are presented in Table 24. As in the concep tual analysis above, the firs t dim ensions from each con d itio n w ere highly co rre lated. Unexpectedly, the second dim ensions were also fa irly highly correlated, but no other dim ensions were sig nifica ntly correlated at all. A canonical analysis analogous to those reported in the previous section was conducted across con ditions. As indicated in Table 25, the tw o canonical correlations w ere highly sig nificant, w hich was prim arily due to the relationship between the firs t pair and second pair of dim ensions. The last tw o dim ensions of the G ratification Con dition did not figure into the canonical solution m uch at all, the to ta l overlap be tw een the sets was only m oderate, w ith the attrib u te dim ension variates explain 105 ing about 35% of the gra tifica tio n dim ension solution, and the gratificatio n dim ension variates explaining about 69% of the a ttrib u te dim ension solution, fo r an overall average of about 52% to ta l shared variance. Sixth grade results Like the second graders, the sixth graders also had fo u r g ra tifica tio n dim en sions and tw o attrib ute dim ensions. The sixth graders' second attrib u te dim en sion, H um or/A ction, seemed likely, fro m a conceptual analysis, to be related to th e ir firs t gratificatio n dim ension, A m usem ent/E xcitem ent, given the attributes and gratificatio ns associated w ith them (see Tables 19 and 20). The correlations am ong the dim ensions (Table 24) showed a strong relationship between the H um or/A ction attribute dim ension and the A m usem ent/E xcitem ent gratificatio n dim ension. A dditionally, there was a fa irly strong relationship between the Realism attribute dim ension and the Learning/D iversion g ra tifica tio n dim ension w hich was expected based on the concepts associated w ith the dim ensions. No other correlations were high enough to be considered strong, though the relationship between the H um or/A ction attrib u te dim ension and the Social Guidance gratificatio n dim ension approached significance, a relationship w hich was expected. In the canonical analysis, the tw o canonical correlations w ere highly sig nificant (see Table 25). The dim ensions' w e igh ts in the canonical co e fficie n t m atrix indicate th a t the firs t pair o f variates (Variate I fo r each condition) were fa irly clear in th e ir com positions. The Realism attrib ute dim ension loaded m ost highly on one, and the Learning/D iversion and (less strongly) the Social Guidance 106 Table 3-24: Correlations between A ttrib u te C ondition Dim ensions and G ratification Condition Dim ensions Attribute Condition Gratification Condition Humor/Action Realism Format Grade 2 Amusement/ Excitement .90* • 0 -fc r 1 Simple Amusement -.05 -.72* Simple Pleasure -.05 .10 Learning/ Displeasure < N O • -.21 Grade 6 Amusement/ Excitement .78* -.13 Learning/ Diversion -.0 6 .86* Learning/ Excitement .06 .17 Social Guidance .40 .20 Grade 10 Amusement/ Excitement .93* C M O • in • 1 Learning/ Diversion .11 .92* .17 Learning/ Displeasure m o • -.06 .70* NOTE: N = 38 Series * p < .01 107 Table 3-25: Canonical Analysis of Attribute Condition Dimensions and Gratification Condition Dimensions Canonical Variate II III Grade 2 Rc .91 .74 F 20.89 13-35 df 8,64 3,33 p .00 .00 Attribute Coefficients Humor/Action 1.00 .04 Realism -.14 .99 Gratification Coefficients Amusement/ Excitement .99 .07 Simple Amusement .07 -.95 Simple Pleasure -.06 .06 Learning/ Displeasure .08 -.21 Redundancy (Variance Explained) Attribute .41 .28 Gratification .21 .14 108 Table 25 (cont'd) Canonical Variate I II III Grade 6 Rc .92 .87 - F 33-52 33.13 - df 8,64 3,33 - P .00 .00 - Attribute Coefficients Humor/Action -.34 .94 - Realism .94 -.35 - Gratification Coefficients Amusement/ Excitement -.39 .79 - Learning/ Diversion .89 • 31 - Learning/ Excitement .21 .16 - Social Guidance .05 .50 - Redundany (Variance Explained) Attribute .43 .37 - Gratification .21 .19 109 Table 25 (cont'd) Canonical Variate I II III Grade 10 Rc .96 .93 .70 F 62.74 45.37 32.87 df 9,78 4,66 1,34 P .00 .00 .00 Attribute Coefficients Humor/Action .94 .28 -.20 Realism -.36 .90 -.27 Format • 13 .21 .98 Gratification Coefficients Amusement/ Excitement .97 .18 -.17 Learning/ Diversion -.16 .99 -.09 Learning/ Displeasure .16 .19 .97 Redundancy (Variance Explained) Attribute • 30 .31 .16 Gratification • 31 .28 .17 110 gratificatio n dim ensions loaded highly on the other. The second variates from each condition w ere som ew hat less clear. One o f the pair was com prised prim arily o f the H um or/A ction attrib u te dim ension, and the o th e r o f the A m usem ent/E xcitem ent g ratifica tio n dim ension, w ith a m oderate w e ig h t fo r the Social Guidance g ratificatio n dim ension. The redundancy analysis indicated that fo r the sixth graders there was greater overlap betw een these tw o sets of dim ensions than fo r the second graders, w ith an average o f about 61% shared variance. Tenth grade results The tenth graders had, o f course, three dim ensions in each condition. Con ceptually, it was possible to see a relationship between each corresponding pair. The H um or/A ction attrib ute dim ension appeared to be sim ilar to the A m usem ent/E xcitem ent g ra tifica tio n dim ension, especially when considering that H um or/A ction also had Family and Problem Solving associated w ith it, w hile A m usem ent/E xcitem ent also had Social Guidance — all concepts w hich one m ight expect to find co -o ccu rrin g . The Realism a ttrib u te dim ension seemed likely to be related to the Learning/D iversion g ra tifica tio n dim ension and possibly the Learning/Displeasure g ra tifica tio n dim ension, w ith the cartoons at one ex trem e and the m ore realistic series w hich deal w ith issues w hich m ig h t involve teenagers at the other. For the third pair o f dim ensions, the Form at attribute dim ension and Learning/D ispleasure gratificatio n dim ension, the possible re la tio n ships were som ew hat less clear, though not opaque. The series at the positive end of the Format attribute dim ension seemed m ost likely to be those closely as 111 sociated w ith learning about the world, w hich could be a com ponent of either of the tw o gratification dim ensions w ith Learning associated w ith them . The correlations am ong these tw o sets o f dim ensions supported these speculations fa irly unam biguously (see Table 24). Each pair of dim ensions described above was highly correlated, and no dim ension was sig n ifica n tly co rre lated w ith any dim ension other than its "opposite." The results of the canonical analysis are presented in Table 25. The three canonical corre latio ns were all high and highly significant, and, as is evident fro m the canonical co e fficie n t m atrix, each dim ension fo r each condition was stro n g ly associated w ith only one variate, and no tw o dim ensions shared strong associations w ith a variate. Thus the dim ensions w ith in conditions were quite independent o f one another, but c o r responding dim ensions across conditions were stro n g ly related. The redundancy analysis confirm ed this interpretation, w ith each set o f variates explaining a very large proportion of the variance in the other set of variables (76%). Recalling th a t Hypothesis 3 predicted th a t the a ttrib u te and g ra tificatio n dim ensions would m ost resem ble each other fo r younger children and least resem ble each other fo r older children, these results are quite straightfo rw ard and run counter to the hypothesis. Rather than the attrib u te and g ra tificatio n dim ensions diverging w ith age, they converge. 3.5 Predictive Validity of the Dimensional Solutions It was hypothesized th a t the view ing frequency o f the series categorized by dim ension loadings w ould better predict social reality beliefs than w ould view ing 112 frequency of series categorized by trad ition al Nielsen categories, w hich in turn w ould be better predictors than overall view ing frequency. To te s t this hypothesis it was firs t necessary to id entify w hat groups o r clusters o f series w ere suggested by the dim ensional solutions fo r each age group and condition. To accom plish this, the chosen solutions fo r each age group in each con d ition w ere subjected to a clu ster analysis using the SAS clu ste r procedure (Ray, 1982), again using the series as cases and the dim ensions as variables. Cluster analysis provides a means of id entifying groups or clusters o f objects based on the sim ila rity am ong the objects considering the values o f m ore than one variable sim ultaneously. This is accom plished by successively finding clusters w h ich m in im ize intracluster differences w hile at the same tim e m axim izing in te rclu ste r d if ferences. Thus the clusters are those series which, as a group, have dim ension loadings w hich are m ost sim ilar to each other and m ost d iffe re n t fro m other groups (clusters). In effect, the clustering procedure partitions the dim ensional solutions represented in Figures 3 -12 into successively fe w e r groups of series based on the sim ilarity o f all possible pairs o f series' loadings. A problem w ith this form o f analysis, as w ith the num ber o f fa cto rs or num ber o f dim ensions problem s in fa cto r analysis and m ultidim ensio nal scaling, is the lack o f universally accepted m ethods fo r choosing the optim al num ber of clusters. Like these other techniques, the usual approach is to exam ine m easures analogous to successive 2 ? 2s and look fo r an "elbow " indicating a leveling o ff in variance explained. Two indices, i?2 and the Cubic C lustering C riterion (CCC) provided by SAS were used in the present analysis to select the appropriate 113 num ber o f clusters fo r each age by co n d itio n group. Graphic representations of the clustering h istory fo r each analysis are presented in Figures 13 th ro u g h 18. The H2 and CCC values fo r the firs t ten clusters fo r each age group and condition are presented in Table 26. Based on the R 1 and CCC criteria, the fo llo w in g num bers o f clusters fo r each grade and condition w ere selected. For te n th graders, each condition yielded seven clusters of series. For sixth graders, there w ere eig h t clusters in the G ratification C ondition and five in the A ttrib u te Condition. For second graders, the G ratification C ondition had six clusters, and the A ttrib u te C ondition had five. The com positions o f the clusters are presented in Tables 27 through 32. It should be noted here th a t fo r the second graders, the CCC was not stable in the A ttrib u te C ondition and never exceeded 2.04 in the G ratifica tion C ondition, suggesting that these solutions are not as nearly optim al as those selected fo r the sixth and tenth graders. To assess the relative predictive v a lid ity o f these clusters as groups of series th at show sim ilar dim ensional structure, fo u r sets of stepw ise regression analyses w ere run, using five social reality scales as outcom e m easures fo r the sixth and tenth graders and fo u r fo r the second graders. Stepw ise regression was chosen over a straight regression approach because no specific m odels were posited a priori, and the stepw ise approach allow s cycling the p re d icto r variables in and out o f the equations in order to arrive at the best overall m odel. To avoid excluding variables w ith som e predictive pow er, the p value fo r inclusion is ty p i cally set higher than one w ould if a specific m odel w ere being tested. For these 114 SERIES L T D G s I H F r f a B I I I R T A A E A R F F H M 3 L 1 T T N A M I L A A F A M S V 3 R c s L S T L I S 0 D C G P E E M E F H P E V 0 A L V y r S I L p C R s u M A I I R I I V S 0 P Y E E S T L A V B A 0 T N A K t L M S E P P D H I N D N V N E N B R L V E M S 0 s C D - S N N L A M A 0 E 0 P C E A E 0 T 0 F B U 0 0 I E 0 N B p p M 0 T H U T I I G G U R C P R U L R 0 S I 0 P I u A 0 N F K G R A S R A 0 A G E I K E M G T G N R I M E M S E E T S A L E G M A c E A N Y 0 E N 0 s E R P E A 0 H 0 U U F C A V A E V D 1 Y L V E S H E T H L S N E S N A V N H s s s s M N T N Y M S H N E N 1 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 2 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 3 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 4 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x S x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x J c x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 6 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 7 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 8 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 9 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 1 0 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x , x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . N 11 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . u 1 2 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x M 1 3 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . B 1 4 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x E 1 5 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x , x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . R 1 6 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . 1 7 x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 0 1 8 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x F 1 9 . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 2 0 x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x c 2 1 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x L 2 2 . x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x u 2 3 . x x x x x x x x x x x . x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x . s 2 4 x x x x x x x x x x x . x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x . . x x x x , T 2 5 . x x x x x x x x x x x . x x x x x x x x . x x x x x x x x x x x x x x x x x x . x x x x x x x . . x x x x . E 2 6 . x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x R 2 7 . x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x S 2 8 2 9 3 0 . x x x x x x x x x x x x xxxx . xxxx xxxx xxxx xxxxxxx xxxxxxx xxxx . xxxxxxx . xxxxxxx . xxxxxxx . xxxxxxx . xxxxxxx . xxxxxxx . . x x x x . . x x x x . x x x x 3 1 . xxxx xxxx . xxxx . xxxx xxxxxxx . x x x x 3 2 xxxx . xxxx . xxxx xxxxxxx . x x x x 3 3 if xxxx . x x x x . x x x x x x x . . x x x x . 3 4 x x x x . x x x x x x x . x x x x 3 5 x x x x . x x x x x x x 3 6 x x x x x x x x . 3 7 3 8 • x x x x . Figure 3-13: Tree Diagram o f Cluster Analysis Results for Grade 2 Attribute Condition Dim ensions SERIES L 1 G 1 S 1 8 B 0 1 F F R R T II T M F L R F A 3 H A I E A T M A V 1 A F L M S A 71 3 S C R A r L E C E S 0 0 1 P 7 G M P II E F 1 s E T R Y S V 1) L 0 L P A S 1 II I 1 M A V A S E 0 T E Y Y 0 A Y B S T L 11 S P 0 E P M 1 1. A K P 1 I! i H S N N B R II H V E Y 0 L S M 0 E A C A It L D - S N E 0 1 71 0 B 0 P 0 E (1 0 1 B M A 1 P F. 0 It 0 1 f U P R R C H G C U r 1 1 0 E P C U R F 0 U 0 N K A 0 A S 1 A R A S 1 G G E R E H 1 M 1 N T G K E M G P 0 L R S E 0 G A V , E H A S 7 F 71 N Y 0 S E A R F Y A 0 A P U 0 U E A 0 II L F. K E A L N II Y II S E T H Y S Y E S II L S 11 S S E N II 11 S M 71 Y S M N T E 1 Y D 1 2 3 M 5 6 7 8 9 1 0 11 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 20 2 1 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ■xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx xxxxxxx xxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx ■ xxxxxxxxxxxxxxxxxxxxxx •xxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx 2 2 X X X X X X X X X X X X X xxxxxxx 2 3 X X X X X X X X X X X X X xxxxxxx 2 4 xxxxxxxxxxxxx xxxxxxx 2 5 ■ . xxxxxxxxxx xxxxxxx 2 6 . xxxxxxxxxx xxxxxxx 2 7 . xxxxxxxxxx xxxx . 2 8 xxxxxxxxxx xxxx . 2 9 . xxxxxxxxxx 3 0 . xxxxxxxxxx 3 1 . xxxxxxxxxx 3 2 . xxxxxxxxxx 3 3 . xxxx xxxx 3 4 . xxxx xxxx 3 5 . . xxxx 3 6 . . . xxxx 3 7 . xxxx 3 8 xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxx xxxxxxx xxxxxxx xxxxxxx xxxxxxx xxxxxxx xxxxxxx xxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxxxxx xxxxxxx xxxxxxxxxx xxxxxxxxxxxxx xxxxxxx xxxx xxxxxxx xxxxxxx xxxxxxx xxxx xxxx xxxx xxxxxxx xxxxxxxxxx xxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxx . xxxxxxxxxx . xxxxxxxxxx . xxxxxxxxxx . xxxx xxxx . xxxx xxxx xxxxxxxxxxxxxxxx xxxx xxxxxxx xxxx xxxxxxxxxx xxxxxxx xxxx xxxx xxxxxxxxxxxxx . xxxx xxxxxxx xxxx xxxxxxxxxx xxxxxxx xxxx xxxx xxxxxxxxxxxxx . xxxx xxxx . xxxx xxxxxxxxxx xxxxxxx xxxx xxxx xxxxxxxxxxxxx . xxxx xxxx . xxxx . xxxxxxx xxxxxxx xxxx xxxx xxxxxxxxxxxxx . xxxx xxxx . . xxxxxxx X X X X X X X xxxx xxxx xxxxxxxxxxxxx . xxxx xxxx . . xxxxxxx xxxx . xxxx xxxx xxxxxxxxxxxxx . xxxx xxxx . . xxxxxxx xxxx . xxxx xxxxxxxxxxxxx . xxxx . . xxxxxxx xxxx . xxxx xxxxxxxxxxxxx . xxxx . . xxxxxxx xxxx . xxxx xxxxxxxxxxxxx . xxxx . . xxxxxxx xxxx . xxxxxxx xxxx . xxxx . . xxxxxxx xxxx . xxxxxxx xxxx . xxxx . . xxxxxxx xxxx . xxxxxxx xxxx . xxxx . . xxxxxxx xxxx . xxxxxxx xxxx . xxxx . . xxxxxxx xxxx . xxxxxxx xxxx . . xxxxxxx xxxx . xxxxxxx xxxx . . xxxxxxx xxxx . . xxxx xxxx . . xxxxxxx xxxx . . xxxx xxxx . . xxxxxxx . xxxx xxxx . . xxxx . . xxxx xxxx . . xxxx . . xxxx xxxx . . xxxx xxxx . xxxx . cn Figure 3-14: Tree Diagram o f Cluster Analysis Results for Grade 6 Attribute Condition Dimensions S E R I E S Figure 3-15: Tree Diagram of Cluster Analysis Results for Grade 10 A ttribute Condition Dimensions Su © C- u > '-J x x x x x x x x x x x x x x x x x x x x x x x x x x x . . . • .................... X XX X XX X X X X X X X X X X X X X X X X X X X X X X X X X XX X X X X X X X X X X X X X X x x x x x x x oo e. — © u j c z s t c i x x x x x x x x x x x x x x x x x x x x x x x x x x x ......................................... XX x x x x x x x x x x x x x x x x x x x x x x Q-CCJfSCZXXXXXXXXXXXXXXXXXX............................................................................................................. x x x x x x x x x x x x x x x x x x X X x x x x x x x x x x x x x x x x y ^ e: — u = — w cc — © c z x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x * * x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x X X X X X X X X X X X X X X X X X X X X x x x x x x x x x x x x x x x x O i S => ST u. C O X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X • . X X X X > - o w c — X x x x x x x x ...................................................................................................................... x x x x x x x x x x x x x x x x u. < s UJ X X X X X X X X ...................................................................................* ..........................* ............................................. X X X X X X X X X X X X £* — Q » _ JU i> X XX XX XX XX XXX XX XXX XXX XX XXX ................................................. x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x *- © < w — SOCCuJGXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX......................................... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x C lj < aJ c-u-ca IUJXXXXXXXXXXXXXXXXXXXXXXX X X X X X X X ........................................ X X X ZIXOGGZXXXXXX XXXXXXX X X X X X X ....................................................................................................... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x , qsujs — a r u * - o z x x x x x x x x x x x x x x x x x x x x x x x x x x x x ..................................... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x X SL — C = I- X X X X X X x X X X X X X X X X X X X X X X X X X X X X • • ................................... x x x x x x x x x x x x x x x x x x x x © =i x w to x x x x x x x x x .............................................................................................................. x x x x x x x x x x x x x x — SOZXXX X X X X X X X X X X X X X X X X X X X X X X X X X X .............................................. x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x X X X X X X X X X X X X X X X X X X X XX X X .............................- ............................................. x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x G>Z<CO>->XXXXXXXXXXXXXXXXXXXXXXX................................................................................. x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x C J3 = — a- xn X X X X X X X X X X X X X X X X ..................................................................... . . . . ............................ X -------1— ! — © < Z X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X ................................... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x z i c i < c s s z o x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x ......................... x x x x x x x x x x x x x x x x x x x x x x © © z t o s - w a c c o x x x x x x x x x x x x x x x x x x x x x x x x .....................* ............................ x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x tn M r - O © © U V5 X X X X X X X X X X X X X X X X X X X X X X X X .............................................. x x x x cs — co u j z c g c g x x x x x x x x x x x x x x .................................................................................................................................... X X X x x x x x x x x x x x x x x x x x x x x x x x x x cr — -J > ui C£ cOa-OCZXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXX............................. X X x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x :=:u j x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x .................... X X X X XX X X X X X X X X X X X X X X X X x x x x x x x x x x x x x x x x x x s < X X X X X X X X X X X X X X X X X X X X X .................................................................................................* XXXXXXXXX x x x x x x x x x ii. < u I- W © **- — iXXXX XXXXX X X X X X X X X X X X x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x Lta < s — -J > t— — UJWXXXXXXXXXXXXXXXXXXXX x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x © — * 4. u. C O ‘-CSQXuJtOXXXXXXXXXXXXXXXXXXXX x x x x x x x x x x x x x x x x x x x x x x x X x x x x x x fiOUiZCOOZXXXXXXXXXXXXXXXXXXXX x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x C - z z - < CSaiUJ<XXXXXXXXXXXXX x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 2<ft.a.>- © O W X X X X X X X X X X X X X X X X X X X X X ............... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x <*>00 O O ZO .<Z>XXXXXXXX XXXXXXXXXXXXXXXXXX x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x OZTUJOOXXXXXXXXXXXXXXXXXXXXXXXXXX X X X X X X X X X X _iO>UJ C O O < I- X X X X X X X X X X X X X . . . . ... .. x x x x x x x x x x x x x x x x x x x x x x x x x x U-<Z*-<C/5> — CO.JXXXXXXXXXXXXXXXXXXXXXX x x x x x x x x x x x x x x x x x x x x x x XXXXXXXXXXX X XX x x x x x x x x J - l-H JU SO=><OUjXXXXXXXXXXXXXXXXXXXXXX 0 0 u z ©© U i C S C C O a- ©c — i < > Uia=sr < C O X X XX X XXX X X XX XX X X X X X XX X X X X X X X X X X X X X X X XX XX X X X XX X XX X X X X X X X XX X X X X X X X X X X X X X X XX X X XX X X X * • * X X X X X XXX X X XX x . X X X X X X X XXX XX X XX XX XX X X X X XX X XX XX X XX X X X Xx X X X XX XXX X X X x XX XX X X X X XX X X X X X X X X XX XX X XX X X X X X X XX XX XX XX X X X X XX X X XX X X *-c\jf>asrir>\O f-eocN © »-<\jr'‘, > .ru '% N O f, , -oo o sO < “ C v < ‘o .3 ,ir\v£> r'»coo\© *-c\/r'0*srirt', © r-co r - * — — f-^ -.-.-.-T -C M C 'U C V IC 'J C 'J C 'J C 'd C S J C M C 'jrO P O ro r'O M r’O r ' X ' O m Z Z 5SC0UJCC O u. o — « © ♦- UJ C C « 0 117 118 SERIES L D S G T I B B F F I I I R H R T R A A F L 3 F M M E A I T A I L N M F V S A A M A T C 3 S R L 0 S 0 T I E C P E L S H G M P F E E V V A L S R C T P L 0 I S I U I V A M E E S a 0 V T 0 S Y A A P I R Y S P E P T L N D I A K L M I H B N V E N R S M V E N I N M A 0 0 L S E 0 c - 0 N L A S N 0 U 0 B N E T F 0 P P 0 D B E 0 C P A U C R P 0 I T R E H T U I G I G M U N U 0 I S 0 I A K 0 A F A R R P R L S R M I E G G E M 0 I E K G G N M T A s C G A s 0 A E M E 0 N Y E N L E E T F A C Y E A R A P A E H U U 0 0 S E H H T L N V S E S N V L S A E E D V V S N H E S N S N 1 S M S T Y M N N H 1 2 3 4 5 6 7 8 9 1 0 1 1 12 1 3 1 4 1 5 1 6 1 7 IB 1 9 20 21 22 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x xxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxxxxxxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxx xxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxx xxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxx xxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxx xxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxx xxxxxxx xxxxxxx . xxxxxxx xxxxxxxxxx xxxxxxx xxxx xxxxxxx xxxxxxxxxx xxxxxxx xxxx xxxxxxx xxxxxxxxxx . xxxx xxxx xxxxxxx xxxxxxxxxx xxxx xxxx xxxx . xxxxxxxxxx . xxxx xxxx xxxx . xxxxxxxxxx . xxxx xxxx xxxx . xxxxxxxxxx xxxx xxxx . xxxxxxxxxx . xxxx xxxx . xxxxxxxxxx . xxxx xxxx . xxxxxxxxxx . xxxx xxxx . xxxxxxxxxx . xxxx . xxxxxxxxxx xxxx xxxxxxxxxx . xxxx . xxxx xxxx . xxxx , xxxx xxxx . . xxxx xxxx . . xxxx . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x XXXXXXXXXXXXX xxxxxxxxxx xxxxxxxxxxxxx x xxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxx XXXXXXXXXXXXX xxxxxxxxxx xxxxxxxxxxxxx xxxxxxx xxxxxxxxxxxxx xxxxxxx xxxxxxxxxxxxx xxxxxxx xxxxxxxxxxxxx xxxxxxx . xxxxxxxxxxxxx xxxx xxxxxxxxxxxxx xxxx xxxx xxxxxxx xxxx . xxxx xxxxxxx xxxx . x x x x xxxxxxx xxxx xxxx xxxxxxx x x x x . x x x x x x x x x x x xxxx . xxxx x x x x x x x xxxx . xxxx x x x x x x x x x x x . xxxx x x x x x x x x x x x xxxx xxxxxxx xxxx xxxx xxxxxxx xxxx xxxx xxxxxxx xxxxxxx xxxx . xxxx . xxxx . xxxx . xxxx Figure 3-16: Tree Diagram of Cluster Analysis Results for Grade 2 Gratification Condition Dimensions Figure 3-17: Tree Diagram of Cluster Analysis Results for Grade 6 Gratification Condition Dimensions H-= < !- C O — s ©X uJa x X X XXX XXX XXX X X X X X X XX X X X X X X X X X X X X X u- < z: •— < co > — C O-1 X X X XXX XXX XXX X X X X X X X X X X X X X X X X X X X X X © — a-_l UJ> X X X XXX XXX XXX X X X X X X X X X X X XX Q>-Z< co u-> X X X XXX XXX XXX X X X X X X X X X X X X X X X X X X X X X _J © > UJ © 0 A r- X X X XXX XXX XXX X X X X X X X X X X X X X X X X X X X X X L l. A M UJ x X X XXX XXX XXX X X X X X XX > -O uj© «- X X X xxx XXX xxx X X X X X X X X X X X X X X X X X X X X X £ < C O© X X X xxx xxx xxx X X X X X X X X X X X X X X X X X X X X X 02 Ijj < -J CL UJ OO. «JUi x X X xx x X X X X X X X X X X X C O C l. — a Ui C 2s <Z X X X xx x xx x x x x X X X X XX X X X X X X X X X X X X X X X a- © a.UJ >UJ X X X xx x xx x x x x X X X X X X X X X X X X X X X X X X X X X p A C H A z x X X xx x x x x x x x XX X X X X X X X X X X X X X X X X X X X CC-QX-UJ C £— o © X X X xx x xx x xx x X X X X X X X X X X X X X X X X X X x X X 07 S © 0 2 u.C O x X X xx x X X X XX XX X X x y z - z o *- o Z X X X x x x xx x x x x X X X X X X X X X X X X X X X X X X X X X £ < c © ©S X X X xxx xxx xxx X X X X X X X X X X X X X X X X X X X X X — — is: ujC O X X X xx x xx x xx x X X X X X X X X X X X X X X X X X 0 0 —s ©© X X X xxx x x x XXX X X X X X X X X X X X X XX X X X X X X U. < J J o ©> X X X x x x x x x xxx X X X X X X X X X X X X X X X X X X X X X < 1 I - uJ < s x X X x x x xxx xxx X X X X XX X X X X X X X X X X X X X X X x z: —© = 1 — X X X xxx xxx xxx X x X X X X X X X X X X X X X X X X X X X o = — a. C O X X xxx X X X X X X X n u N S T UJ X X COX X X xx x xx x xx x XX X X X X X X X X X X X X X X X X X X X G 1 L L 1 o < ©X X X xx x xx x xx x X X X X X X X X X X X X X X X X X X X X X m C O w O OO UJC O x X X xx x xxx xx x X X X X X X X X X X X X X X X X X X X X X -i < >i*JX ©UJ X X X xx x xx x xx x X X X X X X X X X X X su z C O ©© X X X xx x xx x xxx X X X X X X X X X X X X X X X X X X X X X C O C O OCSC- < z > X X X xxx xxx xxx X X X X X X X X X X X X X X X X X X X X X C O— > uj as co a.c ©z x X X xxx xxx xxx X X X X X X X X X X X X X X X X X X X X X O— U. u. CO-XO2C UJC O x X X xxx xx x XXX X X X X X X X X X X X X X XX o - s s u < ©X UJ< X X X xx x xx x xx x X X X X X X X X X X X X X X X X X X X X X x < A. a. >. c A Y C O x X X xx x xxx xx x X X X X X X X X X X X X X X X X X X X X X L<UH(/) © -J X X X xxx x x x xx x X X X X X X X X X X X X X X X X X X X X X 1 - — UJC O X X X x x x x x x x x x X X X X X X X X X X X X X X X X X X X X X o z u a < > x X X x x x x x x x x x X X X X X XX X X ao — oo u z o © a X X X X x x x x x x x x x XX X X X X X X X X X X XX XX X X X X X 03 02 < © > © © « _ w X X X X xx x xxx xxx X X X X X X X X X X X X X X X X X X X -j — •— I— J UJ zz ©© C OUJ Xxx x c c j m © X X tf" N V O X X f— « 0 9 X 10 X »- X X X Xxxx X X x; Xxxx X X X X X X X X X Xxxxx xxx x xx x x x x x . . X X X X Xx x x x x x x x x X X X X X X X X X X x x x x x x x x x x x x x X X X Xx x x x x x x x x x x x x x X X X X X X X xxxxxx X X X X xxxxxx X X X Xxxxxxx X X X X xxxxxx X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Xx X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X • • X X X X X X X X X X X X X X X X X X X X X Xxxx X X X X X X X X X X X X X X X X X X X X X Xxxx X X X X X X X X X X X X X X X X X X X X X Xxxx • • X X X X X X X X X X X X X X X X Xx X Xx X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X • • • X X X X X X X X X X X X X X X X X X X X X X Xx X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X x x x x x x x x x X X X X X X X X x X X X X X x x x x x x x x x x x X X X X X X X X X X X X x x x x x x x x x x x X X X X X X X X X X X X x x x x x x x x x x x • • • . X X X X X X X X X X X X x x x x x x x x x X X X X X X X x X X X X x x x x x x x x x X X X X X X X X X X X X x x x x X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X • X X X X X X X X X X X X X X X X x X X X X X X X X X x x x x • • ♦ * X X X X X X X X X X X X X x x x x X X X X X X X X X X X X X x x x x X X X X X X X X X X X X X x x x x • • • • X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X x X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X x X X X X X * « • X X X X X X X X X X X X >' X X X X X X X X X X X X X X X X X X X X X X X X X s; X X X X X X X X X X X X X x x x x x x x x x x • X X X X X X X X X X X X X X X XXXXXXXX x X X X X x X X XX X XXX X X X X X X X X X X X X X X X X X X x X XX X X X X XX X X XX X X X X X X X X X • X X X X XXXX X X X XX X X X XX X XX X X X X X X X X XXX X XX X XX X XX X X X XX X X XX XX X XX X X XXX X X XX X X X X X X X XX X X XXX X X XX X X X XX X x x x x x x • X X X X X X XX X X X XX X x x x x x x X X X X X X X X XX X X X X x x x x x x X X X X X X XX X X X X X X x x x x x x • X XX X X X X X X X X X X X X X X X XX XX X X x x x x x x x x x X X XXXX x x x x x x x x x X X X X X X x x x x x x x x x X X X X X X X X X X XX x x x Z 2 Z S U S © L. O . J © C O h “ UJC£CO 119 S E R I E S L 0 s G T I I I F I B 8 R H F R T F L F 3 A H M R E A A I T L F V A S M A M A I 3 R A T N C S L 0 E C I P E D S G M E F L S T H P E V D S R T c L P Y L I S U M A V A I I E V 0 T B S 0 V Y A E A L T N I L K A M P R I I S S E P D P H N N R S E M B N V L 0 S c N L N - A S E I N 0 0 Y M A E 0 0 F M B A E 0 P N 0 P T D B U 0 E I 0 T H G I T G I 0 P C E u U R C R P U A A 0 S D K 0 S F A I A R N U R G G E I T G G E N M P L R 0 K I R I M M E s M S A T A E 0 0 N E V E C G N A E R P 0 U H A U 0 L E E E S F C A A V E E H T V V S N N L V S S A H H E N S S S N Y T M M N E V 0 1 S L S H N N E 1 2 3 4 5 6 7 8 9 10 1 1 12 1 3 14 1 5 1 6 1 7 1 8 1 9 20 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 3 8 C 2 1 X X X X . . X X X X X X X X X X X X X L 2 2 X X X X . . x x x x x x x x x x x x x U 2 3 X X X X . . x x x x x x x x x x x x x s 2 4 X X X X . . x x x x x x x x x x x T 2 5 X X X X x x x x x x x x x x x E 2 6 X X X X . . x x x x x x x x x x x R 2 7 . x x x x x x x x x x x S 2 8 x x x x x x x x x x x 2 9 . x x x x x x x x x x x 3 0 x x x x x x x x x x x 3 1 x x x x x x x x x x x 3 2 . x x x x x x x x x x x 3 3 . x x x x x x x x 3 4 . x x x x . x x x x 3 5 x x x x . x x x x 3 6 x x x x x x x x 3 7 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x < x x x x x x x x x x x x x x x x x x x x x x x < x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x . x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x . x x x x x x x x . x x x x . x x x x . x x x x . x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x . x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x -n < o c U I 0 8 < D < D Q > c o O 0 ) Oi 3 3: o o - * > “ Q 5 c = > 2 3 > a 3 -■ M 11 3 ® (D 2 3 £. V ) r * o M - 3 —h o -I 0 ) a ( D TO o Table 3-26: R-Squared and Cubic C lustering C riterion fo r First Ten Clusters fo r Each Condition Attribute Condition Gratification Condition Clusters R2 CCC R2 CCC Grade 2 10 9 8 7 6 5 4 3 2 1 0.981 0.975 0.967 0.953 0.938 0.921 0.873 0.695 0.475 0.000 7.165 6.567 6.056 5.235 4.955 5.236 4.123 - 0.219 0.653 0 .0 0 0 0.854 0.826 0.798 0.766 0.731 0.679 0.589 0.458 0.250 0.0 0 0 2.045 1.533 1.321 1.261 1.339 1.150 0.512 -0.420 -0.960 0 .0 0 0 Grade 6 10 9 8 7 6 5 4 3 2 1 0.978 0.974 0.969 0.962 0.953 0.942 0.899 0.757 0.417 0.000 6.252 6.222 6.499 6.506 6.844 7.452 5.861 1.767 -1.133 0.0 0 0 0.922 0.911 0.897 0.873 0.836 0.781 0.713 0.518 0.300 0 .0 0 0 8.710 8.968 9.074 8.600 7.622 6.397 5.220 1.100 -0.442 0 .0 0 0 Grade 10 10 9 8 7 6 5 4 3 2 1 0.973 0.969 0.963 0.956 0.931 0.897 0.848 0.660 0.382 0.000 12.913 13.274 13.334 13.729 11.466 10.046 8.950 2.557 0.245 0 .0 0 0 0.927 0.918 0.908 0.889 0.861 0.817 0.700 0.572 0.354 0 .0 0 0 4.252 4.543 5.013 4.903 4.604 3-942 0.930 0.353 -0.586 0 .0 0 0 121 Table 3-27: Clusters of Series fo r Grade 2 A ttribute Dimensions and O utcom e Measures They Predict Cluster Series Outcome Measure & Direction 1 DYNASTY FANTASY ISLAND LITTLE HOUSE RIPLEY'S THAT'S INCREDIBLE VIDEO ONE SUPPORT + MEANWLD + 2 THE BRADY BUNCH EIGHT IS ENOUGH THE FACTS OF LIFE FAME FAMILY TIES THE LOVE BOAT ONE DAY AT A TIME REAL PEOPLE 3 BENSON DIFF'RENT STROKES GILLIGAN'S ISLAND GIMME A BREAK HAPPY DAYS LAVERNE AND SHIRLEY M*A*S*H THE MUNSTERS SILVER SPOONS THREE'S COMPANY THE THREE STOOGES 4 THE A-TEAM CHIPS THE DUKES OF HAZZARD THE FALL GUY KNIGHT RIDER MAGNUM, P.I. REMINGTON STEELE SIMON AND SIMON 5 PACMAN POPEYE RICHIE RICH SMURFS SPIDERMAN SUPPORT 122 Table 3-28: Clusters of Series fo r Grade 6 A ttrib u te Dim ensions and O utcom e Measures They Predict Cluster Series Outcome Measure & Direction BENSON THE BRADY BUNCH DIFF*RENT STROKES DYNASTY EIGHT IS ENOUGH THE FACTS OF LIFE FAME FAMILY TIES FANTASY ISLAND GIMME A BREAK HAPPY DAYS LAVERNE AND SHIRLEY LITTLE HOUSE THE LOVE BOAT M*A*S*H ONE DAY AT A TIME SILVER SPOONS THREE'S COMPANY GILLIGAN'S ISLAND THE MUNSTERS THE THREE STOOGES PACMAN POPEYE RICHIE RICH SMURFS SPIDERMAN THE A-TEAM CHIPS THE DUKES OF HAZZARD THE FALL GUY MAGNUM, P.I. REMINGTON STEELE SIMON AND SIMON KNIGHT RIDER MEANWLD REAL PEOPLE RIPLEY'S THAT'S INCREDIBLE VIDEO ONE 123 Table 3-29: Clusters of Series fo r Grade 10 A ttrib u te Dim ensions and O utcom e Measures They Predict Cluster Series Outcome Measure & Direction 1 LITTLE HOUSE THE LOVE BOAT FANTASY ISLAND COMPLY + GERBNERV + 2 BENSON DIFF’RENT STROKES EIGHT IS ENOUGH THE FACTS OF LIFE FAMILY TIES GIMME A BREAK HAPPY DAYS LAVERNE AND SHIRLEY M*A*S*H ONE DAY AT A TIME SILVER SPOONS THREE'S COMPANY 3 THE BRADY BUNCH GILLIGAN'S ISLAND THE MUNSTERS THE THREE STOOGES 4 THE A-TEAM CHIPS DYNASTY THE DUKES OF HAZZARD THE FALL GUY KNIGHT RIDER MAGNUM, P.I. REMINGTON STEELE SIMON AND SIMON SINGERV + 5 REAL PEOPLE RIPLEY'S THAT'S INCREDIBLE 6 FAME VIDEO ONE 7 . PACMAN POPEYE RICHIE RICH SMURFS SPIDERMAN 124 Table 3-30: Clusters o f Series fo r Grade 2 G ratification Dim ensions and O utcom e Measures They Predict Cluster Series Outcome Measure & Direction 1 BENSON THE BRADY BUNCH EIGHT IS ENOUGH FANTASY ISLAND LITTLE HOUSE THE LOVE BOAT 2 DIFF'RENT STROKES THE FACTS OF LIFE FAME FAMILY TIES GIMME A BREAK HAPPY DAYS LAVERNE AND SHIRLEY ONE DAY AT A TIME SILVER SPOONS THREE'S COMPANY 3 DYNASTY MEANWLD + REAL PEOPLE RIPLEY'S THAT'S INCREDIBLE 4 GILLIGAN'S ISLAND THE MUNSTERS PACMAN POPEYE RICHIE RICH SMURFS THE THREE STOOGES 5 SPIDERMAN VIDEO ONE 6 THE A-TEAM CHIPS THE DUKES OF HAZZARD THE FALL GUY KNIGHT RIDER MAGNUM, P.I. M*A*S*H REMINGTON STEELE SIMON AND SIMON 125 Table 3-31: Clusters of Series fo r Grade 6 G ratification Dim ensions and O utcom e Measures They Predict Cluster Series Outcome Measure & Direction 1 THE BRADY BUNCH EIGHT IS ENOUGH LITTLE HOUSE MEANWLD 2 BENSON DIFF*RENT STROKES THE FACTS OF LIFE FAMILY TIES GIMME A BREAK HAPPY DAYS ONE DAY AT A TIME SILVER SPOONS 3 GILLIGAN'S ISLAND LAVERNE AND SHIRLEY THE MUNSTERS THE THREE STOOGES MEANWLD + SINGERV 4 THE A-TEAM CHIPS THE DUKES OF HAZZARD THE FALL GUY KNIGHT RIDER MAGNUM, P.I. REMINGTON STEELE SIMON AND SIMON SUPPORT + 5 PACMAN POPEYE RICHIE RICH SMURFS SPIDERMAN SINGERV + 6 M*A*S*H REAL PEOPLE VIDEO ONE MEANWLD 7 DYNASTY FAME THE LOVE BOAT GERBNERV + SINGERV + 8 FANTASY ISLAND RIPLEY'S THAT'S INCREDIBLE SINGERV SUPPORT 126 Table 3-32: Clusters of Series for Grade 10 Gratification Dimensions and Outcome Measures They Predict Cluster Series Outcome Measure & Direction 1 DYNASTY FAME LITTLE HOUSE THE LOVE BOAT M*A#S*H COMPLY + GERBNERV + 2 BENSON THE BRADY BUNCH DIFF'RENT STROKES EIGHT IS ENOUGH THE FACTS OF LIFE FAMILY TIES GIMME A BREAK HAPPY DAYS ONE DAY AT A TIME SILVER SPOONS THREE'S COMPANY 3 GILLIGAN'S ISLAND LAVERNE AND SHIRLEY THE MUNSTERS THE THREE STOOGES GERBNERV + SINGERV + 4 THE A-TEAM CHIPS THE FALL GUY KNIGHT RIDER MAGNUM, P.I. REMINGTON STEELE SIMON AND SIMON 5 REAL PEOPLE RIPLEY'S THAT'S INCREDIBLE VIDEO ONE 6 THE DUKES OF HAZZARD FANTASY ISLAND 7 PACMAN POPEYE RICHIE RICH SMURFS SPIDERMAN 127 analyses, a p value of .15 was selected, a value su fficie n tly high to allow even w eakly predictive variables in the m odels, but low enough to exclude those w hich w ere unlikely to add much to the overall models. The clusters o f series w hich w ere significant predictors o f social reality beliefs are outlined in Figures 3-12. It was decided th a t pictorial representation of the clusters should be lim ite d to those w hich predicted effects because representation o f num erous clusters in m ore than tw o dim ensions is m ore confusing than clarifying. The firs t analysis em ployed only the mean view ing frequency across all series as a predictor. Three o f the 14 equations produced sig nifica nt results. For second and sixth graders, this score was not related to any o f the outcom e vari ables. For tenth graders, it was positive ly related to one o f the tw o fam ily rela tio n s variables, COMPLY (F(1,76)=2.34, p=.13) and tw o o f the fo u r violence in society variables, GERBNERV (F(1,76)=4.22, p=.04) and SINGERV (F(1,76)=2.98, p=.09) though the relationships w ere very weak (see Table 33). The second analysis em ployed as predictors the mean view ing frequency across series w hich were classified into traditional Nielsen categories (see Table 5 fo r a list o f these categories and the series w ith in them ). Four o f the 14 equa tions produced sig nifica nt results. Only view ing o f General Drama was a sig nificant predictor fo r any age group. As w ith overall view ing, fo r tenth graders it predicted COMPLY (F(1,76)=6.65, p=.01), GERBNERV (F{1,76)=9.89, p«.00), and SIN GERV (F(1,76)=3.78, p=.06). A dditionally, fo r second graders it predicted MEANWLD (F(1,45)=4.63, p=.04). The overall R 2s were higher in each case using the Neilsen grouping o f series than using overall view ing frequency, suggesting th a t e m p lo y- 128 Table 3-33: Social Reality Outcomes Predicted by Average Viewing Frequency of Tenth Graders 2 Outcome B F df P R COMPLY •33 2.34 1,76 • 13 • 03 GERBNERV • 35 4.22 1,76 .04 in 0 • SINGERV • 5*1 2.98 1,76 .09 .04 129 ing viewing frequency of series grouped in some meaningful way is an improve ment over using overall viewing frequency (see Table 34). The th ird and fourth analyses em ployed the clusters of series fo r each age group in each condition w hich were derived from the clu ster analysis described above. The clusters have already been identified in Tables 27-32; they w ill be referred to in the fo llo w in g text as Cluster 1, Cluster 2, etc. It should be noted th a t in som e cases the clusters are identical w ith the Nielsen groupings. All as sociated F statistics in the text are fo r the overall models. In the analysis of the A ttrib u te C ondition data, six out o f the 14 equations produced sig nifica nt results (see Table 35). For the second graders there were tw o significant equations. C luster 1 and C luster 5 predicted SUPPORT (F(2,44)=2.23, p=.12), and Cluster 1 predicted MEANWLD (F(1,45)=6.80, p=.01). For sixth graders, only MEANWLD was predicted, and only by C luster 5 (F(1,82)=2.48, p=.12). For tenth graders, three of the five outcom e variables had significant equations. Cluster 1 predicted COMPLY (F(1,76)=3.20, p=.08) and GERBNERV (F(1,76)=7.37, p=.01). Cluster 4 predicted SINGERV (F(1,76)=2.57, p=.11). For the A ttrib u te Condition, then, the clusters o f series w hich predicted could be considered slig h tly better than the Nielsen groupings fo r second and sixth graders, based on the prediction o f an additional outcom e variable fo r each. A dditionally, fo r the one outcom e variable w hich was predicted by the Nielsen General Drama category fo r second graders, MEANWLD, the overall i?2 was im proved slig h tly by the substitutio n o f the second grade Cluster 1 series. For the 130 Table 3-34: Social Reality Outcomes Predicted by Viewing Frequency of Nielsen Classified Series Outcome Predictor B F df P R Grade 2 MEANWLD GEN DRAMA in CM • 4.63 1,45 .04 .09 Grade 10 COMPLY GEN DRAMA • 35 6.65 1,76 .01 00 o • GERBNERV GEN DRAMA • 34 9.89 1,76 o o • .12 SINGERV GEN DRAMA 00 C O • 3-78 1,76 .06 .05 131 Table 3-35: Social Reality Outcomes Predicted by Viewing Frequency of Attribute Condition Series Clusters Outcome Predictor B F df P R Grade 2 SUPPORT Cluster 1 .39 4.05 1,44 .05 Cluster 5 -•31 2.18 1,44 .15 .09 MEANWLD Cluster 1 .26 6.80 1,45 .01 .13 Grade 6 MEANWLD Cluster 5 -.14 2.48 1,82 .12 .03 Grade 10 COMPLY Cluster 1 .19 3-20 1,76 • o 00 .04 GERBNERV Cluster 1 .23 7.37 1,76 .01 .09 SINGERV Cluster 4 .32 2.57 1,76 .11 • 03 132 tenth graders, on the other hand, no additional variables were predicted, and the overall i?2 fo r each one th a t was predicted was slig h tly lo w e r fo r each m odel based on the A ttribute C ondition clusters than fo r those based on the Nielsen grouping. The data from the G ratifica tion Condition presented a som ew hat different picture. Eight out of the 14 equations produced sig nifica nt results (see Table 36), but of these, half w ere fo r the sixth graders. The second graders had only one significant equation, in w hich C luster 1 predicted MEANWLD (F(1,45)=4.24, p=.05). For tenth graders, the same three outcom es w ere predicted as w ere by the N iel sen and A ttrib u te C ondition clusters. Cluster 1 predicted COMPLY (F(1,76)=5.34, p=.02), Clusters 1 and 3 predicted GERBNERV (F(2,75)=4.Q6, p=.02), and C luster 3 predicted SINGERV (F{1,76)=2.90, p=.09). The i?2s fo r each of the te n th grade m odels w ere slig h tly higher than those fo r the corresponding m odels in the A t trib u te C ondition, but they w ere slig h tly lo w er than those w hich em ployed the Nielsen classified series. The results fo r the sixth graders' G ratification C ondition w ere clearly d if ferent fro m either the A ttrib u te C ondition or the Nielsen groups. For these children, C luster 4 and C luster 8 predicted SUPPORT (F(2,81)=3.48, p=.04). Three clusters, 1, 3, and 6, com bined to predict MEANWLD (F(3,80)=3.72, p=.01). C luster 7 predicted GERBNERV (F(1,82)=2.43, p=.12). Finally, fo u r clusters, 3, 5, 7, and 8, com bined to predict SINGERV (F(4,79)=2.62, p=.04). Only one equation in any other analysis ever predicted an outcom e fo r sixth graders. C luster 5 fo r MEANWLD in the A ttrib u te C ondition, and its R2 was considerably less than the 133 Table 3-36: Social Reality Outcomes Predicted by Viewing Frequency of Gratification Condition Series Clusters 2 Outcome Predictor B F df p R Grade 2 MEANWLD Cluster 3 .18 4.24 1,45 .04 * o Grade 6 SUPPORT Cluster 4 • 31 3.14 1,81 .08 Cluster 8 -.37 5.52 1,81 .02 • o 00 MEANWLD Cluster 1 -.15 5.62 1,80 .02 Cluster 3 .13 3-82 1,80 .05 Cluster 6 -.13 3-35 1,80 .07 .12 GERBNERV Cluster 7 .14 2.43 1,82 .12 .03 SINGERV Cluster 3 -.23 2.60 1,79 .11 Cluster 5 .40 4.03 1,79 .05 Cluster 7 .28 2.60 1,79 .11 Cluster 8 -32 3-52 1,79 .06 .12 Grade 10 COMPLY Cluster 1 • 31 5.34 1,76 .02 .07 GERBNERV Cluster 1 .22 3.94 1,75 .05 Cluster 3 .13 2.44 1,75 .12 .10 SINGERV Cluster 3 .26 2.90 1,76 .09 .04 134 comparable R2 in the Gratification Condition. Overall, then, the sixth graders' clusters of series in the G ratifica tion Con dition can be considered better predictors o f social re a lity beliefs than the clusters in the A ttrib u te C ondition or the Nielsen groupings. For second graders, they were s lig h tly w orse than the A ttrib u te C ondition and the same as the Niel sen groupings. For tenth graders, they were very slig h tly b etter than the A t trib u te C ondition and slig h tly w orse than the Nielsen groupings (based on slightly higher and s lig h tly low er overall R 2s). From these results. Hypothesis 4 can be said to have received mixed sup port. Mean exposure to series w h ich w ere grouped based on th e ir a ttrib u te s as perceived by second graders was a better predicto r of the outcom es o f interest than was trad ition al grouping based on Nielsen categories and was a m uch better predicto r than overall average view ing frequency. Mean exposure to series w hich had been grouped according to the g ra tifica tio n s second graders associated w ith them was not a better predicto r than mean exposure to Nielsen grouped series, though both o f these were better than mean overall view ing frequency. For sixth graders, mean exposure to series grouped according to perceived attributes predicted a little better than mean exposure to Nielsen categorized series and mean overall view ing frequency, in the sense th a t some prediction is better than no prediction at all. When the series w ere grouped according to the g ra tifica tio n s sixth graders associated w ith them , however, a substantial increase in prediction was achieved. For te n th graders, the hypothesis was n o t supported, 135 since exposure to the Neiisen categorized series predicted the outcom es of in te r est som ew hat better than exposure to the series clustered by the gratifications associated w ith them , w hich in turn predicted the outcom es som ew hat better than exposure to the series clustered by the program attrib ute s associated w ith them , all of w hich w ere som ew hat better predictors than overall view ing fre quency. 136 CHAPTER 4 DISCUSSION Survey research conducted to identify the effects of exposure to television on children has usually operationalized "exposure to te le visio n " m easures in tw o ways: overall exposure to televisio n and exposure to tra d itio n a lly labeled genres, usually sim ilar or identical to those em ployed by the prim ary national ratings ser vice, the A.C. Nielsen Company. In in vestigating te levisio n's influence on children's acquisition o f beliefs about social reality, one m ajor program of research, the C ultural Indicators Project (e.g., G erbner et al., 1980) has em ployed the form er approach to m easuring exposure to television, w hile som e critics (e.g., Hawkins & Pingree, 1981) have em ployed the latter. A lthough th is la tte r approach takes into account program m ing differences w hich are rather obvious fro m an adult perspective (e.g., m ystery/suspense program m ing is d iffe re n t fro m situation com edy program m ing w hich is different from cartoon program m ing) and indeed has found genre differences in predicting effects, it had not been dem onstrated th a t these program m ing categories or genres w ere m eaningful to children. Hence it could be argued th a t although they m ight be better than m easures of overall exposure to television, exposure to tra d itio n a lly labeled genres m ig h t not be the best measure fo r predicting social reality effects on children. The present research was designed to provide data w hich w o u ld help resolve this problem in tw o ways. First, it was designed to id e n tify em pirically w hat dim ensions children o f diffe re n t ages em ploy in d istinguishin g am ong 137 television series. Second, it was to determ ine w h eth er grouping series based on perceived sim ilarity was a better w ay to predict te levisio n's effe cts on beliefs about social reality than the tw o m ore com m only em ployed means described above. From previous research on children and televisio n and on cognitive categorization it was concluded th a t children w ere likely to make judgm ents about series based on attributes in trin sic to the series them selves and on gratifications they associated w ith vie w ing the series. Based on general prin ciples of cognitive developm ent (cf., Flavell, 1970, 1977; Mussen, 1979) it was ex pected that older children w ould em ploy m ore content a ttrib u te dim ensions (HIa), m ore com plex content attribute dim ensions (H lb), m ore g ra tifica tio n dim ensions (H2a), and m ore com plex g ratifica tio n dim ensions (H2b) than you n g e r children. Again relying on general principles of cogn itive developm ent, it was expected that attribute dim ensions w ould resem ble g ra tifica tio n dim ensions m ore fo r younger children than fo r older children (H3). Finally, it was expected th a t mean exposure to clusters o f series w hich had sim ilar dim ensional stru ctu re w o uld be a better predictor of social reality beliefs than w ould mean exposure to series grouped by Nielsen categories, w hich in tu rn w ould be better than overall view ing (H4). 4.1 Age Differences in A ttrib u te Dim ensions In general, the results suggested th a t w ith increasing age children use m ore content attributes and m ore com plex content attributes to distin g u ish am ong television series. Based on the MDS analysis, second and sixth graders each 138 em ployed tw o attribute dim ensions to distinguish am ong series, w hile tenth graders em ployed three a ttrib u te dim ensions. U tilizing the content analysis to interpret the dim ensions suggested th a t the second graders' a ttrib u te dim ensions w ere conceptually quite sim ple, w h ile the sixth and tenth graders' w ere much more, though about equally, com plex. It appears th a t all three age groups em ploy an A ctio n/H um or dim ension to distinguish am ong television series. The action end of the dim ension is also associated w ith the concept of "crim e," w hile the hum or end is also associated w ith the concept o f "fam ily." For the older age groups, the hum or end o f the dim ension has the additional characteristic of "problem solving," w hich suggests th a t the older children are able to th in k about series as sim ultaneously funny and structured to illu strate the solving of som e social problem , clearly a m ore com plex conception of the series at th is end of the dim ension than the younger children have. Based on series placem ent on the dim ensions, all three ages can be said to em ploy a Realism dim ension also, though the sim ilarities am ong the ages are not great. Second graders seem unable to describe consiste ntly the characteristics of this dim ension, even though they can group cartoons to g e th e r and separate them from the live action series. Sixth graders make w hat adults w ould consider an accurate, though sim plistic, d istin ctio n in describing th is dim ension, associat ing cartoons w hich are at one end o f the dim ension w ith the concept o f "unreal" and prim arily m ystery/suspense series at the other end o f the dim ension w ith the concepts o f "action" and "crim e." The tenth graders make the same d istin ctio n s as the sixth graders, but they additionally associate cartoons w ith the concept of 139 being appropriate fo r a particular age group (younger children) and series at the other end of the dim ension w ith the concept o f "problem solving." These d if ferences again indicate increasing conceptual co m p le xity o f the dim ension w ith age. Finally, the tenth graders em ploy a third a ttrib u te dim ension w hich tends to separate the m agazine fo rm a t series from the narrative fo rm a t series, w ith the concept o f "novelty" associated w ith the fo rm e r and the conce pt o f "bad" as sociated w ith the latter. Although these results clearly, though not dram atically, su p p o rt the expec tations of age differences, fu rth e r analysis o f these data tends to m itigate this support. Canonical correlation analysis indicated that, based on the series' place m ents on the dim ensions, the second and sixth graders' dim ensional solutions w ere both m ore like the tenth graders' than they w ere like each others'. The reasons fo r this finding are no t entirely clear. It is possible th a t the second graders' data, because they are based on substantially few er cases than the sixth and tenth graders', are less stable than the older children's, and thus the dim en sional solution is not as accurate fo r them as fo r the o ld e r groups. There is som e support fo r th is interpretation in the RSQ results fo r th e d iffe re n t age groups. The second graders' tw o dim ensional solu tio n explained about 83% of the variance in th e ir data, w hile the sixth and te n th graders' tw o and three dim ensional solutions explained about 91% and 94%, respectively. A dditional support fo r th is interpretation com es from the clu ste r analysis, in w hich the CCC criterion fo r determ ining the optim al num ber of clusters w as not stable fo r the 140 second graders' attribute condition data, again suggesting that the second graders' dim ensions should be treated cautiously. A lternatively, or additionally, it may be th a t the sim ila rity of the second graders' solution to the tenth graders' solution is prim arily structural. That is, the canonical analysis results represent the relationship between the series' loadings on the dim ensions and not the conceptual content o f the dim ensions. This in te r pretation also receives som e support from the data. The content analysis label ing of the Realism dim ension suggests th a t the second graders have a d iffic u lt tim e providing an adequate conceptual label fo r th is dim ension, w h ile the sixth and tenth graders use fairly sim ilar, though not identical, concepts. Yet the canonical correlations fo r the second pair of variates fo r the second and sixth graders were only barely significant and accounted fo r only a little o f the overall shared variance between the tw o lo w er grades, w hile the corresponding variates fo r the sixth and tenth graders and second and tenth graders w ere highly sig nificant and accounted fo r alm ost as m uch variance as the firs t pairs of variates. This is due in part to the contribu tion of the tenth graders' th ird dim ension on these variates. Thus it may be th a t the sixth grade children are beginning to apply concepts in ways w hich are m ore like the tenth graders than like the second graders, but both the second graders and sixth graders have som e un articulated sense of series' sim ilarity w hich is like tenth graders', in particular re lated to series' sim ilarity as represented on the tenth graders' th ird dim ension. 141 4.2 Age Differences in G ratification Dim ensions The expectation th a t w ith increasing age children w ould em ploy m ore gratificatio n dim ensions and m ore com plex g ra tifica tio n dim ensions in d istin guishing am ong television series received only weak support. The results of the MDS analysis indicated that the second and sixth graders both em ployed four gratificatio n dim ensions w hile the te n th graders em ployed three. This result was directly counter to the expectation th a t the num ber of dim ensions w ould increase w ith age. The relationship o f the com ple xity o f the three sets o f dim ensions was not as expected either. A lthough the second graders' dim ensions w ere the sim plest conceptually, the sixth graders' rather than the tenth graders' dim ensions were the m ost complex. Based on the content analysis and MDS results, all three age groups em ployed an initial g ra tifica tio n dim ension w hich distinguished am ong those series w hich provide am usem ent and those w hich provide excitem ent. For the tw o older age groups, th is dim ension was also associated w ith o th e r concepts. For both sixth and tenth graders, those series tow ard the am usem ent end of the dim ension were also associated w ith the gratificatio ns "general pleasure" or liking and "social guidance." For sixth graders only, those series tow a rd the negative end w ere associated w ith "general displeasure," or disliking. Beyond th is dim en sion, there w ere none consiste nt enough in th e ir gratificatio ns o r series' place m ents to be considered com m on to all groups. Second graders em ployed tw o dim ensions w hich w ere each associated w ith only a single gratification. These dim ensions were called Simple A m usem ent and 142 Simple Liking. N either of these corresponded to a sim ilar g ra tifica tio n dim ension fo r the sixth or tenth graders. The second graders' fo u rth g ra tifica tio n dim ension, Learning/Displeasure, was related conceptually to the tenth graders' th ird dim en sion, w ith both associating series at one end of the dim ension w ith the g ra tifica tio n "learning" and series at the other end w ith "general displeasure," or disliking. The sixth and tenth graders' second g ra tifica tio n dim ensions were quite sim ilar, both called Learning/D iversion. Both age groups em ployed the g ra tifica tio n s "learning," "involvem ent," and "excitem ent" fo r this dim ension and as sociated them w ith series at the positive end of the dim ension. Tenth graders additionally associated the g ratifica tio n "social guidance" w ith the series at the positive end. Both grades associated "general displeasure" or disliking w ith the negative end of this gratification dim ension; additionally, sixth graders associated the gratifications "passing tim e" and "am usem ent" w ith th is end. W hile the MDS and content analysis results provided no or weak support fo r the hypotheses regarding age differences in g ra tifica tio n dim ensions, the canonical correlation analyses results were m ore in the expected direction, w ith the second and tenth graders' series' placem ents on th e ir dim ensions both resem bling the sixth graders' m ore than they did each others'. Again, these d if ferences were not dram atic, but they were nonetheless clear. The unexpected findings in the g ra tifica tio n cond ition results are som ew hat m ore d iffic u lt to explain than those in the a ttrib u te condition. As w ith the a t trib u te condition results, the num ber o f second graders supplying data is sub 143 stantially less than the num ber of sixth and tenth graders, and thus the dim en sional solution fo r the youngest group may be less stable than fo r the older groups. The relative m agnitudes o f the different groups' RSQs support this pos sibility, as they did in the a ttrib u te condition. However, this explanation cannot apply to the relationship between the sixth and te n th graders, since the num bers of children in these tw o groups are com parable, as are th e ir RSQs. The data un equivocally, though not dram atically, suggest th a t sixth graders em ploy m ore and m ore com plex gratificatio n dim ensions in distinguishing am ong series than do tenth graders. Perhaps the reason fo r th is unexpected finding is the reliance on general principles of cognitive developm ent to form ulate the hypotheses rather than on developm ental changes m ore dire ctly concerned w ith the particular content area — the child's perception o f self v is -a -v is one elem ent of the outside w orld. A l though it is reasonable to expect th a t w ith increasing age children w ill em ploy increasing num bers of increasingly com plex dim ensions in thinking about some aspects of their environm ent, such may not be the case in th in kin g about th e m selves. It may be that m iddle childhood and early adolescence are tim es of ex perim enting w ith ideas about self, and th a t by mid to late adolescence such ideas have settled into a fa irly coherent and sim plified set (Markus & Nurius, 1984). To the extent that gratifications associated w ith television view ing can be th o u g h t of as requiring a heightened awareness o f self, the data from the sixth graders may reflect som e o f this process. That is, the greater num ber o f m ore com plex dim ensions em ployed by the sixth graders as com pared w ith the tenth graders 144 may be due to the younger group being o f an age in w hich they are tryin g to figure out w hy they do things and w h a t makes them act and feel the w ay they do. Asking them their reasons fo r w atching television series, or w h a t the series make them feel or think about, thus may be likely to evoke a greater range of m ore com plicated ideas than identical questions presented to te n th graders. Furthering this interpretation are the differences in view ing patterns w hich typically occur across ages. Lyle and Hoffm an (1972) found th a t youn ger children w atch both less television overall and a narrow er range of types o f program m ing than older children. Children in early adolescence w atch approxim ately the same variety o f program m ing as older teenagers, but w atch m ore to ta l hours o f it. Television view ing thus occupies m ore tim e fo r children of the m iddle age group than fo r either of the other tw o groups, w hich has been in terpreted in the past (cf.. Brown, 1976) as suggesting th a t televisio n is m ore central to the lives of children of this age than to older or younger children. 4.3 Differences Between A ttribute and G ratification Dim ensions W ithin Age The original expectation was th a t attrib ute dim ensions and g ra tifica tio n dim ensions w ould be m ore differentiated w ith age (H3). The results w ere in the exact opposite direction: The older the group of children, the m ore alike w ere th e ir attrib ute and gratification dim ensions. The reasoning behind the original hypothesis again relied on general principles of cogn itive developm ent. In general, the older children become, the m ore able they are to d istinguish objects' intrinsic attributes from the objects' functions fo r children them selves. Though it is no doubt the case that children's ability to do so increases w ith age, it is also 145 the case th a t at least in respect to som e objects, children may tend to m ove tow ard a m ore integrated and stable vie w o f ob je cts' a ttrib ute s and functions. In the case o f television, it may be th a t children develop conceptions o f w hat fu n c tions are associated w ith w hat attributes of program m ing, and th a t these concep tions become relatively stable over tim e. By tenth grade, then, children appear to have developed a conception o f television program m ing in w hich the fun ctio ns a given series has fo r them are system atically related to the series' attributes. 4.4 Content Analysis and M ultidim ensional Scaling as M eans of Representing Children's Judgm ents of Television Series' Sim ilarity The above discussion suggests th a t the findings based prim arily on the content analysis and the findings based prim arily on the MDS analysis were not in com plete agreem ent. That is, the conceptual analysis of each grade's attrib ute and gratification dim ensions based on the content analysis suggested slig h tly d if ferent in ter-gra de relationships of dim ensions than did the canonical analysis of the MDS results. The m ost likely reason fo r this difference is the constraints im posed by each m ethod. Both m ethods relied in itia lly on the same data — ju d g m ent of sim ilarity of the television series. The MDS analysis presum es th a t sim ilar dim ensions are being em ployed by all respondents w ith in age groups, and thus the sim ilarity judgm ents can be represented geom etrically. The dim ensions on which the series are then located are technically, however, content free, since the criteria on w hich the judgm ents are made are not "know n" by the scaling routine. The content analysis picks up at precisely th is point. By requiring the par 146 ticip ants to verbalize the criteria they used in m aking th e ir sim ila rity judgm ents, som e sense o f the content of the dim ensions being em ployed could be obtained. There is nonetheless a problem inherent in tryin g to id e n tify the conceptual aspect o f the dim ensions by em ploying co n te n t analysis. It is th a t content analysis requires th a t the data be partitioned into m utually exclusive and exhaus tive categories. W hile the coding schem e em ployed here m eets these criteria, it w ould o f course be possible to create m ore categories by dividing up existing ones into com ponent parts. It w ould also be possible to create fe w e r categories by com bining existing ones w hich may be argued to be conceptually related. C ontent analysis, then, cannot be expected to supply com pletely adequate in fo r m ation on the conceptual com position o f the dim ensions. It m ight be, then, that by altering the coding schem e som ew hat a better "fit" between the content analysis results and the MDS results could have been obtained. However, given the p o ssib ility o f w idely diverging results, those found here seem quite sensible. Children o f all three ages were able to carry out both tasks (sim ilarity judgm ents and supplying verbal descriptions of w hat made any group o f series sim ilar) and th e ir responses fo r the m ost part make good sense. 4.5 Predicting Beliefs about Social Reality Using the results of the MDS analysis fo r each condition, it was possible to id entify clusters of series fo r each grade level w hich were grouped based on the overall sim ila rity o f th e ir a ttrib u te or g ra tifica tio n dim ension loadings. Average view ing frequency o f the series w hich com prised each clu ster was then used to predict each o f fo u r (for second graders) or five (for sixth and tenth graders) so 147 cial reality beliefs. There was fa irly strong confirm ation o f th e expectation th a t mean ex posures to both Nielsen categorized series and a ttrib u te o r gratification categorized series were better predictors o f the outcom e m easures than was overall view ing (H4). The relative effectiveness o f the Nielsen categorized series, the attribute categorized series, and the g ra tifica tio n categorized series differed fo r each age, however. Based on the num ber o f outcom es predicted and the m agnitudes of the R2s, series clustered by a ttrib u te s w ere best fo r second graders, series clustered by gratificatio ns w ere best fo r sixth graders, and series categorized according to trad ition al Nielsen genres w ere best fo r tenth graders. The reasons fo r these findings are discussed below ; the reader is referred to Tables 27-29, 30-32, and 5 fo r the series w hich com prised the predicting at tribute, gratification, and Nielsen clusters, respectively. Second graders Two out of the fo u r outcom e m easures w ere predicted fro m exposure to one or m ore of the attribute clustered series. Beliefs about the frequency of fam ily m em bers providing support (SUPPORT) fo r one another w ere negatively related to exposure to a clu ste r o f series located at the negative end of the Realism dim ension (all cartoons) and p ositively related to exposure to a cluster located at the positive end o f the Realism dim ension (six series divided equally between w hat w ould be categorized according to Nielsen te rm in o lo g y as General Drama and Variety). This la tte r clu ster was also positive ly related to beliefs about the w orld being a mean place (MEANWLD). A lthough the second graders were 148 not able to provide conceptual labels w hich w ould have clearly described the Realism dim ension, it is not d iffic u lt to provide an explanation fo r w hy this relationship should hold. Some o f the series at the positive end o f the Realism dim ension are dramas w hich often present a fam ily, or a group o f people w ho function as a fam ily, w ho provide support fo r each other in the face o f som e d if fic u lty posed by the outside w orld. Thus increasing exposure to these series be ing related to increasing scores on SUPPORT and MEANWLD seem s to be a plausible relationship. How othe r series (variety series) in this clu ste r operate is m ore problem atic, although the vignettes presented in them ofte n portray people overcom ing obstacles, usually w ith the help o f others, or show a num ber of people m aking m usic to g ether w hich may sym bolize cooperative effo rt. It should be noted th a t th is cluster o f series falls about in the m iddle o f the H u m or/A ction attribute dim ension, suggesting th a t fo r second graders these series possess a m oderate am ount of each, perhaps ju st enough to make them a ttra ctive and in volving w ith o u t being so hum orous th a t they are discounted o r so exciting that th e ir m essages are missed. The cluster of series at the negative end of the Realism dim ension — the cartoons — has a relationship to the SUPPORT outcom e fo r the second graders w hich is less easily understood. Since it is a negative relationship, there should be som ething about th is program m ing fo r w hich increased exposure should be expected to depress beliefs about the frequency o f cooperation and support w ith in fam ilies. Yet fo u r o f the five series portray fam ilies w h ich could be characterized as ranging from m oderately to quite supportive. One possible 149 reason fo r the negative relation to SUPPORT lies in th is cluster's placem ent on the second graders' dim ensions. By being at the extrem e negative end o f the Realism dim ension and close to the "action" end of the H u m or/A ction dim ension (at least closer than the o ther predicting cluster), the series in th is clu ste r may focus second graders' attention on the a ctio n -o rie n te d aspects o f the content to the point where the m ore prosocial messages about fa m ily support are either overlooked or cast in an unbelievable light. Sixth graders For sixth graders, fo u r of the five outcom e m easures w ere predicted by mean exposure to one or m ore groups o f series w hich were clustered by sim ilar gratification dim ension loadings. In addition to the tw o outcom es, SUPPORT and MEANWLD, ju st discussed fo r second graders, tw o d iffe re n t m easures o f beliefs about the am ount o f violence in the world, SINGERV and GERBNERV, were predicted. Beliefs about the am ount o f support in fam ilies were negatively related to one cluster o f three series and positively related to a clu ster o f eight series which was identical to the Nielsen M ystery/Suspense genre. A lth ough the series in the latter cluster do not feature fam ilies, m any of them portray a sm all group of people w ho w ork together, often under the th re a t of physical violence, and count on each other's support fo r survival. This clu ste r o f series rated fairly highly on the Excitem ent and Learning gratificatio n dim ensions, w hich, com bined w ith their typical content, may explain th e ir relationship to SUPPORT. The other cluster of series (which is dom inated by Variety series) does not typ ica lly have 150 content w hich one w ould expect to be negatively related to beliefs about the am ount o f support in fam ilies, nor do the g ra tifica tio n s w ith w hich it is stro n g ly associated suggest the reason fo r its negative relationship w ith SUPPORT. The three clusters o f series w h ich predicted beliefs that the w o rld is a mean place are m ore easily understood. A clu ster com prised o f three fam ily series, tw o of w hich were dramas, and a clu ste r com prised of three seem ingly unrelated series w ere negatively related to th is outcom e. The fam ily series all prom ote positive social values and w ere located at the extreme positive end of the Social Guidance gratificatio n dim ension. Thus increasing exposure to these series could be expected to lead to decreased beliefs that the w o rld is a mean place. The other clu ster w ith the negative relationship to MEANWLD is m ore d if fic u lt to explain. It was at the opposite end of the Social Guidance dim ension from the previous cluster, and the series w hich com prise it do not all have sim ilar content. Possibly th is clu ster operates in the same way as the Variety dom inated cluster did fo r the second graders, in th a t the cooperation portrayed — a prom inent them e in at least tw o of the series — works to decrease beliefs th a t the w o rld is a mean place. The clu ster w hich is positively related to MEANWLD is com prised o f fo u r series w h ich can be characterized as slapstick com edies. This cluster is positioned slig h tly tow ard the negative end o f the So cial Guidance gratificatio n dim ension and tow ard the positive end o f the A m usem ent/E xcitem ent gratificatio n dim ension. These gratifications, in com bina tion w ith the often extrem ely adversarial relationships portrayed in these series, m ight com bine to increase beliefs th a t the w o rld is a mean place. 151 Four clusters predicted one o f the tw o indicators o f beliefs about the am ount of violence in the w orld (SINGERV); tw o were p o sitive ly related and tw o were negatively related. The tw o w ith negative relationships also predicted other previously discussed outcom es. The cluster dom inated by Variety series m ight be negatively related to beliefs about the am ount o f violence in the w o rld be cause of the series' em phases on fantasy or the bizarre, and because they are regarded as providing som e in form ation about the w orld, as evidenced by their location at the "learning" end o f the Learning/D iversion g ra tifica tio n dim ension. The cluster com prised of the slapstick series m ight be negatively related to the SINGERV outcom e fo r sim ilar reasons. Despite its positive relatio nship w ith the belief that the w orld is a mean place, the extreme slapstick nature o f these series may prevent th e ir conveying much inform ation w hich could be regarded seriously about the extent to w hich one w ould be likely to encounter violence in the real world. Of the tw o clusters w hich w ere positively related to SINGERV, one was com prised solely of cartoons and the other of three General Dramas. It may be th a t cartoons are negatively related to SINGERV fo r sixth graders fo r m uch the same reasons they were negatively related to SUPPORT fo r second graders. A l though these cartoons are not extrem ely violent, th e ir placem ent at the extrem e "diversion" end o f the Learning/D iversion gratifica tio n dim ension, and the m oderately high level o f excitem ent associated w ith them based on th e ir place m ent on the A m usem ent/E xcitem ent dim ension could com bine to m inim ize w hatever "prosocial" m essages they may contain but allow w h a t vio le n t content 152 there is to com e through. The clu ste r com prised o f the General Dramas may operate to increase beliefs about violence in the w o rld by portraying real life co n flict in a m anner devoid of extrem e am usem ent or excitem ent, as evidenced by its placem ent in the m iddle of the A m usem ent/E xcitem ent g ra tifica tio n dim en sion and by its relatively high placem ent on th e learning ends o f the Learning/Excitem ent and Learning/D iversion g ra tifica tio n dim ensions. The one gratification cluster w hich predicted the other in dica tor o f beliefs about the am ount o f violence in the w orld (GERBNERV) fo r the sixth graders was the same one as described im m ediately above. It was com prised o f three series w hich w ould be categorized by Nielsen as General Drama and was p ositively re lated to the outcom e, probably fo r the same reasons as it was related to SIN GERV. Tenth graders The clusters o f series in both the attrib ute and gratificatio n conditions resem bled each other noticeably and resem bled the groups o f series categorized according to Nielsen genres to som e extent, too. A lthough the same outcom e measures were predicted by all three of these classification system s, as indeed they were by overall view ing frequency, the mean exposure to the Nielsen cate gory o f General Drama produced equations in w hich the greatest am ount of variance was explained. Unlike the second and sixth graders, neither beliefs about support in fam ilies nor beliefs about a mean w o rld were predicted. In stead, mean exposure to General Drama program m ing was positive ly related to beliefs about the extent to w hich fam ily m em bers com ply w ith others' w ishes and 153 to the tw o measures o f beliefs about the am ount of violence in the w orld. This General Drama program m ing includes series in w hich social problem s are ty p i cally the central dram atic concern and in w hich neither action nor hum or play a central role. W hile the results o f the MDS and content analysis are not d irectly applicable to the Nielsen clusters, exam ination o f the General Drama series' placem ents in the attribute and g ra tifica tio n solutions may provide som e clues as to w hy this genre is the best p re d icto r o f the outcom e measures. If looked at as a cluster, the General Drama series tend to fall tow ard the positive ends o f the Realism attribute dim ension and the Learning/Diversion g ra tifica tio n dim ension, suggesting that they can be taken seriously. On the H u m or/A ction a ttribute dim ension and the A m usem ent/E xcitem ent gratification dim ension, this genre is positioned around the m iddle, suggesting an absence of extrem e hum or or action w hich m ight distract from or lead to devaluing the content. Summ ary Looking at these findings fo r all three age groups, there are tw o general conclusions w hich one can make. First, w hile m ost of the relationships between exposure and effects make sense, there are a num ber o f relationships w hich could have been expected to em erge but were not found in the data. M ost notably, w ith only tw o exceptions out o f eight possibilities, there was no consis te n t relationship between the Nielsen category of M ystery/Suspense and the three mean w orld or violence in society outcom es. There were, however, relationships between a num ber o f other categories and these outcom es. This finding suggests th a t futu re attention needs to be paid to a varie ty of kinds of 154 program m ing, how ever they are categorized. Second, the "age progression" of best predictors goes from series categorized by perceived attributes to series categorized by perceived g ra tifica tions to series categorized by trad ition al adult (Nielsen) genres. Perhaps this should com e as no surprise. A lthough this study has dem onstrated th a t children as young as seven are able to categorize television series by the fu n ctio n s they serve fo r them , it is probably the case, as developm entalists w ould expect, th a t "superficial" perceptual attributes are still the m ost salient aspect o f the series fo r this age level and th a t the categories form ed by grouping the series accord ing to these attributes are the m ost pow erful cogn itive "filte r" through w hich the content may pass. Sixth graders, on the other hand, w hile equally capable of sorting series according to perceptual attributes or gratifications, evidence a clear superiority in the predictive pow er o f exposure to series grouped by gratification. To the extent that children in early adolescence are preoccupied w ith concerns about w ho they are and how they "fit into the w o rld " and th a t television occupies a very im portant place in th e ir w orld, the pow er o f television series grouped by perceived function to predict outcom es should com e as no surprise either. And tenth graders, fo r w hom grouping series by perceived attributes or g ra tifica tio ns produces very sim ilar results, appear to be adult enough to have arrived at a relatively stable connection between television series' attributes, series' g ra tific a tions, and the Nielsen categories w hich can be used to describe the series. Thus treating the television diet of this age child as com prised of trad ition al Nielsen genres seems appropriate. 155 4.6 Lim itations of the Study This study has a num ber of lim ita tio n s w hich should be taken into account when considering its results. The m ost glaring lim ita tio n is its lack of generalizability in the classic (Campbell & Stanley, 1966) sense. Since the children w ho participated in the study w ere a convenience, rather than a random, sam ple o f w hite, m iddle class, English speaking children, the results cannot be extended to all com parable children. Sim ilarly, the televisio n series about w hich the children w ere questioned were not a random sam ple o f all televisio n pro gram m ing or even o f program m ing w ith in the chosen Nielsen categories. Thus it could be that the dim ensions id entified and the clusters of series used to predict social reality beliefs are peculiar to th is study. W hile the lack o f generalizability is certainly a valid criticism , there are several reasons w hy it should not be considered a serious draw back to the present study. First, there is little reason to believe th a t the participants in the study are m arkedly different from children elsewhere. They w ere recruited from both public and private schools w hich are located in m iddle class neighborhoods. The experim enters' interactions w ith them and conversations w ith th e ir teachers and the adm inistrators at th e ir schools all suggested th a t they w ere in no way atypical. Second, although the series chosen fo r the study w ere not a random sample, they do represent the m ost heavily viewed series fo r these age groups. W hile they cannot be presum ed to represent all program m ing in th e ir respective genres, they can be argued to be the p rototyp ical series fo r th e ir genres, and 156 thus in that sense a good representation of children's television diet. A th ird reason w hy the issue of generalizability should not be treated to o seriously is m ore pragm atic. It is d iffic u lt to im agine conducting a study designed to answ er the questions posed in the present one w hich w ould not be open to such criticism . The only com parable study w hich could be asserted to be generalizable was conducted by Frost (1969). In it, a random sam ple o f 750 British adults rated tw o of 61 series on 58 scales; the 61 series w ere v irtu a lly all o f the BBC and ITV w eekly broadcast schedule. O bviously, the num ber of problem s one w ould face fo llo w in g a sim ilar strategy in this co un try at th is tim e w ould be prohibitive. Unlike Britain, the broadcast schedule is not uniform in all parts of the country, the num ber of w eekly series is much greater than 61, the num ber o f children required to respond to enough subsets of the program m ing schedule to provide a com prehensive representation o f it w ould be extrem ely high, and the necessity fo r fa c e -to -fa c e in tervie w s to use th is study's procedures w ould be extraordinarily expensive. In sum, it is unlikely th a t the issue of generalizability could be adequately addressed by recourse to random sam pling of subjects and series. The only practical approach, then, is to approxim ate generalizability through replication using diffe re n t pools of participants and pools of series w hich vary in the extent to w hich they are sim ilar to the series used in th is study. Finally, it could be argued th a t the results fo r second graders are par ticu la rly suspect because of the relatively low num ber of children of th a t age w ho participated. It is d iffic u lt to be certain w hy such a much low er percentage 157 o f second grade parents gave th e ir perm ission fo r th e ir children to participate than did sixth grade parents. Since the second graders all cam e from the same schools as the sixth graders, it is unlikely that there w ere any local differences w hich w ould explain the differential response rates. The only available explana tion is that parents o f second graders were m ore concerned about the loss of class tim e than were the parents o f sixth graders. This was a rem ark freq uently made by adm inistrators, though it was im possible to determ ine w h eth er it applied m ore to second graders than to older children. In any case, it is especially im portant to reassess the stability o f the second graders' results by fu rth e r study. 4.7 Directions for Future Research Categorization of programming: Dim ensions and clusters This study represents an initial atte m pt to describe the dim ensions children use to discrim inate among or categorize televisio n series, and to determ ine the extent to w hich the dim ensions allow characterization o f program m ing in ways sim ilar to or different from adult characterization of program m ing. W hile the results, at least fo r sixth and tenth graders, seem quite clear, it w ould be prem a ture to suppose that the dim ensions described in this study apply to all sixth and tenth graders and all sim ilar series. Further research, em ploying d iffe re n t series and children, should of course be conducted, perhaps using som e o f the results reported here. Scales could be developed w hich could be used to check the validity o f the dim ensions in m ore trad ition al ways, such as asking children to rate each series on the scales a fte r they make th e ir judgm ents o f series' sim ilarity, then regressing the series' scale scores on the series' loadings on the 158 dimensions (cf.. Reeves & Greenberg, 1977). It w ould be especially im portant to pursue this course of research w ith younger children such as the second graders w ho participated in th is study. It was suggested earlier that there may be a progression in predictive pow er o f ex posure to series beginning w ith series categorized by th e ir perceived attributes, through exposure to series categorized by th e ir ascribed gratifications, to the fu sion of these tw o kinds of dim ensions as represented by Nielsen categories. The findings of age differences in the predictive pow er o f the dim ensions w hich sup port this interpretation rely, at least in part, on the degree to w hich second graders' clusters o f series based on a ttrib u te dim ensions are better predictors of social reality beliefs than Nielsen categorized series. This assertion o f course needs testing w ith a larger sam ple of second graders, either alone or in com parison w ith older children. A lternative approaches to these issues could stress the "m em bership" of series in groups or clusters. For instance, children could be presented w ith various groupings o f series such as Nielsen groups, the clusters found here, or other conceptually derived groups, supplied w ith descriptive labels fo r the groups and then be asked to tell w hich o f the groupings made the m ost sense and why. Such an approach w ould have the value of face va lid ity and, to the extent it sup ported the findings o f this study, provide convergent validity fo r them . Predicting beliefs about social reality The results o f the present study suggest an alternative to the tw o m ore 159 com m on ways of predicting social reality beliefs, at least fo r second and sixth graders. C ategorizing program m ing according to perceived attrib ute s or gratifications and using view ing frequency o f the resulting clusters of series may provide a better means than those now used of representing how exposure to television results in effects on children. It m ight be argued th a t the results reported here have inflated R2s due to the use o f stepw ise regression procedures (Hocking, 1983). However, given that the results controlled fo r SES and ethnicity, and in effect controlled fo r age by running analyses separately fo r each age group, the overall R 2s are at least as good as any reported by G erbner et al. (1979, 1980) and probably better than the results reported by Hawkins and Pingree (1981), since they report only correlations. It is thus not inappropriate to take these results seriously and consider em ploying th is m ethod of grouping series to predict exposure effects on beliefs about social reality. An alternative to the strategy proposed above is to use som e form of con te n t analysis to identify series' variations in o u tco m e -re le va n t content in order to w eight m easures of exposure to those series. A few studies have em ployed this approach to studying effects. For instance, D om inick and G reenberg (1972) and McLeod et al. (1972a, 1972b) em ployed rankings o f the violence of program m ing content to predict attitudes tow ard aggressive behavior and reports o f delin quency and aggression. Small to m oderate relationships w ere found in all the studies, and in the tw o McLeod et al. (1972a, 1972b) surveys the relationships varied considerably according to the type o f program m ing (e.g., crim e -d e te ctive vs. situation com edy). In a study carried out at approxim ately the same tim e as 160 those cited above, Milavsky, Stipp, Kessler, and Rubens (1982) used adult rankings of series' violence com bined w ith exposure to these series to predict s e lf-re p o rts o f aggression fo r elem entary school children and teenage boys over the course o f three years. More recently, Buerkel-R othfuss et al. (1982) em ployed an exten sive con tent analysis o f fam ily series to derive rankings on several variables w hich were then used to w eight the view ing frequency o f the series. These vari ables, in conjunction w ith children's perceptions of the realism of the television fam ilies, o f w hat they believed they learned o f fa m ily life from TV, and of several kinds o f parental attitudes and behaviors tow ard television and fam ilies, w ere used to predict reports of four kinds of fa m ily behavior: supporting, com plying, opposing, and ignoring. The firs t tw o o f these outcom es were sig n ifica ntly predicted from exposure, as they were in the present study also. Perhaps the m ost appropriate direction fo r survey research to take in fu tu re studies o f the effects of exposure to television on children is a com bination of the approach taken in the present research and th a t em ployed in studies described above. By em ploying results from content analyses o f series to w e igh t view ing frequency o f series categorized by perceived attributes fo r younger children, by perceived gratifications fo r "m iddle-aged " children, and by trad itional Nielsen genre fo r children in their m id -te e n s and above, even better prediction m ight occur. Such an approach would need to be carefully th o u g h t out, however. To rely on adult created and executed coding of content, as did those studies described above, m ight result in ways of representing the content w hich w ould differ from those that children use. As the present study has dem onstrated, 161 younger children (second and sixth graders) do not classify televisio n series in quite the same ways adults do. Analogous differences m ight also hold fo r representing the am ount o f d iffe re n t kinds of content in series. W hile d e te rm in ing w hether children's judgm ents about am ount or frequency o f d iffe re n t kinds of content is sim ilar to adults' rem ains an em pirical question, it is not unreasonable to expect that both children's ow n categorization o f program m ing and (child or adult independent rating of) variations o f content in th a t program m ing should com bine to enhance prediction o f learning or social reality effects. Testing the efficacy of such an approach is, o f course, up to those interested in continuing the attem pt to build m ore com prehensive m odels of the televisio n effects process. 162 REFERENCES Anderson, J. (1981). Research on children and television: A critique. Journal of Broadcasting, 25, 395-400. Alexander, A. (1980, May). Children's perceptions of television characters: V alidity in the m ultidim ensional scaling approach. Paper presented at the annual m eeting o f the International C om m unication Association, Acapulco. Bandura, A. (1977). Social learning th e o ry. Englewood Cliffs, NJ: P rentice-H all. Bandura, A., 8i Barab, P.G. (1971). C onditions governing nonreinforced im itation. Developm ental Psychology, 5, 244-255. Bantz, C.R. (1982). Exploring uses and gratifications: A com parison of reported uses o f television and reported uses o f favorite program type. C om m unication Research. 9, 352-379. Beals, R., Krantz, D.H., 8t Tversky, A. (1968). Foundations o f m ultidim ensional scal ing. Psychological Review. 75, 127-142. Blumler, J.G. (1979). The role o f theory in uses and gratificatio ns studies. C om m unication Research. 6, 9-36. Blumler, J.G., 8 c Katz, E. (Eds.). (1974). The uses o f mass com m unications. Beverly Hills: Sage. Brown, J.R. (1976). Children's uses of television. In R. Brown (Ed.), Children and te le visio n . Beverly Hills: Sage. Bruner, J.S., Olver, R.R., 8t Greenfield, P.M. (1966). Studies in cogn itive g ro w th . New York: W iley. Buerkel-Rothfuss, N.L., Greenberg, B.S., Atkin, C.K., 8c Neuendorf, K. (1982). Learn ing about the fam ily from television. Journal of Com m unication, 32(3), 191-201. Buerkel-Rothfuss, N.L., 8 c Mayes, S. (1981). Soap opera view ing: The cultivation effect. Journal o f C om m unication. 31(3), 108-115. 163 Campbell, D.T., & Stanley, J.C. (1966). Experim ental and qua si-experim ental designs fo r research. Chicago: Rand McNally. Carlson, J.M. (1983). Crime show view ing by preadults: The im pact on attitudes tow ard civil liberties. C om m unication Research, 10, 529-552. Chatterjee, B.B., 8t Eriksen, C.W. (1962). C ognitive factors in heart rate c o n d itio n ing. Journal o f Experim ental Psychology, 64, 272-279. Comstock, G., Chaffee, S., Katzman, N., McCombs, M., 8*. Roberts, D. (1978). Television and human behavior. New York: Colum bia U niversity Press. Cooley, W.W., 8t Lohnes, P.R. (1971). M ultivariate data analysis. New York: W iley. Cooper, L.G. (1973). A m ultivariate investigation o f preferences. M ultivariate Be havioral Research. 8, 253-272. Cooper, L.G. (1981). Individual and group differences in m ultidim ensio nal scaling: Problems in partially nonlinear structual m odeling (W orking Paper No. 103). Los Angeles: U niversity of California Los Angeles, Graduate School o f Management. Copleston, F. (1964). A history of philosophy (Vol. 5). M odern philosophy: The British philosophers (Part I). Garden City, NY: Image. Dawson, M.E., 8t Furedy, J.J. (1976). The role of awareness in human differential autonom ic classical conditioning: The necessary gate hypothesis. Psychophysiology, 13, 50-53. Dominick, J.R., 8t Greenberg, B.S. (1972). A ttitudes tow ard violence: The in te r action of television exposure, fam ily attitudes, and social class. In G.A. C om stock 8t E.A. Rubinstein (Eds.), Television and social behavior (Vol. 3). Television and adolescent aggressiveness (pp. 314-335). W ashington, DC: U.S. G overnm ent Printing Office. Dorr, A. (1983). No shortcuts to judging reality. In J. Bryant 8i D.R. Anderson (Eds.), Children's understanding o f te le visio n . New York: Academ ic Press. Dorr, A., Kovaric, P., Doubleday, C., Sims, D„ 8i Beizer Seidner, L. (1985, August). Beliefs about the realism o f television program s featuring fam ilies. Paper presented at the annual m eeting o f the Am erican Psychological A ssocia tion, Los Angeles. 164 Doubleday, C.N. (1985). Children's perceptions of televisio n fam ilies. Unpublished doctoral dissertation, U niversity of California, Los Angeles. Elliot, W.A., 8 1 Q uattlebaum , C.P. (1979). S im ilarities in patterns o f m edia use: A cluster analysis of media gratifications. W estern Journal o f Speech Com m unication, 43, 61-72. Fetler, M. (1984). Television view ing and school achievem ent. Journal of Com m unication, 34(2), 104-118. Flavell, J.H. (1970). C oncept developm ent. In P.H. Mussen (Ed.), Carm ichael's manual o f child psychology (Vol. 1). New York: W iley. Flavell, J.H. (1977). C ognitive developm ent. Englewood C liffs, NJ: Prentice-H all. Frank, R.E., 8t Greenberg, M.G. (1980). The public's use o f te le visio n . Beverly Hills: Sage. Frost, W.A.K. (1969). The developm ent o f a technique fo r tv program m e assess ment. Journal of the M arket Research S ociety, 11, 25-44. Fuhrer, M.J., 8i Baer, P.E. (1965). D ifferential classical cond itio n in g : Verbalization of stim ulus contingencies. Science, 150, 1479-1481. Gantz, W. (1980, May). Exploring the m otives fo r vie w ing televised sports. Paper presented at the annual m eeting of the International C om m unications A s sociation, Acapulco. Gerbner, G., & Gross, L. (1980). The vio len t face o f televisio n and its lessons. In E.L. Palmer 8i A. Dorr (Eds.), Children and the faces o f te le visio n — teach ing, violence, selling (pp. 149-162). New York: Academ ic Press. Gerbner, G., Gross, L., Morgan, M „ 8i Signorielli, N. (1980). The "m ainstream ing" of Am erica: Violence Profile No. 11. Journal o f C om m unication, 30(3), 10-29. Gerbner, G., Gross, L., S ignorielli, N., Morgan, M., 8t Jackson-Beeck, M. (1979). The dem onstration o f pow er: Violence Profile No. 10. Journal of C om m unication, 29(3). 177-196. Gibson, E.J. (1969). Principles of perceptual learning and developm ent. New York: A p p le to n -C e n tu ry-C ro fts. 165 Greenberg, B.S. (1974). G ratifications of television view ing and th e ir correlates for British children. In J.G. Blum ler & E. Katz (Eds.), The uses of mass com m unications (pp. 71-92). Beverly Hills: Sage. Greenberg, B.S. (1980). Life o f television: C ontent analyses of U.S. tv dram a. Norwood, NJ: Ablex. Greenberg, B.S., 8t Reeves, B. (1976). Children and the perceived reality o f te le v i sion. Journal of Social Issues, 32, 86-97. Grings, W.W. (1973). The role o f consciousness and co g n itio n in autonom ic be havior change. In F.J. McGuigan 8i R. S choonover (Eds.), The psychophysiology of thinking (pp. 233-262). New York: Academ ic Press. Hawkins, R.P. (1977). The dim ensional structure of children's perceptions of television reality. C om m unication Research, 4, 299-319. Hawkins, R., 8i Pingree, S. (1981). U niform content and habitual view ing: Un necessary assum ptions in social reality effects. Human C om m unication Research, 7, 291-301. H im m elw eit, H.T., Sw ift, B., 8t Jaeger, M.E. (1980). The audience as critic: A con ceptual analysis o f televisio n entertainm ent. In P.H. Tannenbaum (Ed.), The entertainm ent functions o f television (pp. 67-106). Hillsdale, NJ: Erlbaum. Hirsch, P. (1980). The "scary w o rld " of the nonview er and o th e r anom olies: A reanalysis of Gerbner et al.'s findings on cu ltiva tio n analysis, part I. C om m unication Research. 7, 403-456. Hirsch, P. (1981a). On learning from one's own m istakes: A reanalysis o f Gerbner et al.'s findings on cultivation analysis, part II. C om m unication Research, 8, 3-37. Hirsch, P. (1981b). D istinguishing good speculation from bad theory: Rejoinder to G erbner et al. C om m unication Research, 8, 73-96. Hocking, R.R. (1983). Developm ents in linear regression m ethodology: 1959-82. Technom etrics, 25, 219-230. Jain, A.K., 8i Pinson, C. (1976). The effect o f order of presentation of sim ilarity judgm ents on m ultidim ensional scaling results: An em pirical exam ination. Journal o f M arketing Research. 13, 435-439. 166 Katz, E., Blumler, J.G., 8c G urevitch, M. (1974). U tilization o f mass com m unication by the individual. In J.G. Blum ler & E. Katz (Eds.), The uses o f mass com m unications (pp. 19-32). Beverly Hills: Sage. Katz, E., Gurevitch, M., & Haas, H. (1973). On the uses o f mass m edia fo r im p o r tant things. Am erican S ociological Review. 38, 164-181. Kelly, H. (1981). Reasoning about realities: Children's evaluations o f te le visio n and books. In H. Kelly 8c H. Gardner (Eds.), Viewing children thro ugh television (pp. 59-72). San Francisco: Jossey-Bass. Klapper, H.L. (1981). Children's perceptions o f the realism o f televised fiction: New w ine in old bottles. In J.F. Esserman (Ed.), Television advertising and children: Issues, research, and findings (pp. 55-82). New York: Child Research Service. Kovaric, P., Dorr, A., 8c Nicol, J. (1983). W hat's gra tifyin g about d iffe re n t types of television program s? Unpublished m anuscript. U niversity o f Southern California, Annenberg School o f C om m unications. Krantz, D.H., 8c Tversky, A. (1975). S im ila rity o f rectangles. Journal o f M athem ati cal Psychology, 12, 4-34. Krumhansl, C.L. (1978). Concerning the applicability of geo m e tric m odels to sim ilarity data: The interrelationship between sim ila rity and spatial density. Psychological Review, 85, 445-463. Kruskal, J.B. (1964). N o n-m etric m ultidim ensional scaling. Psvchom etrika, 29, 1-27; 115-129. Kruskal, J.B., 8c Wish, M. (1978). M ultidim ensional scaling. Beverly Hills: Sage. Levine, M.S. (1977). Canonical analysis and fa cto r com pariso n. Beverly Hills: Sage. Levy, M.R., 8c Kilburn, H. (1979, May). A m ultidim ensional m essage strategy fo r energy conservation. Paper presented at the annual m eeting o f the Inter national C om m unication Association, Philadelphia. Lom etti, G.E., Reeves, B., 8c Bybee, C.R. (1977). Investigating the assum ptions of uses and gratifications research. C om m unication Research, 4, 3 2 1 - 338. 167 Lyle, J., & Hoffman, H.R. (1972). Children's use of television and other media. In E.A. Rubinstein, G.A. Com stock, & J.P. M urray (Eds.), Television and social behavior (Vol.4). Television in d a y -to -d a y life: Patterns of use (pp. 129-256). W ashington, DC: U.S. G overnm ent Printing Office. Markus, H.J., & Nurius, P.S. (1984). S e lf-understanding and se lf-re g u la tio n in m id dle childhood. In W.A. C ollins (Ed.), Developm ent during m iddle childhood: The years from six to tw elve (pp. 147-183). W ashington, DC: National Academ y Press. M cllw raith, R.D., & Schallow, J.R. (1983). A d ult fantasy life and patterns of media use. Journal of C om m unication, 33(1), 78-91. McLeod, J.M., Atkin, C.K., 8i Chaffee, S.H. (1972a). Adolescents, parents, and televison use: Adolescent s e lf-re p o rt measures from Maryland and W is consin samples. In G.A. C om stock 8t E.A. Rubinstein (Eds.), Television and social behavior (Vol. 3 ). Television and adolescent aggressiveness (pp. 173- 238). W ashington, DC: U.S. G overnm ent Printing Office. McLeod, J.M., Atkin, C.K., 8t Chaffee, S.H. (1972b). Adolescents, parents, and televison use: S e lf-re p o rt and o th e r-re p o rt m easures from the W isconsin sample. In G.A. C om stock & E.A. Rubinstein (Eds.), Television and social behavior (Vol. 3 ). Television and adolescent aggressiveness (pp. 239- 313). W ashington, DC: U.S. G overnm ent Printing Office. Milavsky, J.R., Stipp, H.H., Kessler, R.C., & Rubens, R.S. (1982). Television and ag gression: A panel study. New York: Academ ic Press. Morgan, M. (1982). Television and adolescents' sex role stereotypes: A lo n gitudinal study. Journal of Personality and Social Psychology. 43, 947-955. Mussen, P. (1979). The psychological developm ent of the ch ild . Englewood Cliffs, NJ: Prentice-Hall. Nelson, K. (1973). Some evidence fo r the cognitive prim acy of categorization and its functional basis. M errill-P a lm er Quarterly. 19, 21-39. Nielsen, A.C., Co. (1983). Viewers in Profile: Los Angeles. CA. August. 1983. Los Angeles: Author. 168 Pearl, D., Bouthilet, L., 8t Lazar, J. (1982). Television and behavior: Ten years of scie ntific progress and im plications fo r the eighties (Vol. 1). Sum m ary re p o rt. Rockville, MD: National Institu te o f M ental Health. Pingree, S., 8t Hawkins, R. (1981). U.S. program s on Australian te le visio n : The cu l tiva tio n effect. Journal o f C om m unication, 31(1), 97-105. Ray, A.A. (Ed.). (1982). SAS user's guide: Statistics, 1982 e dition. Cary, NC: SAS Institute, Inc. Reeves, B., 8i Greenberg, B.S. (1977). C hildren's perceptions of televisio n charac ters. Human C om m unication Research, 3, 113-127. Reeves, B., 8t Lom etti, G.E. (1979). The dim ensional structure of children's percep tio n s of television characters: A replication. Human C om m unication Research, 5, 247-256. Reinhardt, P.S. (Ed.). (1980). SAS supplem ental library user's guide: 1980 e d itio n . Cary, NC: SAS Institute, Inc. Ricciuti, H.N. (1965). O bject grouping and selective ordering in infants 12-24 m onths old. M errill-P alm er Quarterly, 11, 129-148. Roberts, D.F., Bachen, C.M., Hornby, M., 8i Hernandez-Ramos, P. (1984). Reading and television: Predictors of reading achievem ent at d iffe re n t age levels. C om m unication Research. 11. 9-49. Rosch, E., & Mervis, C.B. (1975). Family resem blances: Studies in the internal structure o f categories. C ognitive Psychology, 7, 573-605. Rosch, E., Mervis, C.B., Gray, W.D., Johnson, D.M., 8i Boyes-Braem , P. (1976). Basic objects in natural categories. C ognitive Psychology. 8, 382-439. Rubin, A. (1979). Television use by children and adolescents. Human C om m unication Research, 5, 109-120. Shepp, B.E. (1978). From perceived sim ila rity to dim ensional structure: A new hypothesis about perceptual developm ent. In E. Rosch 8i B.B. Lloyd (Eds.), C ognition and categorization (pp. 135-167). Hillsdale, NJ: Erlbaum. 169 Singer, J.L., Singer, D.G., & Rapaczynski, W.S. (1984). Family patterns and te le v i sion view ing as predictors of children's beliefs and aggression. Journal of C om m unication, 34(2), 73-89. Stewart, D.K., & Love, W.A. (1968). A general canonical co rre la tio n index. Psychological Bulletin, 70, 160-163. Takane, Y., Young, F.W., & deLeeuw, J. (1977). N onm etric individual differences m ultidim ensional scaling: An alternating least squares m ethod w ith optim al scaling features. Psvchom etrika, 42, 7-67. Tucker, R.K., & Chase, L.J. (1980). Canonical correlation. In P.R. M onge & J.N. Cappella (Eds.), M ultivariate techniques in human com m unication research (pp. 205-228). New York: Academ ic Press. Tversky, A. (1977). Features o f sim ilarity. Psychological Review, 84, 327-352. Tversky, A., & Krantz, D.H. (1970). The dim ensional representation and the m etric structure of sim ilarity data. Journal of M athem atical Psychology, 7, 572-597. Wish, M. (1976). Com parisons am ong m ultidim ensional structures of interpersonal relations. M ultivariate Behavioral Research, 11, 297-327. W oelfel, J., Cody, M.J., Gillham, J., & Holmes, R.A. (1980). Basic prem ises of m ul tidim ensional attitude attitude change theory: An experim ental analysis. Human C om m unication Research, 6, 153-167. APPENDIX A EXAMPLES OF CARDS USED IN SORTING TASK 171 NJ NJ LITTLE H O U S E O N T H E PRAIRIE C H IPS APPENDIX B ATTRIBUTE CONDITION SCRIPTS 173 B.1 Instructions — 2nd Grade A ttribute Condition INTRODUCTION Hi, m y name is (YOUR NAME), and I'm from (YOUR SCHOOL). I'm here today to ask you som e questions about some TV program s th a t you m ight watch. I'm trying to find out w hat kids like you think about these program s, and the only w ay I can do that is to ask. I th in k you'll find th a t this is both easy and fun, and you m ight even learn som ething from it. W hat you have to say is real im portant, so think carefully about w hat I ask you before you answ er me. No one but me w ill ever see your answers, so please be sure to tell me exactly w hat you think. N othing I ask has any W rong answers — w hatever you say is the rig h t one fo r you, so you ju st te ll me w hat it is. If you have any questions, ju st stop me and ask, O.K.? And if you have to leave, ju st te ll me and th a t w ill be O.K. too. Ready to start? The firs t thing w e're going to do today is to take this pile of cards w ith program names on them and sort them in to piles of program s th a t are SIMILAR. To give you an idea of w hat you're supposed to do, I th o u g h t we'd practice w ith these stacks first. OK, here are three cards. (PASS OUT STACK "A".) One says "m ouse," one says "rat," and one says "bird." W ould you say th a t any o f these are sim ilar to each other? (IF A CHILD SAYS MOUSE & RAT, OFFER PRAISE AND GO ON. IF CHILD OF FERS OTHER COMBINATIONS, ASK FOR REASONS. IF THEY MAKE SENSE — E.G., THEY ARE ALL ANIMALS SO THEY ALL GO TOGETHER — SAY THAT'S OK, AND POINT OUT ALTERNATIVE GROUPINGS. IF THEY ARE TOO IDIOSYNCRATIC TO MAKE SENSE, POINT OUT RAT/MOUSE COMPARISON AND CONTINUE.) So there are a lo t of ways th a t a rat and a m ouse are sim ilar to each other and different from a bird. Mice and rats both have fo u r legs; a bird has tw o. Mice and rats have ears; birds don't. Mice and rats craw l and run; birds hop and fly, Right? OK, now let's try this stack. (COLLECT STACK "A" AND PASS OUT STACK "B") There are five cards here, each one w ith the name o f a d ifferent fru it on it. There's Lemon, Apple, Orange, Lime, and Banana. Now w h y don't you put the cards together th a t you think have sim ilar fru it on them . (WAIT FOR CHILD TO DO SO.) Now w hat makes this group similar? 174 (CONTINUE ASKING ABOUT GROUPS IF THEY MAKE SENSE. PROVIDE AL TERNATIVE GROUPINGS, E.G., BASED ON COLOR (LEMON/BANANA, ORANGE, AP PLE, LIME), “FUNCTION" (GOOD TO EAT BY THEMSELVES) (APPLE/ORANGE/BANANA, LEMON/LIME), PEEL BEFORE EATING (LEMON/ORANGE/BANANA/LIME, APPLE). OR YOU COULD THINK ABOUT ALL THOSE THINGS AT ONCE, SO YOU WOULD HAVE FIVE DIFFERENT GROUPS. MAKE SURE THAT THEY KNOW THAT THEY CAN CONSIDER MORE THAN ONE THING IN ASSESSING SIMILARITY.) There isn't one rig h t way to group these, ju st a w ay th a t makes the m ost sense to YOU. (COLLECT STACK "B") Now w e're going to try this w ith TV program s. Here's a stack of cards w ith the names o f program s on them . I'm going to give you the cards 8 at a tim e, and I'd like you to make piles of program s th a t are sim ilar to each other. You can think about as many things th a t make the program s sim ilar as you w ant to, and you can make as many piles as you w ant to. As I give you m ore cards, you can add them to piles you already have, or, if they're not sim ilar to any of the piles you already have, you can make new piles out o f them . You can have lots o f program s in a pile, ju st a few, or even ju st one, if it's not sim ilar to any others. If you can't read the name o f one o f the program s ju st ask me. If you don't know enough about a program to fig u re out w here it goes, put its card into this blue envelope. (PASS OUT BLUE ENVELOPES.) You can put the envelope over to the side w hile you're m aking yo u r piles until you need it. A fter you've put all the program s into piles, then I'm going to ask you to tell me w hat things about the program s in each pile make them sim ilar. OK, do you have any questions? (ANSWER IF ANY.) If you have any questions w hile you're w o rking on the piles, ju s t ask me. Remember, you can th in k about as many thing s as you w ant when you decide w hich program s are sim ilar. Let's start. (PASS OUT FIRST 8 CARDS. AFTER FIRST 8 ARE SORTED. WATCH TO MAKE SURE THAT CHILD UNDERSTANDS THAT THEY MAY ADD PROGRAMS FROM NEXT 8 TO PILES ALREADY CREATED. AFTER ALL CARDS SORTED, CONTINUE ON WITH GROUP LABELING QUESTIONS. WRITE DOWN ANSWERS ON ENVELOPES AS CHILD 175 GIVES THEM.) Ok, now that we have these piles, I'd like to know w hy you put each group of program s together. I'll w rite your reasons on these envelopes as you put the cards into them . Take as m uch tim e as you w ant. Remember, there are no w rong answers. I ju s t w ant to know w hat you think. Ready? Let's start. (POINT TO FIRST PILE CREATED.) W hat makes these program s sim ilar — W hat things do they all have in common? (REPEAT FOR EACH PILE UNTIL DONE.) That's great. You've been very helpful. Do you have any questions you'd like to ask me? (ANSWER, IF ANY.) Thanks a lot. You can go back to yo u r room now, and have your teacher send the next person. 176 B.2 Instructions — 6th/10th Grade A ttrib u te Condition INTRODUCTION Hi, m y name is (YOUR NAME), and I'm from (YOUR SCHOOL). I'm here today to ask you som e questions about som e TV program s that you m ight watch. I'm trying to find out w hat kids like you th in k about these program s, and the only w ay I can do th a t is to ask. I th in k you'll find th a t this is both easy and fun, and you m ight even learn som ething fro m it. W hat you have to say is real im portant, so th in k carefully about w hat I ask you before you answ er me. Noone but me w ill ever see your answers, so please be sure to te ll me exactly w hat you think. N othing I ask has any W rong answ ers — w hatever you say is the right one fo r you, so you ju st te ll me w hat it is. If you have any questions, ju s t stop me and ask, O.K.? And if you have to leave, ju s t te ll me and that w ill be O.K. too. Ready to start? The firs t thing w e're going to do today is to take this pile of cards w ith program names on them and so rt them into piles o f program s that are SIMILAR. To give you an idea of w hat you're supposed to do, I though t w e'd practice w ith these stacks first. This may seem a little easy and obvious to you, but w e're do ing this w ith 2nd graders too, and we have to do the same thing s w ith everybody. OK, here are three cards. (PASS OUT STACK "A".) One says "m ouse," one says "rat," and one says "bird." W ould you say that any o f these are sim ilar to each other? (IF ANY CHILD SAYS MOUSE & RAT, OFFER PRAISE AND GO ON. IF ANY CHILD OFFERS OTHER COMBINATIONS, ASK FOR REASONS. IF THEY MAKE SENSE — E.G., THEY ARE ALL ANIMALS SO THEY ALL GO TOGETHER — SAY THAT'S OK, AND POINT OUT ALTERNATIVE GROUPINGS. IF THEY ARE TOO IDIOSYNCRATIC TO MAKE SENSE, POINT OUT RAT/MOUSE COMPARISON AND CONTINUE.) So there are a lo t o f ways th a t a rat and a mouse are sim ilar to each other and d ifferent from a bird. Mice and rats both have four legs; a bird has two. Mice and rats have ears; birds don't. Mice and rats crawl and run; birds hop and fly. Right? (COLLECT STACK "A" AND PASS OUT STACK "B.") OK, now let's try this stack. There are five cards here, each one w ith the name of a different fru it on it. There's Lemon, Apple, Orange, Lime, and Banana. Now w hy don't you put the cards to g e th e r th a t you think have sim ilar fru it on them . (WAIT FOR CHILDREN TO DO SO.) 177 Now w h at makes this group similar? (CONTINUE ASKING ABOUT GROUPS IF THEY MAKE SENSE. PROVIDE AL TERNATIVE GROUPINGS, E.G., BASED ON COLOR (LEMON/BANANA, ORANGE, AP PLE, LIME), "FUNCTION" (GOOD TO EAT BY THEMSELVES) (APPLE/ORANGE/BANANA, LEMON/LIME), PEEL BEFORE EATING (LEMON/ORANAGE/BANANA/LIME, APPLE). OR YOU COULD THINK ABOUT ALL THOSE THINGS AT ONCE, SO YOU WOULD HAVE FIVE DIFFERENT GROUPS. MAKE SURE THAT THEY KNOW THAT THEY CAN CONSIDER MORE THAN ONE THING IN ASSESSING SIMILARITY.) There isn't one rig h t way to group these, ju s t a w ay th a t makes the m ost sense to YOU. (COLLECT STACK "B") Now w e're going to try this w ith tv program s. Here's a stack o f cards w ith the names o f program s on them . I'm going to give you the cards 8 at a tim e, and I'd like you to make piles o f program s th a t are sim ilar to each other. You can think about as many things th a t make the program s sim ilar as you w ant to, and you can make as m any piles as you w ant to. As I give you m ore cards, you can add them to piles you already have, or, if they're not sim ila r to any o f the piles you already have, you can make new piles o u t o f them . You can have lots of program s in a pile, ju st a few, or even ju st one, if it's not sim ilar to any others. If you don't know enough about a program to figure out w here it goes, put its card into this blue envelope. (PASS OUT BLUE ENVELOPES.) You can put the envelope over to the side w hile you're m aking your piles until you need it. A fter you've put all the program s into piles, then I'm going to ask you to take each pile, put it in one o f these envelopes, and w rite on the en velopes w hat things about the program s in each pile make them sim ilar. OK, do you have any questions? (ANSWER IF ANY.) If you have any questions w h ile you're w orking on the piles, ju s t ask me. Remember, you can think about as m any things as you w ant when you decide w hich program s are sim ilar. Let's start. (PASS OUT FIRST 8 CARDS. AFTER FIRST 8 ARE SORTED. WATCH TO MAKE SURE THAT CHILD UNDERSTANDS THAT THEY MAY ADD PROGRAMS FROM NEXT 178 8 TO PILES ALREADY CREATED. AFTER ALL CARDS SORTED, CONTINUE ON WITH GROUP LABELING QUESTIONS.) Ok, now that we have these piles. I'd like to know w h y you put each group of program s together. I'd like you to take these envelopes (HAND OVER ENOUGH ENVELOPES FOR THE NUMBER OF PILES THE CHILD HAS CREATED), put one of your piles of program s in it, and w rite on it w hat th in g s about the program s make them sim ilar — w hat things ail the program s in the group have in com m on. Take as m uch tim e as you want. Remember, there are no w rong answers. I just w ant to know w hat you think. Ready? Let's start. (POINT TO FIRST PILE CREATED.) W hat makes these program s sim ilar — W hat th in g s do they all have in common? (REPEAT FOR EACH PILE UNTIL IT'S CLEAR THAT THE CHILDREN UNDER STAND THE TASK.) That's great. You've been very helpful. Do you have any questions you'd like to ask me? (ANSWER, IF ANY.) Thanks a lot. You can go back to your room now, and have yo u r teacher send the next group o f kids. APPENDIX C GRATIFICATION CONDITION SCRIPTS 180 C.1 Instructions — 2nd Grade G ratification Condition INTRODUCTION Hi, m y name is (YOUR NAME), and I'm from (YOUR SCHOOL). I'm here today to ask you som e questions about reasons you m ight have fo r w atching som e TV program s. I'm trying to find out w h at kids like you th in k and feel when they w atch these program s, and the only w ay I ca n do thatask. I th in k yo u 'll find th a t this is both easy and fun, and you m ight even learn som ething from it. W hat you have to say is real im portant, so th in k carefully about w hat I ask you before you answ er me. No one but me w ill ever see your answers, so please be sure to te ll me exactly w hat you think. Nothing I ask has any W rong answ ers — w hatever you say is the right one fo r you, so you ju st tell me w hat it is. If you have any questions, ju st stop me and ask, O.K.? And if you have to leave, ju s t tell me and th a t w ill be O.K. too. Ready to start? The firs t thing w e're going to do today is to take this pile o f cards w ith program names on them and sort them into piles o f program s th a t you w atch fo r sim ilar reasons. So you're going to put all the program s th a t make you feel sim ilar feelings when you w atch them — o r th a t make you th in k sim ilar th o ugh ts — in piles together. To give you an idea of w hat you're supposed to do, I th o u g h t we'd practice w ith these stacks first. OK, here are three cards. (PASS OUT STACK "C"). One says "getting hit by som eone," one says "getting a present from som eone," and one says "getting to go o n a vacation trip." W ould you say you w ould feel pretty much the same feelings about any of these? (IF A CHILD SAYS PRESENT & VACATION, OFFER PRAISE AND GO ON. IF CHILD OFFERS OTHER COMBINATIONS, ASK FOR REASONS. IF THEY MAKE SENSE — E.G., THEY ARE ALL THINGS THAT CAN HAPPEN TO YOU — SAY THAT'S OK, AND POINT OUT ALTERNATIVE GROUPINGS. IF THEY ARE TOO IDIOSYNCRATIC TO MAKE SENSE, POINT OUT PRESENT/VACATION COMPARISON AND CONTINUE.) So there are lots of feelings you can have about getting a present and go ing on a vacation that are sim ilar to each other and different from w hat you'd feel about getting hit. You'd probably be happy, excited, and interested in a present and a vacation, but unhappy, afraid, and maybe sad about g e ttin g hit, right? OK, now let's try this stack. (COLLECT STACK "C" AND PASS OUT STACK "D") There are five cards here, each one w ith the name of a different place on it. There's your room , the playground, a dentist's office, a m ovie theater, and your classroom . Now w hy don't you put the cards to g e th e r that have places where you w ould th in k sim ilar thoughts. You know, not necessarily exactly the Same 181 thoughts, but sim ilar ones. (WAIT FOR CHILD TO DO SO.) Now w hat th o u g h ts do you have about being in th is group o f places that makes them all similar? (CONTINUE ASKING ABOUT GROUPS IF THEY MAKE SENSE. PROVIDE AL TERNATIVE GROUPINGS, E.G., BASED ON FINDING OUT THINGS (ROOM/CLASSROOM/MOVIE THEATER/, DENTIST'S OFFICE, PLAYGROUND), OR THINKING ABOUT THNIGS YOU MIGHT DO WHEN YOU GROW UP (ALMOST ANY COMBINATION). MAKE SURE THAT THEY KNOW THAT THEY CAN CONSIDER MORE THAN ONE THING IN ASSESSING SIMILARITY.) There isn't one rig h t w ay to group these, ju st a w ay th a t makes the m ost sense to YOU. (COLLECT STACK "D") Now w e're going to try th is w ith TV program s. Here's a stack o f cards w ith the names o f program s on them . I'm going to give you the cards 8 at a tim e, and I'd like you to make piles o f program s that make you th in k sim ilar thoughts, or make you feel sim ilar feelings when you watch them . You can th in k about as many thoughts and feelings as you w ant to, and you can make as many piles as you w ant to. As I give you m ore cards, you can add them to piles you already have, or, if a program doesn't make you feel like any o f the piles you've already made do, you can make new piles o u t of them . You can have lots of program s in a pile, ju st a few , or even ju s t one, if it's not sim ilar to any others. If you can't read the name of one of the program s ju st ask me. If you don't know enough about a program to figure out where it goes, put its card into this blue envelope. (PASS OUT BLUE ENVELOPES.) You can put the envelope over to the side w hile you're m aking your piles until you need it. A fte r you've put all the program s in to piles, then I'm going to ask you to te ll me w hat th o u g h ts o r feelings you have when you w atch the program s in each pile. OK, do you have any questions? (ANSWER IF ANY.) If you have any questions w hile you're w orking on the pites, ju st ask me. Remember, you can th in k about as many thoughts and feelings as you w ant when 182 you decide w hich program s go together. Let's start. (PASS OUT FIRST 8 CARDS. AFTER FIRST 8 ARE SORTED. WATCH TO MAKE SURE THAT CHILD UNDERSTANDS THAT THEY MAY ADD PROGRAMS FROM NEXT 8 TO PILES ALREADY CREATED. AFTER ALL CARDS SORTED, CONTINUE ON WITH GROUP LABELING QUESTIONS. WRITE DOWN ANSWERS ON ENVELOPES AS CHILD GIVES THEM.) Ok, now that w e have these piles. I'd like to know w h y you put each group o f program s together. I'll w rite yo u r reasons on these envelopes as you put the cards into them . Take as much tim e as you want. Remember, there are no w rong answers. I ju st w ant to know w h at you think. Ready? Let's start. (POINT TO FIRST PILE CREATED.) W hat makes these program s sim ilar — W hat though ts or feelings do you have when you watch the program s in this group? (REPEAT FOR EACH PILE UNTIL DONE.) That's great. You've been very helpful. Do you have any questions you'd like to ask me? (ANSWER, IF ANY.) Thanks a lot. You can go back to your room now, and have your teacher send the next person. 183 C.2 Instructions — 6th /10th Grade Gratification Condition INTRODUCTION Hi, m y name is (YOUR NAME), and I'm from (YOUR SCHOOL). I'm here today to ask you som e questions about reasons you m ig h t have fo r w atching som e TV program s. I'm tryin g to fin d out w hat kids like you th in k and feel when they watch these program s, and the only w ay I can do th a t is to ask. I think you'll find that this is both easy and fun, and you m ight even learn som ething from it. W hat you have to say is real im portant, so th in k carefully about w hat I ask you before you answ er me. No one but me w ill ever see yo u r answers, so please be sure to tell me exactly w hat you think. N othing I ask has any W rong answers — w hatever you say is the right one fo r you, so you ju s t te ll me w hat it is. If you have any questions, ju st stop me and ask, O.K.? And if you have to leave, ju st te ll me and th a t w ill be O.K. too. Ready to start? The firs t thing w e're going to do today is to take this pile o f cards w ith program names on them and so rt them into piles of program s th a t you watch fo r sim ilar reasons. So you're going to put all the program s th a t make you feel sim ilar feelings when you w atch them — or th a t make you th in k sim ilar thoughts — in piles together. To give you an idea of w hat you're supposed to do, I th o u g h t we'd practice w ith these stacks first. This may seem a little easy and obvious to you, but w e're doing th is w ith 2nd graders too, and we have to do the same things w ith everybody. OK, here are three cards. (PASS OUT STACK "C".) One says "getting hit by som eone," one says "getting a present fro m som eone," and one says "getting to go on a vacation trip." W ould you say you w ould feel pretty much the same feelings about any o f these? (IF ANY CHILD SAYS PRESENT & VACATION, OFFER PRAISE AND GO ON. IF ANY CHILD OFFERS OTHER COMBINATION S, ASK FOR REASONS. IF THEY MAKE SENSE - - E.G., THEY ARE ALL THINGS THAT CAN HAPPEN TO YOU — SAY THAT'S OK, AND POINT OUT ALTERNATIVE GROUPINGS. IF THEY ARE TOO IDIOSYNCRATIC TO MAKE SENSE, POINT OUT PRESENT/VACATION COMPARISON AND CONTINUE.) So there are lots of feelings you can have about gettin g a present and go ing on a vacation th a t are sim ilar to each other and d iffe re n t fro m w hat you'd feel about getting hit. You'd probably be happy, excited, and interested in a present and a vacation, but unhappy, afraid, and maybe sad about getting hit, right? OK, now let's try th is stack. (COLLECT STACK "C" AND PASS OUT STACK "D") There are five cards here, each one w ith the name o f a d iffe re n t place on it. 184 There's your room , the playground, a dentist's o ffice, a m ovie theater, and your classroom . Now w hy don't you put the cards toge th e r th a t have places w here you w ould think sim ilar thoughts. You know, not necessarily exactly the Same thoughts, but sim ilar ones. (WAIT FOR CHILDREN TO DO SO.) Now w hat thoughts do you have about being in th is group of places th a t makes them all similar? (CONTINUE ASKING ABOUT GROUPS IF THEY MAKE SENSE. PROVIDE AL TERNATIVE GROUPINGS, E.G., BASED ON FINDING OUT THINGS (ROOM/CLASSROOM/MOVIE THEATER/, DENTIST'S OFFICE, PLAYGROUND), OR THINKING ABOUT THNIGS YOU MIGHT DO WHEN YOU GROW UP (ALMOST ANY COMBINATION). MAKE SURE THAT THEY KNOW THAT THEY CAN CONSIDER MORE THAN ONE THOUGHT AND/OR FEELING IN ASSESSING SIMILARITY.) There isn't one rig h t w ay to group these, ju st a way th a t makes the m ost sense to YOU. (COLLECT STACK "D") Now w e're going to try th is w ith TV program s. Here's a stack of cards w ith the names o f program s on them . I'm going to give you the cards 8 at a tim e, and I'd like you to make piles of program s th a t make you th in k sim ilar thoughts, or make you feel sim ilar feelings when you watch them . You can th in k about as many thoughts and feelings as you w ant to, and you can make as many piles as you w ant to. As I give you m ore cards, you can add them to piles you already have, or, if a program doesn't make you feel like any of the piles you've already made do, you can make new piles out of them . You can have lots of program s in a pile, ju st a few, or even ju st one, if it's not sim ilar to any others. If you can't read the name of one of the program s ju st ask me. If you don't know enough about a program to figure out where it goes, put its card into th is blue envelope. (PASS OUT BLUE ENVELOPES.) You can put the envelope over to the side w hile you're m aking your piles until you need it. A fter you've put all the program s into piles, then I'm going to ask you to te ll me w hat though ts or feelings you have when you w atch the program s in each pile. OK, do you have any questions? (ANSWER IF ANY.) 185 If you have any questions w hile you're w orking on the piles, ju st ask me. Remember, you can think about as many though ts and feelings as you w ant when you decide w hich program s go together. Let's start. (PASS OUT FIRST 8 CARDS. AFTER FIRST 8 ARE SORTED. WATCH TO MAKE SURE THAT EACH CHILD UNDERSTANDS THAT THEY MAY ADD PROGRAMS FROM NEXT 8 TO PIL ES ALREADY CREATED. AFTER ALL CARDS SORTED, CONTINUE ON WITH GROUP LABELING QUESTIONS.) Ok, now that we have these piles, I'd like to know w hy you put each group of program s together. I'd like you to take these envelopes (HAND OVER ENOUGH ENVELOPES FOR THE NUMBER OF PILES EACH CHILD HAS CREATED), put one o f your piles o f program s in each one, and w rite on it w hat though ts or feelings you have when you w atch the program s in the group. Take as m uch tim e as you want. Remember, there are no w rong answers. I ju s t w ant to know w hat you think. Ready? Let's start. (WALK AROUND ALL CHILDREN AND POINT TO FIRST PILE CREATED.) W hat makes these program s sim ilar — W hat th o u g h ts o r feelings do you have when you watch the program s in this group? (REPEAT FOR EACH PILE UNTIL IT'S CLEAR THAT THE CHILDREN UNDER STAND THE TASK.) That's great. You've been very helpful. Do you have any questions you'd like to ask me? (ANSWER, IF ANY.) Thanks a lot. You can go back to your room now, and have your teacher send the next group of kids. 186 APPENDIX D VIEWING FREQUENCY CHECKLIST — SHORT VERSION 187 ID Children's Ideas About Television Programs Name School Grade Age Girl ( ) Boy ( ) DIRECTIO NS: First fill in your name, school, grade, age, and check whether you're a girl or a boy. On the next two pages is a list of television programs. Most of them are on television right now. A few of them used to be on television but aren't anymore. The boxes to the right are for showing how often you've usually watched each program during the last year. Here's how to mark them. First decide if you've seen the program at all during the last year. If you haven't seen it during that time put a check either in the box which says "never watched" or "have watched, but not last year." If you have seen it during the last year you can check "just a few times," "about once a month," "a couple of times a month," or "about once a week." On the second page you can also check "a couple of times a week or more" since those programs are usually on more than once a week. When you finish the second page, stop and wait for directions. If you have any questions, just raise your hand. 188 ID _______________ How often have YO U watched each program during the last year? never watched have watched but not last year just a few times about once a month a couple of times a month about once a week Dukes of Hazzard Smurfs Diff'rent Strokes Knight Rider Real People Amazing Spiderman and Incredible Hulk Fame Family Ties Dynasty Magnum, P.I. Richie Rich Popeye Gimme a Break Remington Steele Ripley's Believe It or Not That's Incredible Silver Spoons The A-Team Simon and Simon The Love Boat 9 to 5 Pac-Man The Fall Guy GO ON TO TH E N EX T PAGE 189 ID How often have YO U watched each program during the last year? never watched have watched but not last year just a few times about once a month a couple of times a month about once a week a couple of times a week (or more) Little House on the Prairie CHiPs One Day at a Time Happy Days Three's Company M *A *S *H Laverne and Shirley The White Shadow The. Three Stooges The Brady Bunch • The Waltons - The Munsters Benson Eight Is Enough Fantasy Island The Facts of Life Gilligan's Island Video One STOP. W AIT FOR D IR EC TIO N S 190 APPENDIX E VIEWING FREQUENCY CHECKLIST — LONG VERSION 191 ID _ Children's Ideas About Television Programs N am e School Grade fige Girl ( ) Boy ( ) DIRECTIONS: First f i l l in your name, school, grade, age, and check whether you're a g irl or a boy. O n the next three pages is a lis t of television programs. Most of them are on television right now. A few of them used to be on television but aren't anymore. The boxes to the right are for showing how often you've usually watched each program during the Vast year. Here's how to mark them. First decide i f you've seen the program at a ll in the last year. I f you haven't seen i t during that time, put a check either in the box which says "never watched" or "have watched, but not last year." (*) I f you have seen i t in the last year you can check "just a few times," "about once a month," "a couple of times a month," or "about once a week." O n the third page you can also check "a couple of times a week or more" since those programs are usually on more than once a week. W hen you finish the third page, stop and wait for directions. I f you have any questions, just raise your hand. 192 ID __________ How o ften have YO U watched each program during the last year? never watched have watched but not la s t year just a few times about once a month a couple of times a month about once a week Dukes of Hazzard Ozzie and Harriet Smurfs D iff'rent Strokes Knight Rider Real People Amazing Spiderman and Incredible Hulk Fame Fam ily Ties Dynasty Magnum, P.I. Mama Malone Richie Rich Popeye Domestic Life Gimme a Break Remington Steele Two Marriages Ripley's Believe It or Not Webster That's Incredible Silver Spoons GO ON TO TH E N E X T PAGE 193 ID How o ften have YO U watched each program during the la s t year? never watched have watched but not la s t year just a few times about once a month a couple of times a month about once a week A.K.A Pablo The A-Team Simon and Simon The Love Boat 9 to 5 Pac-Man The Partridge Fam ily The Fall Guy GO ON TO THE NEXT PAGE 194 ID How often have YO U watched each program during the la s t year? never watched have watched but not la s t year just a few times about once a month a coupie of times a month about once a week a couple of times a week (or more) L ittle House on the Prairie CHiPs One Day at a Time Happy Days Three's Company M *A *S *H Laverne and Shirley The White Shadow The Three Stooges Family The Brady Bunch The Munsters Benson Leave it to Beaver Eight Is Enough Fantasy Island The Facts of Life Please Don't Eat the Daisies Gilligan's Island Video One My Three Sons The Waltons STOP. WAIT FOR DIRECTIONS 195 APPENDIX F CONTENT ANALYSIS CODING SCHEME 196 F.1 INTRODUCTION This content analysis w ill provide a means fo r id entifying the categories or dim ensions children use in distinguishing am ong a) a set of co m m o m ly viewed television program s, or b) the g ra tifica tio n s ascribed to view ing the same set of television program s. Second, sixth, and tenth grade children w ere assigned to one of the tw o conditions. Program o r G ratification, and were asked to so rt cards w ith the names of fo rty television program s on them into piles or groups o f ei th e r a) program s w hich they considered sim ilar, or b) program s w hich made them have sim ilar feelings or think about sim ilar things. Sixth and tenth graders then put each group or pile in an envelope and w ro te on the envelopes e ith e r a) what made the program s sim ilar or b) w hat the group of program s made them feel or th in k about. Second graders did the same thing, except th a t they to ld an ex perim enter w hat made the program s sim ilar or w hat the program s made them feel or th in k about, and the experim enter w rote these responses dow n. (All children could put as many program s as they were unfam iliar w ith in a blue envelope.) These responses are w hat w ill be analyzed. The Coding Unit The coding unit is either a) an attrib u te used to describe a group of te le v i sion program s (Program Condition), or b) an descrip to r o f the th o u g h ts o r fe e l ings a group of television program s evoked w hich has been w ritte n dow n on an envelope. In either case, they may take several form s. They m ay be single w ords (usually adjectives or nouns) such as "funny" or "happy." They m ay be phrases, such as "just like real life" or "like having a friend in the house." They m ay be clauses, such as "adventure show s w hich have a lo t o f action in them " or "because they all make me learn som ething." Or they may be com ple te sen tences such as "They all are about fam ilies" or "I like to talk about them w ith my friends." In many cases there may be m ore than one a ttrib ute or d e scrip to r on a given envelope. The Coding Process Since the process fo r coding each condition (Program and G ratifica tion) is essentially the same, except that the coding categories differ, they w ill be dis cussed together. Before coding, coders should have all the envelopes to be coded fo r each child, a pencil, scoring sheets, and this manual. For each child's set o f envelopes to be coded, the coder should do the fo llo w in g , in order. 1. Fill in the subject ID, condition, and num ber of "don't know s" on the scoring sheet. 197 2. Indicate "m issing data" fo r all the program s in the "D on't Know" envelopes. 3. Identify all the program attributes or gratificatio n descriptors each envelope by referring to the Coding U nit section above and the Coding C ategories sections below. Each envelope may have m ore than one a ttrib u te or descriptor on it, so special care m ust be taken to make sure every one is identified. Generally, an attrib ute or descriptor is coded each tim e a con ceptually d istin ct Coding U nit is identified. 4. Record all instances of each program attrib ute or g ra tifica tio n descrip to r on the scoring sheet. 5. Record the grade, age, and sex of the child on the scoring sheet. F.2 ATTRIBUTE CONDITION Coding Categories The fo llo w in g categories are to be used in coding each envelope fo r each child. 1. Action 2. Family 3. Aesthetics 4. Time Period 5. Problem Solving 6. Music 7. Novelty 8. Realism 9. H um or/Seriousness 10. Im portance 11. Crime 12. Age Specific 13. Occupation 14. Character A ttributes 15. G ratification 16. N on-C ontent A ttribute 17. Other Code "A ction" if the description o f the program group refers to the concept of action o r activity. Examples o f this concept include: - action - chases - adventure - fig htin g - goodguys vs. badguys 198 Code "Family" if the description of the program group refers to the concept of fam ily, either by its presence or absence. Examples of this concept include: - has parents or kids - takes place in a house or home - about how fam ily m em bers relate - about hom elife - everybody's related Code "Aesthetic" if the description of the program group refers to the con cept o f quality, either by m entioning goodness or badness. This category does NOT include sim ply liking o r disliking, but does include preferences IF they are coupled w ith some sort of evaluative judgm ent as to the program group's worth. This category also includes instances o f program groups w hich are evaluated as good or bad because of adherence to o r violation of aesthetic criteria. Examples o f this concept include: - good, great, best - bad, stupid, dumb - good because they always have good stories - bad because they have dum b plots Code 'T im e Period" if the description of the program group refers to the concept of closeness to the present tim e. Examples o f this concept include: - thing s happening right now o r long ago - current - u p -to -d a te - old-fashioned Code "Problems and Problem Solving" if the description o f the program group refers to problem s or the w orking out of problems. Examples of this con cept include: - they all have problem s - they solve problem s - fam ily problem s - how to deal w ith things - w orking things out Code "M usic" if the description o f the program group refers to the presence o f m usic in the program m ing. Examples of this concept include: - dancing - videos 199 - m usic Code "N ovelty" if the description o f the program group refers to the presence of som ething new and unusual or the absence of som ething old and fam iliar. Examples o f this concept include: - unusual - different - weird - never the same - out o f the ordinary - not very com m on Code "Realism" if the description o f the program group refers to how much like or unlike real life the program m ing is. Examples of this concept include: - tru e -to -life - real o r unreal people - fantasy - cartoon - pretend, m ake-believe Code "Hum orousness/Seriousness" if the description of the program group refers to the presence or absence of hum or or funnyness, or how non -se rious the program s are. Examples of this concept include: - silly - ridiculous - not serious - not dram atic Code "Im portance" if the description o f the program group refers to how im portant the program m ing is. Examples of this concept include: - im portant or not unim portant - not trivial - w o rth seeing Code "Crim e" if the description o f the program group refers to crim e, crim inals, police, or detectives. Examples of this concept include: - crim es - m ysteries - cops - detectives 200 - saving people Code "Age Specific" if the description o f the program group refers to the age of the audience. Examples of th is concept include: - children - adults - kids - teenagers Code "O ccupation" if the description of the program group refers to some com m on job, work, or som e other occupational characteristic, EXCEPT if the job is involved in solving crim es (#11 above). Examples of this concept include: - they w ork together - all do the same thing - share an office Code "Character A ttribute s" if the description o f the program group refers to som e aspect of the characters OTHER than those coded elsewhere. Examples of this concept include: - they're all pretty - they all are happy - they all are m en/w om en Code "G ratification" if the description of the program group refers to some com m on personal function of all the program s. See the next section o f this manual, "G ratification C ondition" fo r an extended list o f gratifications. Some ex am ples of this concept include: - They make me feel good - I find out things from them Code "N on-C ontent A ttrib u te " if the description o f the program group refers to som e com m on attribute w hich does not have to do w ith program co n te nt or w ith the program groups' gratificatio n or personal function (see #13 above). Ex am ples of this concept include: - All same length - All on same station - All on at same tim e or on same day Code "O ther" if the description o f the program group refers to som e a t trib u te or charactersitics w hich cannot be coded in any of the above categories. 201 F.3 GRATIFICATION CONDITION Coding Categories The fo llo w in g categories are to be used in coding each envelope fo r each child. 1. Learning 2. Social Guidance 3. Passing Time 4. Excitem ent 5. Am usem ent 6. General Pleasure 7. Escape 8. Aesthetics 9. Social Interaction 10. Relaxation 11. Involvem ent 12. C uriosity/N ovelty 13. Com panionship 14. Other G ratifications 15. Program A ttribute 16. Other Code "Learning" if the description o f the gratification(s) associated w ith the program group refers to general acquisition of know ledge or in form ation, or to thinking. Examples of this concept include: - find out about things in the w orld - get to see real things - makes me think - th in k about things in life - makes me interested - shows interesting things Code "Social Guidance" if the description o f the g ra tifica tio n s associated w ith the program group refers to the providing of inform ation w hich can inform s the child's behavior in som e way. Examples of this concept include: - get to see how people do things I m ight do - get to see how people solve problem s - get to com pare m y life to others' - find out about ways to act w ith others - find out about things I m ight need to know when I grow up 202 Code "Passing Tim e" if the description of the g ra tifica tio n s associated w ith the program group refers to fillin g in otherw ise "em pty" tim e. Examples of this concept include: - when there's nothing else to do - when nothing else is on - when I'm bored - they're boring but I w atch them Code "Excitem ent" if the description of the gratificatio n(s) associated w ith the program s group refers to increasing the child's general arousal level. Ex am ples of this concept include: - get to see exciting things - makes me excited - get to see action Code "Am usem ent" if the description of the gratificatio n(s) associated w ith the program group refers to am usem ent or laughter. Examples of this concept include: - they make me laugh - they're funny - they're silly Code "General Pleasure" if the description of the gratificatio n(s) associated w ith the program group refers to m aintaining or increasing general w e ll-b e ing. Examples o f this concept include: - they make me feel good - they make me happy Code "Escape" if the description of the gratification(s) associated w ith the program group refers to avoidance of unpleasant thing s in real life. Examples of this concept include: - they make me fo rg e t my troubles - I get to see thing s turn out alright - I don't have to w o rry about anything - they make me fo rg e t I'm a kid Code "Aesthetics" if the description of the gratificatio n(s) associated w ith the program group refer to gettin g to experience som ething o f w o rth or value. Examples o f this concept include: 203 - I get to see good things (actors, stories, etc.) - they're all w ell made - they're much better than anything else on television Code "Social Interaction" if the description o f the gratification(s) associated w ith the program group refers to gettin g to share the process of view ing or the program content w ith others. Examples of th is concept include: - I get to see them w ith my fam ily or friends - I get to talk about them w ith m y fam ily or friends - They make me feel like sharing som ething w ith my fam ily or friends Code "Relaxation" if the description o f the gratification(s) associated w ith the program group refers to general reduction of tension, but w ith o u t im plying escape as in #6 above. Examples of this concept include: - they relax me; they're relaxing - they help me unwind - they make me feel loose Code "Involvem ent" if the description of the gratification(s) associated w ith the program group refers to having feelings in response to characters o r situa tions in the program s, or feeling like you're "in" the program . Examples o f this concept inlcude: - they make me feel (some em otion label other than happy, good, excited, or bored) - I feel like I'm in the show - I get real involved in the story - they make me feel like I w ant to do som ething fo r the characters Code "C uriosity/N ovelty" if the description of the gratification(s) associated w ith the program group refers to experiencing som ething novel, new, or unusual. Examples o f this concept include: - I get to see strange things - they're all wierd - they're all unusual - they make me feel like I'm seeing stu ff fo r the firs t tim e Code "Com panionship" if the description of the gratification(s) associated w ith the program group refers to feelings o f not being alone. Examples o f this concept include: - they make me feel like I'm not alone 204 - they keep me com pany - w atching them is like having a friend over - I watch them when there's nobody to play w ith Code “O ther G ratification(s)" if the description of the gratification(s) as sociated w ith the program group refers to fun ctio n s it serves, feelings it evokes, or kinds o f things it makes the child th in k about w hich are not included in any o f the above categories. Since the above list is supposed to cover the gratifications m ost often reported in the literature, there are no exam ples o f the “Other G ratification(s)" category. Code "Program A ttrib u te s" if the description of the gratification(s) as sociated w ith the program group does NOT refer to any g ra tifica tio n above, or to any other clearly functional reason fo r view ing, but instead refers consiste ntly to program attributes such as those identified in the “PROGRAM CONDITION" section of this manual above. For a com plete list of program attributes, see the previous section of the coding manual. Some examples of these concepts include: - They're all about fam ilies - they're all action shows - they're ail cartoons Code "O ther" if the description of the gratification(s) associated w ith the program group refers to som ething other than any o f the categories listed above. F.4 REVISIONS TO ATTRIBUTE CONDITION CODING If any attribute on the envelope has been coded using either the "Realism" or "Aesthetics" categories, they m ust be fu rth e r broken dow n in to the fo llo w in g categories: Realism - Code "Real" if the description refers to real people, or includes concepts such as "not anim ated" or "not cartoons." - Code "Unreal" if the description refers to unreal people, or includes con cepts such as "cartoons," "anim ated," or "not like real life." Aesthetics - Code "Good" if the description refers to some positive aspect of the pro gram m ing, such as "they're good," "they have good stories in them ," or "they are all better than the rest." - Code "Bad" if the description refers to som e negative aspect of the p ro - 205 gram m ing, such as "they're all bad," "they don 't have very good stories," or they're not as good as som e o f the other s tu ff on tv." 206 APPENDIX G SOCIAL REALITY QUESTIONNAIRE 207 ID Children's Ideas About Television Programs Name DIRECTIONS; First fill in your name. On the next 6 pages are some questions and statements. First read the directions then read the questions or statements and circle or check your answers below each one. Remember, it's important what YOU think. 208 DIRECTIO NS: ID __________ On this page and the next one are questions about some things real life parents and kids do. Each question asks you to tell how many parents or kids you think do something. Below each question are boxes. The boxes on the left are empty. They mean that NONE of all the real life parents or kids do what the question is about. The boxes get more and more full as you go to the right to show more parents or kids. The last ones are completely filled which means that ALL parents or kids do what the question is about. Circle the box that is closest to what you think. 1. How many parents help their kids a lot? D O n u ■ ■ NONE 10% 20% 30% 40% 50% 60% 2. How many husbands and wives help each other out a lot? D 0 0 ■ ■a n D O I 70% 80% 90% ALL NCNE 10% 20% 30% ' 40% 50% 60% 70% 80% 90% ALL 3. How many kids ask their parents for help a lot? NCNE 10% 20% 30% 40% 50% 60% 70% 80% 90% ALL GO ON TO TH E N E X T PAGE. 209 ID Answer these questions the same way you answered the ones on the last page. 4. How many kids tell their parents what to do a lot? D U O B I Q NC*E 10% 20% 30% 40% 50% 60% 5. How many kids do what their parents tell them to do? 70% 80% 90% ALL i NCNE 10% 20% 30% 40% 50% 60% 70% 6. How many husbands and wives tell each other what to do a lot? 80% 90% ALL Q Q Q n □ NCNE 10% 20% 30% 40% 5Q % 60% 70% 80% 90% ALL 7. How many parents tell their kids what to do a lot? o u y ■ ■ i n NCNE 10% 20% 30% 40% 50% 60% 70% 80% 90% ALL GO ON TO TH E N EXT PAGE. 210 ID D IRECTIO NS: Below are three statements. Underneath each one is a place for you to show how much you agree or disagree with it. On the left is a frowning face. As you move to the right, the faces frown less and smile more. Circle a frowning face if you DISAGREE with the statement. Circle a smiling face if you AGREE with the statement. The bigger the smile or frown, the more you agree or disagree. The face in the middle is if you agree and disagree about the same amount. 8. if they got a chance, most people would try to cheat me. DISAGREE agree 9. Mostly people are just looking out for themselves. DISAGREE AGREE 10. You can never be too careful in dealing with people, DISAGREE AGREE GO ON TO TH E N EXT PAGE, 211 ID ________ DIRECTIO NS: On this page and the next two are 13 questions that ask what you think about some things that go on in real life. Below each question are two answers. Check the one that is closest to what you think. 11. Do you think most people are mean or most people are friendly? mean friendly 12. Do you think most people are good or most people are bad? , good bad 13. How important is it to know how to fight? very important not very important 14. Is it all right to hit someone? no yes GO ON TO TH E N EX T PAGE. 212 ID 15. Is it safe to walk alone in a big city at night? no yes 16. Would you be afraid to walk alone in a big city at night? no yes 17. Is it safe to walk around in your own neighborhood? yes no IS. Is it safe to be in your house with the door unlocked? yes no 19. Is it safe to take a walk or ride a bicycle in the park by yourself? no yes GO ON TO TH E N EX T PAGE. 2 1 3 ID 20. During any given week, what are your chances of being involved in some kind of violence? 1 in 10 1 in 100 21. What percent of all American men who have jobs are police officers or detectives? 1% 5% 22. Are most murders committed by strangers or by relatives or people the victim knows? strangers relatives or people the victim knows 23. What percent of all crimes involve violence like murders, rapes, robbery, and assault? 15% 25% TH A P S ALL! 214 APPENDIX H ENVELOPES USED FOR SORTING TELEVISION SERIES 215 The Programs In This Group Are All Similar Because THEY ALL HAVE THE FOLLOWING THINGS IN COMMON: The Programs In This Group Are All Similar Because THEY ALL MAKE M E FEEL: A N D /O R TH IN K ABOUT: UMI Number: DP22393 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. UMI' Dissertation Publishing UMI DP22393 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest’ ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Communicating via electronic mail: Patterns and predictors of use in organizations

PDF

Correlates of children's responses to an affective television and print package designed to reduce sex-role stereotyping

PDF

Communicative performance with media: Effects and implications of television and adult-child interaction on preschooler's language

PDF

Computers and political communication: Innovation in political campaign organizations

PDF

An empirical investigation of the communicative rules which should guide the use of deceptive messages within non-interpersonal and interpersonal relationships

PDF

A dynamic reformulation of perceptions of inequity: Their organizational antecedents and outcomes

PDF

An analysis of the effects of controls imposed on the production environment of a dramatic television series, "The Senator"

PDF

Ethnic identity, language, and mass communication: an empirical investigation of assimilation among U.S. Hispanics

PDF

Building a model of organization acculturation: an interpretive study of organizational culture and stories

PDF

Communicating in organizations: The influence of context, job, task, and channel

PDF

A system of participation and allocation preferences in organizations

PDF

A novel reaction of mismatched cytosine-cytosine pairs associated with Fragile X

PDF

Computer simulation of polar solvents with the SCAAS model

PDF

Calculations of electrostatic interactions in proteins

PDF

A new approach to the problem of measuring the properties of micelles

PDF

Quantizing perception: art, communication, and cognition in the digital age

PDF

A biogeographic analysis of the freshwater fishes of Honduras

PDF

A multidimensional approach to political information seeking and political participation

PDF

Computer modeling and free energy calculatons of biological processes

PDF

¹⁹F-NMR studies of trifluoroacetyl insulin derivatives

Asset Metadata

Creator Kovaric, Peter Michael (author)

Core Title Children's judgments of the similarity of television series and the similarity of gratifications associated with television series

School Graduate School

Degree Doctor of Philosophy

Degree Program Communication Theory and Research

Degree Conferral Date 1985-12

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag mass communications,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Dutton, William (committee chair), Clark, Dick (committee member), Dorr, Almee (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-705432

Unique identifier UC11343859

Identifier DP22393.pdf (filename),usctheses-c17-705432 (legacy record id)

Legacy Identifier DP22393.pdf

Dmrecord 705432

Document Type Dissertation

Rights Kovaric, Peter Michael

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA