THE PERCEIVED REALITY 0F TELEVISION AND AGGRESSIVE PREDISPOSITIONS AMONG CHILDREN IN MEXICO Thesis item the Degree of M A. MICEISRN STATE UNIVERSiTY Felipe Korzenny “£975 ." ‘7' '4 Hz a M . y WSJR a.» 065. 4 {I} ABSTRACT LK THE PERCEIVED REALITY OF TELEVISION AND AGGRESSIVE ‘ ‘ PREDISPOSITIONS AMONG CHILDREN IN MEXICO BY Felipe Korzenny The present study was carried out in Mexico City in order to assess the effectiveness of several independent variables in predicting the perception of television content as real. A further step was taken in the theoretical sequence of media effects, by analyzing the relationship between the perception of television violence as real and aggressive predispositions of young viewers. Two hundred and seventy three Mexican children in the third and sixth grades of elementary schools in Mexico City were administered questionnaires in the Spring of 1975. Eleven hypotheses were tested with respect to eleven in- dependent variables as predictors of the perception of reality of television. The independent variables were: real life ex- periences with television content; socioeconomic status; grade in SChOOl: age; SEX: GPA; the use of television for relaxation, learning and companionship; exposure; and the influence of significant others. A twelfth hypothesis was concerned with the perception of reality of television at three different levels of abstraction: T.V. in general; content areas of T.V.; and six specific characters, groups of characters or behaviors on specific television shows. Felipe Korzenny The final hypothesis predicted that as the perception of reality of television increased, aggressive predispositions in young viewers would also increase. It was found that: l. The perception of reality of television increased with: the use of television for learning and companionship; general T.V. exposure; and with the influence of significant others. 2. The perception of reality of television decreased as socioeconomic status, grade in school, age, and GPA increased. 3. No consistent relationship was found between the per- ception of television reality with real life experiences, the use of television for relaxation, specific T.V. exposure, or the sex of the viewers. 4. Contrary to the hypothesis proposed, as the referent for television became more abstract the children tended to perceive television as more realistic. 5. The perception of reality of television violence did not correlate consistently with two different measures of aggressive predispositions. 6. For those children high in the perception of reality of television violence there was not a consistent relationship between exposure to television violence and aggressive pre- dispositions. 7. By means of multiple regression analysis it was found that the best predictors of the perception of T.V. reality Felipe Korzenny were the influence of significant others, grade in school and age, and to a lesser extent GPA and the use of TV for companion- ship. The best predictor of aggressive predispositions was found to be sex. 8. Sheer exposure to two violent shows was found to be related to aggressive predispositions, while exposure to 13 violent shows and general exposure were not. It was concluded that, among other things, further research should analyze more closely the different referents on television; the sources of experience that the child uses for evaluating what he sees on the screen; the dimensions on which the reality of television is evaluated, if at all; other variables that may precede or interact with the perception of reality of television; and other possible effects of the perception of T.V. as realistic. An analysis over time of the effects of the perception of reality of television and other variables, such as exposure, on attitudes and behavior was suggested. THE PERCEIVED REALITY OF TELEVISION AND AGGRESSIVE PREDISPOSITIONS AMONG CHILDREN IN MEXICO BY Felipe Korzenny A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Communication 1975 Accepted by the faculty of the Department of Communication, College of Communication Arts and Sciences, Michigan State University, in partial fulfillment of the requirements for the Master of Arts degree. Committee Chairman, ii to my wife Sandy, one more time, for her patience, support, help and love iii ACKNOWLEDGEMENTS First of all I must thank Dr. Bradley Greenberg, my advisor and director of this thesis for his teachings, patience, and advice. Dr. Charles Atkin and Dr. Katrina Simmons are also thanked for their advice and guidance as committee members. Dr. Rolf Wigand, Mr. Josep Rota, Ms. Patricia Arriaga, and Ms. Maria Luisa Acuna deserve a special acknowledgement for their active c00peration in the collection of the data for this study, as well as for their advice, and comments. With- out their help, this study would not have been conducted. Special mention should be made of the following persons and institutions without which this enterprise would not have been accomplished: El Consejo Nacional de Ciencia y Tecnologia of Mexico (CONACYT) for their partial sponsorship; The Depart- ment of Communication, Michigan State University, for Graduate Research and Teaching assistantships; Use of the Michigan State University computing facilities was made possible through support, in part, from the National Science Foundation; my parents; my parents in law; my grandparents; Susan Kheder, Steve Prevost, Nelson Wasensky, and Chris Faber for their re- search assistance; and Vicki Conklin for typing the original. iv I want to thank many other persons that directly or indirectly collaborated or gave their assistance in completing this task, however the number of people is too large to make a specific mention of each name. My gratitude goes to all of them. TABLE OF CONTENTS Chapter Page 1 0 INTRODUCTION ....... O .......... 0 O O O O O O O O O O O O O O O O O O l CORRELATES OF PRTV.... ........... . .......... 5 Real life experiences with television content .......... . ...... ............... 5 Level of abstraction of the referent on TV with respect ot PRTV............. 7 Demographics ........ .... ..... ..... ..... 8 Functions and gratifications from television viewing...... ..... .......... 12 Exposure to television................. 15 The influence of significant others.... 17 THE EFFECTS OF PRTV ON AGGRESSIVE PREDISPOSITIONS ......... .... ................ 18 SUMMARY 0000000000000000 A. O O O 0 O O O O O O O O O O O O O I O O 21 2. METHODS.. ..................... ... ..... ...... ..... 23 THE RESPONDENTS AND THEIR ENVIRONMENT....... 23 EXPLORATORY QUESTIONNAIRE................... 25 PRETEST QUESTIONNAIRE..... ....... ........... 26 ADMINISTRATION OF THE FINAL QUESTIONNAIRE... 28 OPERATIONALIZATIONS OF THE DEPENDENT VARIABLES. ........ .......... ........ ........ 29 Perception of reality of television.... 29 Aggressive predispositions towards problem resolution..................... 35 OPERATIONALIZATIONS OF THE INDEPENDENT VARIABLESOOOOOOOOOO.......COOOCOOOOCOOOOOOOO 38 Real-life experiences with TV content.. 38 Functions and gratifications from TV... 40 Exposure to TV................ ......... 42 The influence of significant others.... 45 Demographic or socio-structural variables ...... .............. ........ .. 48 vi Chapter Page STATISTICAL ANALYSIS .................. ..... 51 3. RESULTS.. ..................... . ................. 52 Real life experiences with TV content ...... 52 The level of abstraction for the referent onTV ...... 0.0000.00.00000.0.0000.00....0.. 53 Demographics.. ........ . ........ ............ 59 Functions and gratifications from teleViSion0.0. .......... 0.0.00.0... ...... .. 68 Exposure................................... 73 The influence of significant others (150).. 76 The effect of perceiving television violence as real on aggressive behavioral predispositions...... ..... . ................ 78 4. CONCLUSION... ................................. .. 85 DISCUSSION .............................. ... 91 IMPLICATIONS ............................... 104 APPENDIX A - QUE PIENSAS TU DE LA TELEVISION?.. ...... .. 108 APPENDIX B - INTERCORRELATIONS AMONG ALL PREDICTOR VARIABLES ....................... .... ...... 117 APPENDIX C - ZERO ORDER CORRELATIONS COMMON TO BOTH REEVES' AND THIS STUDY, FOR COMPARABLE MEASURES. . . ............ 0 ..... . ..... 0 ...... 118 BIBLIOGRAPHY ........................................... . 121 vii LIST OF TABLES Table Page 1. Eta correlation coefficients and F values from the test for deviation from linearity for real life experiences with TV content with PRTV....... 54 2. Means, standard deviations, and number of cases for responses to PRTV items at three levels of abstraction............................... ..... .. 55 2a. One factor repeated measures ANOVA for differences in responses to PRTV items at three levels of abstraction............................ 56 3. t tests for PRTV by SES ..... ...... ....... . ..... .. 61 4. t tests for PRTV by grade in school ............. . 62 4a. Zero order correlations for age with PRTV........ 63 5. t tests for PRTV by sex ....... ................... 66 6. Zero order correlations for grade point average With PRTV00000.......0.00.0000...0...0..0....0.0. 67 7. Zero order correlations for relaxation with PRTV. 69 8. Zero order correlations for learning With PRTV.0.0.0.0...00000000000000.0... ....... .0. 70 9. Zero order correlations for companionship With PRTV.... .......... 0 ........ ....0....0.I..O.. 72 10. Zero order correlations for the index of exposure to 22 Shows With PRTVOOOOOOOOI ..... 00.000.00.000. 74 10a. Zero order correlations for specific measures of exposure with PRTV....................... ....... . 75 11. Zero order correlations of the influence of significant others with PRTV...... ............. .. 77 viii Table 12. 12a. 13. Zero order correlations for two measures of aggressive behavioral predispositions with PRTV ............................................ Conditional correlations for exposure to violence of TV with aggressive predispositions at three levels of PRTV of violence. ....... ..... Relationships obtained between the independent variables and the different measures of PRTV.... ix Page 83 84 88 CHAPTER 1 INTRODUCTION The purpose of this study is to try to examine the relationship between several predictor variables and the per- ception of television violence as realistic, and the relation— ship between the perception of reality of television violence and aggressive predispositions in young viewers. It is a common, every-day experience of many of us to hear parents blaming the media for teaching their children undesirable patterns of behavior. This concern has been shared by public and private institutions in different countries. It has resulted in a drive on the part of social scientists for discovering and clarifying the nature of this phenomenon by means of several theoretical perspectives. As evidence of the general concern with this issue, it seems pertinent to cite the major research effort that has been conducted in the United States of America: the Surgeon General, in 1969, appointed a panel of social scientists in order to study the effects of television on young viewers. As a result of this enterprise a technical report of five volumes was presented to him (Comstock and Rubinstein, 1972; Murray, Rubinstein, and Comstock, 1972; Comstock and Rubinstein, 1972; Rubinstein, Comstock, and Murray, 1972; and Comstock, Rubin- stein, and Murray, 1972). In England, the British Broadcasting Corporation has sponsored and initiated several projects which raise questions with regard to the effects of television on children (Greenberg, 1974b). In Mexico, the government banned 37 shows, most of them of U.S. origin, which were considered to be the most violent of those available for broadcasting by commercial networks, after a "survey in five major cities....was undertaken to determine who watched the shows and what they thought of them.. ..the audience was largely children; parents disapproved but complained there was nothing else to watch." (Kiester, 1975) The theoretical interest of generalizing research find- ings to multiple populations is important. It is also very important to note that for countries which struggle in order to achieve an adequate state of development, like Mexico, the potential impact of televised violence on the behavior of peOple in general, and of children in particular, constitutes a major problem. The study reported here was conducted with Mexican children, and it constitutes a systematic replication (Sidman, 1960, pp. 110-139) of an original and pioneer study in the U.S. (Reeves, 1974; and Greenberg and Reeves, 1974). Several intervening variables have been studied which plausibly mediate T.V. violence watching and aggressive be- havior or predispositions (Berkowitz, 1962). However, one of these intervening variables has been systematically overlooked until recently, in spite of its seemingly obvious character, namely the perception of television programming as real. Greenberg and Reeves (1974, p.2) have succinctly stated its importance: If the child perceives program information to be realistic, to be socially useful, to be assimilated equitably with information from non-television sources, then television may blur the child's distinction between real and play. A Mexican government official, after the 37 violent series were banned declared to the press that one of the criteria for that course of action is that television violence "could be associated in the viewer's mind with present-day life" (Kiester, 1975). He implied with that assertion that not only the time context, but the relationship to the im- mediate environment of the viewer, and the impact on the receiver's perception of how the "real world" Operates. This implication has been supported by Gordon (1973, p. 19), as he found that in a sample of American children, "action in the present context was perceived to be significantly’nmuraREALISTIC than the same action in the past or future contexts." It is then the impact or effect of perceiving television content as real that is of ultimate importance, and the pre- diction of the perception of television content as real (PRTV hereafter) acquires meaningfulness accordingly. Reeves in giving suggestions for future research states: In the absence of any effects, it would seem use- less to allocate research effort to defining and predicting the perceptions....if differential effects are found, aggregate measures of various audiences are needed to verify that perceptions exist in the directions that maximize the impact of TV messages....given that audiences perceive television as depicting real life and that these perceptions make a difference in their evaluation of TV information, predicting the perceptions becomes important. (Reeves, 1974, p. 70). Reeves (1974 hereafter*), based on scarce research, addres- sed the question of what are the determinants of the percep- tion of TV programming as real in American children, and he found some evidence for the explanation of the phenomenon: 1. As questions with respect to PRTV move from abstract to more specific, the child tends to report more perception of reality; 2. the influence of significant others of a child was found to be positively related to PRTV; 3. as age decreased, PRTV in- creased; 4. as exposure to TV increased, PRTV increased; and 5. as the use of TV for relaxation, learning, and companionship increased, PRTV increased. He failed to confirm hypotheses with respect to the relationship of PRTV with real life ex- periences, socioeconomic status, intelligence, and sex. The present study attempts (1) to replicate the relation- ships for which Reeves found support; (2) to present a new test, explanations and modifications for the relationships for which Reeves didn't find support; and (3) to introduce aggres- sive predispositions as the criterion variable directly affected by the PRTV of violence. All these are examined in a different sociocultural setting, namely Mexico City. *Reeves' study of 1974 hereafter will be refered to as Reeves, excluding the year, given the continued references to that study. It should be mentioned here that by replicating a study I am trying to enforce the idea that for the formation of solid human knowledge, among other things, the repetition of research endeavors is crucial in the attempt to confirm or revise this same knowledge, and its importance is increased when the popu- lation studied is culturally different from the original one. In order to fulfill the purposes of this study, the hypotheses to be tested will be derived from Reeves', Greenberg and Reeves' (1974) work, and other related literature. CORRE LATES OF PRTV Real life experiences with television content. Symbolic or, in this case, television experiences, and "real life" experiences can be assumed to be cognitively evaluated against each other, when both are available. The question that may arise in the mind of a TV viewer is: are these peOple, or these places, or in general these representa- tions, on this show, like people, places, or in general, ex- periences I have had in real life? To this question, the viewer may give, for example, one of the following categorical answers: yes, some of them, or no. Reeves hypothesized that as real life experiences with TV content increased, PRTV would decrease, since the more elements of experience in real life that exist for comparing symbolic experiences, the less likely one would expect a person to believe what he watches on TV to be true to life. However, this relationship didn't hold in Reeves' study, for families, black people, and policemen. One must note here that families, black people and policemen, are by no means scarce in the environment where this was tested. Some of those classes of characters, or specific characters may in fact correspond to some of the real life experiences viewers had had. This might yield a relationship that says that the more real life experiences the child has the more he will per- ceive television content to be real, as actually found in some of Reeves' results. The "general pattern of findings is such that the category 'high personal experience' generally yields the highest PRTV means" (p. 38). It can be contended, however, that for non-frequent ex- periences, that can never be extreme as indicated above, one may expect the original inverse relationship to hold, e.g., in the case in which a Mexican child in Mexico has most of his experiences with Americans via TV, and only occasional contact with some American tourists, one may expect the child to assert that Americans on TV are not like Americans in real life. De Sola Pool (1965, p. 117) cites some evidence that substantiates this rationale, and says "that the effect of first-hand experience is reduction of stereotyping". It must be noted, however, that stereotyping and PRTV are not equiva- lent. The reduction of a stereotype does not necessarily imply a reduction in the report of PRTV, since what is watched on TV may correspond to what is observed in real life. This may be precisely what one may expect in the cases where the number of real life experiences is extreme or very large. So if the child knows and has interacted with a large number of families, the probability that some of the families he sees on TV are like families he knows in real life is increased. Furthermore, there is substantial evidence that indicates that TV representations are quite selective in general (Gerbner, 1972; Dominick, 1973; Clark and Blankenburg, 1972). This fact leads us to tentatively state that as real experiences with television content increase, stereotyping will decrease, but that PRTV will follow a curvilinear pattern as follows: H1: As real life experiences with TV content increase up to a middle range point, PRTV will decrease, and as real life experiences with TV content increase any further, PRTV will also increase. Level of abstraction of the referent on TV with respect to PRTV. Reeves found that the more concrete the referent for television content the more the child identified such content as reflecting real life. So, for example, the child could more precisely evaluate a specific character than TV in general. This finding is consistent with a general pattern of findings in related research (Greenberg and Dominick, 1969; Dominick and Greenberg, 1970; Greenberg, 1972; Atkin, 1971; and Ward, 1972. See Reeves, 1974 for a graphic summary on p. 8). When children were asked to respond to items such as "People on TV are pretty much like people I meet in real life", and "The same things that happen to peOple on TV happen to me in real life", they typically answer slightly below or on the middle, or not sure, part of the scale. This general pattern of findings makes sense since one would expect children, or respondents in general, to find it easier to evaluate TV content as real or not real as the referent becomes more concrete. So for example, the probability of obtaining a categorical positive answer about PRTV should increase in the following order: "The same things that happen to people on TV happen to me in real life" less PRTV than in "Fist fights on television shows are like fist fights you have participated in or you have seen in real life", which in turn should render less PRTV than in "Fist fights on 'Mission Impossible 'are like fist fights you have participated in or you have seen in real life". Accordingly it is to be expected that: H2: Children will perceive specific television characters or events to be more real than content areas of tele- vision programming and the content areas to be more real than television in general. Demographics Demographic variables, sometimes called socio-structural variables, can be expected to contribute to PRTV, by reflecting the degree and type of socialization that the child undergoes. Socioeconomic Status (SES):-Reeves hypothesized that "as socioeconomic status increases, PRTV will decrease", and -he didn't find support. He points out "that the inability of SES to predict PRTV could possibly be a function of the lack of variance obtained in its measure. Past studies reporting significant findings used economically well defined samples from different geographical areas." Among such past studies one can count Hanneman's (1972), Greenberg's (1971), Greenberg and Dominick's (1969 and 1968), and Greenberg and Gordon's (1971). One can expect PRTV to correlate negatively with SES since it has been repeatedly found that low SES children tend to use the media, and specially television as an important source of information (Feeley, 1974). Greenberg and Reeves (1974, p. 6) point out that: The typical explanation advanced for SES differences in PRTV judgments has been the more restricted Opportunities for both alternative personal and communication experiences among the less advantaged youngsters. A more descriptive rationale based on research findings can be found in Greenberg and Dervin (1972). On these basis the following hypothesis is stated: H3: As SES increases, PRTV will decrease. Agez-Lyle and Hoffman (1972, p. 175) found that: Sixth graders appeared to be considerably advanced over the first graders in rejecting television as an accurate reflection of life, and there was a further increase reflected among the tenth graders. Greenberg (1974) found in his British study that as the child grew older, his PRTV decreased, and Reeves found that in general as age increased, PRTV decreased with the exception of PRTV among black children. 10 This encountered relationship makes sense taking into consideration that the sheer availability of elements of cogni- tive evaluation increases as the child grows older, and con- sequently the tendency to report that TV reflects real life should decrease. Reeves operationalized age as the grade in school the respondents were in, which may be a legitimate Operationaliza- tion when there is a small variance of chronological age with- in grades in school. However, in Mexico it is expected that the variance of ages within school grades will be large, especially for low SES children. Then, grade in school may not reflect the range of experiences that the child accumulates as he grows up. It is then an interesting empirical question to find what is a better predictor of PRTV, grade or age. Both variables will be assessed, and the same relationship is hypothesized: H4: As grade in school increases PRTV will decrease. H4.: As age increases PRTV will decrease. Sex:-Dominick and Greenberg (1970) found some evidence for the notion that girls perceive television as more realistic than boys. Reeves didn't find support for this relationship. Although he didn't offer a rationale for such finding, it is possible to think in terms of socialization practices (Mischel, 1970). Boys and girls in the U.S.A., across differ- ent socioeconomic statuses, ages, and other structural characteristics, can be conceived to have more similar 11 experiences than boys or girls in a country like Mexico, given the larger emphasis that the American society puts on equal opportunities and rights for both sexes. Mexico, follow- ing more traditional patterns of child rearing practices, provides differential opportunities for socialization outside of the home for boys and girls. The place of the girl is at home, and the boy may go play foot ball or do something else outside of his home with friends. One may expect then that for Mexican girls, television can still be a major source of experiences, just because they tend to be more at home, and because the TV set is there. Consequently, H5: Females will perceive television programming to be more real than males. Intelligencez—Reeves hypothesized that as intelligence increased, PRTV would decrease. No past research had dealt with this relationship, but Reeves offered the rationale that "this factor could be important in determining both the amount and the reliability of information about the real world a child has to compare with television content." He adds that even after the information from the real world is gathered by the child, his ability to "compare relevant facts" can be thought also to be a function of intelligence (p. 14). He found moderate support for the hypothesis, and observed that those with the highest IQ scores perceived television to be less realistic, but that differences did not occur across the entire range of IQ scores. Reeves obtained IQ measures for 12 only about 50% (101 respondents) of his sample, where there were no IQ scores at all available for the oldest children (sixth graders). Given the small set of scores, one has reason to believe that if a more complete set of intelligence measures is obtained, some more conclusive evidence may be encountered. For intelligence we may then hypothesize that H6: As intelligence increases PRTV will decrease. Functions andggratifications from television viewing. The analysis of the functions that the media serve to their audience has been a topic of recent concern (Blumler and Katz, 1974). If the gratifications that the audience expects are obtained, the strength of the media usage behavior is ex- pected to be reinforced, and the probability of media usage in the future is increased depending on the frequency of the attainment of satisfactions (Skinner, 1969). Greenberg (1974, p. 88) says that the perception of reality of TV shows "strikes us as an intervening variable in relation to other effects of exposure to television content", and it would theoretically follow the functions sought by the viewer, with behavioral effects being the ultimate ones. Greenberg and Dominick (1969) found, in a preliminary approach to identify the uses that children of different back- grounds made of television, that socio-structural variables made a difference with respect to the reasons that children gave for watching TV. 13 Greenberg (1974) examined the reasons for watching tele- vision that a sample of 726 British children gave, and analyzed them with respect to other variables, PRTV among them. He found seven main reasons for which the children in his sample watched TV: 1. for learning; 2. out of habit; 3. to relax; 4. to forget; 5. for arousal; 6. for companionship; and 7. to pass time. Out of the seven, three of them emerged as strOng correlates of PRTV as follows: learning, r=.28; relaxation, r=.23; and companionship, r=.22. All the correlations were significant beyond the .01 level. Greenberg in the concluding remarks of this same study indicated: We also doubt that these motivations are peculiar to British children. We would expect to find the same kinds of categoriei in similar studies of American children, or 0 any children, for that matter. Indeed, they may be generic across viewing audiences, differing only in emphasis and salience for adults as well as others. Such a major impli- cation obviously requires verification, however, and a follow-up study on such an issue is in order. (p. 89) Reeves proceeded to try to replicate Greenberg's findings as indicated above, and found moderate support for the relation- ship between relaxation and PRTV, and some better support for the relationships between learning and companionship with PRTV. Given Greenberg's and Reeves' findings and suggestions, it seems plausible to search for the relationships encountered in this third study, in a third cultural setting. The use of TV for relaxation:- It has been long argued that the effectiveness of TV for influencing its audience stems 14 from the idea that the audience has its "defenses" down while watching the tube, at home, which is a secure place (see for example Krugman, 1971). There are arguments to the contrary which are highly plausible (see for example Bauer, 1964), however this View is specially tenable when the member of the audience goes to the media specifically for relaxing. In such a case, when the defenses are presumably down, the viewer may tend to believe what he sees more than when he goes to the media for other purposes. Consequently we may hypothesizetfiufl: H7: As the use of TV for relaxation increases, PRTV will increase. The use of TV for learning:-Presumably when one reports going to the media in order to learn, it's because the media, and specifically TV reflect real life in the sense of pro- viding insights into one's own life and the world in general. The viewer would not want to learn misleading ideas, but those that can guide him to a better understanding of things. It seems to be that the person, or in its case child, who goes to the TV set for learning would endorse what he views as reflecting real life, and this same viewer may want to try the "reality" he sees on the screen in his own life. It is hard to conceive of a person trying to learn things that are not perceived as efficacious for dealing with his environment then H8: As the use of TV for learning increases, PRTV will increase. 15 The use of TV for companionshipz—"Young children seek physical contact with or at least seek to be near certain other people." (Maccoby and Masters, 1970, p. 73) However not all children can satisfy this companionship tendency, and their relative deprivation of this social function may result in the substitution of human companion- ship for something that resembles it. This may include the media, and specifically, in this case, television. If tele- vision is conceived of as a friend, it may be thought of as a companion that tells it like it is, at least in a wishful thinking manner. In View of the lack of a more solid theory, the rationale above, as well as the rationale offered for the relationship between relaxation and learning with PRTV, constitute as good a theoretical guess as can be made at this point. With respect to companionship, in this study, it is to be expected then that H As the use of TV for companionship increases, PRTV will 9: increase. Exposure to television. Dervin and Greenberg (1972, p. 200) summarize research evidence that show that low income people watch more tele- vision than the general audience. Greenberg and Dominick (1969) in a study of 392 fourth and fifth graders found that low income children watched longer than high income youngsters, and that children from low income background were more likely to believe that TV 16 content was more realistic than children from higher income families. Furthermore Greenberg in his study of British children found that exposure correlated with PRTV (r=20, see Reeves, 1974, p.17). Greenberg (1972) also found that the frequency of watching ShOWS‘WhiCh feature black people in different roles is related to the perception of reality of those characteri- zations by children. The above findings seem to be consistent. Low income young viewers tend to watch more, and they tend to perceive TV content as more realistic than more privileged children. Those children who watch more seem to be less able to differ- entiate reality from fantasy, since their backgrounds have not equipped them with the kind of information that would permit them to discount as fantasy what they watch. The restricted environment of the economically disadvantaged leads to the seeking for symbolic stimulation in the picture tube, and the lack of other sources of information for the evalu- ation of the information received seems to lead to higher levels of acceptance as real of what is watched on television. Reeves found that as exposure increase, PTRV increased consistently when using a general measure of exposure , although with less consistency when he used a specific measure of ex- posure for content categories (blacks, families , and police) . In this study with Mexican children we expect accordingly that H As exposure to TV increases, PRTV will increase. 10' 17 The influence of significant others. In the sociological literature significant others are conceptually defined as "those persons who exercise major influence over the attitudes of individuals" (see Woelfel and Haller, 1971, p. 75). "Others are significant in direct proportion to the amount of information they convey to an ego about categories he uses to define objects and self... affective factors not withstanding." (Ibid. p. 76) Sociolo- gists have consistently found that the influence of signifi- cant others is decisive in the formation of educational or occupational aspirations (Picou and Campbell, 1975; Duncan, Haller and Portes, 1970; Haller and Woelfel, 1972; etc.), and on other attitudinal or behavioral aspects (Woelfel and Hernandez, 1972, etc.). The work of Woelfel in general assumes a linear model of attitude formation, namely that "an individual attitude equals the vector sum of all information relevant to that behavior an individual receives" (Woelfel, 1972, p. 11). He points out that despite the simplicity of the model, the Woelfel-Haller aggregate expectations show the highest zero order correlations with occupational aspirations (r= .64), and with educational aspirations (r= .66). This model has been recently refined by Woelfel and Saltiel (1975). Reeves found that the information a child receives from significant others about the reality of television program— ming is positively related to the child's perceived reality 18 of such content, at the three levels of abstraction that he studied. It can be argued against this finding that the in— fluence of significant others was measured as perceived by the child. However, recent evidence indicates that children tend to only slightly underestimate the influence they receive from others (Woelfel, 1975). Given the above, it is to be expected that in the Mexican sample H11: The information a child receives from significant others about the reality of television programming, as per- ceived by the child, will be positively related to the child's perceived reality of that programming. THE EFFECTS OF PRTV ON AGGRESSIVE PREDISPOSITIONS Concentrated attention on the perception of television as real, as an intervening variable, could constitute a futile task unless some behavioral effects were anticipated. If such behavioral effects are not to be found, further ex- ploration of the determinants of PRTV are unjustified. Reeves suggested that "it must first be shown that effects of differential reality perceptions do exist. In the absence of any effects, it would seem useless to allocate research effort to defining and predicting perceptions." (p. 70) The present study will attempt to clarify the character of the relationship between PRTV violence and aggressive behavioral predispositions. Given the lack of agreement as to the definition of violence and aggression (Surgeon General's Scientific Advisory Committee on Television and Social Behavior, 1972, 19 p. 5), the definition for both terms will be that used by Wood (1974), given his comprehensive synthesis of the literature. physical behavior which is performed with intent to injure another person. The behavior may or may not result in physical and/or psychological injury. The behavior is not condoned by societal values. (p. 5) Berkowitz (1962, pp. 229-255) in a theoretical review of aggressive effects as a result of exPosure to media violence concludes that certain members of the audience of “the media featuring aggressive content may be relatively likely to accept what they see as true and real", because they may not be able to discount what they see as make believe. "The fantasy world does impinge upon them to a relatively great degree and consequently, if the conditions are right, can excite action." Atkin (1971) found partial support for the relationship of violence viewing to aggressive predispositions when mediated by the perception of the violent content as real. The support is said to be partial since he found that the relationship held only for one of the samples that he studied (Maryland), in which the multiple correlation of a set of variables with aggressive predispositions increased from .32 to .39 when PRTV was taken into consideration. Feshbach (1972) conducted three experiments in order to find out, among other things, if children that were presented with a violent program labeled as real showed more subsequent aggression than children who were presented with a violent 20 program labeled as fantasy. His findings can be said to have corroborated his expectations in general. However, the fantasy or reality of the shows was defined a priori by the experi- menter at the time of the presentation of the show, and there was no manipulation check. Thomas and Tell (1974) found an increment of aggression for angered subjects who were told they watched a real scene of violence, as compared with subjects who were told they watched a fictional scene of violence and others who didn't watch any filmed scene at all. Subjects who watched violence labeled as real who were not angered were also more aggressive than the rest of the groups but not as much as those who were angered in addition. There was a manipulation check in this experiment for the anger condition, but non whatsoever for the reality or make believe of the scenes. Greenberg (1974 b) in a study with British children, found a correlation of .16 (p < .01) between items such as "The shows on TV tell about life the way it really is," and "Sometimes a fight is a good way to settle an argument." This finding suggests one more time that the perception of TV violence as real may contribute to the display of aggressive predispositions and perhaps of actual antisocial behavior. Wood (1974) in an exploratory experiment found a signifi- cant association between PRTV and hypothetical aggression. In comparing a fantasy and a news context for the presentation of violence to children he found that: 21 Children who perceived the fight as very real and viewed the fight in the fantasy context were very aggressive, more aggressive than children in the news condition who perceived the fight as very real, or children in either condition who perceived the fight as less real. (p. 71) This result, although tentative, corroborates previous findings to some extent, and emphasizes the importance of the differential perception of stimuli, regardless of the intention of the source of the message. In general, there is enough reason to expect that children , or receivers in general, who perceive TV violence as real will be more likely to relate such violence to their own lives for problem resolution when problems are encountered. A person may not be expected to try to apply to his own situ— ation methods of problem resolution that are perceived as fantasy, since make-believe may prove to be misleading and to aggravate situational conflicts. To the extent that the media is considered to be a school for real life, one may expect in the Mexican sample that: H : As the perception of TV violence as real increases, 12 aggressive predispositions toward problem or conflict resolution will increase. SUMMARY The present study attempts to constitute a systematic replication of Reeves' efforts to predict PRTV, with a sample of Mexican children in this case, and to explore the attitud- inal consequences of perceiving TV violence as real. The theoretical model to be tested here includes the perception of television content as real as a result of exposure, real 22 life experiences with TV content, functions and gratifica- tions from TV, demographic variables, the level of abstrac- tion for the referent on television, and the influence of significant others as perceived by the child. In turn, the perception of TV violence as real is expected to enhance aggressive behavioral predispositions. In a graphic form the model to be tested can be visual- ized as follows: Demographic .> PRTV variables Influence of Predispositions significant ““9 Level of '———‘——9 toward others abstraction problem for the TV resolution Real life "-_‘——§ referent ‘"_-—“"-'"w'"“ experiences I ”P with TV content [Exposure] Functions gand {gratifications In this model only a correlation is implied by the arrows, and given the nature of this study, the interrelation- ships not explicitely stated will not be considered, since that work should await some more solid theoretical basis. CHAPTER 2 METHODS The data for this study were collected during the winter of 1975 from children in two elementary schools in Mexico City. The presentation in this chapter will follow this order: 1. the respondents and their environment; 2. ex- ploratory questionnaire; 3. pretest questionnaire; 4. administration of the final questionnaire; 5. Operation- alizations of the dependent variables; 6. operationaliza- tion of the independent variables; and 7. statistical analysis The data were physically collected in Mexico City by two students of communication at the Universidad Iberoamericana in that city, under the supervision of two faculty members.* Explicit instructions were sent to Mexico City by this writer, and a continuous flow of feedback took place during the entire period of data collection. THE RESPONDENTS AND THEIR ENVIRONMENT Three hundred children were administered the exploratory instrument and the final questionnaire. In order to maximize *The present study would not have been conducted without the help of Ms. Patricia Arriaga and Ms. Maria Luisa Acuna, and that of the two faculty members: Professors Josep Rota and Rolf Wigand. Their professional collaboration is highly acknowledged and appreciated. 23 24 the differences among respondents, half the respondents were from very low socioeconomic backgrounds, and the other half from very high socioeconomic families. Half of them were third graders, and the other half were sixth graders. There was a roughly equal number of boys and girls in each sub- division. The distinctive characteristics of the high socioeconomic status children were that their parents paid approximately sixty dollars a month for tuition, the school was located in a residential neighborhood, the children had extracurricular activities in school such as painting lessons, and English and French as foreign languages. The school had a swimming pool, a gymnasium, and the children were required to use a class uniform and a gymnastics uniform as needed. This was a private school. The low socioeconomic status children attended a school where no tuition was paid. The school was located in an industrial area of the city. The children had strictly curriculum activities. The use of uniform was not enforced. About ten per-cent of the children did not wear shoes, most of them wore old clothes, did not have a pencil to work with, and even during the winter season some of them did not wear a sweater. This was an official school. It is to be noted that in Mexico private schools are not unusual, and that most well-off families send their children to that kind of school. 25 EXPLORATORY QUESTIONNAIRE In order to obtain a preliminary overview of the viewing habits of the children in this sample, an exploratory question- naire was adminstered to them in the Fall of 1974. Of all the children, 92% said that there was a tele- vision set in their homes. Of the 5% that said that they didn't have a TV set at home, with the exception of one re- spondent, they either watched TV at the house of some friends, some relatives, or a neighbor. When the children were asked whether they watched a black and white or a color TV set, most of the time, 60% said that they watched a black and white set, 37% said they watched a color set, and 3% didn't answer. Since cable television has been introduced in Mexico City, the children were asked whether they had cable TV at home. The shows on CATV are a direct selection of the U.S. programming on the air. Of all the respondents, 11% said they had CATV at home, 82% said they didn't, and 7% didn't answer. The distribution of ages of our respondents was as follows at the time of the administration of the exploratory questionnaire: Ages Number 1 71 54 21 71 42 17 15 2 % .3 24 18 7 24 14 6 5 .7 26 This distribution confirmed our expectation of having a wide spread of ages within grades in school. Grade, SES, and sex were split almost exactly in halves, as indicated in our description of respondents. The children were presented with a list of 75 shows from which they were to check the ones that they watched every week or almost every week, and the ones they watched every day or almost every day. Of the 75 shows, 48 were weekly and 27 were daily. The purpose of this rating was to enable the selection of those shows to which the children most expose themselves. However, since some of the shows on the air changed at the time of the administration of the final questionnaire, some last minute decisions had to be made. It was found that the average child in our sample watched an average of approximately 60 shows a week, excluding nonscheduled events, feature films, contest or game shows, and newscasts. PRETEST QUESTIONNAIRE A pretest questionnaire was administered to 50 children. Half of them similar to those of low SES in the third grade, and the other half were similar to those of high SES in the sixth grade. Both groups had approximately an equal number of boys and girls. The pretest was to test the understandability of the scales, the use of the response categories, and for general improvement of the questionnaire according to the doubts and 27 questions that the children might point out to. The comments of the teachers were also to be taken into consideration. The timing of the administration was an important criterion for the evaluation of the questionnaire as a whole. 1 The administration of the pretest questionnaire was con- ducted by school groups, and the administrator read each question and response category aloud. He waited until the last child finished in order to continue. The administrator asked the children for any questions they had, and took note of all questions and comments. This served for modifying the questionnaire in terms of language, adapting it as closely as possible to the verbal repertoire of the child. The in- structions for answering the questions had to be simplified for understandability and because the younger and poorest children tried to read everything, and according to the administrators of the questionnaire these children read a word per minute". In some cases, response categories had to be eliminated because the children simply did not use them at all. The time of administration was 25 minutes for the 6th graders, high SES, and an hour and 25 minutes for total completion in the case of the 3rd graders, low SES. Since this last administration time was considered to be too long, it was decided to substantially reduce the length of the questionnaire in order to avoid extreme exhaustion on the part of these children. Nine pages, a page per show, were 28 originally assigned to the measurement of PRTV of specific shows and the influence of significant others with respect to those shows. The final number of pages in this section was reduced to six, a show per paws. Other minor reductions were also implemented. ADMINISTRATION OF THE FINAL QUESTIONNAIRE The final questionnaire was administered to the same children to whom the exploratory questionnaire was given. The shortest time of administration was of 25 minutes for the 6th graders, high SES, and the longest was of an hour and 20 minutes for the 3rd graders, low SES. Despite our efforts at simplifying and shortening the questionnaire, the 3rd graders, low SES children still took a period of time com- parable to those in the pretest. This was due mainly to the larger number of children, and to the personal attention that they required for the completion of their task. All children were guided question by question through the questionnaire, and the administrators walked in between the rows of desks in order to make sure that the questions were being considered and answered, as well as for answering all questions the children had. The children were given the assurance that the question- naire was not a test, and that they would not be evaluated in any form for their responses. The children were told to work independently, and most of them did so. However, some children tried to consult with their friends, and the 29 administrators had to intervene. With the exception of the children in the 3rd grade, low SES, the administration was a smooth operation, in general. With the 3rd graders, low SES, the amount of work required was disproportionate but there were no major problems that would at first impression invalidate the questionnaires answered by them. The teachers of the groups did not participate in the administration of the questionnaires. The reader will find a copy of the final questionnaire in Appendix A. The total number of usable respondents was 273. OPERATIONALIZATIONS OF THE DEPENDENT VARIABLES There were two dependent variables studied in this pro- ject: a. the perception of reality of TV by the child at three levels of abstraction, and b. aggressive predisposi- tions towards problem resolution. a. Perception of reality of television. PRTV was measured at three levels of abstraction: tele— vision in general; content areas on television; and specific television characters or behaviors. The items used to measure this variable were the ones used by Reeves, and the same or similar to the ones used by other researchers. Since the questionnaire was administered in Spanish, the questions and scales used will be translated back for this presentation. PRTV in general was measured with the following items: 1. "TV programs show life the same way you see it in reality." 3O 2. "People on TV shows are like people you know in real life." 3. "The same things that happen to people on TV shows, can happen to peOple you know in real life." 4. "The places you see on TV shows are like the places you know in real life." These items were mixed with the content PRTV items. The response categories for these were: That's true I don't know That's not true coded as 3,2 and 1 respectively from tOp to bottom. The last three items intercorrelated significantly: .12, r23= r24= .17, and r34: .32 (p < .05), and they were summed to form an abstract PRTV index which ranged from 3 to 9, with the higher score indicating higher PRTV. The distribution of scores for the abstract PRTV index were: Score on the abstract PRTV index N % 3 9 3.4 4 19 7.1 5 34 12.7 6 45 16.8 7 65 24.3 8 44 16.4 9 52 19.4 Total = 268 100.0% x = 6.78 s.d. = 1.66 For the second level of abstraction questions were generated for three content areas of television programming. These areas were selected according to their availability and the actual or potential interest that they may represent for theoretical work. The content areas selected were 31 families, Americans,and fights. Families represent an area of experience that is presumably abundant in real life and on TV. Americans in Mexico are relatively scarce in real life, but quite abundant on TV, and represent an area of special interest for the study of the formation of stereotypes. Fights, besides being available on TV and in real life, present the opportunity for studying the hypothesized effect of PRTV on aggressive predispositions. The content PRTV items were: 1. "Families on TV shows are like families you know in real life." 2. "The Americans that appear on TV shows are like Americans you know in real life." 3. "Fights on TV shows are like the fights you have been involved in or you have seen in real life." The scales accompanying these questions and the way they were coded were the same as in the case of the abstract PRTV items. Since only the items for families and Americans intercorrelated significantly (412 = .23), it was decided not to form a content PRTV index, and to study each content category separately. The distributions obtained for the three content PRTV items were: Content PRTV of families N i l 78 28.7 2 82 30.1 3 112 41.2 Total = 272 100.0% 2: 2.13 s.d. = .828 32 Content PRTV of Americans N % l 56 20.7 2 87 32.1 3 128 47.2 Total = 271 100.0% X = 2.27 s.d. = .781 Content PRTV of fights N % l 87 32.3 2 52 19.3 3 130 48.3 Total = 269 100.0% X = 2.16 s.d. = .885 At the last level of abstraction PRTV was measured for six specific characters or behaviors on TV. Two of these characters or behaviors were included for each of the three content areas specified above. The questions for specific families were: 1. "Do you think that the family in 'Hogar Dulce Hogar' is like the families you know in real life?" 2. "Do you think that the Partridge Family is like the families you know in real life?" The questions for specific Americans were: 1. "Do you think that Tony Black 'The Magician' is like the Americans you know in real life?" 2. "Do you think that the children in 'Family Affair' are like American children in real life?" And the questions for specific fights were: 1. "Do you think that the fights on 'Mission Impossible' are like the fights you have been in or you have seen in real life?" 2. "Do you think that the fights on 'Hawaii 5-0' are like the fights you have been in or you have seen in real life?" 33 The possible responses to the six questions were "yes" coded as 3, "I don't know" coded as 2, and "no" coded as l. The shows were selected from the exploratory questionnaire described above, according to the criteria of exposure and suitability to the three content areas. The distributions of PRTV for the six shows were as follows: Hogar Dulce Hogar _N .z The Partridge Family N_ _Z l 94 44.1 1 48 19.3 2 60 28.2 2 75 30.1 3 59 27.7 3 126 50.6 Total = 213 100.0 Total = 249 100.0 I = 1.84 i = 2.31 s.d. = .833 s.d. = .777 The Magician E_ .1 Family Affair N_ ‘Z l 59 29.5 1 29 14.1 2 92 46.0 2 92 44.9 3 49 24.5 3 84 41.0 Total = 200 100.0 Total = 205 100.0 §= 1.95 i = 2.27 s.d. = .735 s.d. = .694 Mission Impossible N_ 2_ Hawaii 5—0 N_ .Z 1 114 53.5 1 85 48.6 2 59 27.7 2 48 27.4 3 40 18.8 3 42 24.0 Total = 213 100.0 . Total = 175 100.0 32: 1.65 X= 1.75 s.d. = .778 s.d. = .818 The average number of children who answered these items was 209, or 77% of the total number of respondents. Answers 34 to these items by children who had previously indicated that they don't watch the show at least sometimes were disregarded as missing observations.* With only one exception, no indexes were formed with these items since their intercorrelations were inconsistent as can be seen on the next page. One index was formed by the addition of all six items divided by three, for a-posteriori comparisons among the means of the perception of reality of television at three different levels of abstraction, as reported in the chapter of results. Ten measures of PRTV have been obtained: An abstract PRTV index; Content PRTV of families, of Americans, and of fights; and six measures of PRTV of specific characters or behaviors. The intercorrelations among the different measures of PRTV were as follows: *Previous to these items in the questionnaire, the children were asked to report their frequency of exposure to each of the six shows, as described in the section of exposure to TV in this chapter. Answers to items about a specific show were disregarded if the child said he "never" watched the show. 35 Abstract PRTV index (1) Content PRTV families (2) .06 Content PRTV Americans (3) .03 .23* Content PRTV fights (4) .14*.08 .06 Hogar Dulce Hogar (5) .07 .08 .02 .24* * * * The Partridge Family (6) .10 .29 .22 .12 .19 The Magician (7)-.07 .05 .21*.06-.oo .03 Family Affair (8)-.05-.01 .10-.04-.05 .13* .26* Mission Impossible (9)—.06 .10 .01 .21*.29*.06 .22* .04 Hawaii 5—0 (10) .02 .18*.00 .13*.32*.13 .18* .08 .57* (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) b. Aggressive predispositions towards problem resolution. This was the second dependent variable in this study. It was operationalized in two ways, the first being a measure of predispositions towards problem resolution, and the second a measure of self report of involvement in fights. The first Operationalization was borrowed from Leifer and Roberts (1971), and adapted to the Mexican children after the pretest. The items consist of situations in which the child may find himself in his everyday life, and several pos- sible modes of conflict resolution are given to him to choose from. Leifer and Roberts rationalize that the child modifies his rank ordering of possible responses depending on a set of variables among which exposure to television violence is one. Findicates that the correlation is significant at p:.05 36 The test retest reliability of the items, as reported by Leifer and Roberts was r=.72, and the correlations between the ratings of two teachers and the responses of physical aggression of the children were r=.33, and r=.49 (pp. 52-59) , which can be con- sidered to be a measure of the validity of the instrument. The validity and reliability coefficients mentioned here are all for physical aggression, and the responses were coded as dy- chotomous variables, e.g., physical aggression response = 1; any other choice = 0. After the pretest with the Mexican children, several of the items were altered, and the category of verbal aggression was deleted since the children didn't use it at all. The situations that remained in the final set were as follows: 1. "You are walking down the street. A child is mad at you. He comes and hits you. What do you do?" 2. "You see that somebody is stealing your sandwich. You catch him. What do you do?" 3. "You are waiting on line to drink water. Somebody comes and pushes you. What do you do?" 4. "Somebody is telling stories about you behind your back. You notice it. You see him after school. What do you do?" 5. "When you are leaving school you see two children hitting your best friend. What do you do?" All references to sex in Spanish were avoided when pos— sible. The responses that the children could give to the situations above were: a. for physical aggression, depending on the item: push back, hit back, kick, or pinch; b. for withdrawal, depending on the item: leave them, and go away; and c. for denounce, depending on the item: tell a grown up, tell the teacher, or that's all right. The response categories were alternated, and the items were scored as 1 when the 37 physical aggression response was used or as 0 when any other response was used. The inter-correlations among the five items were: (1) .25 .39 .41 .44 (2) .14 .37 .30 (3) .29 .32 (4) .39 (5) (l) (2) (3) (4) (5) All the correlations were significant (p<.05) and an index of physical aggressive predispositions was formed by adding the responses to all 5 items. The index ranged from 0 to 5, with 5 the highest aggressive score obtainable. The distribution obtained is shown below: Number of physical aggressive responses N i 0 81 30.2 1 52 19.4 2 44 16.4 3 40 14.9 4 37 13.8 5 14 5.2 Total = 268 100.0 E = 1.78 s.d. = 1.59 The second Operationalization of aggressive predisposi- tions consisted of a self report of frequency of engaging in physical fights. The item used for this purpose was "How often do you get into fights?", and the response categories 38 were many times, sometimes, almost never, and never, coded respectively as 4, 3, 2, and 1. The distribution obtained was: Report of engagement in fights N i l 89 32.7 2 70 25.7 3 83 30.5 4 30 11.0 Total = 272 100.0 E: 2.20 s.d. = 1.02 The correlation of the index of situational aggression items with the self report of fights was .24. In summary aggressive predispositions were measured only with respect to physical aggression, and two different measures were used. It should be clarified that the word "fight" in Spanish, in the way and context used here means fist or other type of physical fight. OPERATIONALIZATIONS OF THE INDEPENDENT VARIABLES The independent variables considered here were: 1. real life experiences with television content; 2. functions and gratifications from TV; 3. exposure to television; 4. the influence of significant others; and 5. demographic variables. Real-life experiences with TV content Two questions were generated for each of the three content areas considered here, with respect to real life 39 experiences the child could have had. The questions with respect to real life experiences with families were: 1. "How often do you play with your friends in their homes?" 2. "How often do you talk to the families of your friends in their homes?" For real life experiences with Americans the questions were: 3. "How often do you see Americans in real life?" (not on TV) 4. "How often do you talk to Americans?" For real life experiences with fights the following questions were used: 5. "How often do you see people fighting in real life?" (not on TV) 6. "How often do you get into fights?" The response categories available to the children were: "many times" coded as 4, "sometimes" coded as 3, "almost never" coded as 2, and "never" coded as 1. It should be noted that item 6 is the same one utilized for the self report of engagement in fights, and it was not correlated or otherwise analyzed, with itself or with an index containing it. The correlation for the two items about experience with families was .33; for experience with Americans r=.33; and for experience with fights r=.18 (all coefficients p<.01). Consequently an index of experience was formed for each pair of items by summing them. The scores could range from 2 to 8 for each index with the higher number indicating more real 40 life experiences. The distributions obtained for each of the indexes were: Families 5 g Americans g g Fights g i 2 9 3.3 2 39 14.6 2 22 8.1 3 8 3.0 3 30 11.2 3 33 12.2 4 29 10.7 4 61 22.8 4 57 21.0 5 62 22.9 5 58 21.7 5 55 20.3 6 105 38.7 6 46 17.2 6 62 22.9 7 40 14.8 7 20 7.5 7 34 12.5 8 18 6.6 8 13 4.9 8 8 3.0 Total = 271 100.0 Total = 267 100.0 Total = 271 100.0 2 = 5.62 E = 4.58 x = 4.87 s.d. = 1.31 s.d. = 1.65 s.d. = 1.56 The intercorrelations among the three indexes were as follows: Families with Americans r=.13 (p<.05); families with fights r=.10 (n.s.); and Americans with fights r=-.02 (n.s.). Functions and gratifications from TV Three items were used for measuring each of these uses or functions of television: relaxation, learning and com- panionship. Children were asked to indicate the frequency with which they went to TV for a certain function or gratifi- cation. For relaxation the items used were: 1. "How often do you watch TV in order to be tranquil?" 2. "How often do you watch TV in order to calm down when you are in a temper?" 3. "How often do you watch TV because it's a nice way to rest?" The items used for learning were: 4. "How often do you watch TV for finding out about the 41 things that happen in the world?" 5. "How often do you watch TV in order to learn how to behave?" 6. "How often do you watch TV in order to learn things about yourself?" For companionship the following are the items that were used: 7. "How often do you watch TV because it's like a real friend for you?" 8. "How often do you watch TV in order not to be alone?" 9. "How often do you watch TV when there is nobody to talk to or to play with?" These items are adapted variations of the items used by Reeves and by Greenberg (1974). The response categories were: "always" coded as 4; "many times" coded as 3; "some- times" coded as 2; and "never" coded as l. The intercorrelations for each of the functions were as follows (the subscripts correspond to the item numbers here): Relaxation: r12=.34; rl3=.27; and r23=.24 Learning: r45=.38; r46=.37; and r56=.46 Companionship: r78=.39; r79=.22; and r89=.43 An index was formed for each of the functions since the intercorrelations were moderately high in general and signifi- cant (p<.01). Each index ranged from 3 to 12, with the higher score indicating the higher reported frequency of occurence of a certain type of function or gratification. The distributions for each of the indexes were: 42 Relaxation N g 3 0 0 4 11 4.1 5 7 2.6 6 57 21.4 7 40 15.0 8 51 19.2 9 38 14.3 10 25 9.4 11 15 5.6 12 22 8.3 Total = 266 100.0 )2”: 8.01 s.d. = 2.07 Learning N % Companionship N g 3 2 .7 3 6 2.2 4 16 5.9 4 3 1.1 5 27 10.0 5 6 2.2 6 51 19.0 6 33 12.2 7 28 10.4 7 35 13.0 8 47 17.5 8 53 19.6 9 27 10.0 9 33 12.2 10 34 12.6 10 40 14.8 11 14 5.2 11 24 8.9 12 23 8.6 12 37 13.7 Total = 269 100.0 Total = 270 100.0 X = 7.79 X = 8.64 s.d. = 2.31 s.d. = 2.21 The intercorrelations among the three indexes were: relaxation with companionship = .48; relaxation with learning = .53; and for companionship with learning = .48. Exposure to TV Exposure to TV was measured in two different ways: First, the children were presented with a list of 22 shows on the air at the time of the administration of the 43 questionnaire. They were asked to put a check to the right of the names of the shows that they watched "every week or almost every week". A check was coded as 1, and no check was marked as 0. Then the shows watched by each child were sum- med to render an index of exposure to 22 shows. The names of the shows were randomized throughout the list. The distribution of this measure of exposure was: Exposure to 22 shows 0 2 1 2 0 0 3 l 0 4 l 0 5 2 l 6 7 3 7 10 4 8 15 5 9 14 5 10 18 7 11 17 6 12 27 10 13 23 8 14 18 7 15 17 6 16 23 8 17 14 5 18 15 5 19 10 4 20 ll 4 21 11 4 22 17 6 Total = 273 100 i = 13.78 s.d. = 4.66 The names of the 22 shows were: Wild Wild West, The Pink Panther,The Flintstones, Tom and Jerry, The Monster Family, Police Trilogy (McMillan/McCloud/Columbo), Land of Giants, Cannon, The Adams Family, The Streets of San Francisco, 44 Laurel and Hardy, El Show de los Polivoces, Toma, I Love Genie, La Criada Bien Criada, El Chapulin Colorado, Bonanza, Gunsmoke, Lost in Space, The Name of the Game (Robert Stack), and Disneyland. The second measure of exposure was the shows in the content areas studied here: Hogar Dulce Hogar, The Partridge Family, The Magician, Family Affair, Mission Impossible and Hawaii 5-0. The children were asked to indicate whether they watched each of these shows "almost every week" coded as 2, "sometimes" coded as l, and "never" coded as 0. None of these six shows were included in the prior index of exposure to 22 shows. An index of exposure to these specific six shows was formed by adding up the scores, and the resulting range was from 0 through 12, with the higher number indicating more exposure. The resulting distribution was: Exposure to six shows N % 0 3 1.1 l 2 .8 2 6 2.3 3 11 4.2 4 24 9.1 5 31 11.8 6 35 13.3 7 42 16.0 8 43 16.3 9 34 12.9 10 12 4.6 11 12 4.6 12 8 3.0 Total = 263 100.0 x u n 0‘ 6o b s.d. 45 The correlation coefficient between these two indexes of exposure was .59 (p<.01). The index of exposure to six specific shows is a measure comparable to the abstract PRTV index. However, for comparison with the PRTV of content areas and of specific shows, different subsets of this index had to be used. For comparability with each of the content areas three subindexes were created: one for exposure to families on TV composed of the sum of the specific measures of exposure to "Hogar Dulce Hogar" and "The Partridge Family"; one for exposure to Americans on TV composed of the sum of the measures of exposure to "The Magician" and "Family Affair"; and another one for exposure to fights as the sum of the measures of exposure to "Mission Impossible" and "Hawaii 5-0". All three subindexes could range from 0 to 4, with the higher number meaning more exposure to the content area. For comparability of the PRTV of each of the shows with specific exposure, the single scores on the scales of exposure to the specific shows were used. The influence of significant others For each one of the shows in the specific content areas, the child was asked to report his sources of influence for evaluating the reality of each of the shows. The items used for each of the six shows were: 1. "Do you talk to your friends about (character or behavior) on (name of Show)?" 2. "Do you talk to your mother or father about (character or behavior) on (name of show)?" 46 3. "Do you talk to your brother or sister about (character or behavior) on (name of show)?" The children were asked to respond to these items for each of the six shows with a "yes" coded as l, or a "no" coded as O. The distribution of percentages of affirmative responses was: Friends Parents Brother or Sister Shows % yes % yes % yes 1. Hogar Dulce 54 66 65 Hogar 2. The Partridge 70 61 68 Family 3. The Magician 62 54 56 4. Family Affair 53 58 64 5. Mission 51 50 55 Impossible 6. Hawaii 5-0 53 58 51 It can be seen that the children tended to be consistent across significant others and across shows. No marked pre- ference for one type of significant other is observed, and the influence seems to be generalized for all shows. The scores for the three items were added in order to render a composite measure of amount of interaction. The average bias about the reality of television characters or behaviors from significant others was operation- alized as follows: 4. "Do the people that you talk to about (character or behavior) on (name of show), think that (the character or behavior on the show) are like (peOple or behaviors) in real life?" 47 To this item the children could answer "yes" coded as 1, "I don't know" coded as 0, or "no" coded as -1. The sum of items 1, 2, and 3 was then multiplied by item 4, and the resulting index ranged from -3 to +3, for each of the six shows. This index represents the amount and quality of in- formation the child receives from significant others with respect to the reality of these television shows. A high positive score represents a large amount of in- fluence in favor of TV reality. The distributions for each of the indexes obtained were: For families: ISO* Hogar Dulce Hogar .N .2 ISO The Partridge Family .N .Z -3 32 15.6 -3 20 8.2 -2 24 11.7 -2 7 2.9 -1 23 11.2 -1 10 4.1 O 84 41.0 0 134 54.7 1 5 2.4 1 8 3.3 2 16 7.8 2 11 4.5 3 21 10.2 3 55 22.4 Total = 205 100.0 Total = 245 100.0 3E= —.327 §= .453 s.d. = 1.78 s.d. = 1.71 For Americans: ISO The Magician N_ Z_ ISO Family Affair N. ‘Z -3 17 8.7 -3 12 6.0 —2 5 2.6 —2 4 2.0 —l 5 2.6 —1 5 2.5 0 138 70.4 0 136 67.7 1 9 4.6 1 8 4.0 2 4 2.0 2 11 5.5 3 18 9.2 3 25 12.4 Total = 196 100.0 Total = 201 100.0 32= .026 i = .279 s.d. = 1.37 s.d. = 1.4 *ISO, hereafter stands for the influence of significant others about. 0 0 . 48 For fights: ISO Mission Impossible .N E_ ISO Hawaii 5—0 N_ .Z -3 29 16.4 -3 16 10.3 —2 21 11.9 -2 14 9.0 —1 16 9.0 -1 10 6.4 0 81 45.8 0 97 62.2 1 8 4.5 1 l .6 2 10 5.6 2 6 3.8 3 12 6.8 3 12 7.7 Total = 177 100.0 Total = 156 X=-.458 §=-.237 s.d. = 1.65 s.d. = 1.47 Four more indexes were created from these same data to provide for adequate comparability with the different levels of abstraction of PRTV. The influence of significant others was summed across all shows to form a general index of ISO for comparability with the abstract PRTV index, and each pair of shows within each content area was summed to form an ISO index for families, Americans, and fights. Demographic or socio-structural variables Socioeconomic status (SES):-Socioeconomic status was operationalized according to whether the child attended the high or low SES school. The distribution of the children by SES follows: SES N 3 Low 134 49.1 High 139 50.9 Total 273 100.0 49 Intelligencez-The intelligence of the children was Opera- tionalized as their grade point average for the school year 1974-1975 as reported by the teacher of each group of children. The tests from which the average was computed were uniform tests for all schools, since in Mexico the Ministry of Edu— cation formulates the tests to be given to all children in all grammar schools in the Country. The grading system in the grammar schools in Mexico ranges from "A" to "D", an "A" representing the best perfor- mance, and a "D" the worst. The distribution of grade point averages obtained was: Intelligence (GPA) N g D 25 9.2 C 85 31.4 B 116 42.8 A 45 16.6 Total = 271 100.0 Age and grade in school:- The children were asked to report their age to their last birthday, and the results were: 522 N i 7 1 .4 8 38 14.0 9 55 20.3 10 22 8.1 11 56 20.7 12 52 19.2 13 26 9.6 14 14 5.2 15 6 2.2 16 1 .4 Total = 271 100.0 ’ 10.72 >4 II II s.d. 1.91 50 The children in this study were either in 3rd or sixth grade in school, and they were distributed as follows: Grade in school N g 3 134 49.1 6 139 50.9 Total = 273 100.0 Sexz-When children were asked to report their sex, they were distributed as show below: Sex N % Boys 127 46.5 Girls 146 53.5 Total = 273 100.0 The intercorrelations obtained for all the demographic variables are in this matrix: SES -.07 .54* —.39* .07 GRADE IN SCHOOL .06 .80* —.05 GPA -.19* .18* AGE -.09 SEX 1—1 0 O :L‘. U m Z H m D m a 4 m N m m U m m o o < m * stands for p<.01. 51 The intercorrelations among all the predictor variables can be found in Appendix B. STATISTICAL ANALYSIS The information contained in the questionnaires was transferred to computer cards. The analysis of the data was conducted in a CDC 6500 computer, and the statistical package utilized was SPSS (Statistical Package for the Social Sciences) in its versions 5.8 and 6.0 (Nie, Bent, and Hull, 1970; and Nie, Hull, Jenkins, Steinbrenner and Bent, 1975; respectively). The hypotheses were tested, and the analysis in general was conducted, using Pearson Product Moment Correlations, Eta correlation coefficients, partial correlations, t tests, one way Analysis of Variance for repeated measures, Scheffe tests, and multiple regression. The analysis conducted in each case is indicated in the results chapter for each of the hypotheses. Results were considered to be significant* when they reached the conventional level of probability of .05 or less. Smaller probabilities are also indicated for information to the reader. Where some specific results are described either in the text or in tables as falling short of significance, they are at probability levels .07
.05.
*The word significant in this text only refers to statistical
significance, and no other inferences should be made with
respect to the importance of the results on the basis of that
adjective.
CHAPTER 3
RESULTS
In this chapter the results of the hypotheses presented
in the introduction will be tested in the order specified by
the subscripts of those hypotheses.
The analysis conducted will be described in each of the
twelve sections that encompass the test of each prOposition.
Real life experiences with TV content.
H1: As real life experiences with TV content increase
up to a middle range p01nt, PRTV w111 decrease,
and as real life experiences with TV content
increase any further, PRTV will also increase.
This hypothesis states a nonlinear type of relationship
between PRTV and real life experiences with TV content. The
Pearson product moment correlation coefficient is a measure
of association adequate for linear relationships. The Eta
correlation coefficient is a measure of association for any
kind of relationship, linear or non~linear, and will tend to
be larger than r (see McNemar, 1969, p. 312; and Blalock,
1972, pp. 410-411).
In order to test this hypothesis Eta correlation co-
efficients were computed, and F tests for deviation from
linearity were obtained (for computing formulas see Blalock,
1972, p. 412). The results are shown in Table 1. A relation-
ship was considered to deviate significantly from linearity
52
53
if an F value was significant at p': .05, and would then
warrant further examination through a scattergram. Since
from all the relationships only one was found to significantly
deviate from linearity, no further examination was done, and
this hypothesis is consequently rejected.
It should be mentioned that r coefficients for all
relationships in Table 1 were found to be low and non-signi-
ficant, except for the correlations between real life experi-
ences with Americans and the PRTV of the shows "The Magician"
and "Family Affair", which were .22 and .36 respectively
(p<.001). However, being positive, these correlations go
counter to our expectations that since Americans are rather
scarce as a source of real life experiences in the Mexican
environment, the more real life experiences with them, the
less PRTV. Another exception was found for real life experi-
ences with fights and the content PRTV fights (r=.l6; p<.01).
The level of abstraction for the referent on TV.
H2: Children will perceive specific television
characters or events to be more real than
content areas of television programming,
and the content areas to be more real than
television in general.
According to the Operationalization of the three levels
of abstraction for the referent on TV, the expectation was
that the means for the more specific questions would be
higher than the means for the content questions, and these
in turn would be higher than the means for the abstract
PRTV questions. However, the means as presented in Table 2
54
Table 1. Eta correlation coefficients and F values from the
test for deviation from linearity for real life
experiences with TV content with PRTV.
Independent Dependent Eta F value from the Significance
variable variable test for deviation of F (p<)
from linearity
GENERAL INDEX ABSTRACT .20 .75 n.s.
OF EXPERIENCE* PRTV INDEX
EXPERIENCES CONTENT PRTV .15 .90 n.s.
WITH FAMILIES FAMILIES
PRTV HOCAR .19 1.37 n.s.
DULCE HOGAR
PRTV THE .23 2.58 .05
PARTRIDCE FAMILY
EXPERIENCES CONTENT PRTV .13 .54 n.s.
WITH AMERICANS AMERICANS
PRTV THE .25 .72 n.s.
MACICIAN
PRTV FAMILY .38 .62 n.s.
AFFAIR
EXPERIENCES CONTENT PRTV .19 .67 n.s.
WITH FIGHTS FIGHTS
PRTV MISSION .14 .46 n.s.
IMPOSSIBLE
PRTV .21 1.38 n.s.
HAWAII 5-0
* For comparability with the abstract PRTV index, the three indexes of
real life experiences with families, Americans and fights were added
to form a general index of experience.
55
Table 2. Means, standard deviations,
and number of cases
for responses to PRTV items at three levels of
abstraction.
Item Mean
Abstract PRTV items:
Abstract PRTV index ............ 2.26
People on TV are like .......... 2.27
The same things that happen.... 2.25
The places you see ............. 2.27
Content PRTV items:
Families on TV shows ........... 2.13
The Americans that appear ...... 2.27
Fights on TV ................... 2.16
Specific PRTV items:
The family on "Hogar'Dulce
Hogar" ....................... 1.84
"The Partridge Family" ......... 2.31
"The Magician" is like
Americans .................... 1.95
"Family Affair” are like
Americans .................... 2.27
The fights on "Mission
Impossible" .................. 1.65
The fights on "Hawaii 5—0" ..... 1.75
Standard
deviation
.55
.81
.83
.78
.83
.78
.89
.83
.78
.74
.69
.78
.82
Number of
cases
268
271
272
271
272
271
269
213
249
200
205
213
175
56
.mCOHum>pomno wcfimmaa mo cowumaww omfizumfia umumm muoomnsw no no woman mum mabmu mflnu ca mCOHumusmEou HH< *
mm. maw mm.aa swoop
am. ems 36.88 Haooaaom
Hoo.vo mm.ms o~.m N mm.o monaoaoz amazoom
mm. owa ka.km oaooom oaooaz
am. as mH.oN osooaa ooosoom
mozaoHaHonm a mmaoom zamz so mm oneaHmas to momsom
as. mm. mm. zoceacsmo
@8723
oo.H aN.N mm.~ zamz
Aoo Ame Aav
mmoH>