The Relations of Early Television Viewing to School Readiness and Vocabulary of Children from Low-Income Families: The Early Window Project

by John C. Wright, Aletha C. Huston, Kimberlee C. Murphy, Michelle St. Peters, Marites Piñon
Citation
Title:
The Relations of Early Television Viewing to School Readiness and Vocabulary of Children from Low-Income Families: The Early Window Project
Author:
John C. Wright, Aletha C. Huston, Kimberlee C. Murphy, Michelle St. Peters, Marites Piñon
Year: 
2001
Publication: 
Child Development
Volume: 
72
Issue: 
5
Start Page: 
1347
End Page: 
1366
Publisher: 
Language: 
English
URL: 
Select license: 
Select License
DOI: 
PMID: 
ISSN: 
Abstract:

Child Development, September/October 2001, Volume 72, Number 5, Pages 1347-1366

The Relations of Early Television Viewing to School Readiness
andvocabulary of Children from Low-Income Families:
The Early Window Project

John C. Wright, Aletha C. Huston, Kimberlee C. Murphy, Michelle St. Peters,
Marites Pifion, Ronda Scantlin, and Jennifer Kotler

For two cohorts of children from low- to moderate-income families, time-use diaries of television viewing were collected over 3 years (from ages 2-5 and 4-7 years, respectively), and tests of reading, math, receptive vocab- ulary, and school readiness were administered annually. Relations between viewing and performance were tested in path analyses with controls for home environment quality and primary language (English or Spanish). Viewing child-audience informative programs between ages 2 and 3 predicted high subsequent performance on all four measures of academic skills. For both cohorts, frequent viewers of general-audience programs per- formed more poorly on subsequent tests than did infrequent viewers of such programs. Children's skills also predicted later viewing, supporting a bidirectional model. Children with good skills at age 5 selected more child-audience informative programs and fewer cartoons in their early elementary years. Children with lower skills at age 3 shifted to viewing more general-audience programs by ages 4 and 5. The results affirm the con- clusion that the relations of television viewed to early academic skills depend primarily on the content of the

programs viewed.

INTRODUCTION

Television viewing is consistently blamed for a myr- iad of social and developmental problems, including poor school achievement. ~hesesweeping condem- nations ignore the obvious fact that television con- tains an enormous variety of forms and content; its impacts on children's development probably depend on the content and the degree to which programs are designed to convey that content effectively to devel- oping audiences. Educational programming has be- come a regular, if small, part of broadcast offerings. It is widely assumed that children learn cognitive and social slulls from such programs, but the evidence supporting these beliefs is somewhat scattered.

The purpose of the 3-year longitudinal study de- scribed in this article was to investigate the relations between young children's television viewing experi- ences and their performance on tests of school readiness and vocabulary. The sample consisted of children from families with low-to-moderate incomes because
such families use television extensively (Huston & Wright, 1997) and because many educational pro- grams (e.g., Sesame Street) are targeted at them. As a result, they may experience both the positive and negative effects of television more than do children from higher income families. We approached this topic with two core assumptions. First, different types of television programs can have correspondingly dif- ferent relations to children's intellectual development. Second, such relations are probably reciprocal. Chil- dren's intellectual skills may be affected by viewing, and their acquired slulls may affect subsequent view- ing choices.

Relations of Television Viewing to Achievement

Most investigations of television and school achieve- ment have been conducted with school-age children, and most have not distinguished among different types of program content. Correlational studies show a small but consistent negative relation between con- current total television viewing and various indices of school achievement (e.g., language, cognitive skills), even with controls for confounding variables (Com- stock, 1995).A meta-analysis of 23 studies found that the average correlation of total viewing with school achievement was -.05; the overall relationship was curvilinear, with students who watched about 10 hr a week performing best in school (Williams, Haertel,
Walberg, & Haertel, 1982).

In two longitudinal investigations, by contrast, few relations between viewing and later achievement were found once appropriate confounding variables were controlled. Children in the National Health Ex- amination were assessed twice (between ages 6 and 11and again between ages 12 and 16years). Although total time watching television in each age range was negatively associated with performance on achieve-

O2001 by the Society for Research in Child Development, Inc. All rights reserved. 0009-3920/2001/ 7205-0005

ment tests in adolescence, the relation dropped to zero when controls for grade-school test performance, re- gion, and family characteristics were included (Gort- maker, Salter, Walker, & Dietz, 1990). A study follow- ing three panels of children over 3 years (grades 1-3, 3-5, and 6-8) showed temporal patterns that sup- ported the hypothesis that total television viewing contributed to lowered reading skill over time among the oldest cohort of children studied (Ritchie, Price, & Roberts, 1987).

Some negative effects of total television viewing on reading acquisition in the early years were suggested in a natural experiment in British Columbia. A cross- sectional comparison of towns that did and did not receive television and an examination of changes after the introduction of television produced evidence of weak negative effects of television viewing on read- ing skills in Grades 2 and 3. No effects were observed for Grade 8 children (Corteen & Williams, 1986).

The relation between watching television and school achievement may vary for children living in different environments and from different cultural contexts. In one investigation, reading achievement was associ- ated with low viewing for children from high SES families, but with high viewing for children from low SES families (Searls, Mead, & Ward, 1985). Among students with limited English proficiency, the amount of television viewed was positively associated with achievement in a large survey of California students (Comstock, 1991). Viewing may interfere with school achievement among more advantaged children, but it may contribute to knowledge or language ability for less advantaged children, possibly because of the dif- ferences in alternative opportunities for learning that are available in their homes and neighborhoods. The mixed results concerning the relations of television viewing and school performance may stem in part from the failure to consider the content of the programs viewed and from the wide range of ages included.

Relations of Television Content to Achievement

Educational programming for young children is based on a straightforward assumption that well- designed television can convey academic and social slulls to viewers. Most of the research evaluating the impact of educational programming on language and school-related skills has been conducted with Sesame Street. It is a very popular program, and it emphasizes linguistic and academic skills more than do most other educational programs (Neapolitan & Huston, 1994). During the first 2 years after its inception, multi- site field experiments were conducted to determine its effects on preacademic skills. Children whose par- ents were encouraged by researchers to have them watch Sesame Street improved more than a nonencour- aged control group on such academic and cognitive skills as letter recognition, numerical skills, and clas- sification (Ball & Bogatz, 1970; Bogatz & Ball, 1971). Similarly, a summative evaluation of The Electric Com- pany showed that viewing the program at home fos- tered improved reading performance among children in Grades 1through 4 (Ball & Bogatz, 1973). Survey information collected in the 1990s indicated that Sesame Street viewers had better school-related skills in lundergarten than did nonviewers, but the data were based on parent reports (Zill, Davies, & Daly, 1994).

Longitudinal methods have also been used to ex- amine the relations between viewing and language development and school achievement. In one investi- gation, viewing Sesame Street between ages 3 and 5 pre- dicted improvements in receptive vocabulary at age 5 (Rice, Huston, Truglio, & Wright, 1990). There was also a tendency for children who watched informa- tive programs designed for children and adults to perform better on letter recognition and reading skills (Truglio, Huston, & Wright, 1986). When these same children were studied as adolescents, preschool view- ers of child-audience informative programs had higher grades in high school than did nonviewers, even with controls for family characteristics (Anderson, Hus- ton, Schmitt, Linebarger, & Wright, 2001). Rosengren and Windahl(1989) found that Swedish preschoolers who watched programs specifically designed for them had higher school test performance in Grade 1, which in turn predicted higher marks in Grade 6, com- pared with preschoolers who viewed other programs.

By contrast, viewing general-entertainment content tends to be associated with lower achievement. Among Swedish preschoolers, watching fiction pro- grams in preschool predicted poor school marks at Grades 1 and 6 (Rosengren & Windahl, 1989). These investigators proposed that early viewing of preschool educational programs starts a positive cycle predicting viewing of informational programs and higher grades, while early viewing of fiction starts a negative cycle predicting more fiction consumption and lower grades.

In a longitudinal study of preschool and early school-age children in the United States, exposure to adult entertainment television predicted poor letter recognition at age 5, but was unrelated to later read- ing comprehension (Truglio et al., 1986). There was no relation between viewing cartoons or adult entertain- ment programs and vocabulary (Rice et al., 1990), but one 3-year longitudinal study in the Netherlands showed that viewing entertainment programming was negatively related to subsequent reading com- prehension. There was a positive, although nonsignif- icant, effect for watching informational programs (Koolstra, van der Voort, & van der Kamp, 1997). Cross-sectional studies have shown that viewing sports, cartoons, and music videos predicted low school performance (Neuman, 1988; Potter, 1987; Smyser, 1981), but students who watched news performed better in school than nonviewers (Potter, 1987).

Taken together, these studies paint a picture suggest- ing positive effects of informative programming and small, but consistent negative effects of viewing enter- tainment programs on reading and academic achieve- ment. Nonetheless, there are gaps in our knowledge. Few investigations included careful separation of television content or intended audiences. The early experimental studies of Sesame Street were criticized because the treatment included parental encourage- ment; hence, it was not clear that positive results would have ensued from viewing without adult in- tervention. In particular, the critics raised questions about the value of the program for children from eco- nomically disadvantaged families (Cook et al., 1975). Few investigations of entertainment programs in- clude preschool children, even though it could be ar- gued that television may be a more critical influence in the early years when children are not receiving much formal school instruction than it is after they enter formal schooling.

Finally, if television does affect children's intellec- tual development, little is known about the processes involved. Observational learning is one likely mecha- nism by which children learn the planned messages in educational programs as well as the unplanned messages in many other types of programming. Pro- grams such as Sesame Street and Mister Rogers'Neigh- borhood have planned, age-specific curricula, and they contain linguistic and production techniques designed to enhance learning (Huston & Wright, 1994; Rice, 1984). Learning from television may also contribute to a reciprocal process through which children select their television and play environments to further the interests and motives acquired from television (Bandura, 1978). Long-term effects may come about as a result of entering school with readiness skills that facilitate success and teachers' perceptions of good school ability (Entwisle, Alexander, & Olson, 1997; MacBeth, 1996).

Cartoons, by contrast, contain language and hu- mor that are beyond the comprehension abilities of preschool children (Rice, 1984), as well as high levels of violence. According to one theory, the resulting ag- gression leads to low levels of academic performance because it interferes with the ability to sit still and at- tend to academic content (Huesmann & Miller, 1994).

Wright et al. 1349

Cartoons also have high rates of action, pacing, and visual and auditory effects, all of which may lead to arousal and activity, which are incompatible with ac- quisition of language and academic skills (Singer, 1980). There is some evidence that viewing violence in the early years predicts lowered academic perfor- mance in high school (Anderson et al., 2001).

Although the content of general-audience programming is highly variable, much of it is light enter- tainment that is not designed to convey educational messages. The level of vocabulary and required back- ground knowledge is, for the most part, not geared toward young children. Two processes have been proposed to account for predicted negative influences of entertainment television, including cartoons, on in- tellectual development: television displaces other, more valuable ways to spend time and general-audience television does not engage the child's intellectual effort.

According to the displacement hypothesis, television viewing displaces more intellectually valuable activ- ity, resulting in lowered levels of language and intel- lectual functioning. In another analysis of the data reported in this article, we examined displacement processes by analyzing differences among individuals in time allocation as well as individual changes over time in relation to changes in the time spent in other activities. Both analyses provided support for the hy- pothesis that general-audience programming displaces educational activities (including reading), social interaction, and video game play (Huston, Wright, Marquis, & Green, 1999; Wright & Huston, 1995).

For young children, in particular, general-audience programming may displace or interfere with the child's opportunity to interact with adults. Children usually view general-audience programs in the com- pany of parents or other adults (St. Peters, Fitch, Huston, Wright, & Eakins, 1991). Adults watching television may be unresponsive and inattentive to the child. As a result, children may have fewer social and linguistic interactions with adults whle deriving little from the tele- vision content that is occupying the adults' attention.

The second hypothesis is that entertainment tele- vision does not engage the child's mental effort. The most common version of this hypothesis is that enter- tainment television makes few intellectual demands and, as a result, creates habits of intellectual laziness and disinterest in school (Koolstra & van der Voort, 1996; Salomon, 1984). Some support for this hypothesis appeared in a longitudinal study of second- through eighth-grade Dutch children, which found that televi- sion viewing was associated with a subsequent de- crease in positive attitudes toward reading. Attitudes toward reading, in turn, predicted reading achievement and time spent reading (Koolstra &van der Voort, 1996).

But, for very young children, much of general- audience television may be difficult rather than easy to comprehend. Children are most likely to become actively engaged with television content that is nei- ther too easy nor too difficult to comprehend, that is, content that provides some challenge, but also al- lows an attentive child to gain a sense of mastery (Rice, Huston, &Wright, 1982; Wright & Vlietstra, 1975). A child who spends a lot of time in the presence of television that is not comprehensible may miss such opportunities.

Methodological Issues

For the most part, correlational methods, both cross- sectional and longitudinal, are the only viable designs for studying the relations of television viewing to ac- ademic and intellectual development, because true experiments are virtually impossible to mount. Al- though longitudinal methods provide some basis for inferring causal direction, two issues of interpretation are important. First, it is likely that correlations of viewing with child outcomes may result from bidirec- tional or reciprocal effects; viewing may influence academic performance, and academic skills may in- fluence choices about television viewing. Analyses of longitudinal data rarely have been designed to detect these bidirectional patterns. Second, correlations of viewing and achievement may be a function of other factors. The most well-established influences on both viewing and achievement include a host of demo- graphic and family characteristics (e.g., education, income, and ethnic group), all of which probably rep- resent variations in children's home environments (Clarke & Kurtz-Costes, 1997; Huston & Wright, 1997; Truglio, Murphy, Oppenheimer, Huston, & Wright, 1996). Such characteristics can be statistically con- trolled, but it is always possible that unmeasured variables account for all or part of the associations observed.

Many studies are fraught with problems of mea- suring television viewing accurately. Viewing measures sometimes consist of one or a few questions with du- bious validity asking individuals how much time they spend watching television. Written or oral view- ing diaries are far more accurate (Anderson, Field, Collins, Lorch, & Nathan, 1985), but few studies have used them.

Purposes of the Present Study

The major purpose of the present study was to in- vestigate the relations of television viewing to school- related skills and receptive language development during the preschool and early school years. We chose to study an ethnically diverse sample repre- senting families with low to moderate incomes be- cause these groups have been underrepresented in much developmental research and because television may play an especially important role in their lives. The study represents an advance over earlier longitu- dinal investigations because it began at age 2 to cap- ture very early viewing experience and because the home environment was measured directly rather than relying solely on demographic indicators of fam- ily characteristics. Television viewing was assessed as it occurred naturally, thereby avoiding the problems created by "encouragement to view." Finally, multi- ple 24-hour diaries of children's time use in all activi- ties (not just television) were used to obtain viewing information across time and to provide specific infor- mation about programs watched, and other catego- ries of time use.

METHOD

Design

The design of the study is shown in Figure 1. There were four annual occasions of measurement, or "waves," consisting of a visit in the project office and a home visit. The year elapsing between waves was called a "period; that is, Period 1was between Waves 1 and 2, and so on. Thus, in this study, there were three periods and four waves.

Sample

The volunteer sample comprised two hundred and thirty-six 2- and 4-year-old children and their parents, whose family incomes were low to moderate. Fami- lies were recruited from census tracts in Kansas City, MO, and Kansas City, KS, with poverty rates higher than 10% and from populations receiving some form of means-tested government assistance (such as pub- lic housing). An additional group lived in neighbor- hoods in Lawrence, KS, where the median incomes were among the lowest in the city. Families were re- cruited to participate in a study of television and young children's development through community agencies, Special Supplemental Food Program for Women, Infants, and Children (WIC), Parents as Teachers, preschools and child care centers, posters in supermarkets and laundromats, radio announcements, local-access cable television channels, and word of mouth from other participants.

The path analyses presented in this article are based on 182 children who had complete data or who

Wright et al. 1351

1990- -1991 1992- 1993
Cohort 1Age (Years) 2 3 4 5
Cohort 2 Age (Years) 4 5 6 7
Office and Home Visits Wave 1 Wave 2 Wave 3 Wave 4
Bimonthly Time-Use     f------+
Phone Diaries Period 1 Period 2 Period 3
  Figure 1 Design of the study.

were missing only one test score, one assessment pe- riod, or one viewing period over the 3-year time span. These missing data were imputed by regressing scores for a particular wave or period on scores for the same measure in the other waves or periods, using the sample for the cohort in which the child be- longed. Overall, 2.1% of the scores for the younger co- hort and 1.5% of the scores for the older cohort were imputed values.

The demographic characteristics of the initial and path analysis samples are shown in Table 1. Parents in the analysis sample were less likely to receive public assistance, more likely to be married, better educated, less likely to be African American, and had higher in- comes and higher scores on the Home Observational Measure of the Environment (HOME; Caldwell & Bradley, 1984) than those without complete data (all ps < .01). Nevertheless, the analysis sample was eth- nically diverse and represented families with low to moderate incomes.

Time-Use Diary

In some earlier studies of home viewing, parents kept written diaries of all television viewing over pe- riods of 7 to 10 days (Anderson et al., 1985; Huston, Wright, Rice, Kerkrnan, & St. Peters, 1990). Video- tapes of a sample of families during the diary period demonstrated high validity; the correlation, for pre- schoolers, between total viewing time scores from the two methods was .84 (Anderson et al., 1985). Written diaries are, however, tedious to maintain, and we were advised that asking low-income families to com- plete repeated diaries would risk highly selective at- trition and missing data. Therefore, we elected to col- lect oral time-use diaries, using a method that has been employed extensively by time-use researchers (see Juster & Stafford, 1985).

Throughout the 3 years, families completed peri- odic telephone interviews. The calls served to track any changes in family composition, residence, living arrangements, or child care, and to collect a time-use diary. In these telephone interviews and at each office and home visit after the initial contact, the parent (typically the mother) was asked about the child's ac- tivities on the previous day. The parent described the child's activities sequentially from midnight to mid- night. For each activity, they reported the amount of time spent and who else was present. Ii the child was watching television or the television was on in the same room, the name and type of program being viewed was recorded. The validity of this type of time-use diary has been established through multiple

Table 1 Characteristics of the Total Sample and the Path

Analysis Sample
Whole Sample (N = 236) Path Analysis Sample (n = 182)
Characteristic M SD M SD
Maternal education (y) Income-to-needs ratio Married (%) HOME standard score African American (74) Hispanic American (76) European American (%)        

Native American (%)
Note: HOME = Home Observational Measure of the Environment.

mformants (e.g., husband and wife) and multiple meth- ods (e.g., written and phone) (Juster & Stafford, 1985). This method has an added advantage over most mea- sures of television use; the person is reporting on all activities rather than focusing solely on television. In addition, the report is recent and quite specific in time. Our goal was to obtain six diaries per year-

one about every 2 months. In the early stages of the study, however, we had more difficulty reaching people and completing calls than expected, so staff time de- voted to calling was increased in the later stages of the study, resulting in a higher rate of completed calls. The average number of 24-hr diaries per child was 3.3 in Period 1,4.4 in Period 2, and 5.1 in Period 3.

The first activity named in a time interval was con- sidered the primary activity. If the child was simulta- neously doing something else, that activity was des- ignated "secondary." For example, if the parent said the chld was playing and also watching television, then play was recorded as primary, and television viewing as secondary. If the child was "watching television while she ate lunch," then eating was the secondary activity. If children spent part or all of the day away from the parent (e.g., in child care), the adult in charge was also called to get time-use information for that portion of the day.

Classification of Programs Viewed

All programs viewed were first classified by the in- tended audience (made for children or for general- adult audiences); cluld programs were further grouped by content type using a system developed from ear- lier studies (Center for Research on the Influences of Television on Children, 1983). Four categories of pro- grams resulted: (I) child-audience, informative or ed- ucational programs (regardless of animation); (2) child-audience, fully animated cartoons with no in- formative purpose; (3) child-audience, other pro- grams (neither fully animated nor informative pur- pose); and (4) general-audience programs. Reliability averaged 92% agreement, calculated by comparing codes assigned by two or more independent raters who were familiar with the program.

The average number of minutes of viewing was calculated for each of the four categories of programs: child-audience informative, child-audience cartoons, child-audience other, and general audience. Viewing on weekdays and weekends was calculated sepa- rately, because the kinds of programs offered are different. For each category of child-audience pro- grams, totals included primary and secondary view- ing of live broadcasts and videotapes. For general- audience programs, primary and secondary viewing were separated for descriptive purposes, but com- bined in the path analysis. The number of minutes spent viewing each type of program was calculated for each 24-hour diary. The score for a 1-year period was the average of all weekday or weekend diaries collected during that year up to and including the of- fice visit for the next wave.

Home and Family Characteristics

The demographic characteristics of the families (parent education, occupation, family structure, pri- mary language) were ascertained during the parent interviews at each annual wave. At each home visit, the HOME (Caldwell & Bradley, 1984) was collected. Using observation and unstructured questions, the interviewer awarded points for a list of observed con- ditions, facilities, responses to questions, and social and emotional features of the mother-child interac- tion. A large number of studies show that this instm- ment predicts IQ and school performance of children from a wide range of socioeconomic backgrounds, in- come levels, and ethnic groups (Bradley et al., 1989). The version for infants (ages 0-3) was used for the 2- year-olds; the preschool version was used for chil- dren who were 3 through 5 years of age; and the ele- mentary school version was used for children 6 and over. The Wave 1 HOME score for each child was en- tered as a control in the analyses reported here. Be- cause the infant and preschool versions have different numbers of items, standard scores for each version were calculated for the sample and used in the analyses.

All home visitors during Wave 1 were trained by an expert from the University of Arkansas at Little Rock, where the instrument was devised. Interob- server reliabilities were checked periodically by having two interviewers attend the same visit. One inter- viewer conducted the visit, and both scored it inde- pendently. A minimum criterion of 80% agreement was maintained.

Academic Skills and School Readiness

Reading and number skills. Basic academic skills are at the center of the concept of school readiness. These were measured by standardized tests administered individually during the office visit at Waves 2,3, and

4. Two subtests from the Woodcock-Johnson Tests of Achievement were selected. This battery was stan- dardized on a large, nationally representative sample from age 2 through 79 (Woodcock & Mather, 1989). Both subtests have internal consistencies greater than .90 for children in the age range of 3 through 7. Read- ing and prereading skills were assessed by the letter- word recognition subtest measuring recognition of icons, letters, and words. Math skills were assessed by the applied problems subtest, which involves counting, inequalities, and simple arithmetic.

Vocabulary. The Peabody Picture Vocabulary Test- Revised (PPVT-R) measures receptive English vocab- ulary. It requires the child to point to one of four pic- tures on a page that illustrates or exemplifies a spoken word. The words range from simple objects to com- plex concepts. The test was standardized on an ethni- cally diverse sample ranging from age 2% to 70, but one in which families with low occupational status were somewhat underrepresented. For children ages 3 through 7, internal consistencies range from .70 to .83 (Dunn & Dunn, 1981).

School readiness. General school readiness was as- sessed by the School Readiness Scale of the Bracken Basic Concepts Scales (Bracken, 1984). The test in- cludes knowledge of colors, shapes, letters, num- bers, and spatial and size relations. The standard- ization sample for the test included children from multiple ethnic groups. The internal consistency of the School Readiness Scale is at least .90 for children in the age range of 3 through 5, and test-retest reliability is .98 (Bracken, 1984). It was administered only to the younger cohort, because pilot testing in- dicated that many of the older cohort were likely to reach its ceiling by the third or fourth wave of data collection.

Standard scores based on the published test norms for the child's chronological age were used for all four tests. The Woodcock-Johnson and Peabody Pic- ture Vocabulary tests have standard scores based on a mean of 100 and a standard deviation of 15. The School Readiness Scale scores were transformed to a distribution with a mean of 100 and a standard devi- ation of 15 in order to achieve comparability with the other tests.

Examiners

Because the sample included children from differ- ent ethnic groups, it was important to have diversity among the examiners. The examiners represented various ethnic groups and nationalities, including African American, European American, Filipino American, Mexican, Spanish, Mexican American, Israeli, Iranian, and Singapore Chinese. All spoke fluent English. For children from families in which the primary language was English, there was no effort to match the ethnicity of the examiner with that of parents and children, so many combinations occurred.

For children from Spanish-speaking families, a

Wright et al. 1353

fully bilingual graduate student whose first lan- guage was Spanish administered all tests. Tests that involved written or spoken English (i.e., reading and vocabulary) were presented in English. Written translations of the instructions for the applied prob- lems subtest and some portions of the Bracken School Readiness Scale were prepared by a Mexican graduate student (whose first language was Span- ish). If the child appeared more fluent in Spanish than in English, or if the child stated a preference for speaking Spanish when asked, the tester was allowed to use these translated instructions. Although this procedure ran the risk of undermining the stan- dardization of the tests, it seemed the best option to maximize children's comprehension of the tasks. Norms and procedures standardized on English speakers cannot be assumed to apply to children from other language groups in any case, so it seemed most reasonable to attempt to obtain optimal levels of performance for these multilingual children. By age 4, virtually all of the children chose to speak En- glish during the tests.

Analysis Plan

Path analyses were used to determine the relations of viewing at Periods 1, 2, and 3 to test scores at Waves 2, 3, and 4. The ideal approach would have been a latent-variable, structural equation modeling analysis, but the small sample size rendered this technique inappropriate. The two cohorts were ana- lyzed separately. Path analysis permitted consolida- tion of all time periods in a single analysis for each combination of a viewing category with a dependent variable, and it permitted tests of paths in two directions-from earlier viewing to later test scores, and from earlier test scores to subsequent viewing. For the younger cohort, there were 12 analyses (four test scores X three viewing categories), and for the older cohort, who did not receive the Bracken School Readiness Scale, there were 9 (three test scores X three viewing categories). Because viewing frequen- cies were skewed, each viewing score was transformed to the square root of (number of minutes + 1) to normalize the distributions and eliminate zeroes and values less than one. EQS was used for all anal- yses (Bentler, 1995).

Because of the sample sizes, it was important to limit the number of variables in the path analysis. Two control variables were included: the HOME score from Wave 1and the primary language spoken in the home (1 = English, 2 = Spanish). These were selected on the basis of initial analyses designed to determine which of approximately 20 demographic variables, child characteristics, family attributes, preschool experiences, and television technologies in the home accounted for independent portions of the variance in both the viewing patterns and the test scores. Parent education, income-to-needs ratio, and HOME scores accounted for overlapping variance. The HOME scorewas selected because it is a measure of the proximal quality of the child's environment and because it accounted for slightly more variance in test scores than did either of the other indicators. The primary language spoken in the home was related to educational viewing and somewhat less to viewing other types of television (children in Spanish-speaking families watched more educational televisionthan did children in English-speaking families) and to some test scores. Primary language was statistically independent of HOME scores. In an earlier, unpublished report, ordinary least squares, calculated by regressing each test score in each wave of prior viewing in each television category, were presented (Wright & Huston, 1995). These analyses included multiple controls for demographic variables, child characteristics, child's initial language, family attributes, preschool experiences, and television technologies in the home. The regression approach involved a large number of equations, and accordingly path analysis was chosen as a more parsimonious approach for the present report. It is noteworthy, however, that the results obtained by the two methods were quite similar.

Weekend

RESULTS

Description of Viewing

The number of hours children spent viewing informative and animated children's programs are shown in Figure 2. Available data for all 236 sample members are included. Children watched child-audience informative programs for about 2 hr per week at ages 2 through 4 years, with declines to about 1hr per week at ages 5 through 7 years. Although Sesame Street was by far the most frequently viewed child-audience informative program at all ages (about 80% of all such viewing in 1990-1993), the proportion of viewing devoted to other child-audience informative programs increased as children got older. After Sesame Street, the programs watched most often in Period 3 were Mister Rogers' Neighborhood, Reading Rainbow, Captain Kangaroo, MY. Wizard's World, and 3-2-1 Contact. Children watched cartoons about 7.5hr per week until about age 5, with a decline by age 7 to about 5hr per week. The remaining category of child-audience programs, "childaudience other" (not shown in Figure 2), was viewed slightly over 1hr per week by children at all ages assessed, and there were no significant age differences or age changes.

The average hours of viewing for general-audience programs are shown in Figure 3. Children spent an average of approximately 16 hr per week watching general-audience programs at ages 2 and 3, declining

2-3 3-4 4-5 4-5 54 6-7 2-3 3-4 4-5 4-5 5-6 6-7 Age (Years)
Cohort 1 Cohort 2 Cohort 1 Cohort 2  
Child-Audience Informative Child-Audience Animated  

Figure 2 Average weekly viewing of informative and animated children's programs. Wright et al. 1355

Weekend Secondary

l81

Weekend Primary

16 24 ed Weekday Secondary 8 14 IWeekday Primary

3

K 12 % 10

5

5 8

2 6

z

4

2 4

E

2

n

2-3 3-4 4-5 4-5 5-6 6-7 Age (Years)

Cohort 1 Cohort 2

Figure 3 Average weekly viewing of general-audience programs.

to around 10hr per week by ages 6 and 7. Most of this decline was accounted for by secondary viewing (see Huston et al., 1999, for analyses of age changes); the amount of primary viewing was fairly stable across the age ranges sampled.

Relations of Viewing to School Readiness and Achievement

The relations between viewing and test performance were analyzed using a path model. Child-audience informative, child-audience animated, and general-audience television were each tested separately. The means and standard deviations for HOME scores, television viewing (square root transformation), and test scores for children included in the path analyses are presented in Table 2, and the correlations among variables appear in Table 3. Frequency of viewing child-audience informative programs was negatively related to frequencies of child-audience animated and general-audience viewing in most comparisons, but the correlations were generally quite low. Frequencies of child-audience animated and general-audience viewing were somewhat positively related for the younger cohort, but not for the older cohort. These low intercorrelations suggest that it is appropriate to consider these viewing categories independently of one another. Other children's programming was not tested because it was viewed in-

Table 2 Means and Standard Deviations of Scores Used in Path Analysis

Younger Cohort Older Cohort (N = 90) (N = 92)

Characteristic

HOME standard score Primary language (% English) CI 1 CI 3 CA 1 CA 3 GA 1 GA 3 LW2 LW4 AP 2 AP 4 PPVT-R 2 PPVT-R 4 SR 2 SR 4

Note: All viewing scores were square root of (average weekday miniday + 1).HOME = Home Observational Measure of the Environment; CI = child-audience informative television; CA = child-audience animated television; GA = general-audience television; LW = Woodcock-Johnsonletter word subtest; AP = Woodcock-Johnson applied problems subtest; PPVT-R = Peabody Picture Vocabulary Test-Revised; SR = Bracken School Readiness Scale. The number following each measure indicates the period or wave of data collection.

1356 Child Development
Table 3 Correlation Matrix for Variables in the Path Models  
  PL HOME CI 1 CAI GA1 LW2 AP2 PPV-R2 SR2 CI3 CA3 GA3 LW4 AP4 PPVT-R4
PL HOME CI 1 CA 1 GA 1 LW2 AP 2 PPVT-R 2 SR 2 CI 3 CA 3 GA 3 LW 4 AP 4 PPVT-R 4 SR 4 .13 .49 .23 .02 .07 .06 -.31 -.21 .13 .07 .09 -.22 -.I2 -.30 -.29 -.I8 .33 -.30 -.47 .31 .36 .28 .25 .24 -.I7 -.33 .18 .15 .32 .11           -.58 .46 0 -.I4 -.I8 .51 .56 .57 -.08 -.I7 -.37 .47 .62 .79 .57 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. .16 -.I2 -.29 .61 .66 .61 .64 -.05 -.I1 .44 -.07 .06 .18 .11 .07 N.A. -.08 -.05 0 .03 -.04 .03 .12 -.34 -.I4 .39 -.02 -.10 -.I4 -.27 N.A. -.I5 .25 -.I2 -.I0 -.I5 -.I9 -.I3 -.06 -.06 .08 .29 -.I9 -.I0 -.03 N.A. .04 .04 -.I6 -.21 -.39 -.26     -.I5 .14 .03 .03 -.28 .46 .70 .50 N.A. .03 -.13 -.07 .49 .62 .71 -.36 .39 .11 .03 -.23 .49 .51 .73 N.A. .13 -.26 -.I1 .45 .47 .74

Note: Cohort 1 data are below the diagonal, Cohort 2 data are above the diagonal. PL = primary language; HOME = Home Observational Measure of the Environment; CI = child-audience informative television; CA = chld-audience animated television; GA = general-audience television; LW = Woodcock-Johnson letter word subtest-Revised; AP = Woodcock-Johnson applied problems subtest; PPVT-R = Peabody Picture Vocabulary Test-Revised; SR = Bracken School Readiness Scale; N.A. = not applicable. The number following each mea- sure indicates the period or wave of data collection.

frequently and because its content appeared to be ing. The fit statistics presented include chi square, the highly variable over time. comparative fit index (CFI), Bollen's incremental fit

For these analyses, weekend diaries were excluded index (IFI), and the standardized root mean square re- because viewing on Saturday, Sunday, and weekdays sidual (RMR). A good fit is indicated by a nonsignifi- followed different patterns, and there were insuffi- cant chi square; CFI and IFI.90 or above, with .95 and cient numbers of Saturday and Sunday diaries to gen- above being a very good fit; and RMR .08 or below. erate reliable estimates of viewing for each weekend These are shown in Table 4. The models all reached day. A sizeable number of children had missing data criterion on the CFI, IFI, and RMR criteria, but chi for weekends during at least 1 of the 3 years. Adding square reached p < .05 on 5 of the 21 tests. Accord- weekend data to weekdays for some children and not ingly, these models should be interpreted with some others could distort the ordering of individuals; caution. therefore, we elected to use the weekday diaries that The results of the path analyses are summarized in sampled comparable time periods for all children. Figures 4 through 6. Each diagram contains the path

The path model contains viewing in Periods 1 and coefficients for one of the three categories of televi- 3 and test scores in Waves 2 and 4; Period 2 and Wave sion programs in relation to all test scores. There are 3 were excluded. Models that included all three view- separate diagrams for Cohorts 1 (ages 2-5) and 2 ing periods (1-3) and three annually collected sets of (ages 4-7). In all models, viewing was moderately test scores (Waves 2-4) were also tested, but the fit stable, and test scores were very stable over time. statistics were less than satisfactory. Accordingly, sim- pler models were used for the main analysis. Because

Viewing as a Predictor of Test Scores

relations among variables in path analysis are com- puted in temporal sequence, the models including all Child-audience informative programs. The paths from three periods and waves produced identical results child-audience informative viewing to test scores, for the initial parts of the model (paths from HOME shown in Figure 4, indicate that children who watched and primary language to viewing and test scores, and educational television frequently when they were 2 the path from Period 1viewing to Wave 2 test). There and 3 years old (Period 1) performed better on all four were no substantial differences between the results of tests at age 3 (Wave 2) than did those who were infre- the remaining portions of the two analyses. quent viewers. Period 1 viewing was a significant

A lagged path from Period 1viewing to Wave 4 test positive predictor of letter-word, applied problems, score was included to test for delayed effects of view- PPVT-R, and Bracken School Readiness scores in

Wright et al. 1357

Table 4

Viewing Category

Cohort 1 CI CI CI CI C A C A CA CA G A G A G A G A

Cohort 2 CI CI CI CA C A C A G A GA G A

Fit Statistics for Path Analyses

Test

LW AP PPVT-R SR LW AP PPVT-R SR LW AP PPVT-R SR

LW AP PPVT-R LW AP PPVT-R LW AP PPVT-R Standard x2" CFIb Bollen IFIC RMRd

Note: CI = child-audience informative television; CA = child-audience animated television; GA =
general-audience television; LW = Woodcock-Johnson letter word subtest; AP = Woodcock-Johnson
applied problems subtest; PPVT-R = Peabody Picture Vocabulary Test-Revised; SR = Bracken School
Readiness Scale.
a x2is the significance of the difference between the proposed and observed models.

CFI is the comparative fit index.
IF1 is the incremental fit index from Bollen (1989).
RMR is the standardized root mean square of the typical residual correlation.

*p < .05; + p < .10.

Wave 2 for the younger cohort. Once these initial rela- tions were taken into account, there were no signifi- cant relations of viewing to test scores in Wave 4. That is, viewing in later periods did not contribute to sub- sequent test performance beyond the effects of view- ing in the initial period. The fact that test scores were highly stable over time indicates, however, that the advantage associated with early viewing was still present at Wave 4. For the older cohort there were no significant relations between viewing child-audience informative programs and test performance at any time period (Figure 4).

Child-audience animated programs. The relations of cartoon viewing to test performance are shown in Figure 5. For the younger cohort, children who were frequent viewers of noneducational cartoons at ages 2 and 3 (Period 1)had lower scores than did infre- quent viewers on the letter-word subtest at Wave 2 and on the PPVT-R at Wave 4. Although most of the remaining coefficients were negative, they were not significant. For the older cohort, there was no evi- dence for a relation of viewing cartoons to later test performance.

General-audience programs. The path analyses for general-audience programs appear in Figure 6. For both cohorts, children who watched a lot of general- audience programming in Period 1 tended to per- form more poorly than did less-frequent viewers. Among the younger cohort, heavy viewers of general- audience programs in Period 1(ages 2-3) performed significantly less well than did infrequent viewers on applied problems and the PPVT-R at Wave 2 (age 3); the coefficients for letter-word and the Bracken School Readiness Scale were also negative but non- significant.

For the older cohort, children who watched a lot of general-audience programming when they were 4 to 5 years old (Period 1) performed significantly less well in tests of letter and number skills at age 5 (Wave 2) than did infrequent viewers, and also tended to do

1358 Child Development
  Child-Audience Informative Period 1 .297'** .3462** .3373** .3094** Child-Audience Informative Period 3

Wave 1

a
2
.
a E
a
.l+ 5.
'Letter Word 2AppliedProblems 3Peabody Picture         4 Test Wave 4 s 80 .w
Vocabulary 4BrackenSchool           .l+ 5 z
Readiness       Cohort 1    
            II
            f
            I
Child-Audience       Child-Audience   I
Informative .443l** .4422** .4433** Lnformative  
Period 1       Period 3    

Wave 1

4

'Letter Word Test Wave 2 Test Wave 4 2AppliedProblems 3PeabodyPicture

Vocabulary Cohort 2

Figure 4 Path analyses: Child-audience informative television viewing and test scores for Cohort 1and Cohort 2. * p< .05; ** p < .01.

Wright et al. 1359

Child-Audience Child-Audience Animated Animated

.395l** .4002** .3883** .3994** .

Period 1 Period 3

'Letter Word

2Applied Problems
Test Wave 4

3Peabody Picture
~ocabuiary
4Bracken School
Readiness Cohort 1

Child-Audience Child-Audience

Animated .3811** .3862** .3583**

Animated
Period 1

Period 3

\ \ I'

9

, I,Y

, \

I I , \

,,' \, \

:r,

:0

:01

,+

', \

:'0

It\,%

-

.'... ',','+\

'$,,%

.'.'. '\"..

,

'Letter Word .7043**
2Applied Problems
3Peabody Picture

Vocabulary Cohort 2

Figure 5 Path analyses: Child-audience animated television viewing and test scores for Cohort 1and Cohort 2. * p < .05; ** p < .01.

~ocabuiar~ Cohort2

Figure 6 Path analyses: General-audience television viewing and test scores for Cohort 1 and Cohort 2.

less well on a test of receptive vocabulary skills. View- ing in later periods did not contribute to subsequent test performance beyond the effects of viewing in the initial period.

Test Scores as Predictors of Viewing

The path analyses were also designed to test rela- tions between earlier test scores and later viewing. The path from Wave 2 tests to Period 3 viewing repre- sents a test of the hypothesis that children's academic and language skills (as indicated in their early test scores) affects their later viewing patterns. The reader is reminded that an entire calendar year elapsed be- tween Wave 2 and the onset of Period 3, which lasted for another year. The coefficients for paths predicting viewing from prior test scores are shown in Figures 4 through 6.

Child-audience informative programs. Among the younger cohort, test scores did not predict subse- quent viewing of child-audience informative pro- grams. For the older cohort, children who did well on the letter-word subtest at age 5 (Wave 2) watched sig- nificantly more child-audience informative program- ming at ages 6 and 7 (Period 3) than did those whose letter-word skills were low, and there was a tendency for high math shlls and vocabulary to predict later in- formative viewing as well (see Figure 4).

Child-audience animated programs. Test scores at age 3 were not related to viewing cartoons at ages 4 and 5 for the younger cohort. Among the older cohort, how- ever, children with low PPVT-R scores at Wave 2 (age 5) watched significantly more cartoons in Period 3 (ages 6 through 7) than did those with high vocabu- lary scores (see Figure 5). The relations of reading and math scores to later cartoon viewing were also nega- tive, but nonsignificant.

General-audience programs. For the younger cohort, there was a strong pattern showing that children who performed well on the academic shll and language tests at Wave 2 (age 3) spent less time watching general- audience programs during Period 3 (ages 4 and 5)than did those who did not perform well. Three of the four coefficients were sigruficant, and the fourth was consis- tent in direction (see Figure 6).For the older cohort, test scores were not consistently related to subsequent view- ing of general-audience programs (Figure 6).

DISCUSSION

Child-Audience Informative Programs

The findings of this study provide strong support for the notion that the effects of television viewing

Wright et al. 1361

depend on program content and genre. For very young children, viewing informative programming designed for children was associated with subse- quent letter-word shlls, number skills, receptive vo- cabulary, and school readiness. Individual differences in these skills were fairly stable through the preschool years, suggesting that early effects of viewing can be lasting. These patterns occurred in a multiethnic and multilingual sample of children from low- to moderate- income families-groups that may be especially likely to benefit from educational television. It is notewor- thy that children in families whose first language was Spanish watched a great deal of educational televi- sion. There were too few of them to analyze as a sep- arate group, but parents often commented that the whole family viewed Sesame Street and other programs to improve their English.

If television plays a causal role in these associa- tions of viewing and school readiness, what processes might be involved? One explanation for positive effects of educational programs is relatively straightforward- children learn educational content from viewing. Such programs present information with techniques designed to be comprehensible and appealing to young children. Many of them are based on careful curriculum development and formative research, and, unlike most entertainment programs, they are tar- geted toward a specific age group.

Viewing educational programs may also influence children's motivation and interests in ways that lead them to engage their home and school environments differently. Children who watched a lot of educa- tional television also devoted more time to reading and educational activities away from television than did infrequent viewers (Wright & Huston, 1995). When children enter school, those with good school- readiness skills are likely to have more experiences of success and to be perceived by teachers as more able, perhaps starting a trajectory of better performance and higher levels of academic motivation than chil- dren who enter with poor skills. This snowball effect seems the most likely explanation for the finding in another longitudinal sample that preschool educa- tional viewing was positively associated with official high school transcript grades, with similar controls for third variables (Anderson et al., 2001).

The relations of educational viewing to school readiness were strongest for viewing at ages 2 and 3. It is possible that the earliest years are especially im- portant because very young children are less likely than older children to be in preschool, child care, or other structured settings that provide planned stimu- lation and instruction in a preacademic curriculum. It is also possible that interest in informative programs declines during the preschool and early school years. Certainly, viewing time devoted to informative pro- grams drops with age. Both the drop in viewing and the absence of effects for school-age children, how- ever, may be due to the lack of age-appropriate edu- cational programs, particularly in the early 1990s when these data were collected, rather than to devel- opmental change in interest in informative televi- sion per se. Moreover, we measured only one category of outcomes-academic skills. Children of different ages may learn a wide range of information about sci- ence, society, and the social world from educational programs.

Early viewing may also have lasting effects, in part, because it influences subsequent viewing choices indirectly. Children's test scores did not pre- dict changes in viewing child-audience informative programs in the preschool years, but children with good skills at age 5 tended to view more child-audience informative programs and fewer cartoons in their early elementary years than did those with poorer skills.

General-Audience Programs

This study's findings also support hypotheses drawn from several theories predicting that extensive view- ing of entertainment programming may be a negative contributor to intellectual development. Although cartoon viewing at age 2 and 3 was associated with lower scores on some measures, the patterns were much stronger for general-audience programs. For both cohorts, those who were frequent viewers of gen- eral-audience programs performed less well on tests of school-related skills and /or receptive language than did less frequent viewers. Again, the high individual consistency of test performance over time means that initial effects of viewing tended to be lasting. Once these initial effects were controlled, there were rela- tively few additional effects. One exception was car- toon viewing for the younger cohort; viewing at age 2 and 3 was associated with lowered school readiness at age 5, even though it did not predict school readi- ness at age 3.

Two processes have been proposed by whch general- entertainment programming might lead to lowered intellectual development. The first is displacement. Our other analyses of these data provided some sup- port for the notion that viewing general-audience en- tertainment programs displaces time in educational activities and social interaction. Children who watched a lot of general-audience entertainment programs spent less time using print media than did less- frequent viewers (Wright & Huston, 1995). Moreover, individual patterns of change in viewing over the 3 years were associated with changes in time spent in reading, other educational activities, and social inter- action. When general-audience viewing increased, the time spent in educational activities and social in- teraction decreased, and, conversely, when general- audience viewing declined, cluldren increased the time they spent in educational activities and social inter- action (Huston et al., 1999).

The related hypothesis-that for very young chil- dren, general-audience programming interferes with opportunities for verbal and social interactions with par- ents, child-care providers, and other adults because the adults are watching the television-also found some support in these data. Among the 2-to 4-year-olds in this study, a great deal of general-audience pro- gram viewing was "secondary." In many instances, it is likely that the child was playing, eating, or doing something else while the television was on in the room. Although parents are usually present when young children watch general-audience programs (St. Peters, Oppenheimer, Ealuns, Wright, & Huston, 1991), they may be attending to the program and be unavailable for conversation or interaction with the young child.

The second hypothesis-that entertainment tele- vision may not engage children in effortful mental activity-may also account for the findings. It seems unlikely that general-audience content was easily un- derstood by such young children. In fact, both car- toons and general-audience programs may pose chal- lenges to comprehension for them, and much of this programming may be beyond their level of under- standing. When children perceive programs as in- comprehensible, their attention drops and they are likely to pursue other activities (Anderson & Burns, 1991). Moreover, a child with prolonged exposure to such content may have few experiences of engaging successfully with the material, solving a problem, and getting a sense of mastery. Once children compre- hend some of the general-audience programming available, they tend to watch situation comedies and other light entertainment that may indeed foster a preference for material that makes few intellectual demands.

Early viewing may also have indirect effects on later viewing patterns. Early exposure predicted lower skills at age 3; children with low skills at age 3 increased their viewing by ages 4 and 5; and, in the older cohort, viewing at ages 4 and 5 predicted lower per- formance at age 5. These patterns could reflect the con- tinuing role of home environment as an influence on both viewing and intellectual development, but it may also indicate that more advanced preschool children seek out activities other than television. This pattern did not appear to continue when children entered school.

Selection Issues

The fact that most of the significant relations oc- curred for the first viewing period may mean that early patterns of viewing are established in a stable way that has early effects, but it could also indicate that selection variables affected the initial relations of viewing to test scores. Controlling for the quality of home environment by using the HOME Scale as- sured that there was not a large confound with cog- nitive stimulation and support in the home environ- ment. Controlling for primary language assured that the results were not an artifact of English fluency. Analyses not shown here also excluded a number of other variables including parent education, occupa- tional status, family structure (single or two parents), child-care experience, birth order, parent regulation and encouragement of television viewing, and the child's initial language ability as explanations of the re- lations between viewing and children's performance. It is not logically possible to exclude all possible un- measured variables that might account for the rela- tions between viewing and children's performance, but these results are quite robust in the face of con- trols for many likely explanatory variables (see Wright & Huston, 1995).

Although establishing the independent contribu- tion of television to the outcomes is important, it is not theoretically reasonable to assume that any aspect of a child's experience operates independently of the overall context in which it occurs. It may be that expo- sure to different types of television occurs in different home and child-care contexts. If that is the case, we would not conclude that television has no indepen- dent effects, but that the television content forms a significant part of those varied contexts.

Integrated Model

We stress the importance of content over total amount of viewing because child-audience informative viewing and other types of viewing are associated with opposite patterns of language and school-related skills. Content and hours in front of the television, however, are confounded in the world. In this sample, the average child watched between 1 and 3 hr a week of educational programming; in an earlier study, the av- erages ranged from 3 to 5 hr a week. Our sample watched 10 to 16 hr a week of general-audience pro- grams plus another 5 to 8 hr of cartoons. The issue of content versus total time can be addressed in two

Wright et al. 1363

ways. Would there be negative effects if children spent 15 to 25 hours a week watching only child- audience informative programs? The data did not permit an answer because almost no child did so, and we cannot infer a linear increase in benefits with ex- tremely high amounts of exposure. The cumulative benefits or hazards of viewing would probably de- pend on many factors, including the richness and va- riety of educational content viewed, its appropriate- ness for the child's age and interests, and whether the activities displaced by viewing were more or less valuable for intellectual development than the con- tent of the television programs.

If total viewing time rather than content were crit- ical, then might one expect positive effects from 3 hr a week of watching entertainment programs? Is there some benefit to small amounts of television, regardless of content? No theory or data suggest that sitting in front of the television per se promotes intellectual develop- ment. Positive outcomes might occur, however, if the time spent watching television provided opportunities to learn language or social information (for example, immigrants sometimes use television to learn a new language) or if it displaced time in harmful activities.

Similarly, most theories that predict negative ef- fects of viewing general entertainment suggest that harm is a function of the amount of time devoted to it. One exception may be the "mental effort" hypothesis. If the critical variable is the lack of intellectual de- mands in the content of entertainment programs, then even small amounts of exposure might create a taste for such content, which would, in turn, be car- ried over to other parts of the child's life.

Real children watch both child-audience informa- tive and entertainment programs. How might expo- sure to the two types of programming combine to in- fluence school readiness? We propose that each effect operates relatively independently, and that they are additive. That is, learning content and interests from child-audience informative programs is a function of viewing those programs, relatively independently of the time spent viewing other types of programs. Time with general-audience programs interferes with practicing and acquiring school-related skills through several mechanisms that may be relatively indepen- dent of child-audience informative viewing. In an earlier longitudinal study, we performed "viewing diet" analyses, in which child-audience informative and general-entertainment viewing were considered jointly as predictors. The associations of each viewing category with high school grades in these diet analy- ses were similar to those found using single viewing categories as predictors (Anderson et al., 2001). In the present study, the number of cases did not permit considering viewing diet in the path analysis, but the low correlations among the different types of viewing suggest that the results would be quite similar to those reported for single viewing categories.

These results support the conclusion that, contrary to Marshall McLuhan's (1964) famous dictum, it is the message, not the medium, that matters. Watching ed- ucational programming, at least in small amounts, is associated with high levels of school-related and lan- guage skills, and watching a lot of general-audience programming is associated with low levels of the same skills. These results occurred for children as young as age 2 in low-income families from a range of ethnic identities.

As a society, we are making choices about the pro- gramming offered to children that have consequences for what they derive from it. Public policy in recent years has been designed to require broadcasters to serve the educational and informational needs of chil- dren (Kunkel, 1998). Good educational programs can provide lasting benefits to children at many ages, but it may be especially important to provide such fare for very young children because they are less likely than older children to be exposed to formal preschool instruction, and because stable habits of viewing may be formed in the first few years of life. At the same time, there is a continuing need for age-specific pro- gramming that will interest and teach children be- yond the preschool years.

ACKNOWLEDGMENTS

This research was supported by a grant to the two senior authors from the Children's Television Work- shop, using funds awarded by the John D. and Katherine T. MacArthur Foundation. The conclusions are those of the authors and do not necessarily reflect those of the granting agencies. The authors wish to express their appreciation for the skilled, diligent, and committed efforts of the staff of the Early Win- dow Project: Seanna Donley, Angeles Dowd, Darwin Eakins, Jocelyn Freeman, Cydney Generoux, Curtis Gilyard, Lori Goser, Vicky Hamlin, Paul Jefferson, Janice Jones, Kulkanok Kanoksakharin, Jennifer Klein, Carrie Lamble, Ora Myers, Maria Elena Novak, Sabrina Oppenheimer, Rebeca Perez, Joanne Pickar, Vicki Rawdon, Jane Schneider, Tina Siriboh, Raymond Soh, Jan Anderson Talley, Rosemarie Truglio, Jacky Watson-Jolly, and Laurel Wheeler.

ADDRESSES AND AFFILIATIONS

Corresponding author: Aletha C. Huston, Department of Human Ecology A2700, University of Texas at Aus- tin, Austin, TX 78712-1097; e-mail: achuston@mail. utexas.edu . John C. Wright was also at the University of Texas at Austin; Kimberlee C. Murphy is at the Uni- versity of Kansas, Lawrence, KS; Michelle St. Peters is at the University of Denver, Denver, CO; Marites Pifion is at Psychological Corporation, San Antonio, TX; Ronda Scantlin is at the University of Pennsylva- nia, Philadelphia, PA; and Jennifer Kotler is at George- town University, Washington, DC.

REFERENCES

Anderson, D. R., & Bums, J. (1991). Paying attention to tele- vision. In J. Bryant & D. Zillmann (Eds.), Responding to the screen: Reception and reactiotz processes (pp. 3-25). Hillsdale, NJ: Erlbaum.

Anderson, D. R., Field, D. E., Collins, P. A., Lorch, E. P., & Nathan, J. G. (1985). Estimates of young children's time with television: A methodological comparison of parent reports with time-lapse video home observation. Child Development, 56,1345-1357.

Anderson, D. R., Huston, A. C., Schmitt, K. L., Linebarger,

D. L., & Wright, J. C. (2001). Early childhood television viewing and adolescent behavior. Monographs of the Society for Research in Child Development, 66(Serial No. 264).

Ball, S. J., & Bogatz, G.A. (1970). Thefirst year ofSesame Street: An evaluation. Princeton, NJ: Educational Testing Service.

Ball, S. J., & Bogatz, G. A. (1973). Reading with television: An evaluation of the Electric Company. Princeton, NJ: Educa- tional Testing Service.

Bandura, A. (1978). The self system in reciprocal determin- ism. American Psychologist, 33, 344-358. Bentler, P. M. (1995). EQS structural equations program man- ual. Encino, CA: Multivariate Software, Inc.

Bogatz, G. A., & Ball, S. J. (1971). The second year of Sesame Street: A continuing evaluation. Princeton, NJ: Educational Testing Service .

Bracken, B. A. (1984). Bracken Basic Concept Scales. San Anto- nio, TX: Psychological Corporation.

Bradley, R. H., Caldwell, B. M., Rock, S. L., Ramey, C. T., Bamard, K. E., Gray, C., Hammond, M. A,, Mitchell, S., Gottfried, A. W., Siegel, L., & Johnson, D. L. (1989). Home environment and cognitive development in the first 3 years of life: A collaborative study involving six sites and three ethnic groups in North America. Developmental Psychology, 25, 217-235.

Caldwell, B. M., & Bradley, R. H. (1984). Home observation for the measurement of the environment (Rev. ed.). Little Rock, AR: University of Arkansas.

Center for Research on the Influences of Television on Chil- dren. (1983). CRITC program categorization system coding manual. Lawrence, KS: University of Kansas.

Clarke, A. T., & Kurtz-Costes, B. (1997). Television viewing, educational quality of the home environment, and school readiness. Journal ofEducationa1 Research, 90,279-285.

Comstock, G. (1991). Television and the American child. Orlando, FL: Academic.

Comstock, G. A. (1995). Television and the American child. In

C. N. Hedley & P. Antonacci (Eds.), Thinking and literacy: The mind at work (pp. 101-123). Hillsdale, NJ: Erlbaum.

Cook, T. D., Appleton, H., Comer, R. F., Shaffer, A., Tamkin, G., & Weber, S. J. (1975)."Sesame Street" revisited: A study in evaluation research. New York: Russell Sage.

Corteen, R. S., & Williams, T. M. (1986). Television and read- ing skills. In T. M. Williams (Ed.), The impact of television: A natural experiment in three communities (pp. 39-85). Orlando, FL: Academic Press.

Dunn, L. M., & Dunn, L. M. (1981). Peabody Picture Vocabu- lary Test. Circle Pines, MN: American Guidance Services. Entwisle, D. R., Alexander, K. L., & Olson, L. S. (1997). Children, schools, and inequality. Boulder CO: Westview.

Gortmaker, S. L., Salter, C. A., Walker, D. K., & Dietz, W. H. (1990). The impact of television viewing on mental apti- tude and achievement: A longitudinal study. Public Opinion Quarterly, 54,594-604.

Huesmann, L. R., & Miller, L. S. (1994). Long-term effects of repeated exposure to media violence in childhood. In

L. R. Huesmann (Ed.), Aggressive behavior: Current per- spectives (pp. 153-186). New York: Plenum.

Huston, A. C., & Wright, J. C. (1994). Educating children with television: The forms of the medium. In D. Zill- mann, J. Bryant, &A. C. Huston (Eds.), Media, family, and children: Social scientific, psychodynamic, and clinical per- spectives (pp. 73-84). Hillsdale, NJ: Erlbaum.

Huston, A. C., & Wright, J. C. (1997). Mass media and chil- dren's development. In I. Sigel & K. A. Renninger (Eds.),

W. Damon (Series Ed.), Handbook of child psychology: Vol. 4. Child psychology in practice (5th ed., pp. 999-1058). New York: Wdey.

Huston, A. C., Wright, J. C., Marquis, J., &Green, S. B. (1999). How young children spend their time: Television and other activities. Developmental Psychology, 35, 912-925.

Huston, A. C., Wright, J. C., Rice, M. L., Kerkrnan, D., & St. Peters, M. (1990). The development of television viewing patterns in early childhood: A longitudinal in- vestigation. Developmental Psychology, 26, 409-420.

Juster,F. T., & Stafford, F. P. (1985). Time, goods, and well-being. Ann Arbor, MI: Survey Research Center.

Koolstra, C., & van der Voort, T. (1996). Longitudinal effects of television on children's leisure-time reading: A test of three explanatory models. Human Communicatiotz Research, 23,4-35. Koolstra, C. M., van der Voort, T. H. A., & van der Karnp, L. J. T. (1997). Television's impact on children's reading com- prehension and decoding skills: A 3-year panel study. Reading Research Quarterly, 32, 128-152. Kunkel, D. (1998). Policy battles over defining children's educational television. Annals of the American Academy of Political and Social Science, 558, 39-53. MacBeth, T. M. (1996). Indirect effects of television: Creativ- ity, persistence, school achievement, and participation in other activities. In T. M. MacBeth (Ed.), Tuning in to young viewers: Social science perspectives on television (pp. 149-219). Thousand Oaks, CA: Sage.

McLuhan, H. M. (1964).Understanding media: The extensions of man. New York: McGraw-Hill.

Wright et al. 1365

Neapolitan, D. M., & Huston, A. C. (1994). Educational con- tent of children's programs on public and commercial televi- sion. Lawrence, KS: Center for Research on the Influ- ences of Television on Children.

Neuman, S. B. (1988). The displacement effect: Assessing the relation between television viewing and reading per- formance. Reading Research Quarterly, 23,414-440.

Potter, W. I. (1987). Does television viewing hinder academic achievement among adolescents? Human Commu- nication Research, 14, 27-46.

Rice, M. L. (1984). The words of children's television. Journal of Broadcasting, 28,445-461.

Rice, M. L., Huston, A. C., Truglio, R., &Wright, J. C. (1990). Words from Sesame Street: Learning vocabulary while viewing. Developmental Psychology, 26,421-428.

Rice, M. L., Huston, A. C., &Wright, J. C. (1982). The forms and codes of television: Effects on children's attention, comprehension and social behavior. In D. Pearl, L. Bouthilet, & J. B. Lazar (Eds.), Television and behavior: Ten years of scientific progress and implications for the 80's (Vol. 11). Washington, DC: Government Printing Office.

Ritchie, D., Price, V., & Roberts, D. F. (1987). Television, reading, and reading achievement. Communication Re- search, 14,292-315.

Rosengren, K. E., & Windahl, S. (1989). Media matter: TV use in childhood and adolescence. Norwood, NJ: Ablex.

Salomon, G. (1984). Television is "easy" and print is "tough: The differential investment of mental effort in learning as a function of perceptions and attributions. Journal of Educational Psychology, 76, 647-658.

Searls, D. T., Mead, N. A., &Ward, B. (1985). The relationship of students' reading skills to TV watching, leisure time reading, and homework. Journal of Reading, 29,158-162.

Singer, J. L. (1980). The power and limits of television: A cognitive-affective analysis. In P. Tannenbaum (Ed.), The entertainmentfunction of television. Hillsdale, NJ: Erlbaum.

Smyser, S. 0. (1981). A study of the relationship between television viewing habits and early reading achieve- ment. Dissertation Abstracts International, 41(12-A, Pt. I), 4972-4973. St. Peters, M., Fitch, M., Huston, A. C., Wright, J. C., & Eakins, D. (1991). Television and families: What do young children watch with their parents? Child Develop- ment, 62,1409-1423. St. Peters, M., Oppenheimer, S., Eakins, D. J., Wright, J. C., & Huston, A. C. (1991, April). Media use among preschool children as a function of income and media options. Paper presented at the meeting of the Society for Research in Child Development, Seattle, WA. Truglio, R. T., Huston, A. C., & Wright, J. C. (1986, March). The relation ofchildren's print and television use to early read- ing skills: A longitudinal study. Paper presented at the Southwestern Society for Research in Human Develop- ment, San Antonio, TX. Truglio, R. T., Murphy, K. C., Oppenheimer, S., Huston,

A. C., &Wright, J. C. (1996). Predictors of children's en- tertainment television viewing: Why are they tuning in? Journal ofApplied Developmental Psychology, 17, 474-

494.

Williams, P. A., Haertel, E. H., Walberg, H.J., & Haertel, G. D. (1982). The impact of leisure-time television on school learning: A research synthesis. American Educational Re- search Journal, 19, 19-50.

Woodcock, R. W., & Mather, N. (1989). Woodcock-Johnson Test of Achievement. Allen, TX: DLM Teaching Resources.

Wright,J. C., & Huston, A. C. (1995). EfJects of educational TV viewing of lower income preschoolers on academic skills, school readiness, and school adjustment one to three years

later. Lawrence, KS: Center for Research on the Influ- ences of Television on Children.

Wright, J. C., & Vlietstra, A. G. (1975). The development of selective attention: From perceptual exploration to log- ical search. In H. Reese (Ed.), Advances in child development and behavior (Vol. 10, pp. 195-239). New York: Academic.

Zill, N., Davies, E., & Daly, M. (1994). Viewing of Sesame Street by preschool children in the United States and its rela- tion to school readiness. Rockville, MD: Westat, Inc.

Comments
  • Recommend Us