Academic Achievement and Children's Television Use
ACADEMIC ACHIEVEMENT AND CHILDREN'S TELEVISION USE
The relationships between television viewing and the academic performance of children and teenagers have been the subject of great controversy. Popular opinion and some educators have held that television generally has had a detrimental effect—by taking up time that might be better spent acquiring basic skills or doing homework, by encouraging a preference for quick solutions and entertaining portrayals that is inconsistent with the sometimes frustrating demands of schoolwork, and by creating tastes and enthusiasms that draw young people away from intellectually demanding subject matter. In contrast, very sophisticated statistical analyses of amount of viewing and achievement scores among large samples seemingly have indicated that television has no effect when other contributing factors are taken fully into account. In fact, the actual findings of the many dozens of empirical research studies that bear on the topic do not conform perfectly to either of these perspectives.
Viewing and Achievement
There is absolutely no question that children and teenagers who spend greater amounts of time with television perform less well on standardized tests of achievement. This inverse relationship— the greater the viewing, the lower the achieve-ment—holds for the three basic skills (i.e., reading, writing, and mathematics) and for other subjects as well (e.g., science, social science, and history). The controversy centers on why this should be so.
This inverse relationship has been observed consistently and repeatedly in samples ranging from a few hundred to more than a half million subjects, which taken together can be said to be representative of American children and teenagers. There are several important qualifications, however. The relationship is most severe among young people from households that are higher in socioeconomic status (where parents score higher on education, income, or occupational standing) and among those from households where there are greater educational and cultural resources, such as books, magazines, newspapers, and encyclopedias. The relationship between household socioeconomic status and achievement scores is markedly stronger than the relationship between television viewing and academic achievement, with young people from households of higher status performing much better.
A good example is the data produced by the 1980 California Assessment Program, which was sponsored by the state department of education. Tests and questionnaires were administered to all pupils present on a given day in the sixth and twelfth grades (about 282,000 and 227,000 students, respectively). The pattern among these more than half million young people displayed the inverse relationship between television viewing and scores on standardized tests devised by the department for the three basic skills of reading, written expression, and mathematics. The inverse relationship was less pronounced among students in the sixth grade than among those in the twelfth grade. The inverse relationship was also less pronounced among students from households that had a lower socioeconomic status.
Another good example is the 1990 study by Steven L. Gortmaker (of the Harvard University School of Public Health) and his colleagues. These researchers uncovered a set of very-high-quality data that would allow them to explore relationships between television viewing and achievement scores. The data had been collected by the U.S. government's Health Examination Program from a sample of about 1,750 young people. A first set of data was gathered between July 1963 and December 1965, when the respondents were between the ages of six and eleven, and a second set of data was gathered from the same people between March 1966 and March 1970, when the respondents were between the ages of twelve and seventeen. The advantages of these data were (1) that the sample was very large and representative of the noninstitutionalized population of the United States for the ages covered, (2) that the design permitted the examination of changes in test scores over time, and (3) that the measures included three widely recognized standardized tests of intellectual aptitude as well as amount of television viewing and a variety of background variables. The three standardized tests were the Wechsler Intelligence Scale for Children (WISC), the Wide Range Achievement Test in Arithmetic (WRAT-A), and the Wide Range Achievement Test in Reading (WRAT-R). The fact that the data were collected in the 1960s and early 1970s was not a serious impediment to their use because, in the absence of very large changes in television or in the way in which young people use the medium, one would not expect large changes over time in the relationship between television use and achievement. Certainly, the data would reflect circumstances for young people of those ages at that time.
In this example, Gortmaker and his colleagues embarked on the conventional path of using survey data to investigate the likelihood of a causal link between two variables. The strength of survey data, unlike those produced by laboratory experiments, is that they represent real-world occur-rences, and when the sample is representative of the population as a whole, any outcomes can be said with great confidence to apply to the population represented. The weakness for causal inference is that surveys describe what occurs rather than linking a subsequent outcome to a prior event or treatment, while such a link is provided by an experiment. The logic of making a case for a causal link from survey data is (1) a demonstration that there is a relationship between two variables and (2) the documentation that the relationship persists after as many other variables as possible are controlled statistically so that the possibility can be ruled out that the relationship is actually attributable to another variable. Then, by careful reasoning or statistical analysis, a case must be made that the ostensible cause preceded the effect in time, since the logic of causation insists that a supposed cause cannot occur subsequent to an alleged effect.
The first step of these researchers was to examine the data in the second measurement, when the young people were between the ages of twelve and
The survey is a research method that seeks to describe a population by the use of questionnaires, tests, interviews, and other methods by which the attributes of those making up the population can be recorded. Most often, the population is made up of people, but anything that occurs in aggregates can be surveyed—businesses, housing, manufacturers, radio and television stations, and schools and universities. Usually, a sample is drawn to represent a much larger population because this makes the collection of information about the population much less expensive than if every member were examined. Best known are the opinion polls of presidential choices and other public preferences that receive widespread news coverage and often are sponsored by the media. However, thousands of surveys are conducted each year in the United States under the sponsorship of the federal government, political candidates, businesses, and other organizations, as well as the media.
In the case of academic achievement and television viewing, surveys often have been used to determine whether there is a relationship between the two. For example, both the 1980 California Assessment Program and the 1990 Harvard School of Public Health study by Steven L. Gortmaker and his colleagues matched data from questionnaires about television viewing and other attributes of students with their scores on tests of achievement or mental ability. Both studies produced reliable and valid data for reaching conclusions about the populations represented—California public school students (and particularly those in the sixth and twelfth grades, from whom the data were obtained, but one would expect the findings to be similar for other grades) and children and teenagers nationwide (because the sample of about 1,750 was statistically representative of U.S. children and teenagers).
The three principal criteria by which surveys are evaluated are (1) the sample, (2) the measures, and (3) the analysis. The sample is judged on representativeness and appropriateness for the purpose at hand. A random or probability sample gives every member of a population an equal chance of being included, and it is statistically representative of the population. This means that conclusions can be drawn about the larger population with precise margins of possible error (i.e., in only one out of twenty times would the actual percentage for the population vary from the survey by more than plus or minus a stated number of percentage points). Samples that are not random can still be useful if they are large and varied enough for the comparison of subgroups based on gender, age, socioeconomic status, or other attributes, such as beliefs, attitudes, or test scores of any sort. Measures must be reliable, in the sense that outcomes must not vary unless the variable being measured in fact has changed, and they must be valid, which means that they must represent accurately the intended variable whether it be demographic, a belief, an attitude, or something measured by a test. The analysis will become more useful with a greater effort to relate one variable (such as amount of television viewing) to another (such as achievement in written expression) and to examine the interrelationships of more than two variables at a time.
Surveys have three major uses in research. The first is simply to report on the attributes of a population (e.g., how many respondents come from households whose head is an unskilled worker, or how many respondents approve of the way in which the president is handling the job). The second is to explore relationships between variables (e.g., whether achievement test scores vary with amount of television viewing). The third is to detect evidence of causation by examining whether a necessary condition, an association between two variables, can be explained by the influence of another variable or variables and whether the time-order requirement that a cause must precede a consequence has been met (e.g., whether nontelevision factors explain the inverse association between viewing and achievement, or whether greater television viewing precedes lower academic achievement).
Even when a sample is not unambiguously representative of a much larger population, surveys are useful because their findings may be suggestive of what would be the case for such a population (and it is for this reason that academic researchers legitimately often pull a sample from a convenient population, such as the student body of their university, the enrollment of a nearby school, or the voters in a particular city). Surveys are thus one of the fundamental means of scientific inquiry.
seventeen, to determine the relationship between television viewing and test scores. They found a substantial inverse correlation, with scores for all three tests declining in a linear fashion as the amount of viewing increased. The researchers next addressed whether television viewing should be considered a cause of the lower scores. They turned to the relationships between test scores at the time of the second measurement and the amount of television viewed at the time of the first measurement. This would establish whether the necessary condition was met for an inference that television had a causal role—a time order in which viewing preceded the outcome. Indeed, television viewing stoutly remained inversely associated with the three test scores. The researchers then controlled for the earlier scores on the same three tests, which meant they now would be examining only changes in scores since both the earlier viewing and testing. The inverse associations dropped to a fraction of their original values. Next, the researchers controlled for other variables, such as time of year, region and size of place of residence, race, and household socioeconomic status, all of which have well-documented long-standing relationships with average amounts of viewing. The inverse associations essentially vanished.
Gortmaker and his colleagues concluded that the data "indicate no significant causal relationship between the amount of television viewed and the mental aptitude and achievement test scores of adolescents." However, this conclusion is limited to the type of tests they employed. The three tests, the WISC, WRAT-A, and WRAT-R, are essentially measures of traits that remain quite stable over time, and, in fact, people who take the tests at one point in time usually score about the same when they take the tests at a later point in time. As a result, when the researchers controlled for the earlier test scores, they also reduced strongly the plausibility of any inverse association remaining. Thus, these data only unambiguously confirm that those who score lower on standardized tests of intellectual ability on the average will watch greater amounts of television.
Logically correct in every aspect of its execution, this study by Gortmaker and his colleagues thus does not definitively establish that there is no causal relationship between the amount of television viewing and academic achievement. First, it does not cover the possibility of the displacement of time that might be spent acquiring the three basic skills of reading, written expression, and mathematics, which would occur at earlier ages. Second, it does not employ measures reflective of and sensitive to behavior that might more realistically be negatively affected by greater viewing (e.g., school grades, scores on homework assignments, tests designed to measure progress over a semester or a year, or, outside the classroom, choice of reading matter).
Inverse associations between viewing and achievement scores would not necessarily represent the effects of television use, and there are several reasons for this. First, young people from households that are lower in socioeconomic status on the average watch greater amounts of television. Also, because of the strong positive relationship between socioeconomic status and scholastic performance, young people from households that are lower in socioeconomic status on average score lower on achievement tests. Second, mental ability has a strong positive relationship with achievement, but it is inversely related to television viewing, so that those who on average watch greater amounts of television also on average will score lower on achievement tests. Third, those who are under stress (e.g., with troubling personal, family, or social problems) on average watch greater amounts of television as a means of flight from their difficulties, and these same stress factors are likely to hamper scholastic performance. Similarly, those students who are not performing well in school might, in their frustration, turn to television as an escape.
Thus, there are very good reasons for concluding that those students who watch greater amounts of television have attributes or are experiencing circumstances that are likely to be associated with lower levels of achievement. The explanation suggested by these patterns is that greater television viewing is the outcome of influences that themselves contribute to or are associated with lower achievement, and greater viewing is not a cause of that lower achievement.
However, additional evidence indicates that this explanation is too simplistic. There are many documented ways in which television use may interfere with success at school. First, academic tasks, such as reading and problem solving, have been shown to be less effectively done in the company of television—reading comprehension is lower and right answers are fewer. Second, during the first through third grades when children are learning the basic skill of reading, some students will use time that could be spent mastering this skill for another, less frustrating activity—television viewing. The same applies to the two other basic skills, written expression and mathematics. Third, those students who watch a great deal of television are more likely to scoff at the value of books and are less likely to read outside of assignments, while those who do such reading are more likely to perform well scholastically. Fourth, those who habitually watch a great deal of television also are less likely to expend as much cognitive attention and thought when they do read and are more likely to prefer (to a greater degree) to read undemanding nonfiction about celebrities and light fiction-fare that resembles television programming. As George Comstock and Erica Scharrer conclude in their assessment of the evidence in Television: What's On, Who's Watching, and What It Means (1999), most of these outcomes occur among both children and teenagers.
Thus, the evidence points toward a number of adverse influences—interference with learning the basic skills, lowered quality of effort when reading or completing academic tasks in the company of television, and, among those who watch large amounts of television, desultory concentration while reading, low esteem for books, and the nurturing of tastes for reading matter that resembles television in substance and style. The most reasonable interpretation of the inverse relationships between amount of television viewing and scores on standardized achievement tests is one that incorporates both perspectives; they in part reflect greater attention to television by those who are less likely to do well academically for other reasons and in part reflect the detrimental effects of television viewing on academic achievement.
Effect Size and Who Is Affected
It is important to recognize that the inverse relationships are quite modest, amounting (on the average) to only about a 10 percent decline in scores between those who view the least amounts of television and those who view the most. The detrimental effects that television viewing has on achievement scores are thus quite small. The additional consequences for intellectual activity in general are probably more serious. These include the lowered esteem for books, the reduced concentration while reading, the preferences for undemanding entertainment, and lesser ability in the three basic skills—all of which, for some, may be a consequence of greater television use. Those people who are most likely to be affected adversely are (1) those who voluntarily allow television to displace time that might have been spent learning the three basic skills, (2) those for whom opportunities foregone would have been of greater academic value, and, of course, (3) those who watch television for a very large amount of time.
California Assessment Program. (1980). Student Achievement in California Schools. 1979-80 Annual Report. Sacramento: California State Department of Education.
Comstock, George, and Scharrer, Erica. (1999). Television: What's On, Who's Watching, and What It Means. San Diego, CA: Academic Press.
Gaddy, Gary D. (1986). "Television's Impact on High School Achievement." Public Opinion Quarterly 50(3):340-359.
Gortmaker, Steven L.; Salter, Charles A.; Walker, Deborah K.; and Dietz, William H., Jr. (1990). "The Impact of Television Viewing on Mental Aptitude and Achievement: A Longitudinal Study." Public Opinion Quarterly 54(4):594-604.
Koolstra, Cees M., and van der Voort, Tom H. A.(1996). "Longitudinal Effects of Television on Children's Leisure-Time Reading. A Test of Three Explanatory Models." Human Communication Research 23(1):4-35.
Koolstra, Cees M.; van der Voort, Tom H. A.; and van der Kamp, Leo J. Th. (1997). "Television's Impact on Children's Reading Comprehension and Decoding Skills: A 3-Year Panel Study." Reading Research Quarterly 32(2):128-152.
MacBeth, Tannis M., ed. (1996). Tuning in to Young Viewers: Social Science Perspectives on Television. Thousand Oaks, CA: Sage Publications.
Neuman, Susan B. (1988). "The Displacement Effect:Assessing the Relation Between Television Viewing and Reading Performance." Reading Research Quarterly 23(4):414-440.
Neuman, Susan B. (1991). Literacy in the Television Age. Norwood, NJ: Ablex.
Van Evra, Judith (1998). Television and Child Development, 2nd edition. Mahwah, NJ: Lawrence Erlbaum.
Williams, Tannis MacBeth, ed. (1986). The Impact of Television: A Natural Experiment in Three Communities. New York: Praeger.