Viewing Television Before Age 3 Is Not the Same as Viewing Television at Age 5: In Reply

Abstract

In Reply.—Reading the letter submitted by Christakis and Zimmerman, we were surprised that they chose to focus their attention on our statistical interpretation instead of taking this desirable opportunity to clarify unanswered and important questions related to their own methodology, especially when these questions have also been raised by others.1,2 For example, they highlight concerns about the posited relationships in our model when they fail to even discuss the goodness of fit of their own model.3 Without this information, how were readers to know whether the relationship between early television exposure and later symptoms of attention-deficit/hyperactivity disorder (ADHD) was a statistical artifact or an effect size that indicated a meaningful relationship? They also did not choose to explain why they selected data collected in 1979 rather than more current data that would reflect present patterns of television watching. Finally, they did not explain why they included 2 items in their “attention” measure that were not associated with ADHD: mental confusion and obsessive behaviors.When we referred to measurement error, we were using an accepted standard in describing estimated measurement error in structural equation modeling. As reported in the highly regarded work of Tabachnick and Fidell, “When relationships among factors are examined, the relationships are free of measurement error because the error has been estimated and removed, leaving only common variance.”4(p712) Furthermore, MacCallum et al5 describe the likelihood of multiple models fitting data sets equally well. This work was initially published over 10 years ago, so we assumed that this was an implicit understanding of researchers in the field.We clearly point out that we do not advocate that television viewing is a healthy behavior. As a result, we could never agree that randomly selecting, randomly assigning, and purposely exposing children to high levels of television to conduct a true experimental study to investigate whether certain groups of children would develop ADHD is an ethical means of studying development. Experimental design requires treatment as well as control groups. Any study conducted without random assignment to one of these groups would not be an experiment. Even so, we hope that this inquiry continues with as much enthusiasm as it has already enjoyed and that the information generated will provide parents with practical and meaningful indications.Our intended purpose was not to simply replicate the Christakis et al3 study but to “extend” the investigation to older children. Because research is cumulative, one would expect that we would contrast our findings with their results to better elucidate the connection between television viewing and ADHD symptoms over time, providing more information to contribute to the description and understanding of development. We do not view our findings as contradictory but as complementary. We are continuing to extend work in this field and expect to have results including a younger cohort. “Stay tuned” for further developments.