 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Frederick J. Zimmerman, PhD
Department of Health Services
University of Washington
Seattle, WA 98115
To the Editor.—
We read with interest the article "There Is No Meaningful Relationship Between Television Exposure and Symptoms of Attention-Deficit/Hyperactivity Disorder"1 by Stevens and Mulsow. The authors contrast both their methods and their findings with ours, which were reported earlier in Pediatrics.2 Stevens and Mulsow make much of their use of structural equation modeling (SEM) to reduce measurement error without evidently understanding the distinction between the potential of SEM and its performance in any particular analysis. Their statement that their estimates are free of measurement error because of their use of SEM ("measurement error was estimated and removed") betrays an ignorance of how SEM (or for that matter how any statistical methodology) works. If there were a method that eliminated measurement error, we would have the Holy Grail of statistics. SEM can never produce estimates free of measurement error, and the extent to which it reduces measurement error depends on several key, but unexamined, assumptions, the most salient of which is the appropriateness of the theoretical model. We find their theoretical model (see their Fig 1) to be counterintuitive in its assumptions that (1) parental limits on television would have a direct effect on symptoms of ADHD and yet no effect on television exposure and (2) gender would not be relevant.
It is also curious that Stevens and Mulsow cast their analysis as an attempt to replicate our findings. As the introduction and title of our article clearly state, our intention was to assess the association of early television exposure (at ages 3 and below) with symptoms of attentional problems measured 
4 years later. Not having such data, Stevens and Mulsow instead explored the relationships between television exposure at age 5 and symptoms of attention and behavior problems at age 6. As they note in their discussion, but fail to mention in their introduction, our primary hypothesis was that early exposure to television, because of the special and well-documented neurodevelopmental stage of the newborn brain, poses a unique risk. Although the nature of both of our studies precludes making causal assertions, there are ample reasons to be concerned that early exposure is, in fact, different from subsequent exposure. Their data, by design, cannot rule out this possibility, and our data give reason to be concerned about it. We agree with the authors that more studies are needed and that, ideally, an experimental design would be used. We are surprised that they feel it would be unethical to conduct such a trial given that the title of their article projects a high degree of certainty that there is no meaningful relationship between television viewing and ADHD. In our opinion, given the considerable exposure to television at increasingly younger ages and the Surgeon General's report that emphasized the need to explore environmental factors related to the development of ADHD,3 such studies are extremely important and should be performed. An ethical way to do so would be to actively encourage children not to watch television. We are currently planning just such a study.
REFERENCES
Related articles in Pediatrics:
This article has been cited by other articles:
 |
|
 |
|
  F. J. Zimmerman, D. A. Christakis, and A. N. Meltzoff Television and DVD/Video Viewing in Children Younger Than 2 Years Arch Pediatr Adolesc Med, May 1, 2007; 161(5): 473 - 479. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
