eLetters to:

ARTICLE: Dimitri A. Christakis, Frederick J. Zimmerman, David L. DiGiuseppe, and Carolyn A. McCarty Early Television Exposure and Subsequent Attentional Problems in Children Pediatrics 2004; 113: 708-713 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

eLetters published:

Causal or Casual ?

Andrew gallagher   (2 April 2004)

Fatal Design Flaw: Self-Selection

Gordon E. Finley   (6 April 2004)

An important study but we need to recognize the possible limitations

Robert Kubey   (6 April 2004)

Ivory Tower

Kenneth P. Sullivan, Ph.D.   (6 April 2004)

Re: Causal or Casual ?

Jon Steuernagle IV   (6 April 2004)

Re: Causal or Casual ?

Jan N. Widerman   (6 April 2004)

Detrimental TV watching: Due to concurrent exposure to brominated flame retardants?

Claudette R Bethune   (6 April 2004)

If Psychologists Made These Kinds of Causal Claims, We'd Have to Turn Each Other In!

Wallace E. Dixon, Jr.   (6 April 2004)

Causal or Casual ?

Paul K Brandon   (6 April 2004)

Re: Ivory Tower and Casual or Causal

Maria Markham Thompson   (8 April 2004)

Correlation or cause?

Guy Cumberbatch   (8 April 2004)

High prices limit public discourse

Rebecca J. Carsel   (10 April 2004)

Full text without charge

Charles S. Harris, PhD   (11 April 2004)

A high score of inattentiveness is not ADHD

Søren Dalsgaard   (13 April 2004)

Food for Thought

Paula S. Wojtek   (15 April 2004)

Correlation not implying causation: worse than the authors realize

Bruce Arnold   (15 April 2004)

Re: If Psychologists Made These Kinds of Causal Claims, We'd Have to Turn Each Other In!

Kelly A. O'Grady   (25 April 2004)

Danger of multi-collinearity

Barry Greenberg   (18 May 2004)

Do toddlers really have the attention span to watch televison for hours each day?

Robert D. Cunningham   (16 July 2004)

Alternate design method

Corbin VanDeWege   (16 December 2004)

Another Thought

Kristina E. Clevenger   (12 January 2005)

Statistical problems and interpretation of probability

John Sandberg   (16 March 2005)

Causal or Casual ? 2 April 2004
Andrew gallagher, dr Send letter to journal: Re: Causal or Casual ? andrew.gallagher{at}hotmail.com Andrew gallagher	Having lived with an ADHD child, one of the few things they will do by themselves (so you can get on with domestic chores, see to other kids, talk to friends etc) is watch TV. I don`t like it but you`ve got to get through the day and stay sane. Maybe thats why these kids spend more time in front of the TV - so their parents can cope. It doesnt mean that TV gives them problems, maybe their problem gives them extra TV exposure.
Fatal Design Flaw: Self-Selection 6 April 2004
Gordon E. Finley, Professor of Psychology Florida International University Send letter to journal: Re: Fatal Design Flaw: Self-Selection finley{at}fiu.edu Gordon E. Finley	Front-page media reports (The Miami Herald, 4/5/04) have failed to catch the fatal design flaw of self-selection that requires clarification (Finley, 1998). The present findings must be interpreted cautiously because they are based on subject and/or caretaker self-selection and not experimentation (where subjects are randomly assigned to conditions such as amount of TV exposure). Specifically, rather than TV “causing” ADHD, it is equally plausible that the present results were “caused” by: (a) very young children with dispositions to ADHD self-selecting fast-moving TV as a “self-soothing” or “attention-holding” mechanism; or (b) caretakers self-selecting the context of TV for their ADHD children to facilitate family coping. Reference Finley, G. E. (1998). On individual difference, choice, selection, and complexity in adoption research. Adoption Quarterly, 1 (4), 83 – 91.
An important study but we need to recognize the possible limitations 6 April 2004
Robert Kubey, Prof. of Journalism & Media Studies; developmental psychologist; Director, Center for Media St Rutgers University Send letter to journal: Re: An important study but we need to recognize the possible limitations kubey{at}scils.rutgers.edu Robert Kubey	I've read the article and am very impressed by this study and think it is a very important contribution to the field. I especially appreciate the authors' own excellent caveat, in which they write: "Third, we cannot draw causal inferences from these associations. It could be that attentional problems lead to television viewing rather than vice versa. . . .It is also possible that there are characteristics associated with parents who allow their children to watch excessive amounts of television that accounts for the relationship between television viewing and attentional problems. For example, parents who were distracted, neglectful, or otherwise preoccupied might have allowed their children to watch excessive amounts of television in addition to having created a household environment that promoted the development of attentional problems." The news media will doubtlessly report on this study as if there IS a causal relationship even though the authors rightly warn against doing so and they are to be admired for being candid about this in their write-up. Partialling out SES and numerous other variables is always important in such work, but unless the work was done with a great deal many more questions to parents, it is nearly impossible to know if there are important differences in parental style and approach to television and other activities that might be going on here. But this team has made a very important start and it's important that the study be replicated and different variations of the study be conducted, and in different nations and locales. "Pediatrics" asks contributors if they have an interest in the area and I do indeed as I've long suspected that shortened attention spans are encouraged by a good deal of contemporary media. I commented on this in my work on the "intolerance for unstructured time" (Kubey, 1986, Journal of Communication), in my 1990 book with Mihaly Csikszentmihalyi, "Television and the Quality of Life" (Erlbaum) and in our more recent Scientific American article on so-called "Television Addiction" from Feb. 2002. And many others have as well. Let me add, in my own interviews with leading creators of American television, many of which date back to 1987 and up and through 2003, that the great percentage of the over 100 people I have interviewed have reported to me, when asked, that THEY believed that the attention span of the American public was indeed growing shorter. Forty of these interviews are reported in my book (2004), "Creating Television: Conversations with the People Behind 50 Years of American Television" and I will be happy to make some of relevant quotes available for those who might be interested. Or I could post them on this site if people wish to see them. The Rutgers News Service will be doing a release on this book and its findings probably later this week, or next. But here's the point: most everyone working in the media ardently believes they must pander to a shortened attention span: producers and reporters in print, radio and TV news, and in every other entertainment medium. So, even if the media haven't caused any shortening of the public's attention span, there is no question in my mind that media producers create products and more rapidly intercut their materials in order to hold on to people's waning attention and interest levels. And this might have had an effect itself, even if there was no real effect to begin with. The development of advertising techniques that influenced the style and intercutting ratios in music videos has been a matter of great influence on the popular culture and subsequently on editing styles in film and television and advertising, once again. Byron Reeves and Annie Lang and their colleages at Stanford and at Penn State, respectively, have demonstrated that rapid cuts can activate the primitive, biologically and evolutionarily grounded "orienting response" in people, and this is referred to in the important Jerome Singer article cited in the current piece. Let us not forget the Harvard study that showed that the average length of soundbite for presidential candidates on national TV news in 1968 was about 45 seconds whereas it dwindled down by a factor of 5 to only about 8-9 seconds, I believe as early as 1988. And it's stayed low subsequently. CBS tried to lengthen their soundbites for presidential candidates in 1992, I believe, letting the bites run at least 30 seconds but they killed the policy within about a week's time. I believe this was reported in the New York Times sometime in 1992 or perhaps it was 1996. Moreover, a number of researchers over the years have made observations or commented on the real possibility that children's attentional and impulse control and ability to stay on task, and tolerate frustration, might all be negatively affected by over-exposure to television. Subsequently, though I don't follow it as closely, critics of video games and computers have made similar claims. I do not share in the propensity of all too many researchers to demonize the media. There are wonderful things on television for people of all ages, but everything in life, and especially for young developing minds, needs to be done in moderation. There may be positive effects, as well from rapidly edited media and video and computer play as well. There is reason to believe, for example, that recent increases in IQ scores around the world are partly attributable to video and computer game play (see Patricia Greenfield and others on this). To my knowledge, which is limited here, there has been no better explaination offered for this secular trend. And frankly, I think it possible that even the mind puzzles that kids run into on cereal boxes, in school, and even in restaurants, has prepped them, more than past generations, to do well with certain IQ type questions, especially in spatial relations. But one more intriguing idea that I have seen elsewhere. There was speculative work done some years ago that suggested that the PARENTS of kids with ADHD might view heavily themselves, both as a way to escape their ADHD kid/s and/or that the children might try to act out more, and be more kinetic, like the people on TV in order to gain the attention of their parents who are so absorbed with watching TV. Here, the argument was made that kids are trying to be like what the parents are attracted to on the TV screen. A scary thought and a pretty wild one, but I thought I'd add it as it IS in the literature at least in discussion. Thus, looking at parental level of viewing is an important co-variate to consider in this sort of work as well. At bottom, this is a very important study and should shake up a lot of researchers around the world to closely study the topic, and it should also encourage increased funding by the U.S.P.H.S. in funding, once again, basic research on the influence of media, especially in developing children. R. Kubey, Rutgers
Ivory Tower 6 April 2004
Kenneth P. Sullivan, Ph.D., Psychologist Send letter to journal: Re: Ivory Tower Ksuli{at}hiwaay.net Kenneth P. Sullivan, Ph.D.	I do not like that this article, and many others in scientific literature, are expensive to read. Scientific findings would serve the world better if they could be read freely. Restricting information puts the scientific community in the ivory tower. Oh well, maybe the rest of us can watch this article get summarized on TV.
Re: Causal or Casual ? 6 April 2004
Jon Steuernagle IV, Med Student Send letter to journal: Re: Re: Causal or Casual ? ratphH1nk{at}hotmail.com Jon Steuernagle IV	I would agree with the doctor in the previous post and would like to point out something that I did not see directly addressed in this study. If there appears to be this correlation between X hours of television watched for this age group and an Y% increase in symptoms, that are categorized as attention related at age 7, I would also like to know how the remaining 24 -X hours of the childs day was spent. It would be interesting to see if further correlations can be drawn or perhaps, nullifying activities can be identified. For example: If a child spent 2 hours per day watching television but also engaged in some period of direct interaction with caregivers, or focused activities (such as reading, block activitites, drawing/painting etc..) does this affect the later prevalence of diagnosis of attention problems. Cheers, Jon
Re: Causal or Casual ? 6 April 2004
Jan N. Widerman, Pediatrician Send letter to journal: Re: Re: Causal or Casual ? jwiderman{at}comcast.net Jan N. Widerman	WE know that children with ADHD can focus for long periods of time on things that interest them. Thus, how many times do we hear from the parents that their child cannot have ADHD because they play video games or watch TV for hours. In fact, one theory by Thompson is that one of the characteristics in ADHD is the ability to lose track of time when the excitement of the activity is present. He writes about the days of being hunters and farmers and states that the hunters have the ADHD qualities. We also suspect that there are signs even in the newborn period that may correlate with ADHD (colic?). Lastly, no mention has been made in regard to the parents having ADHD. This article seems to neglect the know genetic occurrence of this disorder.
Detrimental TV watching: Due to concurrent exposure to brominated flame retardants? 6 April 2004
Claudette R Bethune, research scientist National institute of nutrition and seafood research Send letter to journal: Re: Detrimental TV watching: Due to concurrent exposure to brominated flame retardants? claudette.bethune{at}nifes.no Claudette R Bethune	Hello, As a UW alumni from Pharmacuetics, I enjoyed reading your article! While environmental pollutants and childrens health is a complex emerging issue, one class of chemicals in, and emitted from, TV sets is known to cause developmental disabilities. I have been studing brominated flame retardants here in Norway, and research seems to indicate exposure to these thyroid disrupting chemicals can come directly from watching TVs and computers which contain these compounds in the plastic components of these products. A group of flame retardants, PBDEs, are almost identical in their chemical structure to PCB which binds to the thyroid gland which regulates many functions such as behaviour. Flame retardants typically make up from 5-30% of the final product weight. The brominated flame retardants are constantly released from TVs made with them, both new and old sets, and typically the emitted chemicals adhere to dust. Thus the route of flame retardant exposure in children can be through ihalation, dermal,and ingestion of dust particles. Also, viewing TV on a foam sofa increases exposure, as the foam contains flame retardants (typically penta PBDEs). Since TV viewing is now associated with specific chemical exposures, it would be interesting to see a similar study which considers exposure to flame retardants and resuting dioxin and furans which occur during TV time. Claudette R. Bethune, PhD Nasjonalt institutt for ernærings- og sjømatforskning National Institute of Nutrition and Seafood Research (NIFES) P.O. Box 176 Sentrum, N-5804 Bergen, Norway www.nifes.no tlf. +47 55905129 fax+47 55905299 email cbe@nifes.no
If Psychologists Made These Kinds of Causal Claims, We'd Have to Turn Each Other In! 6 April 2004
Wallace E. Dixon, Jr., Developmental Psychologist East Tennessee State University Send letter to journal: Re: If Psychologists Made These Kinds of Causal Claims, We'd Have to Turn Each Other In! dixonw{at}mail.etsu.edu Wallace E. Dixon, Jr.	I am amazed about what passes for scientific claims these days. Not only did the American Academy of Pediatrics some years ago get away with calling for the removal of all television viewing for all children under age 2 with absolutely no evidence other than its own opinion, now it seems to be advocating for making causal attributions from correlational data. I am not challenging the validity of the revealed correlation between childhood attention-related disorders and television viewing in infancy and toddlerhood. I am trusting the authors conducted their analyses accurately. The problem is that the authors ran wild in making causal implications about the direction of effects with absolutely no scientific license to do so. If psychologists did this, I would like to think their peer reviewers would call them on it, and require them to reframe their conclusions in more cautionary language long before publication would ever be permitted. Apparently the same admonitions do not take place among all medical researchers, nor among some of their peer-reviewed journals. Here's the deal. At least two alternative explanations can be made to account for the relationship between TV-viewing and attention- related disorders in childhood. First, it is possible, indeed I would say even likely, that children with attention-span problems are more likely to find television interesting than children without attention-related problems. Children's television programs are made exceptionally attention-grabbing by virtue of their fast-paced, exciting, frequent-scene-changing, musical styles of presentation they employ. Although this is appealing to all children, it's not surprising that it's especially appealing to attention-disordered children who seem to find the normal world especially boring. TV may provide the most stimulating bang for the buck of any environmental stimulus out there. Unlike attention-disordered children, children without such disorders can find solace in lots of stimuli besides TV; but for attention-disordered infants and toddlers, this may be the best they can hope for. The authors dismiss this possiblity because according to them, most experts believe that attention-related disorders don't emerge until later childhood. Unfortunately, the beliefs of experts have nothing to do with reality, unless those beliefs are based on data. To be sure, attention-related disorders are likely to be "discovered" in childhood, as opposed to infancy, because it is during childhood that children with attention-related disorders run smack-dab into the highly structured school setting, which tends not to give children with attention-related disorders much latitude. But the fact that these disorders aren't reported frequently until school age, has nothing to do with whether or not they are present before school age. These disorders may be present from birth, we just don't know! If a child has an attention-related disorder at birth, it strikes me that she or he would likely to keep the disorder until well into adulthood. This leads to a second alternative interpretation of the data reported in this article. That is that parents of infants who have attention-related disorders may turn to TV as a last resort for dealing with their potentially difficult children. The literature on infant and toddler temperament is replete with data linking temperamental difficulty to parenting practices. Children with difficult temperaments tend to have weary, tired, hapless, and frustrated parents. Is this a coincidence? Probably not. It makes sense that children with difficult temperaments would produce parents who develop parenting tactics to deal with their children's difficult temperaments, and TV-viewing might be one of the most successful parenting tactics employed. In this case, TV-viewing wouldn't be responsible for attention-related disorders. Quite the contrary, attention-related disorders would be responsible for parenting tactics that result in TV-viewing. But alas, this possibility is given short-shrift by the authors in this article. The fact is that infants and todders with difficult temperaments also tend to have short attention-spans, so parents who can find a means to successfully maintain the attention of their temperamentally difficult infants and toddlers may have also found a means to attenuate their infants' and toddlers' temperamental difficulty. (Ironically, if I were trying to publish this kind of research in Pediatrics, I might so far as to claim that the difficult temperament caused the parents to become weary, tired, hapless, and frustrated. But as a psychologist, I hold strongly to the old adage "correlation does not imply causation." So I am left with the mere suggestion that it might be so.) The bottom line is that the recieved and, it seems to me, extremely destructive result of this article may be that the general public will declare that when children have ADHD, it is the fault of their parents who let them watch too much TV. If ADHD gets blamed on parents, who knows what's next; maybe health insurance rates for treating ADHD will be determined partially as a function of how many TVs are in the home. Meanwhile, research funding for the real cause of ADHD may become reduced by communities who fail to understand the difference between correlation and causation, and attempts to find a cure for ADHD may become further removed from the gaping arms of waiting, suffering children.
Causal or Casual ? 6 April 2004
Paul K Brandon, Professor Psychololgy Dept, Minnesota State University, Mankato Send letter to journal: Re: Causal or Casual ? paul.brandon{at}mnsu.edu Paul K Brandon	There are many possible contaminating variables here besides the ones that Gallagher mentions. The data are based on retrospective self reports by the parents of the children whose behavior is in question. I see no indication of any systematic validation of the accuracy of these reports. It is quite possible that parents who are more sensitive to or concerned about their children's behavior are more likely to both report behavioral abnormalities and give higher estimates ot the time those children spend watching TV. To conclude, the only way to imply causation to any observed relationship between the reported time spent watching television and reported frequency of problem behaviors would be to randomly assign subjects to TV and NO TV groups, have them log defined behaviors and TV time in real time, and have external observers occasionally make reliability checks on their recording. Failing that, all we have is a vague suggestion that some sort of undefined relationship exists.
Re: Ivory Tower and Casual or Causal 8 April 2004
Maria Markham Thompson, MUP, CPA, CFA parent Send letter to journal: Re: Re: Ivory Tower and Casual or Causal mariasothername{at}aol.com Maria Markham Thompson	It seems particularly inappropriate to limit access to information that has been widely reported and misinterpreted in the general media. The readily available abstract states a recommendation for limiting television viewing in early childhood that is not well supported by the study. Parents who are given such admonishment should have equal access to the study in order to assess the validity of the standard against which their behavior and the choices they make for their children will be measured. A particularly glaring confusion between causality and correlation in the paper is allowing the authors to suggest that "parents who were distracted, neglectful, or otherwise preoccupied might have allowed their children to watch excessive amounts of television in addition to having created a household environment that promoted the development of attentional disorders." The opposite possibility is not considered that the child was born with biochemical or structural brain differences that were stressors for the parents who in turn responded with television as a form of respite care in order to meet their personal and familial obligations. Since the researchers did not actually study children diagnosed with ADHD, it seems somewhat inappropriate for the conclusion to be that they have "added inattention to the previously studied deleterious consequences of excessive television viewing, including violent behavior and obesity." I believe the differences in research methodology between this study and the ones cited are too great to so easily link them. My anecdotal observations of several homes with ADHD children are that at some point parents realize that they cannot provide the high levels of stimuli for the number of hours that our children require. Realities such as gainful employment, household maintenance, and need for sleep become competing imperatives. Lacking unlimited resources to hire teams of people to step into the breach, we often use video-based stimuli (televised, VHS and DVD) as compromises in getting time to do what we must and a few moments of personal solace to maintain our sanity. More research on video-based products that meet the stimulus needs of ADHD children in the early childhood would provide far more benefit to our children and their families. Maria Markham Thompson, CPA, CFA Parent of an ADHD daughter, age 12
Correlation or cause? 8 April 2004
Guy Cumberbatch, Psychologist CRG, Aston Science Park, UK Send letter to journal: Re: Correlation or cause? DrGCumberbatch{at}aol.com Guy Cumberbatch	The amount of television that children watch is well predicted from the viewing behavior of their parents. Moreover, there has been growing support for the idea that children with ADHD tend to have parents with ADHD. Thus it would be risky to assume that the television habits of infants and toddlers is causally related to ADHD rather than merely a correlate. Readers may have noted that under results the figure of 3.6 hours per week watched by 3 year olds should read 3.6 hours per day. I am an independent researcher working for both broadcasters and regulators in the UK and elswhere.
High prices limit public discourse 10 April 2004
Rebecca J. Carsel, M.S. Send letter to journal: Re: High prices limit public discourse bcarsel{at}charter.net Rebecca J. Carsel	I am a member of the public and parent of a three-year-old. After hearing this article widely reported in the news media, I wanted to read it myself to see how the authors address the question of correlation versus causation. I was dismayed to see that it would cost $12 just to read the article. The high cost of this article prevents people from verifying the accuracy of media reports. Surely Pediatrics could make available a widely reported article such as this for free or at a reduced price, at least for a period of time. At $12 a read, Pediatrics is severely limiting public discourse about important scientific issues.
Full text without charge 11 April 2004
Charles S. Harris, PhD, psychologist webmaster, The Nurture Assumption website, http://xchar.home.att.net/tna/ Send letter to journal: Re: Full text without charge xchar{at}att.net Charles S. Harris, PhD	The complete text of this article is available without charge, in Adobe PDF format, courtesy of the American Academy of Pediatrics: http://www.aap.org/advocacy/releases/tvapril.pdf It is also available in HTML format, courtesy of Google: http://216.239.51.104/search?q=cache:TuBCS4w7kGkJ:www.aap.org/advocacy/releases/tvapril.pdf In the published paper, the authors quite properly note, third in the list of limitations of their study: "Third, we cannot draw causal inferences from these associations. It could be that attentional problems lead to television viewing rather than vice versa. However, to mitigate this limitation, we exploited the longitudinality of the data set and focused on television viewing at 1 and 3 years of age,well before the age at which most experts believe that ADHD symptoms are manifest.32,39 It is also possible that there are characteristics associated with parents who allow their children to watch excessive amounts of television that accounts for the relationship between television viewing and attentional problems. For example, parents who were distracted, neglectful, or otherwise preoccupied might have allowed their children to watch excessive amounts of television in addition to having created a household environment that promoted the development of attentional problems." The authors' caveat that causual inferences cannot be drawn from their study, however, did not deter them from drawing causal inferences: "Despite these limitations, our results have some important implications if replicated in future studies. First, we added inattention to the previously studied *deleterious consequences* of excessive television viewing, including violent behavior and obesity.41-43 Second, our findings suggest that *preventive action* can be taken with respect to attentional problems in children. Limiting young children's exposure to television as a medium during formative years of brain development consistent with the American Academy of Pediatrics' recommendations may reduce children's subsequent risk of developing ADHD." (emphasis added) Charles S. Harris, PhD Psychologist Webmaster, The Nurture Assumption website http://xchar.home.att.net/tna/
A high score of inattentiveness is not ADHD 13 April 2004
Søren Dalsgaard, MD, PhD Psychiatric Hospital for Children and Adolescents, University of Aarhus, Denmark Send letter to journal: Re: A high score of inattentiveness is not ADHD sda{at}buh.aaa.dk Søren Dalsgaard	The Pediatric paper is unfortunately currently a big story in Danish news media. Having worked with children with ADHD for the last 6 years I agree with many other comments here on the fact that children with ADHD often want to watch television. One other bias in the study is the lack of assessment of parental symptoms of ADHD. In addition, a high score on a CBCL-subscale is not diagnostic for ADHD, which the authors themselves also mention. Many children with attentive problems may have a high score on such a scale, including children with ODD/CD or children with an attachment disorder.
Food for Thought 15 April 2004
Paula S. Wojtek, adult educator interested party Send letter to journal: Re: Food for Thought pwojtek{at}howardcollege.edu Paula S. Wojtek	I read with interest the above mentioned article and am concerned with the growing number of students with ADD/ADHD. It seemes to me that very young children who ride in cars extended times each day are subject to the same types of rapid movement, changing scenery, distractions, etc. that they are exposed to with television viewing. It might be interesting to include in further studies the time spent in car seats, also.
Correlation not implying causation: worse than the authors realize 15 April 2004
Bruce Arnold, Economist Send letter to journal: Re: Correlation not implying causation: worse than the authors realize brucea{at}cbo.gov Bruce Arnold	It may seem impertinent for an economist to comment on a pediatric research article. However, I received considerable training in multivariate statistical analysis when I obtained my Ph.D., and I have read a considerable amount of advanced material about ADHD because of the fact that my son has been diagnosed with it. I think that the problem of correlation not implying causation that the authors mention may be much more serious than they realize, and that their statistical results may be completely--or almost completely--explained by the substantial genetic component to the causation of ADHD without any resort whatsoever to telvision viewing causing attention problems. Before explaining further, I should say that I had some difficulty interpreting the table of regression results in the paper because in my own field of economics people usually report regression coefficients--not odds ratios--and I was not certain what odds ratios were. If I have puzzled them out correctly, the odds ratio for a given explanatory variable is the multiplicative factor by which a unit increase in that variable affects p/(1-p), where p is the probability of having attention problems. In that case, an odds ratio greater than one means that the explanatory variable is associated with greater attention problems, and a ratio less than one means it is associated with lesser attention problems. Assuming this to be correct, then it appears to me that many of the explanatory variables are serving as proxies for maternal ADHD, which are correlated with attention problems in the child for genetic reason. Thus, the odds ratio for maternal self esteem for age "1" children is 1.36, meaning that low levels of maternal self esteem are associated with a higher probability of attention problems in a child. However, if the mother has ADHD, she likely is more of a failure in life than her equally intelligent friends and family and therefore is more likely than others to have low self esteem. Further, because of the genetic connection, her children are more likely than others to have ADHD. These two facts by themselves could explain the odds ratio, and I cannot think of any other likely explantion of why maternal self-esteem would be stongly associated with attention problems in her children. Similarly for maternal depression, for which the reported odds ratio for age "1" is 1.03. For the same reason that a mother with ADHD would be more likely than others to suffer from low self esteem, I think she would also be more likely to suffer from depression as measured by the variable used in this paper. This plus the genetic transmission of ADHD to the child could explain the reported odds ratio. Similarly for maternal education. The paper reports an odds ratio of 0.91, meaning higher maternal education is associated with lower likelihood of attention problems in the child. If the mother has ADHD, she likely found school to be difficult and not to her taste, so she probably did not get as much education as other people: that is, she was more likely to drop out of high school, or if she finished high school, less likely to go to college and finish a degree, and so on. This plus the genetic transmission of ADHD would explain the reported odds ratio. Similarly for the emotional-support and cognitive-stimulation scores. The reported odds ratios are 0.82 and 0.84, meaning that high emotional support and cognitive stimulation are associated with lower likelihood of attention problems in the child. My guess is that maternal ADHD probably inccreases the likelihood of low emotional support and cognitive stimulation of the child. That plus the genetic transmission of ADHD would explain the reported odds ratios. Finally, and most importantly, is television hours watched per day. The reported odds ratio is 1.09, indicating that increased time watching television by the child at age "1" is associated with increased likelihood of ADHD at age 7. I would guess that a mother with ADHD would be more likely to be, in the words of the authors, "distracted, neglectful, or otherwise preoccupied" and therefore that it would be more likely that there children would watch a lot of television. This plus the genetic transmission of ADHD would explain the reported odds ratio. Note that the fact that the television viewing measure in the paper was at ages "1" and "3"--before ADHD symptoms are known to become manifest--is irrelevant to this explanation. To sum up, the regression results reported in this paper appear to me to show not that increased television viewing causes attention problems but that a large number of proxies for maternal ADHD are associated with attention problems in children. This is exactly what we would expect from the genetic transmission of ADHD, which has already been know for some time.
Re: If Psychologists Made These Kinds of Causal Claims, We'd Have to Turn Each Other In! 25 April 2004
Kelly A. O'Grady, RN Lead Environmental Awareness and Detection (L.E.A.D.) Send letter to journal: Re: Re: If Psychologists Made These Kinds of Causal Claims, We'd Have to Turn Each Other In! lead{at}nrtco.net Kelly A. O'Grady	If Psychologists Made These Kinds of Causal Claims, We'd Have to Turn Each Other In! Psychologists often blame a person’s ability to succeed or fail on upbringing, life experiences or basic personality and now we have a new claim: TV watching. Most researchers would agree that such factors do influence how children function, however researchers now know that character traits are the product of complex neurological processes more heavily dependent upon the balance of neurotransmitter chemicals such as dopamine than on social factors - or pseudo environmental factors such as TV watching. Personality traits once thought a product of a child’s upbringing or social background are now found to be influenced by environmental factors in a very literal sense. For example, procrastination, which used to be considered a character trait, will soon be included as one of the DSM V criteria for Attention Deficit Disorder (ADD). FOr some time, ADD has been linked with early exposure to the poison lead. Lead, an undisputed neurotoxin, may affect numerous cell functions, including the release of neurotransmitters such as dopamine and serotonin. Just how much lead it takes to make this happen, and how one should measure this occurrence seems to be the issue at hand. Deborah Rice, a former Health Canada research scientist, has conducted many animal studies in order to gauge the neurotoxic impact of agents such as lead, mercury and PCBs on learning and behavioural performance. From her work she discovered: Monkeys treated with lead and PCBs show two "hallmark" traits: 1. the inability to learn from the consequences of previous behaviour and; 2. an inability to organize behaviour in a temporal sequence. There are similarities in behaviour between children diagnosed with Attention Deficit Hyperactivity Disorder (ADHD) and her lead and PCB dosed monkeys. ADHD in children is characterized by: • An inability to organize behaviour in a timely manner, or to plan and execute events in a timely fashion; • An inability to persist in the face of distracting or competing stimuli; • An inability to appropriately respond to consequence of past behaviour (learn from past mistakes) • A decreased ability to hold information "online" in order to plan for present or future behaviour; • Impulsiveness.(1) In Rice's studies, laboratory monkeys were given a test where they had to choose between two shapes, for example a cross over a square. Once they mastered this trial, the examiner then reversed the requirement so that the monkeys, in order to get a correct answer, had to change their response, choosing a square over a cross. Monkeys treated with lead, (with fairly low blood lead levels of 20 µg/dL) or monkeys dosed with low levels of PCBs had difficulty with this test. They could not seem to make the necessary switch; they would persist with their old response. Even monkeys with just 10 µg of lead/dL in their blood (our current level of concern in humans) had more difficulty than the untreated group. They too made more errors, and persevered in making the wrong response for longer periods; sometimes dozens of times. Intriguingly children diagnosed with Attention Deficit Hyperactivity Disorder are given a similar test called "The Wisconsin Card Sort Test." In this test, children must switch their response from looking at the pattern in cards to looking at the number, and then switch back again. Children are expected to make this response without any instructions from the examiner. Examiners have found that children with ADHD have difficulty with this test. Is it possible then, that children diagnosed with ADHD are really misdiagnosed lead or PCB toxic children? Like any good scientist, Rice is cautious in making the comparison. She says, "This is not to suggest that ADHD is caused exclusively by neurotoxic agents in the environment. Genetic and other [social] environmental factors play major roles in the etiology of the disability."(1) But she adds, "it seems reasonable to postulate that environmental neurotoxicants contribute to the prevalence of ADHD currently being identified."(1) Kelly O'Grady RN Lead Environmental Awareness and Detection L.E.A.D. Web: www.nrtco.net/~lead 219 Welland St., Pembroke, ON CANADA K8A 5Y5 Tel: 613-735-0717 Fax: 613-732-2859 Lead Environmental Awareness and Detection (l.e.a.d.) is a non-profit organization dedicated to identifying and preventing pediatric neurotoxicity in Canada. We promote the development of optimal social and physical environmental conditions for healthy productive children by recognizing the need for targeted early childhood blood lead screening of at-risk populations, and the need for raising public and professional awareness of this issue. Reference 1. Rice, D.C. Parallels between Attention Deficit Hyperactivity Disorder and behavioural deficits produced by neurotoxin exposure to monkeys. Environmental Health Perspectives 2000;108(Suppl 3):405-408.
Danger of multi-collinearity 18 May 2004
Barry Greenberg, Professor Florida International University Send letter to journal: Re: Danger of multi-collinearity greenb{at}fiu.edu Barry Greenberg	The article fails to provide information relating to the correlations among the set of independent variables. Absent information, it is impossible to determine whether what appears to be due to "television hours watched per day" is really due to "maternal self-esteem". If multi- collinearity exists, we are left in the dark, with no justification for keeping the television set off. As many prior comments have indicated, there is also a great deal of theoretical support for "maternal self- esteem" . Interpretations of the provided logistic coefficients (odds ratios) for "television hours watched per day" indicate that for every additional hour of television viewing per day, there is an increase in the likelihood of ADHD by anywhere from 3 to 15 percent (based on supplied confidence intervals). Given the manner in which the outcome variable was defined and the many other shortcomings in correlational research as identified by previous comments, this "increase in likelihood" hardly seems to justify keeping my granddaughter away from Elmo and Barney.
Do toddlers really have the attention span to watch televison for hours each day? 16 July 2004
Robert D. Cunningham, Neurodevelopmental Pediatrician Send letter to journal: Re: Do toddlers really have the attention span to watch televison for hours each day? rdanacunningham{at}hotmail.com Robert D. Cunningham	One problem with this study that stretched my credibility past the breaking point arose from the reliance on parental report of how much televison toddlers watch each day, without any ability to verify the parents' observations. According to parental report, 1 year olds watch an average of 2.2 hours of televison per day, while 3 year olds watch 3.6 hours per day. I would offer that the average toddler has the attention span to watch television for only minutes at a time before moving onto a different activity. Most of the material presented on television is so far above the cognitive abilitilies of toddlers that there is little reason for toddlers to pay attention to the program. (Toddlers frequently prefer commercials to actual program material.) Toddlers watching television for 2 to 3 hours per day? If this were actually occurring, it would strike me as a group of toddlers with above- average attention spans. But, really, the observation is just too incredible for me to believe. I suspect that what was actually being reported by the parents was how much time the toddler spent in a room, with a parent or sibling, or perhaps alone, with a television on, which is not the same as claiming that a toddler spent 2 or 3 hours each day watching television. This design flaw so terribly undercuts the study that no real conclusions can be made, just tenuous hypotheses.
Alternate design method 16 December 2004
Corbin VanDeWege, Medical Student Indiana University School of Medicine Send letter to journal: Re: Alternate design method emailcamm{at}yahoo.com Corbin VanDeWege	Perhaps a study using children whose parents voluntarily abstain from exposing their children to television/video media would provide a stronger support for etiology. An interesting subset would be families with a strong family history of ADHD.
Another Thought 12 January 2005
Kristina E. Clevenger, stay-at-home mother none Send letter to journal: Re: Another Thought Brian648{at}sctelcom.com Kristina E. Clevenger	I was going to share my experience with my three year old son. I quit my job nine months ago to be a stay-at-home mother. At the time I was seven months pregnant and each day passed my son seemed to demand more and more attention...play with me, hold me, read to me. Before the baby was born I would give in to his demands or I would take him places like story time, Burger King have other children come and play. After the baby was born I could not run all over the place so again I tried to keep up with his demands...nothing worked. On top of that my infant would only nap thirty minutes to forty-five minutes at a time...except when his older brother was gone (he was apparently sensing my stress). So eventually I put him in front of the TV anywhere from one to three hours per day. It gave me a little bit of sanity but only for the time he was actually watching a show. Then when I would turn it off things would be worse. I noticed that when I would read a book to him he lost interest after about fifteen minutes. At two he would have listened for forty-five minutes and entertain himself for longer periods of time. At the beginning of December I was given a book that recommended no television for pre-school children. Since then we have almost cut it off entirely. With that and the use of a kitchen timer, we have made significant progress. I now think that all of his demanding behavior has caused by jealousy and that me putting him in front of the television not only made him dependent on something else to entertain him but actually did shorten his attention span immensely. Only time will tell if we can undo the damage. Until then our TV will be off. One other thought...If the APA recommends no TV watching for children under two years of age (and some studies before age five) because of a possible link, why on earth are they not tellings these things and letting us make the decision!!!
Statistical problems and interpretation of probability 16 March 2005
John Sandberg, Asst. Prof Department of Sociology McGill University Send letter to journal: Re: Statistical problems and interpretation of probability john.sandberg{at}mcgill.ca John Sandberg	Hi; This is so long from the press attention to this article I wonder if anyone will ever read it, but there are at least two potentially serious flaws in the interpretation of the 'significant' coefficients related to TV watching in this piece. First, a 'significant' odds ratio tells us nothing about the change in probability of classification as potentially adhd associated w/tv watching. In this study, the big finding is an odds ratio of 1.09 (table 2) for tv hours per day. What does this mean? Well, you have to have the full regression table which is not in the article to make an exact measurement. I asked the PA for this after the article came out, but never got it. I can't blame him as the press and attention must have been horrific. However if we start w/the average tv viewing times and adhd prevalence given in the article to develop a 'baseline' probability and then use the logit results from table two to generate predicted probabilities for tv watching (which is a reasonable thing to do if the sample used in the paper is representative), the results are striking. What difference in the probability of being classified as adhd under the authors criteria associated with the difference between 0 and 8 hours of tv wathcing a day? .07. That's not a typo. The 95% upper bound on the estimate is only .126, the lower bound .022. THis is not substantively significant. Where is there a big difference in the predicted probability under this model? Out between 10 (where the discrete change in probability from 0 hours as calculated is .18) and 16 hours (where the discrete change in probability from 0 hours as calculated is .27)of TV/day. The question then is what 1 1/2 year olds watch, on average, 10-17 hours of television a day? The answer is almost none. Take a look at the distribution of TV hours given here in figure 1. It's very postively skewed. The analytic sample size is 1278. How many cases are above 10 hours/day are there? My very strong hunch is the entire result here is being driven by a few outliers at the end of the spectrum. Remember this is a 'stylistic' measure of time spent watching TV. Would you be more likely to think the few people who say their kids watch over 10 hours of tv have given accurate answers, or that somehow this is attributable to measurement error (misunderstanding the question, pulling the IW's leg, some kind of coding error)or not. This is a serious question if in fact these results that have gotten a lot of popular attention are driven by the outliers. Second, interpreting the significance from Wald tests of a maximum likelihood estimator (which is done here as TV hours is judged significant) though not necessarily invalid is much less powerfull than a likelihood ratio test vs. the constrained model leaving out TV hours. Though the result of the wald test is highly significant (.0014). It would be more appropriate here to present the LR test. There might be a different result. That's it for anyone still reading this thread!