Re: Babies Affected by Dads' X-Ray Exams, > Affected by Dads

["John Moulder" <jmoulder@post.its.mcw.edu>]

> >.. I believe the real issue has to do 
> >with WHY the Journal of Epidemiology published the article in the 
> >first place. Didn't they review it for scientific/technical content? 
> >One wonders. 
>  
> As one who has reviewed papers for biochemistry journals of both high 
> and low quality (I do not know where J. Epidemiology rates on this 
> scale), please let me answer this. 
>  
> Negative data (i.e., no significant difference) is a valid 
> result. Presumably, as you state, the review process checked the 
> validity of the procedures used to come to this finding

I agree in general that negative data should be often be published.

But:

FIRST:  The authors do not present this as "negative" data in either the paper 
or the press release.  The present a biologically and statistically 
insignificant association as support for a hypothesis.

The press release was titled:  "Babies Affected by Dads' X-Ray Exams"

The press release starts:
"NEW YORK (Reuters) -- Fathers whose testes are exposed to medical x-rays 
close to the time their child is conceived may be more likely to have a baby 
whose growth is adversely affected, results of a new study suggest..." 


SECOND:  There are many kinds of negative studies, and their usefulness to 
world varies:
1)  A negative study that fails to confirm/replicate a previous positive study 
-- These are very important, although often hard to publish
2)  A negative study that fails to confirm/support a previously untested 
hypothesis  --  The importance of these depend on the importance of the 
hypothesis
3)  A negative study that fails to find evidence for an effect that has never 
been reported to exist, or hypothesized to exist -- I fail to see why we 
should attach any importance to these.

I see this birth weight study as being group 3.  No one has ever reported that 
paternal irradiation adversely affects birth weight; and no one, to my 
knowledge, has ever seriously hypothesized that it does.  Thus to show that 
paternal irradiation has no biologically significant, or statistically 
significant, effect on birth weight isn't very useful.


John Moulder (jmoulder@its.mcw.edu)
Associate Editor
Radiation Research