From: Robin Hanson (hanson@econ.berkeley.edu)
Date: Tue Sep 01 1998 - 12:44:02 MDT
Curt Adams writes:
>>>>http://homenet.andrew.cmu.edu/progress/research.html
>>>I can't find any good indications of direction of causality. Indeed, the
>>>path analysis in Figure 1 explicitly assumes that internet usage
>>>affects followup psychological status and not the other way around.
>>
>>Huh? Isn't there an arrow from the the lower left social/psych box
>>straight to the internet use box?
>
>That's from the pre-trial psychological conditions. Changes in
>psychological state after the screening test don't affect internet
>use, by their model.
They are trying to use the timing of variables to infer causality;
things that are measured earlier likely cause, and are not caused by,
things that are measured later. So to disambiguate cause, they measure
soc/psych both before and after internet use. Of course there's always
the possiblity of unmeasured variables causing everything, such as
unmeasured "during" soc/psych. But it seems to me they have done the
reasonable thing with the data the have.
>The details of the study were better than the summary implied. I'd like
>to see the results without the internet usage hours truncated. If there
>is a causal relationship, extremely heavy users should be very messed up.
>One hypothesis would be internet addiction, a la gambling/alcohol/etc
>addiction. In this case, light and moderate users should be fine and
>the heavy users will be the ones with the problems.
Huh? They used log of internet hours, without truncation. They truncated
social circle and minutes with family, but undoing that wouldn't seem to
help your addiction scenario. It does seem they should have logged those
variables without truncation though.
Robin Hanson
hanson@econ.berkeley.edu http://hanson.berkeley.edu/
RWJF Health Policy Scholar, Sch. of Public Health 510-643-1884
140 Warren Hall, UC Berkeley, CA 94720-7360 FAX: 510-643-8614
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:49:32 MST