IP: Clarification concerning Internet Paradox Paper

[Dave Farber <farber () cis upenn edu>]

Mike is a grad student who was involved in the study and will base his PhD on parts of it. Dave


Reply-To: <patterso () interval com>
From: "Mike Patterson" <patterso () interval com>
To: <farber () cis upenn edu>


Hi Dave,


I spoke to you before about writing a response to misconceptions about the
Internet Paradox study.


First I present 8 points with more detailed explanations below.


1] When you ask heavier users (who became more depressed) about the
Internet's effect on
their lives, they often say it has improved their lives.
2] The change in depression and loneliness was small (but significant).
3] Our sample demographics are generally representative of the U.S.
population
4] Our results are based on home Internet use, it is unclear what the
effects would
have been in a classroom or other settings
5] Our primary results are based on statistically significant comparisons,
not solely
qualitative research.
6] We compared peoples differences in amount of Internet use, rather than
the simple presence
or lack of Internet access.
7] People with higher depression scores were not more likely to use the
Internet initially, but
those who used the Internet more, reported more depression.
8] The results are not limited to teenagers.


----more detail on points----
1]
We were initially surprised and skeptical of our findings because they were
contrary to what we expected.  We found that people with heavier use, who
also became more depressed, were often the most enthusiastic about the
Internet and offered colorful examples of social events and benefits of
their Internet use.


We did not directly ask "Does Internet use make you depressed?"  It is
difficult for people to accurately answer this type of self-evaluation
question.  We use several psychological scales to capture depression at
two time points and associated this with their online usage.  Note that
the initially depressed people were not more likely than less depressed
people to use the Internet more.  We also controlled for the effects of
race, income, and gender.


A recent critique of our results is based on
the methodology of asking users questions like "Does the Internet expand
interactions with friends and family?" (www.activemedia.com).  Our findings
and observations are not counter to people self-reporting positive effects.


2]
It is also important to note that the changes in depression and loneliness
were small, but significant.  Clinical psychologists would prefer that we
use the term "dsythmia" which connotes depression to a lesser degree than
the more serious clinical depression.  It's much more like feeling in a funk
than feeling suicidal.


3]
One advantage of our research project is that we loaned computers and gave
Internet access to families that were not already connected in 1994.  This
study was an attempt to look into the future when technological and economic
barriers to access would be lowered.  Our sample is generally representative
of the U.S. population in terms of demographics, although the household
income is slightly higher.  This sample is however more representative
of the U.S. population than a sample of only Internet users, who tend to be
white males, with higher income, and education.  We do not know how well the
effects would generalize to long-time users, who may have developed more
meaningful Internet connections.


4]
Our results are based on home Internet use.  It is unclear what the effects
would be
in a classroom or other settings.  It could be that Internet in classrooms
actually
have positive social effects.  It would be inappropriate to use these
results to argue
against getting Internet access for inner-city schools.


5]
Our primary results are based on statistically significant comparisons, not
solely qualitative research.  Some who only read the NY Times article
believed our results were only qualitative.


6]
Using a control group would have allowed us to determine whether or not
people given Internet access have different effects from those who were not
given access.  It would have been desirable to test that question.  Having a
control group has certain advantages, particularly when comparing blocks of
people.  A control group would allow us to rule out an extraneous event
(e.g. Clinton caught lying) which could
have lead people in both groups to become more depressed, not just those
with Internet access.
But having a control group would not explain why increases in Internet use
are related to changes in depression, loneliness, family communication, and
size of social circle among those with Internet access.


7]
People with higher depression scores were not more likely to use the
Internet initially, but those who used the Internet more, reported more
depression.  Thus Internet use preceded depression.


8]
Teenagers tended to both use the Internet more and become more depressed.
However, we controlled for whether or not they were a teen.  Internet use
was still related to depression both among teens and adults.


I am very interested in the questions that you and others would like
answered in future research, because I am currently developing a related
dissertation topic.


Thank you,
Mike Patterson
mp72 () andrew cmu edu


Reference: Robert Kraut, R.; Patterson, M.;  Lundmark,V.; Kiesler, S.
Mukophadhyay, T. & Scherlis, W. (1998) Internet paradox: A social technology
that reduces social involvement and psychological well-being? American
Psychologist, 53 (11).
Draft Online: http://homenet.andrew.cmu.edu/progress/HN.impact.10.htm