Internet Paradox

"Internet Paradox" (American Psychologist, September 1998, pages 1017-1031) is an interesting "experiment". They are testing to find if internet involvement is likely to contribute to social problems such as loneliness, and depression. But I'm not convinced that it's an experiment.

The sample that they study is of 93 families from 8 diverse neighborhoods in Pittsburgh, Pennsylvania. They use controlled variables like education, race, and other demographic variables, for the researchers found no differences between the Internet users and the nonusers memberships in religious, leisure, and community organizations or the time they spent with family or friends.

These variables represent the population if you were to be summing up the data to reflect the Pennsylvania region. However if you were trying to get an idea if the internet generally affects the world, or nation, I do not think this experiment fully covers a full sample, because picking all of the samples from Pittsburgh, Pennsylvania makes the data almost dependent.

This experiment did not meet a good experimental design, in that it had treatments that included a control (on two or more levels), but it dealt with replication (many people from one area), randomization was met, as they picked the people randomly from one area, and blindness was not met, because the people knew what they were being tested for, and so did those testing them. So I do not think this experiment met these guidelines, and can't really be called an experiment after all.