Comparing the results of laboratory and World-Wide Web samples on the determinants of female attractiveness

Correlating the means from the Internet samples to the laboratory samples on important interactions yielded values near 1.0, indicating that the two data sets are driven by the same psychological variables.

Abstract

The Internet provides a new method to obtain subjects. Arguments supporting the use of the Internet as an experimental medium include the ease of data collection, limited resources needed, and ability to gain a sample more representative of the population. However, it has not been demonstrated that the Internet provides the conditions necessary to yield valid data. This paper compares results from experiments conducted over the Internet and in a laboratory. Correlating the means from the Internet samples to the laboratory samples on important interactions yielded values near 1.0, indicating that the two data sets are driven by the same psychological variables. Further evidence of the validity of these Web studies comes from regression analyses. However, more studies of this sort are needed to further identify what types of experiments may be validly conducted over the Internet.