An assessment of the generalizability of internet surveys

The Internet has become an increasingly popular form of data collection bec ause it permits complex questionnaires to be administered more quickly, fle xibly, and inexpensively than conventional survey methods. However, the Int ernet is restricted to individuals with access to computer networks. Thus, causal inferences to the general population from analyses of Internet sampl es necessarily test on two untested assumptions: (a) that the decision-maki ng processes of Internet users are similar to those used by the general pop ulation, and (b) that representative samples of Internet users can be drawn . The authors provide mixed support for these assumptions. They find that c urrent Internet sampling techniques only permit the generation of diverse, not representative, samples. However, comparing samples drawn simultaneousl y using the Internet and probabilistic telephone methods, the authors demon strate that the psychological mechanisms underlying common political decisi ons do not differ between Internet users and the population. They discuss t he implications of these findings for future survey research.