An empirical assessment of data collection using the internet

Identical questionnaire items were used to gather data from 2 samples of employees. One sample (n = 50) responded to a survey implemented on the World Wide Web. Another sample (n = 181) filled out a paper version of the survey. Analyses of the 2 data sets supported an exploration of the viability of World Wide Web data collection. The World Wide Web data had fewer missing values than the paper and pencil data. A covariance analysis simultaneously conducted in both samples indicated similar covariance structures among the tested variables. The costs and benefits of using access controls to improve sampling are discussed. Four applications that do not require such access controls are discussed.