With the rise in popularity of computer assisted survey instrumentation it is important to determine whether respondents are influenced by the medium of data collection. Two hundred forty-five staff members at a large health-care related organization, diverse across education, race/ethnicity and gender, were surveyed for the present study.  Participants were asked to complete two identical satisfaction questionnaires, administered once on paper and once on-line. The Rasch probabilistic model was used to assess the collected information. Analysis suggested the existence of several statistically significant differences between the on-line and on-paper surveys. Additional study will be necessary, but the assumption of online and paper equivalence should likely not be taken for granted.