The U.S. Food and Drug Administration (FDA) commissioned a Web survey about food labeling with two independent samples: an online consumer panel and shopping mall visitors. The purpose of the survey was to compare and contrast respondent behaviors and survey data between the two sample sources and to help us assess potential advantages and disadvantages of replacing online consumer panels for shopping mall visitors in data collections aimed at quantifying causal relationships rather than at developing population estimates. The 15-minute survey was conducted in 2006 concurrently with 765 online participants and 600 mall participants, using the same mode of collection and the same questionnaire. Participants were randomly assigned to view different food labels and to answer questions about the label and the food they viewed. Among other things, they were also asked about their awareness of several food ingredients, including a fictitious one, and their height and weight as well as their self-perceived weight status. This presentation will first illustrate the use and purposes of Web surveys at the FDA. The presentation will then report some of our observations regarding instrument design, panel characteristics and composition, sample disposition, respondent behaviors, collected data, and inferences from the data. Overall, our observations suggest that (1) the causal relationships produced by the two samples were qualitatively comparable, (2) the online sample appeared to be less vulnerable to social desirability bias, satisficing, and privacy concerns, and (3) participants in the online sample appeared to have put in more effort in performing required tasks.