One of the frequently taken for granted effects of using self-administered mail and web surveys is that they produce primacy effects, i.e., a greater likelihood of choosing the first offered categories. However, considerable research has made it clear that primacy does not occur consistently. In this paper we report experimental results from a web survey that was designed to isolate conditions in which primacy effects are expected to occur as a result of satisfying. Building upon previous theoretical work that has identified three regulating factors thought to influence the occurrence of satisfying—task difficulty, respondent ability to respond, and respondent motivation for responding—several specific tests were developed. A series of questions were constructed varying assorted characteristics of question stems and response options in ways hypothesized to affect primacy. The experimental manipulations were applied to four panels of a web survey completed by a random sample of Washington State University students surveyed in the winter of 2006-7, for which a final response rate of about 60% is expected, amounting to approximately 300 completed surveys per panel. This study builds upon previous surveys conducted with the some population and general topic that have previously demonstrated situations in which primacy did and did not occur. The overarching goal of this research is to contribute to our understanding of the circumstances under which primacy occurs in order to help researchers identify when primacy is likely to be a problem in web surveys.