This study describes the design, implementation, and analysis of three related surveys launched to investigate E-Learning in undergraduate humanities classes at a public technological research university. Examined are the independent variables of learning style, general expectation, and interaction. The surveys were administered in the summer of 2006 (Survey 1, n = 108 [41.53% return], Survey 2, n = 89 [34.23% return], Survey 3, n = 62 [23.85% return] and fall of 2006 (Survey 1, n = 68 [50.75% return], Survey 2, n = 58 [43.28% return], Survey 3, n = 58 [52.25% return]. Modest to high correlations and statistically significant relationships identified among survey questions designed to explore the three independent variables led the researchers to create constructed scales based on those variables. When the dependent variable of expected course grade was used to hypothesize a relationship based on performance, the model was weak (R[superscript 2] = 0.17, "F"(5, 51) = 2.092, p = 0.1). However, when a new dependent variable was constructed, based on Zhang and Sternberg's (2005, 2006) Type II learning styles--those expressing a preference for structure, cognitive simplicity, conformity--the model was strengthened (R[superscript 2] = 0.227, "F"(4, 52) = 3.816, p = 0.01). It appears that learning style, general expectations, and interaction are more related to the instructor's online presence than to standard criterion variables such as course grade. Because little is known about E-Learning of humanities classes, examining complex variable relationships allows researchers to explore a new educational field. (Contains 6 tables and 1 figure.)