This study investigates factors affecting the degree of novice physics students application of the naive impetus theory. Six hundred and fourteen first-year university engineering physics students answered the Force Concept Inventory as a pre-test for their calculus-based course. We examined the degree to which students consistently applied the naive impetus theory across different items. We used a 2-way repeated measures ANOVA and linear regression to analyze data coded from "incorrect" student responses. It was found that there were statistically significant main effects for item familiarity and item requirement for explanation vs. prediction on the measured degree of impetus theory application. Student course grades had no significant effect on impetus theory application. When faced with items that were unfamiliar and predictive, students appeared to rely on non-theoretical, knowledge-in-pieces reasoning. Reasoning characteristic of naive theories was more frequently applied when students were completing familiar problem tasks that required explanation. When considering all the above factors simultaneously, we found that the degree of naive impetus theory application by students is attributable to variables in the following order: familiarity, prediction, and explanation.