Anxiety and classroom examination performance

Studied classroom examination performance over the course of an entire semester as a function of (a) students' anxiety level (high, medium, low); (b) difficulty level of exam question (more vs. less difficult); (c) type of exam question (rote memory vs. generalization); (d) sex of S; and (e) intelligence level of S. As expected, there was a strong main effect of anxiety and a significant interaction between anxiety and item difficulty. The latter was consistent with the Hull-Spence formulation of learning and performance--as the habit strength of incorrect responses increases (i.e., as the items become more difficult), the superiority of low over high anxious Ss also should increase. When item difficulty was held constant, anxiety did not interact with type of exam question.