The temporal characteristics of evaluation anxiety are not well-defined by previous research. We examined the effects of length of the pre-evaluation interval (3, 6, or 12 minutes) and stage of the pre-evaluation interval at which evaluation anxiety was measured (start, middle, or end) on evaluation anxiety while participants performed an activity that was the focus of the impending evaluation. Participants wrote their opinion on a controversial social issue while anticipating a subject matter expert's judgment of their social maturity, and evaluation anxiety was measured by a battery of state anxiety measures. Higher levels of evaluation anxiety were detected on the Worry-Emotionality Questionnaire (WEQ) Worry subscale at the end of the pre-evaluation interval than at earlier stages, regardless of interval length, although individual difference variables exerted an important influence. Individuals with high trait self-presentation concerns experienced particularly high state anxiety (State-Trait Anxiety Inventory) at the start of the longest pre-evaluation interval. Low self-efficacy individuals showed a U-shaped pattern across pre-evaluation stages on both the WEQ Worry and Emotionality subscales, while high self-efficacy participants showed either no change (worry) or an inverted-U pattern (emotionality). Implications for the experimental measurement of evaluation anxiety were discussed.