A 3×3×3 factorial design, N=135, is used to evaluate the effects of three conditions of feedback (zero, low extrinsic, and high extrinsic), three conditions of response pattern restriction (no restriction dictated, restriction to three patterns, restriction to one pattern), and three levels of repetitive-figure manipulation diversity determined by scores on Torrance Tests of Creative Thinking, Form A, Activity III: repetitive parallel lines. The experimental task was to turn nine circularly arranged knobs repetitively in a white visual, broadband audio environment over three 6-minute periods separated by rest pauses. O'Brien (1976) found significant differences for measures of response variability including pattern changing activity, pattern shifting, error factors, and other measures. An extended ad hoc analysis of that exploratory study is accomplished in which a comparison of percent common patterns emitted evidences that Ss who score differently on the pretest perform with different degrees of uniqueness when they are restricted to repetition of previous patterns, but not when they are free to vary types of patterns emitted. Also, comparison of initial trials indicates that high pretest scorers are more likely to begin the task with rarer patterns, low scorers use more common patterns, and intermediate scorers fall in between.