EFFECTS OF STIMULUS COMPLEXITY IN COMPUTER GRAPHICS AND PREDICTION OF GRAPHED CONCEPTS IN STATISTICS INSTRUCTION

omputer administered instruction makes possible sophisticated graphic display modes and novel student-graphic interaction. This study investigated the effects of varying the stimulus complexity in computer graphics and students' prediction of graphed concepts on achievement and lesson completion time.;Fifty-two subjects, enrolled in either an introductory or intermediate applied statistics course, viewed one of four CAI lessons on the fundamentals of linear regression: (1) static display-no prediction; (2) static display-prediction; (3) evolving display-no prediction; and (4) evolving display-prediction. (Note: Prediction treatments were 15 display screens longer than No Prediction treatments.) Following the lesson students took a 24 item multiple-choice paper-and-pencil posttest. All items contained graphics and measured knowledge of graphed regression concepts.;The analytical strategy consisted of MANOVA followed by ordered regressions for both dependent variables. Criterion scores were adjusted for the students' course affiliation. The significance level for all tests was.05.;Regarding achievement, students who viewed the evolving graphics scored one point higher than students who viewed the static graphic. Students who predicted graphed concepts scored three points higher than students who did not predict. Only the later result, which accounts for 12% of the variance in achievement, is statistically significant. The effect of the level of the statistics course and all interactions are also non significant.;The design was a 2 by 2 factorial. The two levels of graphic display were: (1) static display; and (2) evolving display. The two levels of prediction were: (1) students predicted which of two statements graphically described a concept prior to viewing that concept graphed; and (2) students did not predict.;Regarding lesson completion time, students who viewed a lesson enriched with evolving graphics finished three and one-half minutes sooner than students who viewed lessons augmented with the static graphic. Likewise, students who predicted graphed concepts completed the lessons four minutes faster than students who did not predict. However, these results, the effect of the students' course affiliation and all interactions are statistically non significant.;The partial correlation between achievement and lesson time with treatment effects controlled is...