| This study sought to determine if expectations measured by the "Perceptions, Expectations, Emotions, and Knowledge about college" (PEEK; Weinstein, C. E.; Palmer, D. R.; & Hanson, G. R., 1995) and fourteen Additional Questions improved the prediction of college student academic performance or retention over other predictors such as high school percentile rank and ACT and math placement scores. It also examined whether the actual experiences of engineering students were consistent with their expectations.; Expectations did not improve the prediction of academic performance. The goodness-of-fit X2 from the logistic regression model of retention was significant (p < .0001), but the Hosmer-Lemeshow statistic suggested the overall model did not predict retention accurately. A trigonometry placement exam score was the only significant predictor variable (p < .01). Significant MANOVA interactions were found on the PEEK Academic scale for Time x Gender (p < .05) and for Time x Gender x Ethnicity (p < .05). |
