PSYC 204 Lecture Notes - Lecture 22: Lincoln Near-Earth Asteroid Research
Document Summary
Y y" = error of estimate & sum to = 0. Y = deviation of score from mean. Variance of observed always higher than variance of predicted or variance of error. Best prediction= when data actually on regression line so variance & standard deviation of error of prediction = 0. Lower r2 proportion of variance that can be. Higher r2 proportion of variance that can be. Standard error = 0 when regression line fit perfectly but. Proportion of variance - accounted for by linear regression. *amount of variance that can be predicted by regression line. How much change/variance in y can be predicted just by x. 65% of total variance = accounted for by linear regression between gpa & cegep score. = 1: if lower r2 = cannot predict well variance of y by just looking at x, if higher r2 = can predict well variance of y by just looking at x.
