TY - JOUR
AU - Watley, Donivan J.
AB - Analyzing College Effects: Correlation vs. Regression CHARLES E. WERTS Educational Testing Service and DONIVAN J. WATLEY National Merit Scholarship Corporation A commonly used procedure in studying college effects involves an input-output model in which student input is controlled by using regression analysis to compute an "expected" output (e.g., Astin, 1963, 1964; Thistlethwaite & Wheeler, 1966). The corre­ lation of a school environment variable with the residual output (i.e. actual minus "expected" output) is interpreted as a measure of the school's influence on the output. Although sometimes labeled a partial correlation, it is more accurately described as the part correlation (McNemar, 1962, p. 167) of the school with the output variable when the influence of the input variable has been re­ moved from the output. A potentially serious interpretational problem is that part cor­ relations may severely underestimate the magnitude of the true college effect. This possibility was noted previously by Richards (1966): "suppose that a real effect of small colleges is to encourage students to develop warm personal relationships with the faculty, and that the socio-economic status of college students has no inherent relationship to their tendency to develop warm relations with the faculty. Suppose 1. This study is 
TI - Analyzing College Effects: Correlation vs. Regression1: 
JF - American Educational Research Journal
DO - 10.3102/00028312005004585
DA - 2016-06-23
UR - https://www.deepdyve.com/lp/sage/analyzing-college-effects-correlation-vs-regression1-cHlmVjNUP8
SP - 585
EP - 598
VL - 5
IS - 4
DP - DeepDyve
ER -