skip to main content skip to footer

Effects of Coaching on the Validity of the SAT: A Simulation Study SAT

Author(s):
Baydar, Nazli
Publication Year:
1990
Report Number:
RR-90-04
Source:
ETS Research Report
Document Type:
Report
Page Count:
68
Subject/Key Words:
Coaching, SAT, Test Preparation, Test Validity

Abstract

This report presents the results of a simulation study of probable effects of trends in student coaching on the 10-year decline in the predictive validity of the SAT for the college first year performance (Morgan, 1989). Coaching could lead to three different types of score gains: Gains associated with test taking familiarization; with improvement in reasoning and comprehension skills; and with learning test-specific tricks and strategies. The simulations account for the latter type of gains only. A set of equations are shown that demonstrate the relationship between the score gains due to learning test-specific tricks and the predictive validity of the SAT for the first year grade point average under simplifying assumptions. Next, simulations are presented that account for complex associations between various background characteristics of the students, their choices for being coached and their freshman year performance. Data from four colleges from the College Board's Validity Study Service database were used. These four colleges include a highly selective, a moderately selective, a less selective, and a religious-affiliated college. Coaching behavior was simulated at three levels of probabilities of being coached (5, 15 and 25%) and at four levels of effectiveness of coaching. Under the assumptions that the probabilities of being coached increased by 1% every year and that coaching effects remained constant, it is concluded that the trends in coaching probably did not account for more than a 1% decline in the validity of the SAT in ten years in highly selective and less selective colleges alike. (69pp.)

Read More