Three Ordered Sets of Factors and Their Relationship to ACT Scores

Thomas J. Russo, Utah State University

Abstract

There has been in recent years a marked decline in college entrance examination test scores. Declines have been documented both on the Scholastic Aptitude Test (SAT) and the American College Testing Program (ACT). In turn there has been an increasing interest concerning the test score declines as well as possible sources of influence on test scores. These sources or factors seem to be functions of three main "contexts." These contexts are: (a) school-related factors, (b) student-related factors and (c) family-structure related factors. It was of interest to attempt to explain the relative association of each variable to composite ACT scores and of each "context" to composite ACT scores. In turn, it was of interest to attempt to assess the extent to which combinations of two, or all three sets of variables aid in explaining the variance on composite ACT scores.

The sample population consisted of entering college freshmen at Utah State University for the fall quarter of 1976. All S's were from one of the six major feeder high schools to Utah State University. Information was gathered through the use of a questionnaire distributed at freshmen registration and by mail. Other sources of information included school principals and official student ACT transcripts and school records. Using composite ACT scores as the dependent variable both stepwise and hierarchical multiple regression analyses were done.

As a result of stepwise multiple regression using all ten factors, it was found that academic course taking had the greatest partial regression coefficient. Next in predictive power was the level of educational aspiration of the student. Size of school entered the prediction equation on Step 3 and was negatively associated with composite ACT scores.

Student-related factors helped to explain 38 percent of the total variance on composite ACT scores and as such comprises the most significant "context" of association. More modest support has been demonstrated for school-related factors. Family-structure factors do not appear significantly related to composite ACT scores. Recommendations were made for a larger sample size from a more diverse geographic region.