College Board Study: Predictive Validity of the SAT & the Relationship with IUPUI Admission Standards Gary Pike, Ph.D., Executive Director, Information.

College Board Study:Predictive Validity of the SAT & the Relationship with IUPUI Admission Standards

Gary Pike, Ph.D., Executive Director, Information Management &

Institutional Research

Chris J. FoleyDirector of Undergraduate Admissions

October 2013

Guiding Principles for Setting Standards

• Consider the impact on the size and the composition of the incoming class;

• Set standards that select students who have a reasonable chance of success;

• Retain academic performance in high school as the primary element;

• Use standardized test scores only in collaboration with another predictor of success (e.g., academic performance in high school);– Supported by research on use of test scores in predicting student

success when used with other factors;• Construct guidelines that would be easily communicated to

audiences to ensure transparency of decisions.– Consider both internal and external audiences

Current Factors for Freshman Admission Decisions(from most to least important)

• Curricular Rigor– Students with Academic Honors are admitted regardless of

GPA or test score• Cumulative High School GPA– We take the best of either the weighted or unweighted GPA– Core 40 students with a 3.5 HS GPA or higher are admitted

regardless of test score• SAT or ACT score– We “super-score” both the SAT and ACT– We take the best of either the SAT or ACT

Admission Standards for Freshman Applicants with Core 40 Diploma Only

Measures Related to First-Year GPA

• Strong Predictors– High School GPA– SAT (Critical Reading + Math)

• Moderate Predictors– Number of AP Exams Taken– Number of Honors/AP Courses Taken

• Weak Predictors– Years of math in High School– Years of English in High School

Predictive Accuracy of the Model

Student Performance N Percent

Better than predicted 744 34.6%

As well as predicted 1112 51.6%

Lower than predicted 297 13.8%

TOTAL 2153

Predictive Accuracy for GroupsDemographic Group Predicted Actual Adjusted Difference

Gender Males 2.52 2.57 2.49 -0.03

Females 2.61 2.76 2.62 0.01

Race/Ethnic Asian 2.63 2.71 2.70 0.07

Afr. Amer. 2.31 2.25 2.11 -0.20

Hispanic 2.46 2.70 2.61 0.15

White 2.62 2.75 2.63 0.01

Language English 2.58 2.67 2.56 -0.02

English + 2.52 2.81 2.73 0.21

Enrollment Full-time 2.58 2.69 2.58 0.00

Part-time 2.48 2.59 2.47 -0.01