A Multidimensional Perspective of College Readiness: Relating Student and School Characteristics to Performance on the ACT

ACT Research Report Series 2015 (6)Daniel M. McNeishUniversity of MarylandJustine Radunzel, PhDACTEdgar Sanchez, PhDACTA Multidimensional Perspective of College Readiness: Relating Student and School Characteristics to Performance on the ACT® © 2015 by ACT, Inc. All rights reserved. ACT®, ACT Explore®, and ACT Plan® are registered trademarks of ACT, Inc. 4653Daniel M. McNeish is a PhD candidate in the Measurement, Statistics, and Evaluation Program at the University of Maryland College Park and interned at ACT during the summer of 2014.Justine Radunzel, a principal research scientist in the Statistical and Applied Research Department at ACT, works on postsecondary outcomes research and validity evidence for the ACT® test. Edgar Sanchez, a research scientist in the Statistical and Applied Research Department at ACT, works on predictive modeling of student educational outcomes such as enrollment, persistence, and graduation.AcknowledgmentsThe authors thank Jeff Schiel and Mike Valiga for their input and handling of the online questionnaire, and Jeff Allen, Krista Mattern, Jim Sconing, Richard Sawyer, and Karen Zimmerman for their helpful comments and suggestions on earlier drafts of this report. ContentsAbstract ............................................................................... ivIntroduction ............................................................................ 1Data ................................................................................... 3Data Collection ....................................................................... 3Instruments .......................................................................... 4High School Characteristics .......................................................... 6Method ................................................................................ 7Weighting ........................................................................... 7Missing Data ......................................................................... 8Principal Components Analysis of the Online Questionnaire ............................ 8Clustered Nature of the Data ......................................................... 9Modeling Technique ................................................................ 10ACT Score Models .................................................................. 11Comparing Differences in ACT Scores among Student Demographic Groups ......... 14Relating Noncognitive Variables and HSGPA ........................................ 14Results ............................................................................... 15Descriptive Statistics ............................................................... 15ACT Score Models ................................................................. 17Unadjusted and Adjusted Mean Differences by Student Demographic Characteristics ... 28HSGPA Predicted from Noncognitive Student Characteristics ........................ 30Discussion ........................................................................... 32Academic Factors .................................................................. 32School Characteristics .............................................................. 33Noncognitive Characteristics ........................................................ 33Student Demographics ............................................................. 35Application for Alternative Statistical Techniques ..................................... 36Limitations and Future Research .................................................... 36Implications ........................................................................ 37References .......................................................................... 38Appendix A .......................................................................... 45Appendix B .......................................................................... 46Model-Based Methods ............................................................. 46Design-Based Methods ............................................................ 46Appendix C .......................................................................... 48 AbstractThis study examined the contributions of students’ noncognitive characteristics toward explaining performance on the ACT® test, over and above traditional predictors such as high school grade point average (HSGPA), coursework taken, and school characteristics. The sample consisted of 6,440 high school seniors from 4,541 schools who took the ACT in the fall of 2012 and completed an online questionnaire about their high school experience, study and work habits, parental involvement, educational and occupational plans and goals, and college courses taken and/or credits earned in high school. Twelve percent of the total sample responded and met the study inclusion criteria. A blockwise regression model with cluster-robust standard errors was used to assess the relationships be