Higher Education Brief: Classic Learning Test Concordance with the SAT

UPDATED: May 1, 2024

College Board has received questions from our higher education members  related to the Classic Learning Test (CLT). We’ve put together this fact sheet to address these inquiries, including those related to the concordance published by CLT.

Executive summary

The College Board recognizes the role that the Classic Learning Test and other assessments can play in a diverse educational landscape.  

In April 2023, Classic Learning Initiatives, LLC, published a paper that includes a concordance between the Classic Learning Test and the SAT—The Concordance Relationship Between the Classic Learning Test (CLT) and the Scholastic Aptitude Test (SAT). While the paper makes statements about comparability between the SAT and CLT, the study does not meet industry standards (as developed jointly by the American Educational Research Association, National Council on Measurement in Education, and the American Psychological Association) for an acceptable concordance. Furthermore, new external research shows that the CLT is not predictive of college performance or success and has no relationship to student outcomes.  

College Board was not involved in the study. Without addressing several methodological and comparability issues, we cannot validate the interpretation and use of concorded CLT-to-SAT scores for high-stakes decisions, like admissions and scholarship awards, based on the published study.  

Iowa Board of Regents findings  

In April 2024, the Iowa Board of Regents staff and the state’s public universities  recommended to not allow the use of the Classic Learning Test in the Regent Admissions Index at Iowa's public universities.  

This recommendation was made based on the following conclusions from the report (see Appendix 2 for more information):  

  • The CLT is not predictive of college performance or success.
  • There is no evidence of a relationship between CLT and student outcomes (i.e., enrollment, graduation, retention).
  • There is no evidence that CLT can be used for course placement.
  • The concordance tables published by CLT  were developed without a representative sample of students, have not been sufficiently tested, and therefore should not be used.

Why valid and accurate test concordance matters

When scores from different tests are being used to consider students for scholarships or college admission, the scores must be able to be reliably compared to each other to ensure that students taking different assessments aren't advantaged or disadvantaged. A well-formulated concordance links two different tests and allows scores on different exams to be used in comparable way. College Board and ACT have completed concordance studies together for years using technically sound best practices.

The concordance study published by CLT does not meet psychometric and industry standards on multiple dimensions.  

Core methodological issues with the published CLT study

The study does not use a representative sample of test takers.  

  • The study uses 4 different samples totaling about 5,000 students. The report does not cite the demographics of the participants or controls used to ensure representation of all test takers. For comparison, the most recent concordance between the ACT and SAT included 589,753 students, and the sample was nationally representative.  
  • In addition, the study did not control or minimize the time between when the CLT and the SAT were taken—students could have had multiple years of learning between the 2 tests, which negates the valid comparability of scores.  

The SAT score data were not verified by College Board.  

  • The study indicates students largely self-reported SAT scores, and over 20% of scores used in the study were different than the scores colleges had on record.  
  • For the most recent concordance study between ACT and SAT, ACT and College Board provided official, college reportable scores obtained in official, secure administrations of the exams.  

Comparability between the SAT and CLT  

A core requirement for a valid and reliable concordance between two tests is that they measure similar content at similar levels of rigor. Analysis of available information about the development and design of the CLT raises questions about the comparability between the two assessments.  

  • The SAT is an achievement test that measures the knowledge and skills students learn in high school that are needed for college and career success. The CLT is aligned to a “classical curriculum.” When 2 tests are constructed around different curricula and standards, a rigorous joint study must be performed to determine if there can be concordance.
  • The SAT is a proven, valid predictor of college performance, based on years of published and accessible research and data. CLT has not published evidence of validity or predictiveness of college performance.  

 

Our preliminary analysis of the skills and knowledge tested on the CLT indicates that the CLT and SAT do not test math at the same grade level. In reviewing a published CLT practice test, we found that 25% of questions were below high school grade level. For example, statistical concepts are not tested on the CLT. As a result, scoring very well on the CLT may not necessarily indicate an equally high score on the SAT, which is aligned to high school grade levels represented in state standards. 

Appendix: Comparison of concordance study methodologies

 ACT-SAT (2018) CLT (April 2023) 
Participants NCAA, ACT, SAT CLT 
Sample size 589,753 4,375 for total score; 1,550 for section scores. 
Testing population Targeted to all high school students; SAT cohort usually of 1.5M–2.2M students Targeted to homeschooled students and students enrolled in private schools and religious charter schools. Between 2016 and 2023, CLT reported 32,615 scores to 24,362 unique students. 
Sample representativeness Representative of SAT/ACT test takers, which is generally representative of high schoolers nationally. No demographics reported.  

Official scores  

  

ACT and College Board provided official, college reportable scores that were obtained in official, secure administrations of the exams. Official scores were not used and most students self-reported their SAT score. For students with scores at a partner college, >20% did not have SAT scores that matched the students’ self-reported scores.  
Shared content map/standards   Content of both tests selected to assess the most commonly tested HS standards that are predictive of success in first-year college coursework. No alignment published. Mission of organization is focused on testing classical curriculum.  
Subgroup evaluations Subgroup concordance evaluated and matches total group. None reported. 
Time between assessments and testing order Sample controlled to minimize for time between assessments and to account for the order assessments were taken. No controls cited. 
Joint guidance on use of tables CB-ACT publication includes lists of potential uses and limitations.  No joint statement or guidance on when appropriate to use.  

Appendix 2: Evidence from Iowa Board of Regents Report

  1. Finding: The CLT is not predictive of college performance or success. 
    “In examining the research on CLT, the team located no evidence to support the predictive efficacy of the CLT, particularly for students at institutions like ISU, SUI and UNI. The sample upon which the psychometric properties were derived included only high school students who were either home schooled or attended private or charter high schools, thus is not representative of the student bodies of Iowa’s public universities.”
  2. Finding: There is no evidence of a relationship between CLT and student outcomes (i.e., enrollment, graduation, retention).
    “To date, the team has identified no peer-reviewed journal articles addressing the relationship to student outcomes, and the CLT technical guide does not include any study of the relationship between the CLT or its subscales and student outcomes. The relative youth of the CLT does not allow for a robust study of student learning outcomes or retention to graduation. Additionally, any studies released in the immediate future will be limited to students at private, primarily religiously affiliated institutions of higher education. Even with the recent acceptance of CLT as an admissions test at the University of New Mexico, all public institutions in Florida, and select other public institutions, any study related to student outcomes at public colleges and universities will not happen for years.”
  3. Finding: There is no evidence that CLT can be used for course placement.
    “The team identified no evidence that a criterion score from the CLT could be used for placement purposes for students in mathematics or English courses as is current practice with the SAT and ACT. The inability to use CLT for course placement is problematic.”
  4. Finding: The concordance tables published by CLT were developed without a representative sample of students, have not been sufficiently tested, and therefore should not be used.
    “CLT concordance tables with SAT and ACT are not fully tested. … The absence of a representative sample of CLT scores makes direct comparisons with SAT or ACT scores, both of which have been validated and analyzed for decades, impossible.”