Comparison of the Incremental Validity of the Old and New MCAT

Sep 4, 2026

Upload by Mali

The study compares the predictive and incremental validity of the Old and New Medical College Admission Test (MCAT) using a sample of over 300 medical students. Zero-order and incremental validity coefficients, along with prediction models from all-subsets regression using Mallows Cp, are evaluated through cross-validation by splitting two medical school classes into screening and calibration samples. Evidence supports incremental validity for both tests, with larger coefficients for the New MCAT. Prediction models for NBME Part I and II are cross-validated, while clerkship performance models are equivocal.

LanguageEnglish

CategoryExam

FormatPDF

Pages35

Reading Time1 h 28 min

AccessFree to download

Comparison of the Incremental Validity of the Old and New MCAT

Incremental Validity

ED 22941l AUTHOR TITLE PUB DATE NOTE PUB TYPE 下 EDRS PRICE DESCRI PTORS I DENTIFIERS ABSTRACT DOCUMENT RESUME Wolf,Fredric M.;And Others Comparison of the Incremental Valrdity of the Old andNew MCAT. Apr 83 39p.;Paper presented_at the Annual Meeting of theAmerican Educational Research Association(67th,Montreal,Quebec,April 11-15,1983). Speeches/Conference Papers(150)--Reports -Research/Technical(143) MF-01/PC02 Plus Postage Academic Achievement;Admission Criteria;Biology;Chemistry;*College Entrance Examinations;,Comparative Analysis;*Competitive Selection;Correlation;Medical Education;*Medical Students;Predictive Measurement;*Predictive Validity;Regression(Statistics);*Test Validity *Cross Validation;*Medical College Admission Test;National Board of Medical Examiners;Test Revision The predictive and incremental validity of both the01d and New Medical college Admission Test(MCAT)was examined andcompared with a sample of over 300 medical students.Results of zeroorder and incremental validity coefficients,as well as prediction models resulting from all possible subsets regression analyses usingMallow's Cp criterion,were subjected to cross-validation analyses byrandomly dividing two medical school classes into screening andcalibration samples.Results supported the incremental validity ofboth the Old and New MCAT。Coefficients were generally larger for theNew than_for the Old MCAT.Prediction models of National Board ofMedical Examiners Part I and Part II performance,comprised of theNew Biology and Chemistry subtests and the 01d Science and GeneralInformation subtests were cross-validated.Prediction models ofclinical evaluation clerkship performance were equivocal.(Author) ****************★****★****★*★***★★**★**★**********★**★*★★*★*★★*★★★★*★★★ 11462203 A(M) 7M83 U.3.DEPARTMENT OF EDUCATIONNATIONAL INSTITUTE OF EDUCATIONEOUCATIONAL RESOURCES INFORMATION CENTER(ERIC) X Ths document has boen roproducod arecorvad from tho person or organzationonginating ! Minor changos have been made to improvūroproductuon quality ● Points of yiew or opinions stoted in this document do not necessardy repres0nt officol NIEposition or policy — # Comparison of the Incremental Validityof the Old and New MCAT Fredric M.Wolf,Judith G.Calhoun,Bruce R.Maxim,Wayne K.DavisUniversity of MichiganMedical School "PERMISSION TO·REPRODUCE THISMATERIAL HAS BEEN GRANTED BY F m.心:16 TO THE EDUCATIONAL RESOURCESINFORMATION CENTER(ERIC)." Alan L.Hull Case Western Reserve UniversitySchool of Medicine Running Head:Incremental Validity Presented at the meeting of the American Educational Research Association,Montreal,April 1983.Correspondence may be addressed to the authors at theDepartment of Postgraduate Medicine and Health Professions Education,University of Michigan,G1208 Towsley Center,Ann Arbor,MI 48109. # Comparison of the Incremental Validityof the Old and New MCAT 虚 The predictive and incremental validity of both the Old and New Medical CollegeAdmission Test(MCAT)was examined and compared with a sample of over 300 medicalstudents. Results of zero order and incremental validity coefficients,as well asprediction models resulting from all possible subsets regression analyses using Mallow'scriterion,were subjected to cross-validation analyses by randomly dividing two medicalschool classes into screening and calibration samples.Results supported the incrementalvalidity,of both the Old and New MCAT.Coefficients were generally larger for the Newthan for the Old MCAT.Prediction models of NBME Part I and I performance,comprisedof the New Biology and Chemistry subtests and the Old Science and General Informationsubtests were cross-validated. Prediction models of clinical evaluation clerkshipperformance were equivocal. Incremental Validity 1 A number of studies recently have compared the ability of the Qld and the NewMCAT to predict student achievement in medical school.Since the New MCAT wa

FAQs

What is the main purpose of this study about the MCAT?

To examine and compare the predictive and incremental validity of the Old and New MCAT for medical students.

How did the researchers test the prediction models?

They computed validity coefficients and built prediction models using all possible subsets regression with Mallows Cp, then performed cross-validation by dividing two classes into screening and calibration samples.

What did the results show about the Old versus New MCAT?

Both tests showed incremental validity, and the coefficients were generally larger for the New MCAT than for the Old MCAT. Prediction models for clerkship performance were equivocal.

Sep 4, 2026

Upload by Mali