New Perspectives on the Validity of the GRE ® General Test for Predicting Graduate School Grades

Scores from the GRE® revised General Test provide important information regarding the verbal and quantitative reasoning abilities and analytical writing skills of applicants to graduate programs. The validity and utility of these scores depend upon the degree to which the scores predict success in graduate and business school in specific contexts. To assess the predictive validity of the GRE test for graduate business programs, we collaborated with a number of universities and obtained data from each university's admissions office and registrar. We focused specifically on part‐time and full‐time students in master's of business administration (MBA) degree programs. Given the nested structure of the data, we used a 2‐level (representing students and institutions) hierarchical linear model (HLM) to estimate regression models with first‐semester MBA grade point average (GPA) or cumulative MBA GPA as the dependent variable and GRE scores and undergraduate GPA (UGPA) as independent variables. For predicting cumulative MBA GPA, the pseudo R‐squared (R2) value was .04 using UGPA, cohort year, and program type as the only predictors; this value increased substantially to .19 with the addition of GRE scores. The HLM results show that both GRE‐Quantitative and GRE‐Verbal scores were statistically significant predictors of both first‐semester MBA GPA and cumulative MBA GPA.

Section: Research On the Validity Of Gre Scores For Predicting Graduamentioning

confidence: 59%

Section: Discussionmentioning

confidence: 79%

Section: Research On the Validity Of Gre Scores For Predicting Graduamentioning

confidence: 99%

See 1 more Smart Citation

The Validity of Scores from the GRE^® revised General Test for Forecasting Performance in Business Schools: Phase One

Young

Klieger

Bochenek

et al. 2014

“…Consequently, we do not evaluate whether the adaptations and changes, or lack thereof, are good or bad, right or wrong. However, we do wish to note that large‐scale studies show that cognitive assessments like such as the GRE revised General Test and Subject Tests are robust predictors of performance, both absolutely and relatively (Klieger et al, ; Kuncel et al, ; Kuncel & Hezlett, ; Liu et al, in press). We therefore encourage graduate and professional programs to at least consider these empirical findings when evaluating how to view and use the GRE revised General Test and Subject Tests.…”

Section: Discussionmentioning

confidence: 99%

Perceptions and Uses of GRE^® Scores After the Launch of the GRE^® revised General Test in August 2011

Klieger

Belur

Kotloff

2017

Self Cite

This survey study investigated how graduate school admissions committees perceive and use the GRE® General Test and GRE® Subject Tests after the launch of the GRE® revised General Test in August 2011. These perceptions and uses impact the validity of the tests. Prior research about the perceptions and uses of the General Test and Subject Tests was last conducted in 2002 and, prior to that, in 2000 (for writing) and 1984. Therefore, even without test revisions, perceptions and uses of the GRE tests may have changed. Overall, by graduate discipline area, by graduate degree level sought, and by future career track preparation (research vs. professional), we examined online survey responses from 163 individuals involved in graduate school admissions. We did not find major changes in the use or perceived utility of the revised General Test or Subject Tests for admissions or funding decisions. General Test scores are valued in relation to other admissions information. Sometimes, valuation of these scores leads to practices proscribed by published GRE Program guidelines, such as the use of cut scores. Scores on the Subject Tests continue to be valued less than other admissions information. Perceptions and uses of the GRE tests often vary when decision makers consider applications from international applicants and underrepresented racial or ethnic minorities. That admissions committees can receive and compare scores reported on the former General Test score scales to scores reported on the revised General Test score scales has resulted in various practices, including the common use of percentile ranks to compare scores in ways discouraged by GRE Program guidelines. To the extent that decision makers use concordance information recommended by Educational Testing Service (ETS) to compare scores, they do so with mixed success.

“…While predictive validity evidence for CR format tests is more limited, the English Placement Test essay has been demonstrated to correlate r = .35 with English grades and r = .21 with fall semester GPA (Michael & Shaffer, ). For the GRE, Klieger, Cline, Holzman, Minsky, and Lorenz (in press) reported small but significant correlations ( r = .16) between GRE Analytical Writing scores and graduate‐school GPA for samples of more than 24,000 graduate students. The highest correlation of GRE‐AW with GPA at the master's level was observed for English language and literature students ( r = .28), suggesting that this test was most predictive for fields requiring considerable reading, critical analysis, and writing of texts.…”

Section: Review Of Existing Assessments and Design Challengesmentioning

confidence: 99%

Assessing Written Communication in Higher Education: Review and Recommendations for Next‐Generation Assessment

Sparks

Brantley

et al. 2014

Written communication is considered one of the most critical competencies for academic and career success, as evident in surveys of stakeholders from higher education and the workforce. Emphasis on writing skills suggests the need for next‐generation assessments of writing proficiency to inform curricular and instructional improvement. This article presents a comprehensive review of definitions of writing proficiency from key higher education and workforce frameworks; the strengths and weaknesses of existing assessments; and challenges related to designing, implementing, and interpreting such assessments. Consistent with extant frameworks, we propose an operational definition including 4 strands of skills: (a) social and rhetorical knowledge, (b) domain knowledge and conceptual strategies, (c) language use and conventions, and (d) the writing process. Measuring these aspects of writing requires multiple assessment formats (including selected‐response [SR] and constructed‐response [CR] tasks) to balance construct coverage and test reliability. Next‐generation assessments should balance authenticity (e.g., realistic writing tasks) and psychometric quality (e.g., desirable measurement properties), while providing institutions and faculty with actionable data. The review and operational definition presented here should serve as an important resource for institutions that seek to either adopt or design an assessment of students' writing proficiency.