Constructs, Tape Measures, and Mercury.

Area of Science:

• Psychology
• Sociology
• Educational Measurement

Background:

• The Graduate Record Examination (GRE) is a widely used standardized test for graduate school admissions.
• Concerns exist regarding potential biases and the validity of GRE scores across diverse demographic groups.
• Existing literature debates the role of GRE scores in graduate admissions, with some advocating for their removal.

Purpose of the Study:

• To critically examine the construct validity of the GRE within the context of racism.
• To challenge the definition of what the GRE measures and propose alternative interpretations.
• To contribute to the ongoing discussion on whether GRE scores should be a factor in graduate admissions.

Main Methods:

• Application of Lewinian field theory to understand the GRE's measurement limitations.
• Analogy drawn from genome-wide association studies to illustrate predictive model biases across populations.
• Critique of the psychometric perspective on GRE bias and its implications for admissions.

Main Results:

• The GRE's underlying construct definition is questioned, suggesting alternative interpretations of measured abilities.
• An analogy from genetics highlights how models trained on one population may underpredict outcomes in others.
• The validation of GRE scores using first-year grades is discussed in light of potential measurement biases.

Conclusions:

• The GRE's measurement of abilities may be influenced by systemic racism, impacting score validity.
• Relying solely on GRE scores for admissions may perpetuate inequities.
• Alternative approaches to assessment and admissions processes are needed to ensure fairness and equity.