Frederic M. Lord spent his entire career at Educational Testing Service (ETS), where he developed the mathematical and statistical foundations of modern test theory. His work on item response theory, test equating, and the statistical analysis of test scores transformed psychometrics from an informal collection of practical techniques into a rigorous mathematical discipline.
Item Response Theory
3PL: P(X=1|theta) = c + (1-c) / {1 + exp[-a(theta - b)]}
theta = latent ability
a = item discrimination
b = item difficulty
c = pseudo-guessing parameter
Lord was among the first to develop the mathematical theory of item response models and to show how they could solve practical testing problems. His work on item and test information functions showed how to quantify the precision of measurement at each ability level, enabling the design of tests that are maximally informative for specific populations. His development of the two-parameter and three-parameter logistic models (extending Birnbaum's work) provided flexible tools for modeling item responses that account for differences in item discrimination and the effects of guessing.
Lord developed the theoretical foundations for test equating -- the process of placing scores from different test forms on a common scale. His IRT-based equating methods, which use item parameters estimated on a common scale, provided mathematically principled solutions to a practical problem of enormous importance in educational assessment, enabling fair score comparisons across different test administrations.
Statistical Theories of Mental Test Scores
Lord's 1968 monograph Statistical Theories of Mental Test Scores (with Novick) unified classical test theory and item response theory within a single mathematical framework. The book provided rigorous treatments of true score theory, reliability, validity, and the estimation of latent traits, establishing the mathematical standards for psychometric research that remain in force today.
Legacy and Impact
Lord's work made possible computerized adaptive testing, item banking, test equating across forms and populations, and the statistical techniques used in every major educational assessment program worldwide. His combination of mathematical sophistication and practical orientation -- always driven by real testing problems -- exemplifies the best of applied mathematical psychology.