Home Articles Abstract
Research Article

Statistical Misdiagnosis in Business Administration Research Papers

Sung, Samgyeong · Choi, Jonghu · Lee, Jaechang

Published: January 1993 · Vol. 22, No. 2 · pp. 163-198
Full Text

Abstract

Examining the research papers published in the Korean Management Review—the most comprehensive and longest-running management research journal in Korea—statistical techniques were rarely noticeable in its early period, but univariate statistical techniques appeared in 1977, multivariate statistical techniques began to be used from 1983, and in recent years the proportion of papers utilizing statistical techniques has sharply increased with the application of diverse and advanced techniques. Considering that numerous errors have been identified even in Western academic papers where rigorous reviews of statistically-derived conclusions are institutionalized, the need for statistical evaluation of Korean management papers was also recognized. Therefore, this study developed a comprehensive checklist for evaluating papers spanning the 20-year period from the journal's founding in 1971 to 1990, and developed technique-specific checklists for univariate statistical techniques including t-test, chi-square test, analysis of variance, and regression analysis, as well as multivariate statistical techniques including principal component analysis, factor analysis, discriminant analysis, cluster analysis, canonical correlation analysis, and multivariate analysis of variance. Comparative evaluation against the checklists identified an unexpectedly large number of problems, ranging from issues of unification and localization of statistical terminology to statistical errors that could call into question the conclusions drawn in the papers. Given the rapid increase in research papers using statistical techniques, this study aimed to present a standard for reflection and future improvement in the proper use of techniques not only in the management field but also in other disciplines.