Key result
The study reintroduces the correct formula for the biserial-phi correlation coefficient in mixed dichotomous cases and presents a correction for unequal sample sizes.
This paper provides a corrected formula for the biserial-phi correlation coefficient in mixed dichotomy cases, including an adjustment for unequal sample sizes.
Answers are grounded in this paper and cite their sources.
Corrected biserial-phi formula may enhance accuracy in mixed dichotomous analyses; leaves open whether prior cardiovascular findings require reanalysis.
The correlation coefficient appropriate in a fourfold situation resulting from the crossing of two dichotomous variables depends on whether the dichotomies are true or arbitrary. Almost all sources dealing with the fourfold situation discuss only the pure cases (both dichotomies true or both dichotomies arbitrary). The mixed case (one dichotomy true and the other arbitrary), introduced at least four decades ago and accompanied in more than one source with the wrong formula for the coefficient, is reintroduced in this study. In addition, a correction for unequal sample size respecting the true dichotomy is presented.
No takes yet. Share an insight, caveat, or question.
Becker et al. (1988) studied this question. Biserial-Phi Correlation Coefficient was evaluated. The study reintroduces the correct formula for the biserial-phi correlation coefficient in mixed dichotomous cases and presents a correction for unequal sample sizes.
Synapse has enriched 10 related papers. Consider them for comparative context: