The statistic λ is compared with the usual biserial‐r with respect to efficiency for estimating the true correlation in the population from which the sample was drawn. Both statistics are designed for the case where one of the observed variables is assumed to result from the dichotomizing of a continuous variable. When the true correlation exceeds .50, the statistic λ appears to be more efficient than is biserial‐r.
No takes yet. Share an insight, caveat, or question.
Frederic M. Lord (1961) studied this question.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: