A study was made of the r-point biserial coefficient using four non-normal distributions for the continuous variable: rectangular, bimodal-normal, bimodal-peaked, and bimodal-peaked and skewed. Ns of 10,30, and 100 were used. It was argued that linearity was the main assumption required when using Pearson correlations and that the usual maximum r -point biserial of .798 could be exceeded when the shape of the continuous variable more nearly approached that of the dichotomized variable. Correlations were found over .80 with rectangular distributions, and over .90 with bimodal-peaked distributions.
Get full access to this article
View all access options for this article.
References
1.
Adams, J. F.The effect of non-normally distributed criterion scores on item analysis techniques. EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT, 1960, 20, 317-320.
2.
Bowers, J.A note on comparing r-biserial and r-point biserial. EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT, 1972, 32, 771-775.
3.
Glass, G. V and Stanley, J. C.Statistical methods in education and psychology. Englewood Cliffs, N. J.: Prentice-Hall , 1970.
4.
McNemar, Q.Psychological statistics (3rd ed.). N.Y.: John Wiley and Sons, 1962.
5.
Nefzger, M. D. and Drasgow, J.The needless assumption of normality in Pearson's r . American Psychologist, 1957, 12, 623-625.
6.
Walker, H. M. and Lev, J.Statistical inference. N. Y.: Holt, Rinehart and Winston, 1953.
