In terms of testing the assumption of normality for a Point Biserial Correlation; I have a dichotomous variable (1 = microsleep, 0 = no microsleep) and a continuous variable (number of driving events occurring per drive)

My continuous variable has alot of 0 values (i.e. no event occurring that drive)

My shapiro wilk statistic was significant for no micro-sleeps, but not significant for micro-sleeps (skew and kurtosis all within + - 2). Homogeneity of variance was non-significant.

My question is, how robust is the point-biserial correlation including its significance test for normality violations? References for the finding would be greatly appreciated.

  • $\begingroup$ Maybe you can find something here $\endgroup$ Oct 7, 2018 at 14:13


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.