Date and Time: May 9th, 2017, 10:00 - 11:30 am
Room: A 101 in the Economics Building (Museum)
Real data in social, behaviroal, education and economic research are often not normally distributed. Geary (1947) has long pointed out that "Normality is a myth; there never was, and never will be a normal distribution."
In our own meta-analysis of 1595 univariate distributions and 249 multivariate distributions from data published in top journals, we found that 75% of the univariate distributions and 68% of the multivariate distributions are non-normal. Since non-normality affects statistical inference, I will present both frequentist and Bayesian methods to deal with non-normal data. For illustration, I will show how to handle non-normal data in structural equation modeling of multivariate data and in growth curve modeling of longitudinal data. I will also show free software we have developed to carry out non-normal data analysis discussed in this talk.
About the speaker
Zhiyang Johnny Zhang is an associate professor at the Department of Psychology of the University of Notre Dame, IN, US.