ABSTRACT
All of the inferential statistics discussed in this book share a set of assumptions. Regression, ANOVA, correlation, and t tests all assume that the data involved are scores on some measure (e.g., IQ scores, height, income, scores on a measure of depression) calculated from samples drawn from populations that are normally distributed, and everything is fine in the world of research. Of course, as discussed in Chapter 1, these conditions are often not met in social science research. Populations are sometimes skewed rather than normal. Sometimes researchers want to know about things besides those that can be measured. Research is often messy and unpredictable rather than clean and easy.
