Unformatted text preview: be used when one or both of the assumptions about equal population standard deviations and normal distributions are violated. When the observed data are skewed, or when extreme outliers are present, it usually is better to analyze the median rather than the mean. One test for comparing medians is the Kruskal‐
Wallis Test. It is based on a comparison of the relative rankings (sizes) of the data in the observed samples, and for this reason is called a rank test. The term nonparametric test also is used to describe this test because there are no assumptions made about a specific distribution for the population of measurements. Another nonparametric test used to compare population medians is Mood’s Median Test. See Section 16.3 for an example of each of these alternative methods. Two‐Way ANOVA So far we have focused on the one‐way ANOVA procedure. The "one‐way" referred to having only one explanatory variable (or factor) and one quantitative response variable. Section 16.4 presents a brief overview of two‐way ANOVA, which examines the effect of two explanatory variables (or factors) on the mean response. The researcher is interested in the individual effect of each explanatory variable on the mean response and also in the combined effect of the two explanatory variables on the mean response. The individual effect of each factor on the response is called a main effect. If one of the factors does not have...
View
Full Document
 Summer '10
 Gunderson
 Statistics, Variance, Mean squared error

Click to edit the document details