Chapter 17

Non-Parametric Statistics


Introducing non-parametric statistics

Testing independent samples

Testing related samples

Correlating ranks

The statistical methods I cover in earlier chapters have a couple of things in common. First, we assume ratio (or at least interval) data. (If you don’t know what that means, reread the section entitled “Types of data” in Chapter 1.) Second, we can use sample statistics to estimate parameters of the sampling distribution, and we use the central limit theorem to characterize the nature of the distribution so that we can test hypotheses.

Sometimes we have to analyze nominal data or ordinal data (again, Chapter 1). And sometimes we can’t specify the distribution of the variable we’re working with.

To deal with these cases, statisticians have developed non-parametric statistical tests. The list of these tests is long, and growing all the time. Many of them require special lookup tables for hypothesis tests.

I want to avoid those special tables. So, to make the cut for this chapter, a test had to either (a) test hypotheses via a well-known distribution built into Excel or (b) work with a well-known distribution when samples are large. Of the non-parametric tests that fit the bill, I chose six classical ones.

Because Excel has no special data analysis tools or built-in functions for these tests, the general plan is to translate the test formulas into Excel formulas and then use an Excel statistical function (NORM.DIST or ...

Get Statistical Analysis with Excel For Dummies, 4th Edition now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.