Parametric and Nonparametric Statistics
Now that we have the basic idea about statistical inference, and just before we leap in with both feet, a further bit of terminology. Statisticians often refer to two different kinds of statistical tests by using the obscure labels “parametric statistics” and (naturally enough) “nonparametric statistics.” The former includes the stuff of Chapters 4 to 9—t tests, analysis of variance, regression analysis, analysis of covariance, and time series. The latter covers stuff like chi-square, logistic regression, and log-linear analysis, which are discussed in Chapters 10 to 12.
The basic difference is that, in parametric statistics, the DV is some measured quantity (a ratio or interval level variable), so it makes sense to calculate means and SDs. With nonparametric stats, the DV is usually either a count or ranking (such as of dead bodies, which happen often, or cures, which happen rarely) for which it makes no sense to calculate means (e.g., “The average religion of Americans is 2.67”).
Why the fancy names? Well, if you can calculate means and SDs, then you can draw bell curves through the data, and the normal curves, as we saw, are defined by two parameters—the mean and SD—hence, parametric statistics. Conversely, you can’t put a bell curve through a histogram displaying the percentage of Americans who are Protestant, Catholic, and so on (you can, but it would make no sense). So these data have to be analyzed using some other nonparametric way, and the name follows.5343
Content on this page was last changed on March 19, 2009.
© 2002 BC Decker Inc. Show Disclaimer
| 5343. | Norman GR, Streiner DL. PDQ Statistics . 3rd ed. Hamilton, Ontario: BC Decker Inc.; 2003. |
Next Page: Sample Size Calculations »