Lecture 9

# Lecture 9 - 2 Homogeneity of variance in groups Groups have...

This preview shows page 1. Sign up to view the full content.

Lecture 9 2/06/08 Some elementary statistics. A statistical test tells us the probability due to chance alone that we can get that effect in the experiment. Different statistical tests have different sets of assumptions that the data must meet in order for the test to give reliable results. We can break the statistical test into two large groups. Parametric and nonparametric tests. Parametric – Require two important two assumptions. 1) Data must be normally distributed (normal distribution curve). Recall mathematical description. So much data has to be in between certain standard deviations. Can’t use test on data clumped at one end.
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 2) Homogeneity of variance in groups. Groups have to have relatively same variance. Variance is a measure of the amount of differences? 3) We want to use parametric test. Higher power. Capability to show two means are different even though they are relatively the same. Nonparametric – Neither assumption needed. Use F test to test for homogeneity of variance. We try to avoid it when we can due to low power. Difference between means has to be pretty big usually for this test to work. Chi-Square Test Use this test in genetics because when the data comes in ratio form, this test is well suited to those assumptions....
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online