The Concept of ANOVA
Class, the concept of ANOVA is difficult to grasp at first, but if we understand what it is we are
looking for, then it becomes easier. These comments are a bit long, but take each point at a
time. Recall that when we are interested in whether the difference between two means (a
population and a sample mean) is statistically different, we use the t or z test. What if we have
more than two means? What do we do then? The answer is to use analysis of variance
(ANOVA). We simply calculate as many means as we have and test the difference among them
with the analysis of variance procedure. An example is comparing the costs of three different
health insurance programs that are being piloted tested by a city. The independent variable in
this case is the type of health insurance program, which has three categories: HMO1, HMO2,
and HMO3 (a nominal variable). The dependent variable is the cost of each program (an
intervalratio variable). If we are interested in comparing the costs of the three insurance
This is the end of the preview. Sign up
to
access the rest of the document.
 Spring '10
 Regis
 Statistics, ANOVA Class

Click to edit the document details