Linear Statistical Models

Looking at Anova Data

Updated for Stata 11


Displaying Raw Anova Data

Althought it isn't done for every analysis, there will be times that you want to display the raw data for an anova. The tabdisp command allows you to display the anova data in tabular form.

General Descriptive and Exploratory Data Analysis

In general, it is important to look at your data, to try to understand it as best you can. You can use all of the tools in descriptive statistics and exploratory data analysis that were covered in the regression part of the course.

Looking at Data by Group

In addition to exploratory data analysis and general descriptive statistics you will want to look at data separately for each group in order to check on how well our data meet the assumptions of analysis of variance, in particular, the assumptions of normality and homogeneity of variance (homoscedasticity).

In analysis of variance, we often want to look at the variability and shape of the distribution within each cell of the design. Say that we wanted to look at the anova model for write with female and prog as our categorical variables. There are two levels of female and three levels of prog resulting in a total of six cells. Here are some commands that we can use to look at the data at the marginal and/or cell level.

Graphing Cell Means

In addition to looking at the data to check on assumptions, it is often useful to graph the cell means as a way to help understand interactions. In this example, we will be using the anovaplot command.


Linear Statistical Models Course

Phil Ender, 17sep10, 18mar03; 27mar02; 23feb01