Analysis of Tables
Version 7

An introduction to Anota

Anota (ANalysis Of TAbles) is a statistical technique to explore possibly existing relationships between categorical (nominal) variables. One of the variables is assigned a special role. It is called the dependent variable. All other variables, the explanatory variables, are considered to be predictors of the dependent variable.

ANOTA resembles linear regression analysis. The main difference is that the dependent variable in linear regression analysis must be numerical. In Anota dependent as well as explanatory variables are categorical. The estimated coefficients have the same interpretation as regression coefficients. They measure the effect of the categories of the explanatory variables on the categories of the dependent variable. The coefficients are corrected for possible effects of other explanatory variables and therefore present 'pure' effects.

Due to the specific nature of the model, it is not necessary to have the raw data. If suffices to input all possible bivariate tables. This reduces the amount of data which have to be processed.

For carrying out an analysis on a set of variables, the program needs the bivariate tables for each combination of two variables. So, for an analysis on three variables A, B and C, the following tables should be prepared: A x B, A x C and B x C. You can either enter these tables from the keyboard, or let the program read a specially prepared data file.

You specify an Anota-model by assigning one variable the role of the dependent variable, and one or more variables the role of predictor variables. The program computes coeficients and standard errors of coefficients.

You can see the output of the analysis on the screen. It is also possible to generate an HTML-file or a text-file with the results of the analysis.

The program has the following limitations: