Home | 50-50 MANOVA | Rotation Tests | Software | Fisher's Exact Test | Publications

Fisher's
Exact Test











INTRODUCTION

Created by
©
Øyvind Langsrud

German version


INTRODUCTION

This page can be used to test statistically whether there is any relation between two categorical variables (with two levels). Fill inn the table and press COMPUTE.

The output consists of three p-values:

NOTE: Decide to use Left, Right or 2-Tail before collecting (or looking at) the data.
An example of such data (not real) are 100 persons classified after sex and e-mail address:

female male
e-mail address 3 15
no e-mail address 37 45

I.e. 3 of 40 women have e-mail address and 15 of 60 men have e-mail address.

The data could have been collected in different ways:

  1. We have asked 60 men and 40 women. I.e. the total number of men and women is fixed.
  2. We have asked 100 persons about their sex and whether they have e-mail address. I.e. only the total number of persons is fixed in advance.
  3. We have asked all persons in Norway born 24/11-68. I.e.: The total number of persons is not fixed in advance.

The 2-Tail p-value is calculated as defined in Agresti (1992) Sec. 2.1. (b).
REFERENCE: Agresti A, (1992), A Survey of Exact Inference for Contegency Tables, Statitical Science,7,131-153