Contributor
Posts: 21

# Yates correction

I tried to do chi square test for the following observed frequencies

---------------------------------------------------------------------------------------------
A  B
A  8  4
B 12 10

SAS gave the following output which value and probability should be taken in this case

The FREQ Procedure

Statistics for Table of TRTGRP by ERR_ORD

Statistic                     DF       Value      Prob
------------------------------------------------------
Chi-Square                     1      0.4710    0.4925
Likelihood Ratio Chi-Square    1      0.4769    0.4898
Continuity Adj. Chi-Square     1      0.1035    0.7477
Mantel-Haenszel Chi-Square     1      0.4571    0.4990
Phi Coefficient                       0.1177
Contingency Coefficient               0.1169
Cramer's V                            0.1177

WARNING: 25% of the cells have expected counts less
than 5. Chi-Square may not be a valid test.
Super User
Posts: 23,683

## Re: Yates correction

Chi Square.

You should check the Fisher statistic as well, as per the warning.

Contributor
Posts: 21

## Re: Yates correction

Reeza thanks for the response i did Fisher's exact test too,so what will be the probability value in this case

The FREQ Procedure

Statistics for Table of TRTGRP by ERR_ORD

Fisher's Exact Test
----------------------------------
Cell (1,1) Frequency (F)         8
Left-sided Pr <= F          0.8533
Right-sided Pr >= F         0.3766

Table Probability (P)       0.2300
Two-sided Pr <= P           0.7170

Super User
Posts: 23,683

## Re: Yates correction

Last number 0.7170

By the way, what does you subject line have to do with your question?

Contributor
Posts: 21

## Re: Yates correction

Sorry if subject line caused some sort of confusiion.My primary doubt was whether to take  (Yates) Continuity Adj. Chi-Square or normal Chi-Square.

Discussion stats
• 4 replies
• 600 views
• 0 likes
• 2 in conversation