Your data analysis solution
Name is required.
Email address is required.
Exceeding max length of 5KB

# Chi Sq residual graph output in contingency tables

By Sue | Dec 04, 2017 11:04PM CET

Hello
There is a 3D graphical output listed as 'residuals' and a table of 'residuals' and 'adjusted residuals' in the contingency table analysis function, but there is no explanation of how these tables are calculated. It doesn't' look like the residuals of observed and expected observations, so what are these tables calculating? I can't find any documentation on this in Help and it's not covered in your tutorial.

By Sébastien | Dec 05, 2017 10:06AM CET | XLSTAT Agent

Hi there,

Residuals = (observed – expected) / sqrt(expected)
Adjusted residuals = (observed – expected) / sqrt[ expected*(1 – ( rowtotal / N ) )*(1 – ( coltotal / N ) )]
rowtotal being the sum of frequencies of the corresponding row and coltotal the sum of frequencies of the corresponding column.
Roughly, when residuals are > 2 or < – 2, then the observed frequency is significantly different from the theoretical one for alpha = 5%

Jean Paul

Up

0

Down

By Sébastien | Dec 05, 2017 10:06AM CET | XLSTAT Agent

Hi there,

Residuals = (observed – expected) / sqrt(expected)
Adjusted residuals = (observed – expected) / sqrt[ expected*(1 – ( rowtotal / N ) )*(1 – ( coltotal / N ) )]
rowtotal being the sum of frequencies of the corresponding row and coltotal the sum of frequencies of the corresponding column.
Roughly, when residuals are > 2 or < – 2, then the observed frequency is significantly different from the theoretical one for alpha = 5%

Jean Paul

1c26995d494fb3061dd0ae8571ffc0a4@xlstat.desk-mail.com
https://cdn.desk.com/
false
desk