K
if we have more than two rows or two columns in contingency table, then how can we compare between them to find whether which variable is significantly different from others ?
see attachment
Statistical Steven
thx for response
kindly look into this article
Page no. 3 and 4
table 5 and 6
how chi square is applied there?
It does not give the original data. The Chi-Sq analysis is just a contigency table so no transformation would be necessary. I must be missing an assumption the author made in getting their conclusions.Hi Steven
This is 'shooting from the hip' as I haven't read the attached pdf in any great detail. However, you mention that you are unable to reproduce the chi square values.
"Data were initially recorded and stored in Excel spreadsheets. Log-transformed
data were analysed using the Minitab statistical computer package (release 12.21,
Minitab Inc.) and Genstat (Version 8, SAS Institute). The chi-square (v2) test and
analysis of variance (ANOVA) were used to determine possible associations be-
tween udder/teat confirmation or teat-end lesions and quarter SCC and bacterial
IMI." (my emphasis).
Does the paper report the original data, along with the results of the analysis on the log transformed data?
Also, Genstat has an option for generalised loglinear models which can be used as a test of association in multi-way tables. It's possible that some of the values were generated using this option in Genstat.
NC