Reading Crosstab Output of Z-Score Bonferonni Test

[hansonac]

Can someone help me read this? From reading the SPSS documentation if the subscript letter is different, than there is a significant difference. But what is confusing me is that this output seems to suggest that even though there are two different subscript letters, that colums one and two are no different because the last three have two letters. I am really confused.

Thanks in advance!!

[GauravSrivastava]

Hey Hanson,

If first 2 column are showing different subscript (a & b), that means both column are significantly different. whereas if rest 3 column are showing both letters (a,b) that means they are neither different from 1st column nor with the 2nd column.

Hope this make sense.

Thanks!

Gaurav

On Thu, May 30, 2013 at 12:25 AM, hansonac <[hidden email]> wrote:

Can someone help me read this? From reading the SPSS documentation if the
subscript letter is different, than there is a significant difference. But
what is confusing me is that this output seems to suggest that even though
there are two different subscript letters, that colums one and two are no
different because the last three have two letters. I am really confused.

Thanks in advance!!

<http://spssx-discussion.1045642.n5.nabble.com/file/n5720468/output_help.jpg>



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Reading-Crosstab-Output-of-Z-Score-Bonferonni-Test-tp5720468.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Ruben Geert van den Berg]

If I understand correctly, proportions that don't have any subscripts in common are different.

So a proportion with [a,b] does not differ from a proportion with [a],[b] or  [a,d,e] (they still have [a] in common).

a proportion with [a,b] does differ from a proportion with [c],[c,d] or [e,f]

I'm not sure whether the Bonferroni correction applies to all tests related to the table or per row but perhaps someone else knows?

So for a 4 X 4 table, there can be 6 pairwise comparisons for each of 4 rows. I'm not sure whether alpha is divided by 6 or by 24 here... And in case of empty cells I guess only the comparisons that are actually tested count towards this total number?