Automatic Critical Residual Calculation for Contingency Tables

Coelen

Is there an option to automatically highlight (bold) the critical (or cell dependence) values, based on the DF and found signficance? Now I need to manually verify these results using such a table.

https://statisticsbyjim.com/hypothesis-testing/t-distribution-table/

patc3

If you're talking about cells resulting from a chi-square test (seeing as you're mentioning contingency tables in the title), then standardized (adjusted Pearson) residuals are interpreted as z scores, which carry no degrees of freedom. A z test is significant (p <= .05) when z is above +1.96 or below -1.96.

In this example (a 2x3 table), my omnibus chi-square is not significant (p = .205), and there are no cells with a significant standardized residual (above +1.96 or below -1.96):

Coelen

Sorry, I've must have phrases myself unclearly. I'm aware of the standardised residuals. However, to save space, instead of showing the standardised residuals, I would like the count (or percentage) to be printed in bold. Now it's a manual task, but was wondering if it's possible (ignoring earlier mentioned table).

patc3

Currently I doubt this is possible. Also some fields use different alpha values (cutoffs to declare significant) so I assume it would need to be customizable. You could make a feature request here: https://jasp-stats.org/feature-requests-bug-reports/

In the meantime I personally don't know of an automatic solution for this within JASP (you could copy-paste the table from JASP into Excel and use conditional formatting in Excel)

Coelen

Thanks again for your fast response. I'm currently doing the manual version, but it's not important enough for a feature request :)