help with chi squared

css

New Member
#1
Hi everyone,

I am analyzing data coming from a memory test. The test involves remembering words belonging to 4 different categories. After confirming that my two subjects' groups have the same degree of accuracy (number of correct responses), I would like to see whether the errors each subjects' group makes are similarly distributed or not (that is, if they do the same number of mistakes but in different words' categories)

I have started by calculating the chi-squared goodness of fit in spss for each group separately. Since all stimuli categories has the same number of items, I assumed as null hypothesis that errors' frequency should be equivalent in each of the fourth categories. That seems the case in one of the subjects' groups (group A), but not in the other (group B). Now I would like to know in which stimuli category the number of errors displayed by group B are significantly different from that expected. Of course, I can "guess it" from the residual values but I would like to statistically "prove it" and assign a p value...any suggestion?

On the other hand, I would also like to compare whether the number of errors in these categories are different or not between groups. It "should" because total number of errors is not different but their distribution is. However, I am unsure on how to conduct this between group comparison

thanks in advance for any help or suggestion