Comparing two percentages for significance

[breiteyes55]

Hello...this was posted but was lost recently.

I am trying to figure out the best method to compare two percentages for significance. I am analyzing data from open ended surveys which have been coded and categorized. If I have 75% of individuals in Group A (n=140) answer one way and 14% of individuals in Group B (n=140) respond the same, how do i determine whether or not the difference is significant?

I was originally doing chi-square tests which seemed accurate if the the samples were the samesize, as in the above example. But, I have some cases where individuals in group A (n=140) respond one way and Group B (n=75) respond the same. In these cases, I was hoping for a test that would fully take into account the differences in sample size.

I recently found the following two websites suggesting z-tests. Would this be more appropriate? Also, if yes, how do I clearly represent the significance in text? ie; the difference was significant (n=(total sample or separate samples?), p<.05, z=2.34).

Thank you in advance!

[quark]

I think the z-test for two proportions will work. There's a nice calculator at:

http://www.dimensionresearch.com/res...ors/ztest.html

Your representation is quite clear.

Hope this helps.