Monday, December 12, 2011

Comparing overlapping confidence intervals to test significance?

Hi,


I have some survey data with both state and national level data. I want to test if there are significant differences between a state and the national percentages (for variables such as race, sex, maternal education, as well as other mostly categorical and binary variables). So, one population is a subset of the other. I compared 95% confidence intervals to determine significant differences (overlapping CI's are not significant). But what would be the statistical test I would use to do this? Chi-square? thanks.|||You have two samples from two "populations" (state and national).


You want to compare differences in mean, I suppose.


Therfore a two sample Student-t test will be appropriate.


You can find the procedure fx here:


http://www.fon.hum.uva.nl/Service/Statis鈥?/a>





Helps?

No comments:

Post a Comment