How to compare Probit means between two groups.

#1
I have calculated probit median (mean) ages of attainemnt of various stages (skeletal maturity) for two independent groups of subejcts.
How can I compare the two means obtained for each stage between the two groups, to see if the difference is statistically significant?
In a few articles that I read they used pair-wise comparison, but I couldn't understand. Is there any special test to use? Should I compare the slopes of the probit lines only. I use SPSS and Minitab programs.

Any help provided at this point would be much appreciated.

Thanks in advance.

Kostas
 
E

elnaz

Guest
#2
Hello
i will say you a thing that before any thing you should test of normality your data if they are normal for comparing means of two normal and independent groups you should use parametric test ,that is independent- sample t test but for comparing means of two normal and dependent groups you should use parametric test ,that is paired- sample t test.
if they are not normal for comparing means of two subnormal and independent groups you should use nonparametric test ,that is mannwhitney
if they are not normal for comparing means of two subnormal and dependent groups you should use nonparametric test ,that is wilcoxon.
i hope i could help you .
 
#3
Thank you Elnaz for replying to my question, but I think you didn't notice that the means I obtained are Probit means.The mean(or more correct median with the 95% fiducial limits) values that I want to compare are the results of PROBIT analysis, after I obtained cumulative frequency tables for my data.
Any ideas how this comparison can be done are welcomed?

Kind regards,

Kostas