Could someone clarify this please? (on alternative hypothesis')

C

commons

Guest
#1
Hello,

- firstly, i have TWO INDEPENDENT samples (174 and 348) of peoples ages,

- i've checked for normality, and as they're not normal i'm carrying out a Mann-whitney test,

- When carrying out this test on the computer, it asks for the Alternative hypothesis and it lists three options: less than, not equal, or greater than.

RIGHT! Here's what i need to know.....

Which Alternative hypothesis do i use?? I always just thought it was 'equal' for the Null Hyp. and 'not equal' for the Alternative Hyp.

When i looked in my textbook (Devore - Probability and Statistics), there was an example on this and so i tried it on my computer testing each possible alternative and it turned out that they used the 'greater than' hypothesis!!!!

Can someone please tell me why!!

Thank you thank you thank you,

C
 

quark

Administrator
#2
The alternative hypothesis depends on the research question. If you test whether the mean age of one group is greater than that of the other group, then Ha is "greater than...".

A one-tailed test looks for an increase (Ha: greater than) or decrease (Ha: less than) in the parameter whereas a two-tailed test looks for any change in the parameter (Ha: not equal).