Categorical variable ; pre and post intervention

[vmp]

Hi all, I really need help with this one.

I am trying to see if my population change their alcohol habits between before and after an psychosocial intervention.

We asked the participants : how many days per week do you consume alcohol ?
a. Never
b. 1 to 2 days per week
c. 3 to 4 days per week
d. 5 to 6 days per week
e. every days

We asked the same question a month later to the same participants after giving them an intervention.

I thought of using chi 2, but my groups are not independent. Which test do you think I should use ?

Thank you in advance.

 

[hlsmith]

Did everyone get the intervention? Can you match all the pre and post data?

 

[vmp]

Yes. Only some missing data in the post

 

[Karabiner]

So you have an ordinal variable which was measured twice in the same subjects.
The sign test could be used here.

With kind regards

Karabiner

 

[vmp]

Thank you, I will look into this test.
Best,

 

[hlsmith]

Agreed. To help in your searches, it can be called the Wilcoxon sign test and you typically apply it to the differences of the two measures to reveal a non-null difference.

 

[Karabiner]

Agreed. To help in your searches, it can be called the Wilcoxon sign test and you typically apply it to the differences of the two measures to reveal a non-null difference.

Just to add a bit more detail, if we find information about the Wilcoxon signed rank test
(which requiers interval level) , it must not be confused with the sign test (for ordinal
variables).

With kind regards

Karabiner

 

[hlsmith]

OK!