Need help choosing statistics test for dissertation

#1
Hi everyone, really struggling with choosing a right statistical test for my thesis. will really appreciate your help.


My study aims to see if there are more short sighted people in patients with headache with migraine, tension headache, compared to the non-headache group.

I want to prove that patients in the non-headache group have the lowest prevalence of short sightedness (categorical variable).

Which statistical test is most suitable?

Thank you.
 

Karabiner

TS Contributor
#2
It depends on how the variable "short sightedness" was measured (just a yes/no variable?) and how large the sample sizes are and whether additional headache-related characteristics of the patients (such as gender, age, exposition to risk factors etc.) should be included into the analysis.

With kind regards


Karabiner
 
#3
It depends on how the variable "short sightedness" was measured (just a yes/no variable?) and how large the sample sizes are and whether additional headache-related characteristics of the patients (such as gender, age, exposition to risk factors etc.) should be included into the analysis.

With kind regards


Karabiner
Hi. Thanks for your response.
The short sightedness is divided into mild, moderate, and severe (categorical).
the different headache groups are tension type headache, migraine, and non-headache. There’s also an overall headache group (combination of tension type headache and migraine).

the question is whether there is a significant difference between the overall headache groups va non-headache group, and also btw migraine vs non-headache, and tension type headache vs non- headache.

no need to include risk factors or gender.

your help is really appreciated.
Thank you.
 

Karabiner

TS Contributor
#4
If it is not possible to use the actual diopters, then you could use a Kruskal-Wallis H test for
the global effect of the factor "group" (3 levels), and Mann-Whitey U-test for pairwise
comparisons between groups.

With kind regards

Karabiner
 
#5
If it is not possible to use the actual diopters, then you could use a Kruskal-Wallis H test for
the global effect of the factor "group" (3 levels), and Mann-Whitey U-test for pairwise
comparisons between groups.

With kind regards

Karabiner
Thanks so much Karabiner. Much appreciated.

Can I use the Z-score calculator for 2 population proportions if I want to compare just Mild short-sightedness alone (not including moderate and severe short sightedness) between the migraine and non-headache groups?

Thank you.

Warm regards
 

Karabiner

TS Contributor
#6
That would be a fourfold table (migraine / non-headache versus mild short-sightedness / other short-sightedness) with a Chi² test.

With kind regards

Karabiner
 
#7
That would be a fourfold table (migraine / non-headache versus mild short-sightedness / other short-sightedness) with a Chi² test.

With kind regards

Karabiner
Hi Karabiner,

what if I would like to analyse just mild short sightedness on its own, in the migraine vs. non-headache groups only?
Which test would suit?

Thank you.

Warm regards.
 
#9
What exactely would be the research question or hypothesis?

With kind regards

Karabiner
Hi Karabiner,

The question is whether mild short sightedness is less common in non-headache groups. It is basically to prove that mild short sightedness is more prevalent in headache groups.

Overall it is to see if short sightedness in general is more common in headache groups, but I would also like to analyse if the subgroups of mild/moderate/severe short sightedness are also more common in headache populations.

Hope I am making sense.

Thank you.

Warm regards
 
#11
The proposed crosstabulation with Chi² test can do that.

With kind regards

Karabiner

Thank you.

Another question.

Is there an alternative to the Wilcoxon Signed-Rank Test? It works fine for the overall headache group, as I have 40 patients, but there are only 5 patients in my Migraine group. I would like to compare the short sightedness before and after the intervention in just the migraine group (Spectacles to correct short sightedness)

Thank you.

Warm regards