Hypothesis Test method Query

GKK

New Member
#1
What would be the appropriate statistical method to investigate whether patients who had a history of having drug “X” on a regular basis had significantly different systolic blood pressure (SBP), diastolic blood pressure (DBP) and cholesterol level (CHL) comparedto that for patients without drug “X”. Will I be correct if I used Independent t-test to investigate? Also, to check the normality I have used histogram. Is it right? Or should I use box plots?

The total study population is 1429.The objective of the study was to investigate whether drug “X” (having “X” on a regular basis) was protective for various risk factors of cerebrovas
 

Karabiner

TS Contributor
#2
Independent t-test looks o.k. With such a large sample size,
normality within groups is not required (central limit theorem
applies). If sample sizes are markedly different, variance should
be equal between groups (otherwise, use t-test with Welch correction).

With kind regards

K
 

GKK

New Member
#3
Hi K,

Thank you for your response. I have worked out the study using Kruskal-Wallis test for the above query.

Could you please have a look and advise me if I am on the right track? Or do you still reckon I use the Independent t-test?

Regards,

GKK
 

Karabiner

TS Contributor
#4
I suppose it won't change the results if you used t-tests instead of H-tests.
Mind that you do not have 3 independent groups. You have 2 independent
groups and 3 dependent variables, i.e. you have three analyses/tests. It
would be more common to use Mann-Whitney U tests for group comparison.

With kind regards

K.
 

GKK

New Member
#5
Hi Karabiner,

First of all Thank you so much for your time and help, much appreciated.

I am sorry if I am being a pain, but I really am finding hard to understand statistics. Hence, I am asking more and more doubts which may sound foolish.

Could you please advise me if the attachment (solution) I sent is right or wrong? Do you reckon I need to change it fully or can I stick on to that solution?

Thank you once again.

Kind Regards,

GKK
 

Karabiner

TS Contributor
#6
I think it's ok. Just mind the wording with regard
to dependent/independent groups. And be prepared
that someone might ask you why you performed the
H-test instead of the usual U-test.

By the way, if your studies include living persons
or creatures, it would on the long run be better
if you looked for support in real life rather than
in some web forum.

With kind regards

K.