Is independent t-test still justified?

I am kinda stuck and undecided whether I can use the independent samples t-test in a specific case I'm about to describe, so I'm hoping for some help.
The study involves 39 patients, all of whom have undergone a biopsy. 3 of them have undergone a different biopsy (different lesion, location, referral etc), also included in the study. So in the end we have 42 biopsies made in 39 patients. I am asking whether it is ok to treat the 42 procedures as independent (which I think they are, as they are not repeated, there is no before-after component or anything), and use the independent samples t-test, for instance? If not, what would be the best thing to use or to do?

thank you for your help!


TS Contributor
I am asking whether it is ok to treat the 42 procedures as independent
No, it isn't. Procedures undertaken on the same subject are not independent from
each other.
use the independent samples t-test,
You did not state your research question, or what groups you would like
to compare, or what your dependent measure looks like. Therefore your
question cannot be answered.

With kind regards

thanks for your reply!
I am trying to compare continous variables, for instance the amount of radiation received, between certain groups, in case of a t-test, 2 groups (like groups 1: soft tissue locations of biopsy and group 2: bony locations).

Is there a test I could use for such purposes thaat would let me include all procedures despite 3 of them being done on same indivduals? Or is the only option to omit the repeated ones?


Omega Contributor
Everything points towards excluding the 3 biopsies. You could try to do cluster analysis, which is complicated, but you would only have three individuals with greater than 1 biopsy, and this would not be great at all.

It is easy for us to say exclude them, I realize this. But to outsiders it seems like an easy choice, other option would be to not exclude them and have others point this out to you later as well.