Which statistical test suits better for my datas ?

#1
Hello !
I'm a student in biology, and I have a quite basic statistic question.
I need to compare the mean volume of a cellular structure in two groups (treated vs non treated), 5 subjects per group. I have hundreds of mesurements for each subject. Currently I'm just considering the mean for each subject, and I'm doing a t-test to compare the two groups, but I don't think it's the best way. Is there another test that I should consider ?

Enclosed is a set of factice data just to illustrate


THank you very much by advance !
 

Attachments

Miner

TS Contributor
#2
Why do you have hundreds of measurements per subject? Was this to reduce measurement error on a single cellular structure? Are there hundreds of cellular structures per subject? Or was it done over time (as in a repeated measures design)?

You could use a 2-way ANOVA with groups and subjects as factors allowing you to investigate the group x subject interaction. However, this may not make sense depending on the research question and why you made multiple measurements.

BTW, I am not in the biology field.
 
#3
Thank you for your answer !
It's hundreds of structures per subject. In fact, I quantified in microscope images (one image per subject) the number and the volume of these structures (synapses) on a given area. For exemple this kind of images . Each mesurement is the volume of one synase. The question is, is synapse mean volume different on the treated and control conditions ?
 

Miner

TS Contributor
#4
Given that answer, I would be inclined to use the 2-way ANOVA as described above. The group effect would answer your specific question. While the 2-sample t-test would do the same, you would not be able to test the interaction term.