Paired vs unpaired t test


Following is my experimental condition -

I have chronically treated normal cell line with smoke condensate for 12 months. Over the time the cells have changed their morphology and characteristic. They have transformed into cancerous cells - the cells morphology has changes (shape/size), they have gained all the characteristic of cancerous cells. To characterize the changes I have carried out genomic and proteomics analysis. But I am confused now - whether to use paired or unpaired t test to compare the parental with cancerous form of the cells. Could you please tell me whether to use paired or unpaired analysis. I am inclined towards unpaired as over the continuous exposure of cells to smoke for 12 months the characteristic properties (which define cells as normal/cancerous) have changed.

Thank you for your time.


Well-Known Member
When you use dependent samples you use paired t. For example you compare the same person before treatment and after treatment. Can you paste some data for better understanding?


TS Contributor
Can you better describe your research question with highly specific details? What is the dependent/Y variable of interest? What are the independent variables? Do you have only before and after measurements at time zero and the final time? Do you have multiple measurements over the time period between time zero and final time (such as 3 month intervals)?