Question: What sample size do I take in a thesis to determine the correlation between 2 variables?
Answer: There’s not enough information here to answer. You can find correlation when N = 3. The bigger the sample, the more accurate your estimate will be. To say more, we’d need to know the approximate correlation you are going to find and how accurate you need your estimate to be.
If you have N = 2, the correlation will automatically be 1.0. That’s not useful. But with N = 3, you can run a correlation. However, the confidence interval around the correlation will be very wide and the correlation is likely to be close to 1 or -1, even with random data.