I have two methods of processing certain data, where one is a more advanced version of the other. After processing the data, the output is scored from -1 to 1. I have a set of 7 distinct data points that I can process and score.

I wish to see if these methods of processing the data significantly differ in score from one another. I therefore thought to turn to the paired sample t-test. However, I am unsure if my case fits the criteria for doing such a test. According to https://statistics.laerd.com/stata-tutorials/paired-t-test-using-stata.php, there are 4 assumptions that need to hold. Assumption 1 is true, since the score is a continuous interval. For assumptions 3 and 4, I have tested my data and the scoring, and found that they are roughly true.

Assumption 2 is where I need clarification. I think this assumption holds as well, since the calculations I make on the data are independent (the computer obviously does not change or take into account how I got my output when performing the scoring). Am I therefore "clear" to perform a paired sample t-test on this data, or am I missing something?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.