I have a list of values for a particular test and a single value from another test. How do I prove that the value for the single test is statistically significantly different or similar to the values on the list.

1) I got one suggestion that it is possible to generate a standard deviation measure for the list and then prove that the singular value is atleast two standard deviations away from my singular value of interest.

2) I got another suggestion to calculate the 5th percentile and 95th percentile of the list I have and then check to see if my value falls between the range.

I do all my stats on PRISM. I would be deeply obliged if somebody could suggest an easy way to do this comparison

  • 1
    $\begingroup$ "he said that t-tests are used to compare lists against lists and not for lists against a single value. " He is wrong, and would benefit from an introductory statistics course. The one-sample t test is precisely for comparing (the mean of) a list of values to a single value. $\endgroup$ – Alexis Feb 8 '18 at 18:13
  • $\begingroup$ Suggestion 2, if applied to results of two identical tests, would tell you there is a "significant" difference 10% of the time. Would you really trust the result of such a procedure? Suggestion 1 is less predictable but in many circumstances would have even worse performance. $\endgroup$ – whuber Apr 2 '18 at 15:13
  • $\begingroup$ @whuber What do you suggest. Do I go with a one sample t-test. $\endgroup$ – The Last Word Apr 2 '18 at 15:27

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.