T-test

T-test – statistical analysis that compares the means of two sample sets to determine statistical difference between them.

• T-tests are parametric, meaning they rely on certain assumptions:
  • The data is independent
  • The data is normally distributed
  • The data sets have similar variance
• The type of t-test needed must be selected based on the data itself:
  • Paired (data sets come from a single population)
  • Two-sample (data sets come from two different populations)
  • One-sample (data from a single population is compared to a standard)
• T-test tails must also be selected based on the desired outcome of analysis:
  • Two tailed test should be used to determine whether two populations are different.
  • One tailed test should be used to determine whether one mean is greater than the other.