Sign Test

• When to use

  To determine the significance of results when:

  • Researcher is looking for a difference between two groups (not a correlation)
  • A repeated measures or matched pairs design was used (not independent groups)
  • The data is nominal (or can be converted into nominal data)
• How to calculate:

  1. Allocate each participant (or pair if matched pairs) a sign: + if score improved, - if it decreased, or 0 if stayed the same.
  2. Add up number of + and number of -
  3. Take smallest number (least frequent sign). This is our calculated value, known as S
  4. Calculate number of participants - do not include anyone who's score stayed the same. This number is known as N
  5. Compare the calculated value (S) to a critical value table. S must be equal to or less than the P value to be significant.