Parametric vs nonparametric? It’s the age old question, because many often wonder which statistical test to use for their data analysis. To begin, first determine if your data fits into the parametric or nonparametric categories.

Parametric

Parametric procedures include ANOVAs, independent t-tests, and paired t-tests. Each parametric procedure requires explicit assumptions on the populations, and the parametric procedures use the mean to determine significance. In addition, assumptions of parametric tests include the following as basic requirements:

  • A normal population distribution, and
  • Equal variances within groups

Nonparametric

Nonparametric procedures include the Mann-Whitney, Mood’s Median, and Wilcoxon tests. These nonparametric procedures make limited assumptions of populations, and they use the median to determine significance. Nonparametric procedures are categorized by:

  • A population distribution that is not normal

References

Austin, E. (2005). Handbook of parametric and nonparametric statistical procedures. British Journal of Mathematical & Statistical Psychology, 58, 382.

Harwell, M. R. (1988). Choosing between parametric and nonparametric tests. Journal of Counseling and Development: JCD, 67(1), 35.

Norusis, M.  (2004). SPSS 12.0: Guide to Data Analysis. Upper Saddle River, NJ: Prentice Hall, Inc.