1

Suppose that I have two or more groups of data (e.g. two or more cities with 10 sampled population each) and I want to test if they behave differently for a variable (e.g. specific crime) through the years.

The obvious parametric choice is one way ANOVA for repeated measures, but if the assumptions (normality, outliers) are not satisfied then what is the non parametric equivalent? Friedman test does not take into account more than one group (e.g. city) concurrently.

amoeba
  • 93,463
  • 28
  • 275
  • 317
AchiPapakon
  • 80
  • 5
  • "two or more cities with 10 sampled population each" - what would that mean? You measure your outcome ("specific crime") for each of the 10 people in each city? And these 10 people are then tracked over the years? – amoeba Feb 08 '18 at 10:12
  • Yes, 10 people from each city that satisfy some criteria. I "look back" in time, not in the future, but statistically it's the same. – AchiPapakon Feb 08 '18 at 11:06
  • OK. Anyway, this is not a one-way RM-ANOVA, this is mixed design ANOVA with one RM factor and one between-subject factor. – amoeba Feb 08 '18 at 11:13

0 Answers0