I know there is Friedman's ANOVA for dependent variables, but this it does not include between-subject factors. I am looking for a non-parametric analysis method that can be applied to a mixed experimental design (one within-subject factor, two levels; one between-subject factor, three levels).
You could set this up as a resampling or a permutation test. See
http://hbanaszak.mjr.uw.edu.pl/TempTxt/Good_2005_Resampling_Methods__A_Practical_Guide_to_Data_Analysis.pdf
see pages 90 onwards especially.
and see http://hbanaszak.mjr.uw.edu.pl/TempTxt/Good_2005_Resampling_Methods__A_Practical_Guide_to_Data_Analysis.pdf
especially p 138 onwards.
In essence, resampling methods involve shuffling data and comparing the actual data obtained with shuffled data to get a probability of obtaining the data you actually got. How you shuffle the data depends on whether it is paired or not. Once you understand the principles you can analyse complicated configurations of data, paired non paired whether it is stratified in layers or not. You need to learn the principles first and then you can design tests for your own configurations of data.
see
http://www.resample.com/excel/available-online/
look at the user manual for how to set up stratified data in order to do a permutation test.
You could also have a look at the excellent book by Rand Wilcox - Introduction to robust estimation and hypothesis testing (3. ed.), chapter 8.6
He also provides R codes for many applications.
http://www.sciencedirect.com/science/book/9780123869838
you follow the path in spss:
analyze....non parametric tests.....lagacy dialogs....k independent sample
Yes, it is Friedman test. I recently used it.
This book explains it nicely.
Mayers, A. (2013). Introduction to Statistics and SPSS in Psychology. Pearson.
Generalized Estimating Equation (GEE) assumes a linear distribution. I don't think is the right solution to analyze non-linearly distributed data. Kruskal-Wallis H test is the equivalent of a one-way ANOVA, hence it is going to compare the two independent groups but can't account for the within subject factors. I still can't find a solution on SPSS but I will let you know as soon as I find the answer!
nparLD routines in R. https://cran.r-project.org/web/packages/nparLD/index.html
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA