nonparametric equivalent for a between subjects ANOVA

[dwatts]

Hello,


I am currently conducting analysis for a 2x2 independent ANOVA with each independent variable having two levels. I need to conduct a nonparametric equivalent as I have a significant result within the levene's test.
I have tried a Kruskall-Wallis, however, this only works, as far as I have researched and tested on SPSS ,with a one-way independent ANOVA.
I have two grouping variables to define each with two levels, however, nowhere does SPSS permit me to do this.

Any help would be greatly appreciated.

[Dragan]

Hello,


I am currently conducting analysis for a 2x2 independent ANOVA with each independent variable having two levels. I need to conduct a nonparametric equivalent as I have a significant result within the levene's test.
I have tried a Kruskall-Wallis, however, this only works, as far as I have researched and tested on SPSS ,with a one-way independent ANOVA.
I have two grouping variables to define each with two levels, however, nowhere does SPSS permit me to do this.

Any help would be greatly appreciated.

In the context of a 2x2 ANOVA it is legitimate to use the ranks of the original data points and apply the usual OLS procedure on the ranks. This is known as a Rank Tranformation (RT) ---Conover and Iman (1981).

Note: that this is only legit for a 2x2 design and not beyond this.

[dwatts]

Im not entirely sure if that would work, as I want to see whether there is an interaction between the independent variables. I have conducted a rubber hand illusion with varying coloured arm stimuli.

1st independent variable: (light rubber hand vs. dark rubber hand)
2nd independent variable: (synchronous brush stroking vs. asynchronous brush stroking)

The dependent variable is how strongly they feel the illusion. Converting my data into ranks still inhibits me from seeing how the interaction works in nonparametric terms.