Correction needed for an non-parametric Anova ?

#1
Hello !

I'm trying to analyze the behavioral data of my research experiment and I'm a bit lost... o_O
Briefly, subjects undergo 4 different conditions (of increasing complexity 1<2<3<4).
At the end of each conditions, they perform a little test.
These are the results :
Code:
.    Cond 1    Cond 2    Cond 3    Cond 4
Subject 01    96    100    88    80
Subject 02    100    100    92    72
Subject 03    100    100    80    72
Subject 04    100    100    92    76
Subject 05    96    100    80    68
Subject 06    100    96    92    76
Subject 07    100    100    96    92
Subject 08    100    96    100    72
Subject 09    100    96    92    84
Subject 10    100    88    88    72
Subject 11    100    92    100    84
Subject 12    100    96    92    76
Subject 13    100    100    88    80
Subject 14    100    96    88    80
Subject 15    100    96    88    76
I want use a statistical test to affirm when the difficulty increases, scores decrease.
I think I should use a non-parametrical ANOVA (Friedman).
But, the results of condition 1 and 2 caps at 95-100.
Do I need to perform any correction before applying the Friedman anova ?
Someone told me about a ArcSin correction but I don't know if it's valid or not...


Thanks a lot for reading ! :)


Finn'
 

Karabiner

TS Contributor
#2
Friedman is a test based on ranks. So what sense would any transformation make?

But even if a transformation would make sense here (say, if you performed a parametric repeated measures ANOVA), it would not solve the problem that you have a massive ceiling effect.

Friedman provides no tests of your a>b>... hypothesis. Why don't you use repeated measures Anova with planned contrasts?

With kind regards

PonderStibbons