Simple effects to understand interaction between repeated variables in mixed ANOVA

#1
Hi there,
I am having problems with the statistical analysis for a linguistics paper I am writing. So this study is a mixed ANOVA design with 3 independent variables:

Within Subjects Variables:
Noun_Type (Count vs Mass)
Plural_Morphology (Singular vs Plural)

Between Subjects Variable:
Language_Background (English, Chinese, Spanish)

Dependent variable was response accuracy. There were several significant effects. Of interest here is a significant interaction between the two Within-Subjects variables: Noun_Type * Plural_Morphology. Please see this interaction graph:



Based on this graph, it seems that while there is a strong effect of Plural_Morphology at Count (Level 1 of Noun_Type), there is no effect at all of Plural_Morphology at Mass (Level 2 of Noun_Type). So, I decided to carry out a simple effect analysis, to look at the effect of Plural_Morphology at each level of Noun_Type. I used the following SPSS syntax:

MANOVA
CountSingular CountPlural MassSingular MassPlural
/WSFACTORS Noun_Type (2) Plural_Morphology (2)
/WSDESIGN=MWITHIN Noun_Type (1) MWITHIN Noun_Type (2)
/Print
SIGNIF(UNIV MULTIV AVERF HF GG)


Please let me know if the above syntax is erroneous in any way, because somehow I got the Output (see below) apparently showing that there is a significant effect of Plural_Morphology at both levels of Noun_Type!, which is strange, given the interaction graph.

What went wrong?


Thanks!!!

 
Last edited: