In: Statistics and Probability
When doing a multi-factor ANOVA, why do you first check the "corrected model" p-value before checking the p-value for the interaction and main effects? Why do you check the single p-value for the entire model before checking multiple p-values?
When you analyze a factorial design, you are testing for the main effect of each of your independent variables as well as the interaction between those independent variables. In this case, you are testing for the main effect of Commitment (whether ratings of the target differed based on the subjects’ level of commitment to their current relationship, ignoring the effects of target attractiveness), the main effect of Attractiveness (whether ratings of the target differed based on the target’s attractiveness, ignoring the effects of subjects’ commitment to their current relationship), and the interaction between Commitment and Attractiveness (whether the effects of one factor are different depending on the level of the other factor).
General Linear Model: Univariate
To get started, select Analyze → General Linear Model → Univariate.
The “general linear model” is a broad category of analysis tools that includes the tool we want: factorial ANOVA (ANalysis Of VAriance). “Univariate” refers to how many dependent variables we are looking at: one.
Put “rating” into the “Dependent Variable” box. Put "Commitment" and “Attractiveness” into the “Fixed Factors” box.
A factor is another name for an independent variable, especially one that is nominal-scale (categorical). A fixed factor is a factor that takes on only a small number of values that represent levels that the researcher is interested in comparing. Random factors are complicated and beyond the scope of this course.