Degrees of Freedom for ANOVA with Random and Fixed Effect

#1
Hi,

I have completed a 2x2 experiment testing the effect of level of information (high vs. low) by spokesperson gender (male female). For each condition (e.g. high info-male) there are four replications.

As much as I understand, it is recommended that replications are treated as random effects. I included replication as a random effect and included the interaction between the conditions and replication into the model.

However, the ANOVA results provide degrees of freedom with decimals for the main effects of Info amount and spokes person gender. Am I doing something wrong?

Thank you
 
Last edited: