Add new comment

Hi Roos! I'm glad the study notes helped you understand the topics :) Regarding your question: I think the covariate has to be continous (interval or ratio level) because the slope has to represent actual ordered numbers to be able to identify the relationship between the covariate and the dependent variable. However, I found on internet that a covariate can be categorical as well, but then the analysis method you use is not an ANCOVA:

https://statistics.laerd.com/spss-tutorials/ancova-using-spss-statistics.php#:~:text=Note%3A%20You%20can%20have%20more,is%20not%20often%20called%20ANCOVA.&text=If%20you%20have%20two%20independent,run%20a%20two%2Dway%20ANCOVA.) 

I hope this answers you question!