ANCOVA Question

Question

Do you have to use ANCOVA for treatments? Can I use them for public data? For example... DV=School performance scores, IV = High/Med/Low Group Rates of drug use by county(groups, not continuous rates), CV = Free and reduced lunch rates as a measure of poverty. Even through I just have the data reported through the dept of health on drug use can I use that at the treatment/DV?

Answer 1

Hi Kgd,

 

You don't need to use ANCOVA for treatments - you can certainly use ANOVA and completely ignore the presence of your CV. The benefit of ANCOVA is that it gives finer resolution when trying to detect treatment effects. Whenever ANOVA supports a treatment effect, so will ANCOVA, but there are instances where ANCOVA supports a treatment effect but not ANOVA. This would happen if some portion of the overall variability is coming from the CV.

 

The treatment is the same thing as the independent variable when you talk about ANOVA or ANCOVA, so I'm not entirely sure what the second part of your question is. If you want to take drug use as the DV, you cannot use ANOVA or ANCOVA because both of these methods require a continous or approximately continuous dependent variable.

 

I have no idea how your data were measured or what school performance score is, but my intuition tells me that higher rates of drug use would be associated with lower school performance score, but lower school performance score would not necessarily indicate higher rates of drug use. In other words, my intuition tells me drug use probably shouldn't be a DV even if it were measured continuously.