Tests for Discrete Variables?

[LDubs]

I have a basic stats question, followed by a basic SAS question 🙂

I'm working my way through analyzing data for an assignment. For my dependent variable(motor and cognitive recovery post-stroke), I have data from a questionnaire, with possible scores between 18 and 126. My independent variables (positive affect and social support) are also rated on scales of 0 to 12 and 11 to 55.

I'm struggling with what type of tests to run, because I'm not sure what to consider the data.

Are they continuous, discrete, or could they be considered ordinal since the higher the scale, the more positive the result?

I might be overthinking, but any help is appreciated.

My second question is, what tests work best for these type of data?

[Rick_SAS · Accepted solution]

I think the variables can be considered continuous. Although the person filling out the survey will probably use integers, values like 8.5 or 12.3 make perfect sense for these data. It's a bit like measuring weights to nearest pound or kilo: most people report integer values although the scale is continuous.

As to "what tests work best," that depends on your research question. Are you trying to predict or model values of the dependent variable based on values of the indep variables? If so, consider starting with a linear model, perhaps by using PROC GLM.

[LDubs]

Thank you for your response!  That is exactly what I'm hoping to do, so a PROC GLM it is.