1
$\begingroup$

for a university project I'm doing a literature review and I have a dataset where each line represents one paper and then I scored the data for two variables which each come in three levels: level of uptake where each paper can have either no uptake, desired uptake, or actual uptake and level of empowerment where each paper can either have no empowerment, some empowerment, or strong empowerment (not what they're actually called but you get the idea).

Now I want to find out if papers with e.g. a high level of uptake are also more likely to have a high level of empowerment. So I'm guessing I'd have to transform my three categories into numbers 1-3 and then do some kind of correlation test/test of independence? But I'm very unsure which one would be the right one and how to go about it. After a lot of time spent on google to try and figure it out, I'm now more confused than I was before. Mostly because every post I find about similar cases says something different and I lack the statistical knowledge to know what applies to my case and what doesn't. So, if anyone could help me figure out how to do something like this in R (keeping in mind that I'm a complete novice to R and statistics in general) I'd be eternally grateful!

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ I concur with what is said, if your question is more approach/methods related, SO might not be right place. .... If you look at it from a practical side: you mentioned already to code your levels as numbers or categories. With this you can construct combinations (e.g. no uptake / no empowerment, no uptake / strong empowerment). Doing a basic count (aka histogram) over these combinations gives you an idea of their "occurrences"/frequencies. That is a very basic measure ... but could be a first step to get a feel for your (paper) data set. Good luck! $\endgroup$
    – Ray
    Commented Jun 25, 2021 at 17:47
  • $\begingroup$ This is exactly what a Chi-square test for independence is designed for. For a more complete answer, ask on the stats site. $\endgroup$
    – user2554330
    Commented Jun 25, 2021 at 19:07

1 Answer 1

Reset to default
1
$\begingroup$

You have categorical variables, so treat them as such. Build a contingency table and use a chi-squared test of independence. That will treat the two variables as nominal, that is, there is no ordering relationship between the levels.

But one of your variables empowerment could be seen as ordinal. If you want to treat it like that, see Correlation coefficient between a (non-dichotomous) nominal variable and a numeric (interval) or an ordinal variable which proposes ordinal regression.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.