Picking the right statistical test

Hi all,

My data is structured as follows - it's all means:

Urban/Rural - nested within Urban/Rural are White-Nonwhite... I don't want to compare Urban White against Rural White, but instead Urban White against Urban Nonwhite, and Rural White against rural Nonwhite.

Is there a better way other than doing a bunch of T tests?

This is a forum mainly for questions about R, not a stats advice forum.

If it is at all possible you need to consult with a statistician one--on--one. They can discuss what the research questions are, any assumptions you are making and so on and possibly do some basic exploratory analysis.

In a pinch, you are probably much better off asking for advice on https://stats.stackexchange.com/ which is a stats advice forum.

Probably there are any number of things that are better than t-tests but it is hard to say what without a lot more information.

Appreciate it! I'll ask there, thanks!

As a simple thought Bayesian Hierarchical or multilevel models might be a good approach but I am just trying to master the simplest Bayesian principles so I am no help.

See Bayesian multilevel models | Stata and A Primer on Bayesian Methods for Multilevel Modeling — PyMC3 3.10.0 documentation for a bit of information on them. Again, you need a statistician if at all possible. If you are at a university or work for a large organization they should have a stats consultant stashed in a back room somewhere. University departments such as Sociology, Psychology, and Political Science will normally have faculty with a good background in statistics.

Oops, that A Primer on Bayesian Methods for Multilevel Modeling — PyMC3 3.10.0 documentation is Python. I think this video gives a bit of an intro to tho subjert -- in R An introduction to Bayesian multilevel modeling with brms - YouTube

Just a clarification about this, it is true that this forum is mainly R focused but the forum scope is data science in general so it is completely fine to make a non programming stats question, although, as Jhon said, most likely your would have better luck asking on a specialized forum.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.