Search results

You can't treat the same team as "different" teams because they're correlated and mess up your analyses. You'd need to do some type of longitudinal modelling line time series or mixed-effects models to account for the repeated instances over time.

Have you considered posting on cross-validated? https://stats.stackexchange.com/ Sometimes questions that aren't answered here get an answer there!

Well... you need to use the quantile function or inverse CDF of the chi-square distribution for that and I don't think there are closed-form expressions for it because of the gamma functions involved.

Let's do this step by step. We'll start with something simple and build towards a full HLM. First, you have a few important-yet-ambiguous statements that need more precise definitions. What does "general characteristics of the attachment", "predicts the burnout risk" and "higher the attachment...

Perfect. That means my "help" had the intended effect.

This seems very relevant to these questions Particularly to Question 10.

You mean there is no model in the Hayes PROCESS macro that matches yours?

Can you post the matrix on here?

For models like yours the only way to approximate power is through a computer simulation. This can get you started: https://www.tandfonline.com/doi/abs/10.1207/S15328007SEM0904_8

What's with the groupings in the "Allowed" plots? Do you have a categorical predictor there?

Well, the Harvard Data Science Review is a good place to start: https://hdsr.mitpress.mit.edu/

I'm sorry but what is the question here? What you're reporting seems pretty normal with respect to how power works.

You're running into a combinatorial explosion problem, which is very typical of discrete mathematics. But it shouldn't be too difficult to simplify the expression to something you can evaluate. If you're unfamiliar with how to simplify factorial ratios, here's something to get you started:

This seems useful and relevant to your question

Definitely heteroskedastic. Is your dependent variable discrete?

Thank you. I was a tad bit surprised that there would be a proc for power in the case of generalIZED linear models (glims) as opposed to generaL linear models (glm). Anyhoo, with this out of the way the answer is simple: there is no such proc because there are no closed-form expressions for the...

Do you mean "glm power" or "glim power"? Believe it or not, the "i" can make a world of difference here...

Ok, this makes more sense now. Sure, the p-values should change because you're messing the the standard errors in one analysis (using the robust options) and in the other you're letting them be what they are. And you have a small number of clusters (11) which implies a small sample at the...

This is to be expected because both approaches are doing different things to the data. But different regression coefficients? Nah, that seems buggy. Although...wait. When you said "I re-ran the model using robust standard errors" does that mean you ran OLS regression with robust/sandwich...

Couple of things to keep in mind: The Spearman correlation captures a very specific type of association, monotonic. So a small rank correlation does not imply a lack of relationship. If you have two variables X and Y and define Y = X^2 you can see that the rank correlation is very small (it's 0...