New answers tagged interaction
2
votes
Accepted
How to interpret dummy variables and interactions terms on dummy variables in a regression?
Parameter interpretations come from parameter equations (from the true regression)
The true regression function in this model is:
$$u(x_1,x_2,x_3,z) \equiv \mathbb{E}(\log Y|x_1,x_2,x_3,z) = \beta_0 + ...
- 130k
0
votes
Ceiling effect in mediation AND/OR moderation
I don't think that having the results in hand and rewiring a post-hoc solution to the problem should drive how you construct models. You were already testing a model, now you have to deal with that ...
- 16.1k
1
vote
Accepted
Should I Include Visit as Both a Fixed Effect and Random Effect in a Longitudinal Mixed-Effects Model with Interactions?
One issue in your code is that the way you have tx and menscat variables created, they end up perfectly correlated. I adjusted that bit of code, below, with a few other changes:
...
- 5,525
1
vote
Interpretation of interaction results when regression coefficients are significant but average marginal effects are not
It's extremely hard to interpret individual regression coefficients when there are interactions. With the default treatment/dummy coding in R, the model summary() ...
- 97.5k
0
votes
Can I use simple slopes with AFT models?
There should be no problem with using simple slopes (outcome estimates as a function of a continuous predictor at specified levels of an interacting predictor). You just have to use appropriate post-...
- 97.5k
1
vote
Two way ancova violation: Do I have to include significant interaction term?
ANCOVA is just a particular form of a linear multiple regression model. There's no reason to limit your analysis to specific rules classically associated with ANCOVA.
If you have enough data then it's ...
- 97.5k
1
vote
interactions in non-negative least squares
First, binning a continuous predictor is not a good idea. You shouldn't break up your g variable that way.
Second, non-negative least squares makes sense if the ...
- 97.5k
0
votes
How do I interpret this non-significant simple moderation analysis?
It means that there is insufficient evidence that the effect of X depends on M, or that X’s effect on Y varies across levels of M.
- 1,113
2
votes
Accepted
Is it okay to have two separate interactions with the same variable in one model?
In theory, there should be no problem with this. It's possible that it's overfit, but this depends on your data - you can use information criteria or other model comparison techniques to evaluate if ...
- 343
2
votes
Splines of interactions vs interaction of splines
A simple multiplication is only one form of interaction. Forming splines of such oversimplified expressions often don’t fit the data and are dependent on the variables’ measurement origin or of how ...
- 96.4k
0
votes
Splines of interactions vs interaction of splines
Probably a typo but you have y on both sides of the equation. That will not work well. Beyond that:
I don't think it's good to classify these as "correct" or "incorrect"; I'd use &...
- 125k
0
votes
Hierarchical logistic regression interaction
Besides the coefficients for x and z and their interaction, your HLM software presumably can return the corresponding variance-...
- 97.5k
Top 50 recent answers are included