Accounting for other variables when looking for main effects (regression)

I am doing a negative binomial regression, to look for interactions between categorical and continuous predictors. I am currently reporting main effects for both predictors, but am wondering whether I should account for the other predictor when doing so.

So, if lang2 & BDCSBI are the predictors and sdq_conduct is the outcome, should my code look like:
svy: nbreg sdq_conduct lang2 BDCSBI
svy: nbreg sdq_conduct lang2

So effectively, should I include both predictors in the code when searching for the main effect of just one?

Also, what is the difference between simple and main effects?
Thanks for any help.
Not interested in any interaction effects? Main effect - the effect of an IV on the DV ignoring all other IVs. Following an interaction, a simple effect is the effect of an IV on a particular level of another IV.


Less is more. Stay pure. Stay poor.
If you have significant interaction terms in your model, you no longer look at the main effects of the terms that are in the interaction. This is because those main effects are conditional on another term. Multiplicative interaction term model should look like:

Y = X1 + X2 + X1*X2