the Difference between effect and const_marginal_effect

[PetrichorSjj]

I assumes that all of the comparisons will be True，why will this happen?

[kbattocchi]

At first glance that does seem surprising. Could you provide a complete repro?

[juandavidgutier]

effect()

-Estimates heterogeneous (individualized) effects.

-Calculates the treatment effect for each specific observation in X.

-The effect may vary according to the characteristics of each observation.

const_marginal_effect()

-Estimates a constant (homogeneous) effect.

-Calculates a single effect value that applies to all observations.

-Assumes that the treatment effect is the same for everyone.

[PetrichorSjj]

-Assumes that the treatment effect is the same for everyone.

actually it's not, it will return a number for every sample, and each value is not the same. For example, if I have 100 samples, it will return 100 numbers.

[PetrichorSjj]

At first glance that does seem surprising. Could you provide a complete repro?

Sorry I can't provide it because I used the data of my company. I think it's a bug though.😅