Hello this is my first post here.

So I am making a model to try and show the asymmetric effect in my independent dummy variables but with the model being linear obviously the coefficients are just mirrors of each other. So with my limited econometric knowledge beyond your typical inference and linear model these are my best attempts at rectifying my issue and I need your judgment on whether they are of any use or not.

Firstly PS is the dummy variable that I am attempting to find an asymmetric effect for. My first idea is to treat it as a continuous variable and find the marginal effect when PS=1,0.5 and 0. My plan is to use the standard errors instead of the standard deviations (due to standard deviation being identical for PS=0 and PS=1 seeming that they are the same variable.) and multiply these against the margins to provide a consistent scale. I would then find the difference between the scaled margins of PS=1 and PS=0.5 and then subtract the difference of PS=0.5 and PS=0 to determine if there is an asymmetric effect. i.e if <0 PS=0 is weighted more than PS=1, if it is >0 then the reverse is the case, and finally if the result of this is 0 then there is no asymmetric effect.


Because of the issues I faced trying to upload everything I have attached further explanations and regressions to this pdf file. sorry for any inconvenience.

Thanks for the support.