Interpretation of elasticity for a binary or dummy variable in a logit model

[Gebretsadik Shibre]

Dear Statalisters,

I would like to kindly ask your assistance in the interpretation of an elasticity of a dummy variable in a logit model.

global X mat_ageQ2 residenceQ2 reegQ2 reegQ3 reegQ4 reegQ5 reegQ6 reegQ7 reegQ8 reegQ9 reegQ10 reegQ11 education_partnQ2 education_partnQ3 education_partnQ4 mediQ2 mediQ3 mediQ4 weaalthQ2 weaalthQ3 weaalthQ4 weaalthQ5 mom_occuQ2 part_occQ2 surveyyyQ2 surveyyyQ3 surveyyyQ4 educatt_momQ2 educatt_momQ3 educatt_momQ4 religion_poQ2 religion_poQ3 religion_poQ4 swp_decc2 swp_decc3 wanted2 wanted3 

qui sum anc_quality_binary [aw=weight_b] 
sca m_y=r(mean)

svy: qui glm anc_quality_binary $X, family(binomial) link(logit) 
qui margins , dydx(*) post

foreach x of varlist $X {
sca b_`x'=_b[`x']

}
foreach x of varlist $X {
qui {

sum `x' [aw=weight_b]
sca m_`x'=r(mean)
sca elas_`x' = (b_`x'*m_`x')/m_y                                                 
}

The output is too much. I just need interpretation of one of the regressors: rural setting compared to the urban area.
The elasticity of rural area is 0.23. Urban is reference group. You can also comment on whether it is an elasticity or semi-elasticity.

Thank you in advance.

[Carlo Lazzaro]

Gebretsadik:
this kind of queries are way too unclear to be replied positively.
please read/act on the FAQ before/upon your posting. Thanks.

[Gebretsadik Shibre]

Thank you. I have just updated the previous post.