Hi,
I have conducted propensity score matching using working hours as my treatment variable and the WHO-5 Well-being Index (scale from 0 to 25) as my outcome variable using the following command:
psmatch2 workinghour age worklife partner sex education countryid loginc jobsatisfaction, out(mentalhealth) logit ate no replacement
I then conducted sensitivity analysis using the command rbounds:
gen diff2= mentalhealth-_mentalhealth
rbounds diff2, gamma(1(0.05)1.5)
I have attached the output of this to the post.
Would it be correct to say that my results are robust to negative bias but not robust to positive bias?
My question is related to whether working excessive working hours leads to poor mental health and I find a significant result (negative ATT so mental health is lower for excessive working hour group). But I'm not sure what direction of bias any unobserved factors would have so I'm not sure whether to focus on the negative bias or positive bias.
I would really appreciate any help that can be provided.
I have conducted propensity score matching using working hours as my treatment variable and the WHO-5 Well-being Index (scale from 0 to 25) as my outcome variable using the following command:
psmatch2 workinghour age worklife partner sex education countryid loginc jobsatisfaction, out(mentalhealth) logit ate no replacement
I then conducted sensitivity analysis using the command rbounds:
gen diff2= mentalhealth-_mentalhealth
rbounds diff2, gamma(1(0.05)1.5)
I have attached the output of this to the post.
Would it be correct to say that my results are robust to negative bias but not robust to positive bias?
My question is related to whether working excessive working hours leads to poor mental health and I find a significant result (negative ATT so mental health is lower for excessive working hour group). But I'm not sure what direction of bias any unobserved factors would have so I'm not sure whether to focus on the negative bias or positive bias.
I would really appreciate any help that can be provided.
Attached Files
