what do non-significant treatment effects in PSM estimates imply? How to check for sensitivity of the result?

[ram shrestha]

I am running Propensity Score Matching (PSM) to get treatment effects. But, my results are not statistically significant.
1) Does this imply my results are affected unobserved confounders?

I checked for the sensitivity analysis by using mhbounds (Becker and Calenado 2006) [Sensitivity analysis for average treatment effects] for hidden bias.
However, the results are not robust even for the value T (tau)= 1.
2) Is my analysis has problem? How do I check for sensitivity of the analysis?

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a helpful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

Results that are not statistically significant can mean many things - your variables really don't associate with the dv, your model fit due to omitted variables is so poor that you have a large error variance, outliers in the dv, specification problems of all sorts, etc.

I don't use mhbounds so I can't help you with that.