IV Regression

[Natalia Smith]

Hello everyone. May I please have some assistance on what command I should use for my IV regression?

These are my variables:

y1 - primary binary dependent variable
y2 - endogenous ordered categorical variable (rating scale)
z - binary instrumental variable
x1 to x10 - a mix of explanatory and control variables

So basically the regressions would look something like this:

• First Stage: reg y2 z x1 x2 x3 x4 x5 x6 x7 x8 x9 x10
  • predict y2hat
• Second Stage: reg y1 y2hat x1 x2 x3 x4 x5 x6 x7 x8 x9 x10

I just want clarity about whether I should use ivregress 2sls, ivprobit, or something else. May you also explain how to interpret the results for whichever command I should use?

Thank you,

Natalia

[Natalia Smith]

Any bit of information will help.

[Muhammad Ramzan]

Hi,

you need to apply ivprobit as your dependent variable is binary.

Thanks

[Muhammad Ramzan]

For interpretation you need to compute the mfx and interpret them please see any published paper, that have applied the inprobit thanks

[Nick Cox]

Reposted at https://www.statalist.org/forums/for...-iv-regression