+ Reply to Thread
Results 1 to 2 of 2

Thread: Instrumental variable apporach: two stage least squares

  1. #1
    Points: 6, Level: 1
    Level completed: 11%, Points required for next Level: 44

    Thanked 0 Times in 0 Posts

    Instrumental variable apporach: two stage least squares

    Hello everyone,

    I want to test cross sectional data with a household survey with 60k plus of respondents (however missing data result in overall data loss). The survey asks questions that answer either 0 or 1 or are in likert scale. I already estimated a probit model.

    My simplified model would be:

    y = X1 + X2 + X3 + controls + e

    where X1, X2 and X3 are sociability measures but can all be considered endogenous.

    Therefore, I would like to estimate the model using instrumental variable approach. I have found (possibly weak) instruments, but still want to continue. I have several questions:

    1) Should I instrument all my variables X1, X2 and X3 or should I instrument for example only X1?
    2) Should I leave the original model mostly intact or is it ok to leave out some control variables (to minimize further data loss)?
    3) Should I standardize all variables and then run the IV model or is there some kind of probit version of instrumental variable approach I should use?

    All the best,

  2. #2
    Omega Contributor
    Points: 38,396, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Not Ames, IA
    Thanked 1,186 Times in 1,147 Posts

    Re: Instrumental variable apporach: two stage least squares

    Are you able to graph you model and the suspected relationships? Are the control variables confounder or on the propose causal path? Have you done instrumental variable analysis before and understand the logic behind it?
    Stop cowardice, ban guns!

+ Reply to Thread


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts

Advertise on Talk Stats