Tweetovi

Blokirali ste korisnika/cu @pedrohcgs

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @pedrohcgs

  1. Prikvačeni tweet
    17. svi 2019.

    Here is something I do not quite follow: Why do people almost always refer to two-way fixed effects (TWFE) models as synonymous to Difference-in-Differences (DID)?? TWFE is an *estimation method*, while DID is a *research design*. These are very different things!

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    31. sij
    Odgovor korisnicima i sljedećem broju korisnika:

    In case others here were interested: The R package by Shea & Torgovitsky for "Using IVs for Inference About Policy Relevant Treatment Parameters" (Mogstad, Santos, Torgovitsky) has arrived R: Doc:

    Poništi
  3. 28. sij

    Overall, that is a really cool paper, full of new ideas that can be used in many other setups!! Gret job, and Alex! We are now waiting for your R package so we can use these tools in our papers! 9/9 end

    Prikaži ovu nit
    Poništi
  4. 28. sij

    They illustrate how these four ways of salvaging falsified models can be used in practice by analyzing two IV models:(a) one with homogeneous treatment effects (classical IV setup); (b) one with heterogeneous treatment effects. They apply these results to 4 empirical studies 8/9

    Prikaži ovu nit
    Poništi
  5. 28. sij

    4 - As a sensitivity analysis, you can present identified sets for points beyond the falsification frontier. This is interesting because, although the points on the falsification frontier are consistent with the data, this does not mean one of them is true. Cautious Optimism 7/9

    Prikaži ovu nit
    Poništi
  6. 28. sij

    3 - You can report a specific point at the falsification frontier. This is like "zooming" at particular point of the falsification frontier. Say, the point where you keep one assumption fixed at the baseline, and relax the other until the model is consistent with the data. 6/9

    Prikaži ovu nit
    Poništi
  7. 28. sij

    2 - You can report the identified set for the parameter of interest assuming that the true model lies somewhere on the falsification frontier. This is like getting the "most informative" set of parameters that are consistent with the observed data. 5/9

    Prikaži ovu nit
    Poništi
  8. 28. sij

    1 - You can report the falsification frontier. The idea of the falsification frontier is to relax the assumptions "little-by-little" until your model is consistent with the observed data. If you strengthen the assumption a bit more...bang..you're back to square one. 4/9

    Prikaži ovu nit
    Poništi
  9. 28. sij

    Matt and Alex propose four constructive answer to the "what should I do" question. I try to describe these below. 3/9

    Prikaži ovu nit
    Poništi
  10. 28. sij

    The idea of the paper is pretty cool. It goes like this. In many situations, our econometric model has some testable/refutable implications. We then move on and test them, and sometimes we reject these implications. So what should we do next? 2/9

    Prikaži ovu nit
    Poništi
  11. 28. sij

    It's late but still Monday, so here we go with the Econometrics paper of the week! This time I'd like to highlight "Salvaging Falsified Instrumental Variable Models", by Matt Masten () and Alex Poirier. Link: 1/9

    Prikaži ovu nit
    Poništi
  12. 27. sij
    Poništi
  13. proslijedio/la je Tweet
    27. sij

    Since indirectly roped me into this conversation, I'll add a few thoughts here. 1/15

    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    26. sij
    Poništi
  15. proslijedio/la je Tweet
    26. sij

    Vanessa has gained 2 beautiful angels🕊🕊 May their legacy stay alive forever💙 RIP Kobe & Gianna💙🕊💙

    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet

    Sometimes when I'm really lacking in courage because of anxiety or whatever, I will put on this video. It's the last three minutes of Kobe's last game, the greatest retirement game by any athlete ever. Everyone should watch it.

    Prikaži ovu nit
    Poništi
  17. 26. sij

    Kobe was one my first inspirations—not the tallest, not the fastest, but the one who were always working very hard, whose mentality was so strong that made him rise to a different level. Very sad to see him leave this world like this...Thoughts and prayers to his family

    Poništi
  18. 22. sij

    This paper looks very interesting. Looking forward to read it more closely! Link is here:

    Poništi
  19. 21. sij

    I find the paper very nice and well motivated. But that should not be too much of a surprise as I always learn a lot from Brant, either by co-authoring with him or just reading his papers!

    Prikaži ovu nit
    Poništi
  20. 21. sij

    Brant also makes life a bit easier for us by providing an easy-to-use package that implement his proposed tools:

    Prikaži ovu nit
    Poništi
  21. 21. sij

    Brant note that these extra conditions to get point identification can be too strong in many diff-in-diff applications. To avoid these drawbacks, he then propose alternative assumptions that you can use to *bound* these distributional parameters.

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·