Matt Graham

@Matt__Graham

PhD candidate, Yale political science. Onetime policy analyst in immigration () and education ().

Brooklyn & New Haven
Vrijeme pridruživanja: lipanj 2011.

Tweetovi

Blokirali ste korisnika/cu @Matt__Graham

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

  1. Prikvačeni tweet
    9. sij

    POLL WATCHERS: Ever wondered if people were guessing on a survey question? Here's a new rule of thumb: "we don't know" means "they're not sure." Even when people who offer an opinion aren't asked their confidence level, you can still learn about it. [1/9]

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    31. sij

    Interested in measuring conspiracy beliefs? New paper out at POQ with grad students and Brian Sullivan. We argue common measures inflate conspiracy belief and we propose a better measure. Preprint: 1/6

    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    30. sij

    Folks have noted this disturbing statement on Harvard's website, which must be a violation of journal policies "Scholarly journal editors tell us that...PPR reviews...will often speed up publication decisions and in some cases even skip the journal peer review process entirely."

    Prikaži ovu nit
    Poništi
  4. 29. sij

    Real quote... the extraordinary vanity it takes to repeatedly claim historical firsts, without any effort to engage the historical record, runs strong in the First Family.

    Poništi
  5. 29. sij

    This is a great conference! Everyone apply!

    Poništi
  6. proslijedio/la je Tweet
    24. sij

    I also said that competence breeds enjoyment. We're taught to believe enjoyment breeds competence (I "like" reading so I read a lot). It's often the other way round. The more you read the more competent you are, and the more you can enjoy it. Thx for coming to my TED talk. 4/4

    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    19. sij

    Shocked but probably shouldn’t be by POTUS statement—“who the hell cares about the budget? We are going to have a country.” Sure a country built on debt like his failed casinos.

    Poništi
  8. 14. sij

    This report is the most even-handed take on immigrants' use of welfare I've ever seen. Bravo. Both sides have facts to support their case. Values largely determine one's preferred slice of the data. cc:

    Poništi
  9. 11. sij

    Less-known candidates should note this new research: You have more room to make an impression than your rivals. Tell the world (& donors) your ceiling is high!

    Poništi
  10. proslijedio/la je Tweet
    11. sij

    Thoughtful supporters of democratic revolution in Iran would not support maximum pressure, airstrikes and other aggressive tactics, which historically heighten nationalist support for Incumbent governments.

    Poništi
  11. proslijedio/la je Tweet
    9. sij

    Very cool new work from -- extent of don't know responses correlates quite well with the average certainty level of the DO know responses!

    Poništi
  12. 9. sij

    In summary: new research shows that when a lot of people are say DK, it's a sign that EVERYONE is relatively unsure of what they think about the question --- even the people who ended up answering it! [9/9] Check out the paper:

    Prikaži ovu nit
    Poništi
  13. 9. sij

    To explain why this happens, the paper proposes a new model of DK responding. It's a simplification, but it's consistent with many existing findings on variation in the percent DK between questions, question formats, and survey modes. [8/9]

    Prikaži ovu nit
    Poništi
  14. 9. sij

    This can change how we interpret public opinion polls. Candidates who are less-known don't just have fewer people offering an opinion. Those who offer an opinion are also less sure of what they think. [7/9]

    Prikaži ovu nit
    Poništi
  15. 9. sij

    For example, the % DK explains more than 90% of the variation in average confidence in evaluations different 2020 candidates. (Data from August 2019; candidates selected for variance in % DK, not just top contenders.) [6/9]

    Prikaži ovu nit
    Poništi
  16. 9. sij

    Fortunately, this liability is also an asset. If you account for the fact that different types of questions produce different types of confidence distributions, the heuristic becomes much more accurate. [5/9]

    Prikaži ovu nit
    Poništi
  17. 9. sij

    For example, questions about public figures tend to produce fairly bipolar "know it or don't" distributions (top row), while questions about economic statistics produce flatter distributions that reflect high rates of guessing (bottom row). [4/9]

    Prikaži ovu nit
    Poništi
  18. 9. sij

    A proxy that explains 50 percent of what you care about is good but imperfect. A key source of error is that different types of questions produce different-shaped confidence distributions. [3/9]

    Prikaži ovu nit
    Poništi
  19. 9. sij

    My new paper () assembles 100+ questions from extant and original surveys. The % saying don't know explains about 50% of the variation in average confidence among the people who offered an opinion (i.e., those NOT saying DK). [2/9] Paper:

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    7. sij

    Well-researched, insightful and provocative. If you're interested in immigration and identity in this era of intense polarization, it's a must read.

    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:

    ·