Jonathan Mummolo

@jonmummolo

Political scientist at researching policing, bureaucracy, political behavior. Former reporter .

Princeton, NJ
Vrijeme pridruživanja: veljača 2009.

Tweetovi

Blokirali ste korisnika/cu @jonmummolo

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

  1. Prikvačeni tweet
    21. sij

    published a study last year claiming no racial bias in police shootings. The study's central claim was mathematically unsupported. & I submitted critique to PNAS, which was rejected. We appealed. Today PNAS published our critique.1/n

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    3. velj

    Happy to share that my first published paper is out at APSR! and I show that when looking at the distribution of pork to legislators’ districts, the data you choose to use can alter your results in important ways.

    Poništi
  3. proslijedio/la je Tweet

    Wow, everyone really is claiming victory in the Iowa caucuses

    Poništi
  4. proslijedio/la je Tweet
    3. velj

    Does he have any idea what a multilevel model is?

    Poništi
  5. proslijedio/la je Tweet
    2. velj

    "It took us months to contest a flawed study on police bias," write and . "Here’s why that’s dangerous."

    Prikaži ovu nit
    Poništi
  6. proslijedio/la je Tweet

    A study claimed to show that "White officers are not more likely to shoot minority civilians than non-White officers." But as Dean Knox and write, it was based on a logical fallacy and erroneous statistical reasoning

    Poništi
  7. proslijedio/la je Tweet
    30. sij

    Professor Susanne Ditlevsen talking about et al.s important criticism of the failed police shooting study in PNAS. At the yearly professor symposium at the University of Copenhagen — a good day for science

    Poništi
  8. proslijedio/la je Tweet

    A study claimed to show that "White officers are not more likely to shoot minority civilians than non-White officers." But as Dean Knox and write, it was based on a logical fallacy and erroneous statistical reasoning

    Poništi
  9. 29. sij
    Prikaži ovu nit
    Poništi
  10. 29. sij

    When research makes mistakes, even blatant ones, it’s almost impossible to set the record straight. and I discuss this in a new op-ed in Science needs to do better.

    Prikaži ovu nit
    Poništi
  11. 28. sij

    Social science is very lucky to have .

    Poništi
  12. 28. sij

    Very clear & succinct summary by of a large literature on police violence, & an impartial assessment of methodological errors therein, including critiques of studies by the author's friends & colleagues. Insight and integrity on display here.

    Poništi
  13. proslijedio/la je Tweet
    27. sij
    Odgovor korisnicima i sljedećem broju korisnika:

    One bad conservative study doesn't balance one bad liberal study-- they're both bad. Adding more bad studies makes it worse.

    Poništi
  14. 26. sij

    I would be eager to read research on researcher biases if it was well designed. But the idea that you can see a single critique, look at the direction of result being criticized and infer bias is silly. Ironically, it's a terrible research design, very prone to confirmation bias.

    Poništi
  15. 26. sij

    Being called liberally biased for critiquing research claiming no racial bias in police violence. 2 yrs ago I was accused of conservative bias for critiquing a Voter ID study. Both studies made indisputable errors on matters of great policy importance. I call balls and strikes.

    Poništi
  16. proslijedio/la je Tweet

    In a letter appearing in , faculty members and discuss the logical fallacy of a 2019 study on racial bias and policing:

    The word "police" with a target in the letter "O"
    Poništi
  17. 26. sij

    This isn’t about motivations. Those are unknowable. And I don’t endorse attacks on people’s motives, even if I disagree with them. The great thing about science—real science—is that motives are irrelevant. We can all look at the evidence and make a judgement based on that.

    Tweet je nedostupan.
    Poništi
  18. 25. sij

    That methodological study already exists: Bayes (1763). Pr(race|shot) != Pr(shot|race). Given this, your study can’t test, as it claims, “whether officers are more likely to shoot” anyone. We should seek the data we need, not estimate the wrong quantity & hope for the best.

    Tweet je nedostupan.
    Poništi
  19. 25. sij

    Ok, but you would still want to get data on *all* officers, not just those involved in shootings. Suppose you found 5% of officers involved in shootings had those traits. Is that informative? If 5% of those not involved in shootings also had those traits, you'd probably say no.

    Poništi
  20. 25. sij

    If you mean Fryer (2019) & Johnson et al. (2019), then yes. We outline problems in Fryer (2019) here: We outline problems in Johnson et al. (2019) here:

    Poništi
  21. 25. sij

    It is deeply troubling that peer-reviewed journals across disciplines continue to allow fundamentally flawed research of this kind into print. This is a basic and indisputable statistical error applied to the study of a life and death policy issue. This needs to stop. (n/n)

    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:

    ·