Dan de Kadt

Reminder: last day to submit to the JEBO issue in honor of @cfcamerer.bsky.social We are excited about all the great submissions we have gotten thus far.

What are we talking about here, no internet or pure ethernet speed?

Uh have you even tried being an influencer without WiFi for 2 hours??

Evergreen

Just to put a fine point on it: I agree that the data visualisations were useful because they highlighted this issue. Unfortunately I think they are also very persuasive to many readers.

I am a proponent of data visualisation, but I think that they do carry risks in this regard.

Agree 100%. You are cited in the replication!

Just to confirm, I received confirmation of the archive’s publication half an hour ago. But the materials all on GitHub too.

Imagine sitting in your office writing one of your general exams and Bibi Netanyahu brings you lunch.

Do they have your faith? Having accidentally mis-pasted things in excel before, this terrifies me.

For the avoidance of doubt: does svyglm() use robust standard errors? Yes!
notstatschat.rbind.io/2025/12/01/d...

No replication game or umbrella org. This is a piece of work I did on my own with no external prompting. Happy to talk more about it.

‼️ Two weeks left to apply for the postdoc in Political Text Analysis in the MULTIREP project 👇.

Are Chelsea too reliant on set pieces?

Yes you are correct (about time investment being a barrier).

But then we end up in a very funny place, don’t we?

I’m not really sure what you mean by this?

Hey Tom, happy to talk on a call if you would like.

I’ve also made available as much information as I can about the process in this thread and in a repo linked at the end.

(P.S. I personally would not classify my replication as a “gotcha,” but appreciate perspectives may differ on that.)

GitHub - ddekadt/instrumentally_inclusive_replication_public Contribute to ddekadt/instrumentally_inclusive_replication_public development by creating an account on GitHub.

To see all the different versions of the replication, the various papers and responses, the replication materials (for the original paper and for my replication), please check out this repo.

github.com/ddekadt/inst...

# Initial Events: Sept '24: I wrote my report 2 Oct '24: I sent it to the original authors as a courtesy, and told them I would send it to the APSR team on the 23rd as a courtesy, before posting on my website on 1 Nov. 10 Oct '24: The original authors replied. 23 Oct '24: I determined that their email did not change my perspective on the original paper or my replication, and sent the report to both the APSR editorial team and CUP's publishing ethics team in an email (as I had indicated I would). # Formal Review Process: 1 Nov '24: The APSR team invited me to submit on the reappraisals/replication track. For personal reasons I decided not to post publicly on 1 Nov. [Note: I was unaware of the existence of this new publication track at the APSR until I received this invitation.] 10 Jan '25: I received an invitation to revise and resubmit the paper after external review. 2 June '25: I submitted a revised version of the paper. # Post-Acceptance Process: 9 July '25: I received a conditional accept subject to some internal APSR steps that required I submit initial replication materials for review [I do not know the details of this process.] July '25: The original authors wrote to the APSR to suggest a corrigendum, and were informed that the replication was forthcoming [here I rely on the original authors' account of the process and have no further details.] 26 September '25: The paper was officially conditionally accepted in editorial manager, and replication materials were formally submitted to dataverse. 28 November '25: The paper was released on FirstView.

I will end by laying out my "process" as clearly as I can.

I obviously do not know everything - this is all from my perspective (plus what the authors have stated publicly).

In the end, I waited 15 months before "going public," and the authors had a version of the report for 14 of those months.

We could discuss how we as a community of scientists might build better incentives and guardrails to make sure that the papers being published in our ("top" or not) journals aren't full of errors.

Instead we're discussing how to secure the authors' right of reply. Broken incentives? There you go.

2. We could discuss the "process" that led to the publication of the paper in our discipline's "top" journal.

Whatever you think of my replication, the paper contained many errors:

- wrong std. errors
- wrong models/tables
- missing conf. intervals
- made up data on plots
- incorrect math
- etc.

To me this is a depressing theme in modern academia.

There is so much work being produced, and so many competing demands on our time, that people rarely seem able to just closely read work and frankly say "yes, I believe this" or "no, I don't."

If we aren't doing this, what _are_ we doing?!

1. We could interrogate and discuss the merits of the original paper, the mertis of the critique, the merits of the response, and the merits of the response-to-the-response.

This is the stuff that is fundamentally at stake! I encourage everyone to read them closely and evaluate as they see fit.

Good morning. A coda (from me) on this replication business.

While I understand the impulse, it is disheartening that most of the discussion generated by this replication is about the replication "process" -- both the APSR's process and my own process.

What could we be talking about instead?

My full response is available here:

instrumentally_inclusive_replication_public/brief_response_nov29/dekadt_brief_response_nov29.html at master · ddekadt/instrumentally_inclusive_replication_public Contribute to ddekadt/instrumentally_inclusive_replication_public development by creating an account on GitHub.

Direct link to the response:

github.com/ddekadt/inst...

bsky.app/profile/dand...

instrumentally_inclusive_replication_public/brief_response_nov29/dekadt_brief_response_nov29.html at master · ddekadt/instrumentally_inclusive_replication_public Contribute to ddekadt/instrumentally_inclusive_replication_public development by creating an account on GitHub.

If you’d like to read my response to the authors‘ posted response, it’s at the link below. Just download the .html (top right) and open in any browser.

(Note: I deleted a few skeets this morning as my links were dead or pointing to the wrong place.)

github.com/ddekadt/inst...

instrumentally_inclusive_replication_public/brief_response_nov29/dekadt_brief_response_nov29.html at master · ddekadt/instrumentally_inclusive_replication_public Contribute to ddekadt/instrumentally_inclusive_replication_public development by creating an account on GitHub.

If you’d like to read my response to the authors‘ posted response, it’s at the link below. Just download the .html (top right) and open in any browser.

(Note: I deleted a few skeets this morning as my links were dead or pointing to the wrong place.)

github.com/ddekadt/inst...

May write more about this at length in the future:

good causal inference training leads to better descriptive research

Ok but turkey shakes?