iiinova — where your digital self keeps evolving

Alex Coppock @aecoppock.bsky.social · 30/09/2026

New from Hainmueller and Laitin on the causal effects of becoming a citizen on wages. Short story: the observational ests overstate the premium relative to the experimental ests. Authors conclude "observational averages should not be used as causal forecasts for policy" doi.org/10.1073/pnas...

screen shot of the article with "observational averages shouldnot be used as causal forecasts for policy" highlighted

12512

Alex Coppock @aecoppock.bsky.social · 12/08/2026

Northwestern is hiring in quant methods (assistant level, tenure track) Please repost and share with those you think would be good candidates, thank you! polisci.northwestern.edu/about/facult... #psjobs #psjminfo #poliscijobs

Assistant Professor of Quantitative Political Methodology (Tenure-Track)
The Department of Political Science at Northwestern University invites applications for a full-time position in Quantitative Political Methodology at the rank of Assistant Professor, tenure-track, to begin Fall 2027. We emphasize methodology with a strongly applied perspective: successful candidates will maintain an active, independent research program that includes methodological contributions in their own right, while also teaching quantitative research methods courses at the undergraduate and graduate levels and applying those methods to substantive questions in their area of expertise. We are especially interested in candidates whose work bridges data-driven/computational approaches with design-based inference, though we welcome applicants making cutting-edge contributions in either of these traditions or in other areas of political methodology. Successful candidates' substantive research agendas may be in any area of political science or may cross boundaries; as with all Northwestern faculty, successful candidates will be expected to provide service to the department, the university, and the profession.

Aerial photo of Northwestern's beautiful lakeside campus with political science's new building circled

04133

Alex Coppock @aecoppock.bsky.social · 24/06/2026

A favorite part was the "estimand census", in which we enumerated all of the estimands estimated in the published papers, but also all the the estimands that *could have been* estimated given the design. We (a) add many new published quantities and (b) show how little overlap there is for most qs

180

Alex Coppock @aecoppock.bsky.social · 24/06/2026

The headline finding is that the effects of religious messages average about a percentage point. The effects vary widely, from precisely estimated 0-point effects to precisely estimated 10-point effects! We do not understand why, though we try to sort through some possibilities.

screenshot with this text:

Our goal with this exercise was to evaluate the
generalizability of three main claims in this literature:
(1) religious treatments are especially effective because
of their religious character, (2) effects are larger for
religious affiliates than nonaffiliates, and (3) effects are
larger for high religiosity types than low religiosity
types. Our summary conclusion is that none of these
claims is correct in general.
While it is true that some estimates of the effect of
religious messages (both compared with pure control
and compared with a secular alternative) and the effect
of religious endorsements are strongly positive, these
patterns do not hold across the many contexts we study.
Most of the estimates we report cannot be distinguished
from zero, nor can the meta-analytic averages. For this
reason, we are skeptical of the claim that religious
treatments are especially effective because of their
religious content. Our analysis of heterogeneity
strongly contradicts main claims (2) and (3). In just
3 of the 66 opportunities we study do we find statistically significant evidence of effect heterogeneity by
religious affiliation or religiosity. This lack of heterogeneity appears to be quite generalizable across contexts and across the three classes of estimands we focus
on here.
These are not the conclusions one would draw from
reading the abstracts of the papers we include. We
hand-coded each abstract for whether they make a
positive claim about an average treatment effect, a null
claim about an average treatment effect, or a heterogeneous effects claim. This exercise, while somewhat
coarse, gives us a sense of the headline claims in the
field: 64% of the abstracts make positive treatment
effect claims, 32% make null claims, and 46% make
heterogeneous effects claims.

screenshot with this text:

While our meta-analytic average estimates were all null,
we did also document substantial heterogeneity across
studies. A natural question is, what explains why some
effects are large and others are not? We do not know.
Poring over the forest plots in Figures 2–4 like tea
leaves, we came up with a few post hoc possibilities for
the heterogeneity across studies. First, it could be that
treatments that are especially naturalistic, in the sense
of being similar to real-life religious communications,
are more effective than shorter, seemingly artificial
treatments. We find that religious message treatments
that are higher on this dimension are more effective
(3.4 points, SE: 1.5 points, p= 0.036) than messages that
are lower. Second, it could be that some outcomes are
more malleable by religious message than others; tolerance toward immigrants, refugees, former combatants, and religious out-groups might be one such
domain. However, we do not find evidence that messages that target tolerance are more effective (4.1
points, SE: 2.1 points, p ¼ 0:067 ) than messages that
target other outcomes. Third, we attempt to characterize which messages “stick to religion” in the sense of
emphasizing the religious foundations of the persuasive
attempt, under the theory that such messages will be
more persuasive (Djupe and Neiheisel 2022). As it
happens, we find no differences in message persuasiveness (−0.5 points, SE: 1.7 points, p ¼ 0:785 ) by this
variable, though we readily admit that our own coding
of treatments as “sticking to religion” or not might
differ from others’ judgments. We do not put great
faith in the success or failure of these explanations,
since the studies vary in many uncontrolled dimensions,
and since we coded these messages’ style and substance
after having seen the results of the meta-reanalysis.

120

Alex Coppock @aecoppock.bsky.social · 24/06/2026

Our (with Radha Sarkar) meta-reanalysis of the experimental religion and politics literature is now out at @apsrjournal.bsky.social Open-access link: doi.org/10.1017/S000... Brief thread

title page of journal, that reads:

The effects of religious messages and endorsements on political attitudes: A meta-reanalysis
Sarkar, Radha, Coppock, Alexander. 2026. The effects of religious messages and endorsements on political attitudes: A meta-reanalysis. American Political Science Review.

Abstract
The experimental study of the effects of religious messages and endorsements by religious leaders on political attitudes has a relatively short history yet has coalesced around three main claims: religious treatments are especially effective because of their religious character, effects are larger for religious affiliates than nonaffiliates, and effects are larger for high religiosity types than low religiosity types. Here we meta re-analyze the experimental record (58 treatment-outcome pairs drawn from 43 studies reported in 26 papers) to probe the generalizability of these claims. Our findings indicate that these three headline claims do not generalize straightforwardly across contexts: effects are large in some cases and close to zero in others, and we find no evidence in favor of the claimed heterogeneities by religious affiliation or religiosity. Based on a census of the estimands in this literature, we offer suggestions for future research that would enhance commensurability and synthesis.

12715

Alex Coppock @aecoppock.bsky.social · 19/03/2026

This scatterplot shows the correlation of pre-treatment party with hypothetical party ID; off diagonals show imagined "party switching". Not *zero* in the SQ condition, but substantially larger in the partisan hypothetical conditions.

120

Alex Coppock @aecoppock.bsky.social · 19/03/2026

Here's an example of the status-quo, pro-D, and pro-R hypotheticals, after which people were asked what their 7 point party ID would be.

SQ:
For the next few years, the Russia–Ukraine war remains at a stalemate. Neither side advances far beyond where they are today. The fighting is constant but does not escalate.

Pro-D
Donald Trump wins the 2024 election and immediately withdraws all US support for Ukraine. Putin easily conquers Ukraine, slaughtering hundreds of thousands of people in the process. He then invades Poland, again killing many innocent people in his quest for domination.

Pro-R
Donald Trump wins the 2024 election and immediately negotiates a ceasefire between Russia and Ukraine. Russia withdraws from most of Ukraine, and hundreds of thousands of lives are spared because of Trump.

110

Alex Coppock @aecoppock.bsky.social · 19/03/2026

New paper with Don Green and @ethanvporter.bsky.social in the QJPS. After much deliberation, we went with a title that just states the result. 📝 journal: www.emerald.com/qjps/article...

abstract: While attempts to change Americans’ partisanship via persuasive treatments largely fail, partisanship can and does change over time. In this paper, the authors first confirm, via survey and field experiments, that typical campaign messaging in the United States does not budge partisanship. The authors then present experiments in which participants encounter extraordinary hypothetical scenarios (e.g. one party causes economic collapse) before reporting what their partisanship would be under such circumstances. Twelve percent of partisans imagine switching parties in the pro-out-party hypothetical conditions, compared with 5% in the control hypotheticals in which the status quo persists, for a seven-percentage point (SE 1.5 points) difference. These hypothetical shifts are on par with the largest changes in American macropartisanship ever recorded. While the act of ruminating on hypothetical scenarios is not followed by changes in partisanship measured post-treatment, the evidence suggests that extraordinary world events may be able to shift partisan affiliation.

24620

Alex Coppock @aecoppock.bsky.social · 27/02/2026

Writing simulations in DeclareDesign just went from "I should do that, but it's kind of a lot of work" to extremely easy

Screenshot of claude just writing a design no trouble

46110

Alex Coppock @aecoppock.bsky.social · 13/02/2026

Penelope Van Grinsven and Lilly Zuckerman curate Above Board Ceramics -- this year's show is now live and is fabulous. www.aboveboardceramics.com [disclosure Penny and I are married!]

160

Alex Coppock @aecoppock.bsky.social · 08/02/2026

Rough. it doesn't have to be this way! Check out what the the Saint Paul Chamber Orchestra is up to. www.thespco.org/concerts-tic...

020

Alex Coppock @aecoppock.bsky.social · 27/01/2026

🎺 Call for proposals 🎺 1️⃣ replicate an existing experiment 2️⃣ run a novel experiment on repdata.com 3️⃣ coauthor with Mary McGrath and me to meta-analyze the replications and existing studies 4️⃣ publish your study details: alexandercoppock.com/replication_... applications open Feb 1 please repost!

Call for Proposals: Data Collection for
Replication+Novel Political Science Survey Experiments
Alexander Coppock and Mary McGrath
January 27, 2026
We invite proposals for a survey experiment replication+novel design competition. Se-
lected replication+novel design survey experiments will be conducted on large samples of
American respondents, quota sampled to match U.S. Census margins and filtered for quality
and attention by the survey sample provider Rep Data (repdata.com).
Each proposal consists of two parts: (1) a replication study of an existing, previously
published survey experiment, and (2) a novel experimental design on a topic of the authors’
choosing.
The replication studies and reanalyses of the existing studies will be combined into a
meta-paper to be co-authored by all authors of accepted proposals along with the princi-
pal investigators (Coppock and McGrath). As a condition for acceptance, authors commit
to sharing the data and producing a write-up of the findings from their novel design for
submission to a scholarly journal, and public posting of a working paper pre-publication.

08071

Alex Coppock @aecoppock.bsky.social · 07/01/2026

Do we have any information about the study design? The publicly available deactivation RCTs I've seen do also show that deactivation improves well-being, e.g. www.nber.org/papers/w33697

We estimate the effect of social media deactivation on users’ emotional state in two large randomized experiments before the 2020 U.S. election. People who deactivated Facebook for the six weeks before the election reported a 0.060 standard deviation improvement in an index of happiness, depression, and anxiety, relative to controls who deactivated for just the first of those six weeks. People who deactivated Instagram for those six weeks reported a 0.041 standard deviation improvement relative to controls. Exploratory analysis suggests the Facebook effect is driven by people over 35, while the Instagram effect is driven by women under 25.

140

Alex Coppock @aecoppock.bsky.social · 02/12/2025

Thanks to the authors for compiling this information; this view on our science is truly fascinating. 2023: 40% design-based, 40% model-based, 20% other, among quantitative studies aimed at "explanation" which I understand to mean causal inference. Where we will be in 2043?

Figure 4: Trends in the use of design-based methods among explanatory empirical quanti-tative papers

0111

Alex Coppock @aecoppock.bsky.social · 19/11/2025

(prohibition) persuasion in parallel Had fun reconstructing the data from this 90-year old persuasion experiment that shows that "wet" and "dry" college students update their attitudes in the direction of counterattitudinal persuasive information. paper: doi.org/10.1080/0022...

The Speech of man is one of the most obvious of his social activi- tics. O u r educational systems lean heavily on the presumption that speech is a vital tool in the stimulation of individual development. T h e lawyer and theologian carry on their work almost entirely
through the use of speech. The zealous protection of the right of freedom of speech in democratic societies may be said to be based on the premise that where men do not enjoy freedom of speech they .soon become slaves. Millions of dollars are spent annually by advertisers with a naive faith in the power of the word to convince men and move them to action. The use of speech by man is a psy- chological phenomenon unique in the world of subjects for scientific rtudy. Yet in spite of all these facts, one finds surprisingly few re- ports in the scientific literature of objective measurements of the actual effects of the use of speech on the social environment.

1120

Alex Coppock @aecoppock.bsky.social · 05/11/2025

I'm one of those academics who measure message persuasiveness via survey experiment. I think it's absolutely the right method for doing so. I think the response to this good point from @anatosaurus.bsky.social is not to abandon survey exps but instead to measure "getting heard" (attention) also.

At first glance, Pollingism espouses a so-simple-a-child-gets-it truth: ask people what matters to them and repeat it back. It often gets short-handed as “say popular things” and, by extension, eschew unpopular ones. However, this notion is built off of assumptions that do not bear out in the real world.

Let’s start with an irrefutable claim: a message that nobody hears cannot persuade them. Nor can it get them to act in opposition to the present regime, vote in an election, or select a specific candidate.

For years, we have heard entirely accurate lamentations that Democrats are failing to reach people. To be sure, the right’s far better media system merits every alarm bell, not merely for electoral outcomes but also for our hopes of a functional society, free of lies about everything from vaccines to wind turbines to the dining habits of newcomers. But even setting that imbalance aside, Pollingism’s message creation plus testing model fails to consider, let alone optimize for, how we would get a message to be heard.

Presently, the dominant approaches to campaign message crafting and evaluation involve (1) randomized-controlled trial (RCT) testing of messages and/or ads, and (2) issue-based polling to figure out which topics voters rate most vital and therefore merit airtime. But these tools cannot tell us what would break through in the cacophony of political noise to which voters are subjected.

3142

Alex Coppock @aecoppock.bsky.social · 05/11/2025

lol at the MPSA shade in the last line of the FAQ

Will the deadline be extended a week, like other conferences?
No!

1150

Alex Coppock @aecoppock.bsky.social · 04/11/2025

Maybe your texts tell you to "BE A VOTER" like mine. It all started because of a PNAS paper that claimed that the noun form it increased voter turnout (relative to the verb form ) by 11 to 14 percentage points. It keeps not replicating, obviously. Most recently doi.org/10.1017/bpp....

1357

Alex Coppock @aecoppock.bsky.social · 30/10/2025

Thank you for sharing -- I was unaware of this framework for choosing among replication targets. I think I stand with the critics. Of course I'm left still not knowing how to choose among empirical estimands! Separately, this figure is amazing:

a good figure that skewers an "anonymous" reviewer for requiring many citations to (presumably) their work

121

Alex Coppock @aecoppock.bsky.social · 29/10/2025

chat's personalized flattery is getting out of hand

chat really told me: Excellent — that’s exactly what Hadley would do next.

141

Alex Coppock @aecoppock.bsky.social · 27/10/2025

encountered this article by David Weimer from 1986 (cited by 7) with three good ideas for improving science that just a few short decades later, we've started to implement. Collective Delusion In The Social Sciences: Publishing Incentives For Empirical Abuse doi.org/10.1111/j.15...

title page of article:

There is a skeleton in the social science closet. Almost all of us have heard it rattled at one time or another.! Nevertheless, we collectively ignore it. Some ignore it because they do not think it is very scary: others because its putting to rest would require a major restructuring of our scholarly journals. Perhaps others even fear that admitting its existence would undermine belief in the empirical progress? the social sciences have apparently enjoyed since the widespread diffusion of statistical training and the ready availability of the computer.
I will argue that the
skeleton, the bias social science journals exhibit for publishing articles reporting statistically significant results, is dangerous

The results of statistical tests should not be submitted to or als until after articles have been accepted for publica-
tion..
would base their decisions on
theory under
discussions of the theory under consideration, the specific hypotheses to be tested, and the data sample to be used.?
This would divert attention from the final result to the a priori specification of hypotheses and the appropriateness of data and statistical technique for testing them. Unfortunate-ly, this approach would not work unless adopted by all the relevant journals. If a single journal adopted it, researchers

Any journal that publishes an article with statistical models should be required (by the disciplines?) to provide a page of space to anyone who wants to report the results of applying the authors' models to different sets of data. This approach would provide a less biased sampling of research results. It would also allow us to better gauge the robustness of our theories.8 Some journals now provide some space for veri-fications; doing so should be standard editorial practice.

Researchers should be required to submit a statement with
indicating
their articles indicating whether or not the model being presented is the one actually first estimated from the data.
A general description of statistical techniques would also be included in the submission. Is the presented model the first specification or is it the end product of stepwise regression or some other questionable data search? Where appropriate, the editor would affix a warning along the following lines next to results:

1135

Alex Coppock @aecoppock.bsky.social · 14/10/2025

👀 this new meta-analysis on edutainment by @bardiarahmani.bsky.social, Montano, @dylanwgroves.bsky.social, and Green doi.org/10.1017/bpp.... 377 ests in 77 exps: edutainment moves attitudes, norms, beliefs 📊 Effects persist ⏳ Many reasonable theories about effect heterogeneity are not supported 😇

1196

Alex Coppock @aecoppock.bsky.social · 14/09/2025

love this -- so very useful. btw you *can* write your own handler as you did here (which is v. flexible and extensible) but you can also do many post-estimation things with the .summary argument if you like!

020

Alex Coppock @aecoppock.bsky.social · 03/09/2025

whaaaat!? that's so cool! literally just: weird_trick = D * X iv_robust(weird_trick ~ D | Z) to get the mean of X among compliers!

1193

Alex Coppock @aecoppock.bsky.social · 27/08/2025

This post seriously conflates "polling" with "survey experimentation" Also any paragraph that starts this way is a big 🚩

I know that correlation does not equal causation. But

051

Alex Coppock @aecoppock.bsky.social · 26/06/2025

With @mollyow.bsky.social, we wrote up an email exchange about testing for causal effect heterogeneity with the grf package. In short, a recommended approach had v. bad properties (wild over-rejection of true nulls of homogeneity) but Molly had a fix. alexandercoppock.com/testing_with...

2134

Alex Coppock @aecoppock.bsky.social · 06/05/2025

This new package for mediation is great! The user interface is *so intuitive* but the unsung hero of this software is the documentation:

Arguments
data
Not needed, wouldn't make your analysis and more correct.

Details
We have:

X: a binary treatment variable, that is affected by an unmeasured confounder (U).

M: a mediator, that is mostly driven by U, not X

Y: an outcome, that is influenced by both M and U, but you can’t see U.

What you don’t have:

Any valid identification strategy

Temporal ordering

Careful causal reasoning

Value
Nothing, because you shouldn't do mediation analysis.

1255

Alex Coppock @aecoppock.bsky.social · 29/04/2025

whoa! Some mild dependence of the SE of the "both" estimator over the full range of rho...

130

Alex Coppock @aecoppock.bsky.social · 29/04/2025

had some fun simulating this one in DeclareDesign. outcome model is Y ~ 0.5 * D + 0.5*X_1 + 0.5*X_2 + U with D confounded by X_1 and X_2. gotta control for both or BIAS. Simulation agrees with you, no precision loss (on ATE estimate) in this setup when X_1 and X_2 are v. correlated.

2322

Alex Coppock @aecoppock.bsky.social · 24/04/2025

I appreciate your perspective here Dan and agree that data viz can easily obscure inferences With my visualizations, *what* I'm trying to convey is the experimental design and the results *to whom* I'm trying to communicate is the community of scientists

120

Alex Coppock @aecoppock.bsky.social · 22/04/2025

Congratulations @awilke.bsky.social! Anna appears to a tables > figures kind of scholar so no screenshots of results -- you'll have to click through to actually read this paper to learn what happened :) doi.org/10.1111/ajps...

I implemented a ﬁeld experiment inSouth Africa that randomly varies the capacity of police to locate households.Findings from surveys conducted several months later suggest households that have become legible to police are more willing to rely on police and less willing to participate in vigilantism

130

Alex Coppock @aecoppock.bsky.social · 16/04/2025

We don't need new software for most of these recommendations, we just need to use existing tools in ways that respect the design. A major exception is how to deal with overplotting, since experiments often feature discrete treatments and discrete outcomes (binary or likert scales), e.g.:

190

Alex Coppock @aecoppock.bsky.social · 16/04/2025

Like all good package names, this one needs some explanation: vayr stands for Visualize As You Randomize, which is a philosophy for design-based graphs. The basic idea is that the visualization should both convey and respect the experimental design. alexandercoppock.com/coppock_2020...

A good statistical graph for a randomized experiment simultaneously conveys the study's design, analysis, and results. It reveals the experimental design by mapping design elements to aesthetic parameters. It illuminates the analysis by plotting the statistical model in “dataspace.” When the design and analysis of an experiment are encoded in a plot, the interpretation of the experimental results is clarified. “Analyze as you randomize” is a dictum attributed to Fisher that guides interpretations of experimental data. This chapter extends that principle to visualizations of randomized experiments. While not every experiment requires a visualization, those that do should be visualized in ways that communicate the design and results together.

1100

Alex Coppock @aecoppock.bsky.social · 16/04/2025

New visualization tool alert! The vayr package version 1.0.0 is now on CRAN. It contains position adjustments for ggplot2 that help with overplotting in pleasing ways. My favorite is position_sunflower(). - install.packages("vayr") - alexandercoppock.com/vayr #rstats #ggplot2 #dataviz

319353

Alex Coppock @aecoppock.bsky.social · 08/04/2025

TIL, the original meaning of "conjoint" was about the "conjoining" of multiple attributes, not multiple profiles! www.jstor.org/stable/31495... That said, I agree there's no serious design distinction between single profile conjoint experiments and factorial experiments.

120

Alex Coppock @aecoppock.bsky.social · 07/04/2025

One picture I've made shows the 21 possible random assignments in Chapter 3 of Gerber and Green like the left figure (treat 2 of 7). the right figure shows the sampling distributions by other assignment schemes. kinda neat?

030

Alex Coppock @aecoppock.bsky.social · 27/03/2025

Left: frequency of finding het fx by starsign when there is none (declaration and diagnosis in alt-text) Right (inspo): classic R2 clapback -- you want het fx? Ok fine, here's fx of aspirin by starsign. (doi.org/10.1016/S014...

declaration_9.1 <-
declare_model(N = 1000,
# make birthdates and obtain starsign
birthdate = sample(seq(as.Date("1940-01-01"), as.Date("2005-03-11"), by = "day"), N),
starsign = DescTools::Zodiac(birthdate),
# potential outcomes unrelated to starsign
potential_outcomes(Y ~ Z + rnorm(N))) +
declare_assignment(Z = complete_ra(N)) +
declare_measurement(Y = reveal_outcomes(Y ~ Z)) +
declare_estimator(Y ~ Z*starsign,
term = TRUE)

simulations <- simulate_design(declaration_9.1, sims = 1000)

For example, subdivision of the patients in ISIS-2
with respect to their astrological birth signs appears to
indicate that for patients born under Gemini or Libra there
was a slightly adverse effect of aspirin on mortality

050

Alex Coppock @aecoppock.bsky.social · 25/03/2025

Really great study, thank you for sharing it! Not to take anything away from the great work you've done in this project, but I'm especially excited about this next part...

Future research will be able to use similar techniques in a causal setting to measure both attention and any downstream effects on e.g., political attitudes and behavior. Doing so will shine a light on whether and how attending to content online (with which one may or may not disagree) has an effect on subsequent attitudes and behavior such as negative outgroup affect and online hostility.

120

Alex Coppock @aecoppock.bsky.social · 10/03/2025

My spouse Penelope Van Grinsven is a ceramic artist and co-curator of the Above Board Ceramics exhibition, now in its fifth year! the show: www.aboveboardceramics.com the shop: www.aboveboardceramics.com/theshop Sale ends Tuesday, March 11 @midnight eastern

0152

Alex Coppock @aecoppock.bsky.social · 26/02/2025

That *is* odd! I download the replication archive and I think I subsetted to the right estimates... and it looks like the are the same pvalues to 3 digits!!

110

Alex Coppock @aecoppock.bsky.social · 26/02/2025

Just read this lovely experiment on the (null) persuasive effects of "listening" by Erik Santoro, @dbroockman.bsky.social @jkalla.bsky.social and @ronip.bsky.social www.pnas.org/doi/epub/10.... Left: survey of prior beliefs Right: Results It's the persuasive appeal that does the work.

1142

Alex Coppock @aecoppock.bsky.social · 25/02/2025

We're still banging on about not using the "change format" in public opinion surveys. And Matt makes the good point that we shouldn't share such surveys either. The counterfactual format is a good replacement!

Change format condition Counterfactual format condition
Belief Q: “As you may know, former
President Donald Trump was recently indicted for removing classified information from the White
House, including sensitive nuclear
secrets. Does the indictment make
you think it is more likely or less
likely that Trump mishandled nuclear secrets?” [More likely/
No change/Less Likely]
Q1: “As you may know, former president Donald Trump was recently
indicted for removing classified information from the White House, including sensitive nuclear secrets. In
your opinion, how likely is it that
Trump mishandled nuclear secrets?”
[0-100 scale]
Q2: “Suppose you did not know
about the indictment. How would
you have answered the following
question: In your opinion, how likely
is it that Trump mishandled nuclear
secrets?” [0-100 scale]
Support Q: (for Republican primary voters) “Does the indictment make
you more or less likely to vote for
Trump in the primary?” (for all
others) “Imagine that the 2024
presidential election is between
Joe Biden and Donald Trump.
Does the indictment make you
more or less likely to vote for
Trump?” [More likely/No
change/Less Likely]
Q1: (for Republican primary voters)
“How likely are you to vote for
Donald Trump in the 2024
Republican primary?” (for all others)
“Imagine that the 2024 presidential
election is between Joe Biden and
Donald Trump. How likely are you
to vote for Donald Trump?” [0-
100 scale]
Q2: “Suppose you did not know
about the indictment. How would
you have answered the following
question: How likely are you to vote
for Donald Trump

0110

Alex Coppock @aecoppock.bsky.social · 25/02/2025

Thanks for your interest! Here's the wording in Table 1:

100

Alex Coppock @aecoppock.bsky.social · 13/02/2025

Such a banger. The rain DAG has made its way around socials a few times, but check out the extremely lucid and incisive discussion section.

It is tempting to interpret instruments as being “used up” by multiple uses. However, it is important to distinguish the fact of whether an exclusion-restriction violation exists from our knowledge of that fact (which new research can change). This may seem unfair to authors who used the best available information, but the goal of this method is to assess empirical claims not to cast judgment on researchers.

Causal knowledge in social science has few easy shortcuts. Apparently simple solutions such as IVs still rely on strong substantive claims about the world. To build cumulative knowledge, we must engage directly with the messy causal web of the social world.

This paper shows that a widely used instrument is flawed because it systematically violates the exclusion-restriction assumption for most plausible applications. These findings suggest our standards for accepting instruments are overly lenient and atheoretical. The causal revolution in social science has revealed the wishful thinking underlying selection-on-observables. The same rigor must be applied to IV assumptions.

26416

Alex Coppock @aecoppock.bsky.social · 30/01/2025

spare a minute to read a reflection on abandoning a false theory you built a career on

For a kid who grew up with empty bookshelves, this should have been a moment of triumph [1].

Instead, I felt like a fraud.

At that same conference, I had to confront an uncomfortable truth: the foundation of our celebrated paper was crumbling. Ego depletion—the once-famous idea that self-control relies on a finite resource that can be depleted through use—wasn’t real. That award? It was like winning a Nobel prize for developing the frontal lobotomy as a treatment for mental illness; and, yes, that really happened.

This isn’t just another story about failed replications or p-hacking (though those shenanigans will make an appearance). It’s a story about what happens when we fall in love with our theories more than the truth. The replication crisis didn’t just shake the foundations of psychology; it shook those of us who had built our careers on ideas that no longer held up to scrutiny.

For me, this reckoning wasn’t just professional. It was personal

081