Lightnews — Scholar-powered news

Dave Siegel

@daveasiegel.bsky.social

1.8K followers 520 following 38 posts

Professor of Political Science and Public Policy at Duke University. Associate Editor (Formal Theory) of the AJPS. daveasiegel.com

Posts Media Videos Starter Packs

Pinned

Dave Siegel @daveasiegel.bsky.social · Sep 13

I’m pleased to announce to #polisky our forthcoming APSR article, “Measurement that Matches Theory.” In it, we introduce a Bayesian IRT framework that allows users to identify latent dimensions with substantive theoretical meaning. The paper is open access for now here: tinyurl.com/48ktz5jc. (1/8)🧵

Measurement That Matches Theory: Theory-Driven Identification in Item Response Theory Models | American Political Science Review | Cambridge Core

Measurement That Matches Theory: Theory-Driven Identification in Item Response Theory Models

www.cambridge.org

2 9 26

Dave Siegel @daveasiegel.bsky.social · May 22

Probably should've tagged this thread: #Polisky #PolScience

Dave Siegel @daveasiegel.bsky.social · May 22

This was very much a group effort. Most of my co-authors aren't here, but I'll tag the amazing @margaretfoster.bsky.social. We hope this tool can be widely useful, and are happy to answer any questions you might have! (7/7)

1 2

Dave Siegel @daveasiegel.bsky.social · May 22

Right now one needs to convert to binary input data to use IRTM, but we're also working on extensions to other data types. We provide a vignette in the package to walk the user through IRTM's use. (6/7)

1 2

Dave Siegel @daveasiegel.bsky.social · May 22

Because you can run IRTM on large datasets with complex loading constraints, it opens up theoretically-driven measurement to a lot more contexts. We focus on the importance of matching measurement to theory in the paper, but are working on making clearer the range of uses in ongoing work. (5/7)

1 2

Dave Siegel @daveasiegel.bsky.social · May 22

It's also user-friendly, requiring little coding or package knowledge: just give it input data and a constraint matrix for the pre-specification, and it will provide individual-level posterior distributions over latent dimensions. No need to learn any package-specific syntax. (4/7)

1 2

Dave Siegel @daveasiegel.bsky.social · May 22

IRTM trades off some generality in specification, as packages such as brms or blavaan can achieve, for substantially increased speed: it runs on data with 10k respondents and 200+ items with 4 latent dimensions in about 10 minutes, regardless of the complexity of the pre-specified connections. 3/7

1 3

Dave Siegel @daveasiegel.bsky.social · May 22

IRTM is a Bayesian implementation of an Item Response Theory model in a similar vein to other constrained approaches such as BCFA or BSEM: it allows the user to pre-specify connections between items/questions and theoretically-coherent latent dimensions. (2/7)

1 2

Dave Siegel @daveasiegel.bsky.social · May 22

Happy to see that this is now out. The R package associated with the project, IRTM, can now be found on CRAN. Or can you still download it from my github. We'll continue to update it over time. This is a brief thread about what IRTM can do. 🧵 (1/7)

American Political Science Review @apsrjournal.bsky.social · May 20

From our new issue: "Measurement That Matches Theory: Theory-Driven Identification in Item Response Theory Models" by Marco Morucci, Margaret Foster, Kaitlyn Webster, So Jin Lee, and David Siegel. #APSRNewIssue www.cambridge.org/core/journal...

2 6 25

Reposted by Dave Siegel

American Political Science Review @apsrjournal.bsky.social · May 20

2 8

Dave Siegel @daveasiegel.bsky.social · Dec 27

🚩

Dave Siegel @daveasiegel.bsky.social · Nov 18

Great news!

Dave Siegel @daveasiegel.bsky.social · Nov 16

Thanks for this! I'd love to be added, if possible.

Dave Siegel @daveasiegel.bsky.social · Nov 13

Thanks for putting this together! Would love to be added.

Reposted by Dave Siegel

Jon Green @jongreen.bsky.social · Nov 1

if you want a break from the election, we've posted a revision of our paper on pundits and ideological coalitions. bigger emphasis in this draft on the point that political ideologies aren't quite the same thing as political philosophies osf.io/xfy8r

3 8

Dave Siegel @daveasiegel.bsky.social · Sep 27

I said this in my first-year grad research methods class, but it bears repeating. There are people who will tell you that you are not smart enough to understand something. Those people either don't understand the nature of intelligence, don't have your best interests at heart, or both. Ignore them.

1 1 9

Reposted by Dave Siegel

Jon Green @jongreen.bsky.social · Sep 17

Just accepted at JOP w/ Daniel Naftel @kshoub.bsky.social, @jarededgerton.bsky.social, Mallory Wagner, and Skyler Cranmer. We find that patterns of gender inequality observed in formal deliberative settings extend to informal televised political discussions. www.journals.uchicago.edu/doi/10.1086/...

Paper title page (too long for alt text)

Figure 2: Effect of Women's Numbers on the Female-to-Male Ratio of Words Spoken

1 8 16

Dave Siegel @daveasiegel.bsky.social · Sep 13

It's here: github.com/dasiegel/IRT-M

GitHub - dasiegel/IRT-M: IRT-M R package

IRT-M R package. Contribute to dasiegel/IRT-M development by creating an account on GitHub.

github.com

1 1

Dave Siegel @daveasiegel.bsky.social · Sep 13

Forgot the figure!

Distributions of latent perceptions of threats, grouped by attention to media.

Dave Siegel @daveasiegel.bsky.social · Sep 13

This is NSF-funded work, joint with Marco Morucci, Margaret Foster, Katie Webster, and So Jin Lee. None of whom are on Bluesky yet. (8/8)

1 2

Dave Siegel @daveasiegel.bsky.social · Sep 13

We think the framework has the potential to help a lot of measurement problems. We’re working right now to extend the input data from dichotomous to multichotomous and continuous as well. It may be of interest to #econsky as well. (7/8)

1 1

Dave Siegel @daveasiegel.bsky.social · Sep 13

There’s a lot more in the paper. For example, the figure shows how IRT-M can produce estimates of abstract concepts (sense of "threat", by the media sources that they trust) in data not designed to measure the concepts (the Eurobarometer survey). (6/8)

1 1

Dave Siegel @daveasiegel.bsky.social · Sep 13

The package does the rest. The latent dimensions it captures can be correlated, and IRT-M discovers any such correlation from the data. The supervised steps ensure that the measures remain consistent across time and space. (5/8)

1 1

Dave Siegel @daveasiegel.bsky.social · Sep 13

Then, researchers identify data sources and assess how the latent dimensions would show up in the data, which can depend on context. Then, they construct a constraint matrix that encodes dependencies between items in the data and the latent dimensions. (4/8)

1 1

Dave Siegel @daveasiegel.bsky.social · Sep 13

We call this framework IRT-M, and it makes it easier for researchers to construct, measure, and present subtle or abstract concepts in their data. It’s a semi-supervised method. First, researchers identify theoretically meaningful latent dimensions. (3/8)

1 1

Dave Siegel @daveasiegel.bsky.social · Sep 13

There’s an accompanying R package with a vignette as well here: github.com/dasiegel/IRT-M. (2/8).

GitHub - dasiegel/IRT-M: IRT-M R package

IRT-M R package. Contribute to dasiegel/IRT-M development by creating an account on GitHub.

github.com

1 1