sociologist/criminologist AP @HamlineU. PhD @UMNsociology. Crime, punishment, inequality, quant.
Minneapolis
Joined February 2011
- Tweets1.2K
- Following489
- Followers387
- Likes2.1K
Ryan Larson retweeted
People dunking on this saying that TWFE is just fine are omitting an important detail that these are from “top 5” -- papers that have been vetted for identification, passed several robustness checks, gone through several rounds of reviews, and then gone through.
Ryan Larson retweeted
This paper extends difference-in-differences methods to settings with non-binary treatments and #heterogeneous effects. The #framework allows researchers to estimate #dose-response relationships when treatment intensity varies over time and across units. spkl.io/60177Vc8R
Ryan Larson retweeted
Please enjoy!
Writeup: dropbox.com/scl/fi/ago5vsi89…
Repo: github.com/peterdhull/twfe-r…
Planning to expand to other journals soon
Ryan Larson retweeted
Nature research paper: Banking records reveal characteristics of financial abuse
go.nature.com/46HkdT7
Ryan Larson retweeted
He fit the same statistical models with three different software and got much different estimates. It’s another dimension of the multiverse.
statmodeling.stat.columbia.e…
Ryan Larson retweeted
New in Criminology! Punishment as a racialized package of confinement, cash, and surveillance. By @ryanplarson, Rob Stewart, @Veronica_L_H, and some old guy. This detailed thread by Ryan show ‘n tells the story.
New paper out in Criminology w/ Rob Stewart, @Veronica_L_H , & @chrisuggen
We model punishment more closely as it is experienced by defendants - a constitutive package of incarceration, probation, and monetary sanctions.
onlinelibrary.wiley.com/doi/…
Overall, we hope our study illustrates the benefits of modeling punishment more closely as it is experienced: as a constitutive package of different forms and intensities of punishment.
non-paywall link here: ryanplarson.github.io/person…
These results highlight how punishment is racialized in complex ways depending upon the axis and context. Accounting for this mix is important for the estimation of racial disparities in sentencing - with both prob/LFOs racialized punitive terrains above and beyond incarceration.
Not only are the packaging effects conditioned by sentencing context, but so is the racialization in sentencing - we see more robust race effects among non-presumptive prison cases for both probation and LFOs.
How does this change the picture of the race effects? For LFOs, the IV race effects are substantively similar, but probation for Black and Native American defendants is higher as compared to White (~7.4% and 66%) - highlighting how packaging can mask important racial disparities.
We find that a 1% increase in instrumented incarceration results in a 0.27% decrease in LFOs and a 2.3% decrease in probation. Importantly, sentencing context moderates these effects - with more packaging malleability at lower levels of sentencing (i.e., non-presumptive prison).
We model the interdependency of incarceration and probation/LFOs using IV models leveraging county jail capacity as an instrument. In the first stage, capacity interacts with sentencing context - pushing felony towards prison and GM to lower confinement to reduce overcrowding.
However, this classic approach assumes that each axis of punishment is independent of the others. We conceptualize punishment as a constitutive "package" where punishment modalities are endogenous to one another. This carries both substantive and methodological implications...
These "extra-legal" race effects hold up in standard multivariate models that adjust for legal, defendant, and case factors, with BIPOC defendants associated with 13-39% more incarceration, 19-29% lower probation, and 7-46% lower LFOs compared to white defendants.
First we descriptively show that the overall racialized pattern in sentenced punishment reveals a bifurcation along the lines of race, with White defendants receiving longer community surveillance and larger monetary punishments, and BIPOC defendants receiving
longer confinement.
New paper out in Criminology w/ Rob Stewart, @Veronica_L_H , & @chrisuggen
We model punishment more closely as it is experienced by defendants - a constitutive package of incarceration, probation, and monetary sanctions.
onlinelibrary.wiley.com/doi/…
Ryan Larson retweeted
New in Nature Human Behaviour: How Deceptive Online Networks Reached Millions in the US 2020 Elections
-Reached at least 37M Facebook and 3M Instagram users
-3 networks out of 49 responsible for >70% of users reached
-Exposed users older, more conservative
Ryan Larson retweeted
@Nature has published three groundbreaking papers on reproducibility, analytical robustness, and replicability across the social sciences. Sincere thanks are due to the many folks who contributed to these projects. It’s painstaking work, and a great service to social science.
Ryan Larson retweeted
New Acemoglu et al. paper: AI may boost short-term decision quality but can erode the learning incentives that sustain collective knowledge in society, creating a risk of long-run knowledge decline or collapse.
