Professor in Political Science @tcddublin | PI of ERC project PaCE @LabConflict @ERC_research. Conflict forecasting, machine learning, game theory.
Dublin City, Ireland
Joined November 2023
- Tweets214
- Following2.9K
- Followers3.3K
- Likes311
Thomas Chadefaux retweeted
📊 Check out the last @Bangladesh's fatalities pattern that mostly led to a decreasing conflict risk in history.
Explore more:
🗺️ Webapp thomasschinca.github.io/Pace…
📄 Monthly Report forecastlab.org/predictionma…
💻 Github repo github.com/ThomasSchinca/Pac…
Thomas Chadefaux retweeted
📊 Check out the last @Mozambique's fatalities pattern that mostly led to a slightly increasing conflict risk in history.
Explore more:
🗺️ Webapp thomasschinca.github.io/Pace…
📄 Monthly Report forecastlab.org/predictionma…
💻 Github repo github.com/ThomasSchinca/Pac…
📘 New publication in Journal of Forecasting, with T. Schincariol
We show that migration flows exhibit recurring temporal patterns. Leveraging these temporal shapes dramatically improves forecasting accuracy, even with limited data.
🔗doi.org/10.1002/for.3209
Check out our latest article in Political Analysis with Jian Cao! We show how units can react at varying speeds to shocks, which can distort causal estimates in Synthetic Controls.
We introduce a more efficient dynamic estimator to correct these biases bit.ly/3O3EsB2
Thomas Chadefaux retweeted
📊 Check out the last @Russia's fatalities pattern that mostly led to a stable conflict risk in history.
Explore more:
🗺️ Webapp thomasschinca.github.io/Pace…
📄 Monthly Report forecastlab.org/predictionma…
💻 Github repo github.com/ThomasSchinca/Pac…
Thomas Chadefaux retweeted
📊 Check out the last @Brazil's fatalities pattern that mostly led to a decreasing conflict risk in history.
Explore more:
🗺️ Webapp thomasschinca.github.io/Pace…
📄 Monthly Report forecastlab.org/predictionma…
💻 Github repo github.com/ThomasSchinca/Pac…
Thomas Chadefaux retweeted
Going straight into the syllabus, where it will appear on IV Is A Party Trick, Not A Real Method Day.
Out now open access at @AJPS_Editor. 194 potential exclusion-restriction violations for studies using weather as an instrumental variable onlinelibrary.wiley.com/doi/…
Thomas Chadefaux retweeted
For those of you doing PhDs that need to hear it: your dissertation is an exam *not* a published book & totalising perfection *is* the enemy of the good & the great. If this reminder is helpful, do pass it on.
Thomas Chadefaux retweeted
"We show that Black families whose ancestors were enslaved until the Civil War have considerably lower education, income, & wealth [today] than Black families whose ancestors were free before the Civil War."
Forth @QJEharvard by @AlthoffLukas & Reichardt
lukasalthoff.github.io/jmp/a…
Thomas Chadefaux retweeted
📊 Check out the last @Burundi's fatalities pattern that mostly led to a decreasing conflict risk in history.
Explore more:
🗺️ Webapp thomasschinca.github.io/Pace…
📄 Monthly Report forecastlab.org/predictionma…
💻 Github repo github.com/ThomasSchinca/Pac…
How (not) to publish a paper:
1. Submit paper
2. Rejected, R3 demands X
3. I'll do X later
4. 5 years later, f***k it, resubmit the exact same paper without X
Thomas Chadefaux retweeted
Desperately seeking a 🇫🇷 government. With no bloc holding a majority, 3 scenarios :
-minority left-wing government (short life guaranteed)
-cross-party coalition (differences hard to overcome)
-technical government (last resort)
In @TheEconomist
economist.com/europe/2024/07…
Thomas Chadefaux retweeted
This is basically wrong. The reason economists publish less is that the min publishable unit (MPU?) in econ is big. Fields make diff choices on this. The finance MPU is smaller than econ. The CS MPU is tiny. The MPU in some humanities/soc sci fields is a good which is longer
Thomas Chadefaux retweeted
The tactical cross-party candidate withdrawals in 🇫🇷 to stop Le Pen’s party winning a majority may be working. The first seat projection by a polling group for the final vote on July 7 shows the RN getting 190-220 of 577 seats, well short of the 289 needed
Source: @harrisint_fr
The idea of a geographic divide also fuels the far-right's narrative of a disconnect with "elites," which is both misleading and dangerous.
This graph from a 2017 book is problematic because it ignores education as a controlling factor, likely explaining the observed effect.
