Julian Gerez
@juliangerez.com
Assistant Professor of Criminology, Law and Society and Political Science (by courtesy) at the University of California, Irvine. I study the political economy of security, crime, and law enforcement.
These are great—it would be cool to facet the plots so that they share the same y-axis, because I think these are even understating the shift (subject to turnout differences and ecological inference, etc.)!
November 5, 2025 at 5:07 PM
These are great—it would be cool to facet the plots so that they share the same y-axis, because I think these are even understating the shift (subject to turnout differences and ecological inference, etc.)!
And conditioning on those other characteristics isn't necessarily helpful (hence the DAG point), because many of those variables lie downstream of moderation in the causal process...
November 1, 2025 at 6:23 PM
And conditioning on those other characteristics isn't necessarily helpful (hence the DAG point), because many of those variables lie downstream of moderation in the causal process...
Put differently, none of this evidence can tell you what would happen if you could flip a switch and make a candidate moderate, all else equal. It only tells you the difference between candidates who are moderate—along with all other characteristics that come bundled with that—and those who are not.
November 1, 2025 at 6:23 PM
Put differently, none of this evidence can tell you what would happen if you could flip a switch and make a candidate moderate, all else equal. It only tells you the difference between candidates who are moderate—along with all other characteristics that come bundled with that—and those who are not.
To expand on this, my view on this debate is that part of the issue is that the Nate Silver-types misunderstand causality.
Even if the moderate advantage is X number of percentage points that doesn't imply that *moderation* will be successful. These are all just conditional correlations (at best).
Even if the moderate advantage is X number of percentage points that doesn't imply that *moderation* will be successful. These are all just conditional correlations (at best).
November 1, 2025 at 6:23 PM
To expand on this, my view on this debate is that part of the issue is that the Nate Silver-types misunderstand causality.
Even if the moderate advantage is X number of percentage points that doesn't imply that *moderation* will be successful. These are all just conditional correlations (at best).
Even if the moderate advantage is X number of percentage points that doesn't imply that *moderation* will be successful. These are all just conditional correlations (at best).
Even a very low p-value only indicates that, if the null hypothesis were true, observing a test statistic that large would be unlikely. It does not, on its own, establish causality. Causal inference depends on credible research design and identifying assumptions, not statistical significance.
November 1, 2025 at 6:04 PM
Even a very low p-value only indicates that, if the null hypothesis were true, observing a test statistic that large would be unlikely. It does not, on its own, establish causality. Causal inference depends on credible research design and identifying assumptions, not statistical significance.
Genial Sergi! Lo compartiré con mis redes.
October 25, 2025 at 2:30 PM
Genial Sergi! Lo compartiré con mis redes.
This is great! I've always wanted to do this but never had the time—and am grateful you did so at the shot level. Can you provide some descriptives on the variance parameters? That was always my big curiosity from this type of approach.
www.juliangerez.com/blog/2025/02...
www.juliangerez.com/blog/2025/02...
Utilizing variance in xG to understand finishing ability: a concept
www.juliangerez.com
October 22, 2025 at 7:17 PM
This is great! I've always wanted to do this but never had the time—and am grateful you did so at the shot level. Can you provide some descriptives on the variance parameters? That was always my big curiosity from this type of approach.
www.juliangerez.com/blog/2025/02...
www.juliangerez.com/blog/2025/02...
I'm a true believer and I checked out the book from the library... but I will order a copy to have on my shelf and help proselytize!
September 17, 2025 at 2:47 PM
I'm a true believer and I checked out the book from the library... but I will order a copy to have on my shelf and help proselytize!
Do you think this would be an issue if he played a different position?
August 14, 2025 at 5:58 PM
Do you think this would be an issue if he played a different position?
Congratulations Ana Paula, fantastic!
August 4, 2025 at 12:15 PM
Congratulations Ana Paula, fantastic!
Indeed. And Columbia has made it harder for everyone else by demonstrating that the federal government's coercive strategy works. They've narrowed the space for other institutions to resist this type of extortion.
August 2, 2025 at 2:25 PM
Indeed. And Columbia has made it harder for everyone else by demonstrating that the federal government's coercive strategy works. They've narrowed the space for other institutions to resist this type of extortion.