Sönke H. Matthewes
@sonkem.bsky.social
Assistant Professor of Applied Economics at Utrecht University. Research Associate at CEP/LSE and IZA. Fields: education, labor, & applied metrics.
https://sites.google.com/view/sonkematthewes/
https://sites.google.com/view/sonkematthewes/
Haha very recognisable!
July 24, 2025 at 10:15 AM
Haha very recognisable!
Shouldn’t genes have been cited as confounders (instead of mediators) by those doubting a causal relationship or what am I missing?
July 21, 2025 at 10:57 PM
Shouldn’t genes have been cited as confounders (instead of mediators) by those doubting a causal relationship or what am I missing?
What’s a firm choice course?
July 15, 2025 at 9:42 AM
What’s a firm choice course?
I don’t know what the plan is, but would be nice to see you!
May 29, 2025 at 4:34 PM
I don’t know what the plan is, but would be nice to see you!
Better LATE than nothing ;)
May 29, 2025 at 1:41 PM
Better LATE than nothing ;)
* omtrent brede scholen: tenminste niet als ze moeten concurreren met vwo scholen.
February 4, 2025 at 11:23 AM
* omtrent brede scholen: tenminste niet als ze moeten concurreren met vwo scholen.
for (i.e. compete with) non-academic schools, though.
Finally, also note that (in contrast to VHS) we have few CS openings in our data, limiting stat power. I stylized our findings a bit for this thread: probably better to say we don’t reject the null of no effect for CS rather than there is no xf.
Finally, also note that (in contrast to VHS) we have few CS openings in our data, limiting stat power. I stylized our findings a bit for this thread: probably better to say we don’t reject the null of no effect for CS rather than there is no xf.
February 4, 2025 at 11:12 AM
for (i.e. compete with) non-academic schools, though.
Finally, also note that (in contrast to VHS) we have few CS openings in our data, limiting stat power. I stylized our findings a bit for this thread: probably better to say we don’t reject the null of no effect for CS rather than there is no xf.
Finally, also note that (in contrast to VHS) we have few CS openings in our data, limiting stat power. I stylized our findings a bit for this thread: probably better to say we don’t reject the null of no effect for CS rather than there is no xf.
2. To your second point: our observation period purposefully excludes the recent reforms (starting in 2010) that relabeled non-academic track schools as comprehensive schools. Virtually all CS in our sample should be genuine CS from the start. In practice, they still seem to mainly substitute 3/n
February 4, 2025 at 11:12 AM
2. To your second point: our observation period purposefully excludes the recent reforms (starting in 2010) that relabeled non-academic track schools as comprehensive schools. Virtually all CS in our sample should be genuine CS from the start. In practice, they still seem to mainly substitute 3/n
I see why you prefer to call it increasing flexibility in the case of VHS (that's how we call it in the first sentence of the abstract). To me it doesn't matter.
Our paper doesn't speak to whether or not students benefitted from the early selection / attending a less demanding track first. 2/n
Our paper doesn't speak to whether or not students benefitted from the early selection / attending a less demanding track first. 2/n
February 4, 2025 at 11:12 AM
I see why you prefer to call it increasing flexibility in the case of VHS (that's how we call it in the first sentence of the abstract). To me it doesn't matter.
Our paper doesn't speak to whether or not students benefitted from the early selection / attending a less demanding track first. 2/n
Our paper doesn't speak to whether or not students benefitted from the early selection / attending a less demanding track first. 2/n
Thanks!
1. To your semantic point: we call it de-tracking *at the margin* bc both new school types de-track *upper-sec* school by offering the academic upper-sec grades and awarding Abitur. Indeed they don't de-track lower-sec school (though CS might have that ambition but de facto it doesn't). 1/n
1. To your semantic point: we call it de-tracking *at the margin* bc both new school types de-track *upper-sec* school by offering the academic upper-sec grades and awarding Abitur. Indeed they don't de-track lower-sec school (though CS might have that ambition but de facto it doesn't). 1/n
February 4, 2025 at 11:12 AM
Thanks!
1. To your semantic point: we call it de-tracking *at the margin* bc both new school types de-track *upper-sec* school by offering the academic upper-sec grades and awarding Abitur. Indeed they don't de-track lower-sec school (though CS might have that ambition but de facto it doesn't). 1/n
1. To your semantic point: we call it de-tracking *at the margin* bc both new school types de-track *upper-sec* school by offering the academic upper-sec grades and awarding Abitur. Indeed they don't de-track lower-sec school (though CS might have that ambition but de facto it doesn't). 1/n
Welcome. Thanks for engaging! The paper tackles a rather small q tbh, given that we only observe aggregate Abitur attainment. For me it was interesting to see that (given early tracking) the sometimes overlooked upgrading options like VHS play a larger role for attainment than more discussed CS.
February 3, 2025 at 8:56 AM
Welcome. Thanks for engaging! The paper tackles a rather small q tbh, given that we only observe aggregate Abitur attainment. For me it was interesting to see that (given early tracking) the sometimes overlooked upgrading options like VHS play a larger role for attainment than more discussed CS.
Ah, time dimension here is already cohorts (not years). So the 1st treated cohort is the 1st who had a CS available when transitioning from primary to secondary. Their attainment is recorded a few years down the line. So, there is no mechanical reason for CS to have more lagged effects than VHS.
February 3, 2025 at 8:25 AM
Ah, time dimension here is already cohorts (not years). So the 1st treated cohort is the 1st who had a CS available when transitioning from primary to secondary. Their attainment is recorded a few years down the line. So, there is no mechanical reason for CS to have more lagged effects than VHS.
I’ve abstained from making too much of it so far, but yes that’s a good point. Thx! (Though, as a more general point, the parallel trends assumption gets increasingly stronger the further away from the event date, right?! It’s something I’ve wondered about often but that rarely gets I mentioned.)
February 3, 2025 at 8:13 AM
I’ve abstained from making too much of it so far, but yes that’s a good point. Thx! (Though, as a more general point, the parallel trends assumption gets increasingly stronger the further away from the event date, right?! It’s something I’ve wondered about often but that rarely gets I mentioned.)
(We have many more treatment events for VHS, so the “natural experiment” for VHS is clearly much stronger.)
February 3, 2025 at 8:02 AM
(We have many more treatment events for VHS, so the “natural experiment” for VHS is clearly much stronger.)
Thanks! Short answer is that we lack the power to detect this. There is only *very* few CS openings in our time frame that we can observe 9+ years later. I stylized our findings a bit for this thread. Prob it’s better to say we don’t reject the null of no effect for CS rather than there is no xf.
February 3, 2025 at 7:59 AM
Thanks! Short answer is that we lack the power to detect this. There is only *very* few CS openings in our time frame that we can observe 9+ years later. I stylized our findings a bit for this thread. Prob it’s better to say we don’t reject the null of no effect for CS rather than there is no xf.
Much more in the paper ofc. Check it out here in case you're interested. It's open access.
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
De-tracking at the margin: How alternative secondary education pathways affect student attainment
This paper estimates how marginal increases in the flexibility of between-school tracking affect student attainment by exploiting the addition of non-…
www.sciencedirect.com
February 3, 2025 at 7:09 AM
Much more in the paper ofc. Check it out here in case you're interested. It's open access.
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
In contrast, CS don't raise Abitur attainment likely bc they don't provide a better peer environment than lower tracks bc the academic-track still cream-skims the best students. Despite their name, CS fail to deliver a truly comprehensive alternative within an otherwise tracked system. [n/n]
February 3, 2025 at 7:09 AM
In contrast, CS don't raise Abitur attainment likely bc they don't provide a better peer environment than lower tracks bc the academic-track still cream-skims the best students. Despite their name, CS fail to deliver a truly comprehensive alternative within an otherwise tracked system. [n/n]
Likely reasons for the success of VHS are that they lower the costs of educational upgrading for late-bloomers (highlighting the importance of 2nd-chance options in tracked systems!) and their hybrid curriculum, which may retain students in general school who'd otherwise enter VET. [11/n]
February 3, 2025 at 7:09 AM
Likely reasons for the success of VHS are that they lower the costs of educational upgrading for late-bloomers (highlighting the importance of 2nd-chance options in tracked systems!) and their hybrid curriculum, which may retain students in general school who'd otherwise enter VET. [11/n]
Finding 2: For vocational high schools (VHS), in contrast, we find robust evidence for positive attainment effects not only in cross-sectional and TWFE panel regressions, but also in an event-study design that exploits the quasi-random timing of new school openings. [10/n]
February 3, 2025 at 7:09 AM
Finding 2: For vocational high schools (VHS), in contrast, we find robust evidence for positive attainment effects not only in cross-sectional and TWFE panel regressions, but also in an event-study design that exploits the quasi-random timing of new school openings. [10/n]
Finding 1: comprehensive school (CS) supply correlates positively with Abitur attainment, but this is not robust to the inclusion of regional controls, suggesting that it reflects regional differences in educational demand rather than causal supply-side effects. [9/n]
February 3, 2025 at 7:09 AM
Finding 1: comprehensive school (CS) supply correlates positively with Abitur attainment, but this is not robust to the inclusion of regional controls, suggesting that it reflects regional differences in educational demand rather than causal supply-side effects. [9/n]