Popular in Social Sciences

Books

City Time: On Being Sentenced to Rikers Island

Independent Scholar David A. Campbell

Culture and Society

The City as a Growth Machine: Toward a Political Economy of Place

New York University Harvey Moltoch

Economics

Gender and Economics

University of California Santa Barbara Shelly Lundberg

American Politics

Keeping the Republic

Boston College Marc Landy

Banks and Policy Organisations

Greening the financial sector

European Investment Bank Matteo Ferrazzi

British Politics

Entrepreneurial thinking: the politics and practice of policy impact

University of Sheffield Matthew Flinders

Culture and Society

Carceral Entrapments: Views from the Prison/Street Interface in India

Indian Institute of Technology Delhi Mahuya Bandyopadhya

Trade-offs of Attending Better Schools

View more insights in Social Sciences - Economics - Labor Economics

View more figures from The University of Texas at Austin

View the page of Raissa Fabregas

Read the Study

Raissa Fabregas discusses how better middle schools affect marginally admitted students’ test scores, on-time graduation, socio-emotional skills, aspirations, and high school track selections.

Image courtesy of interviewee. February 6, 2024

Generating Workspace

Export as PDF

This video video describes a study exploring the impact of attending better middle schools on children's long-term success, particularly focusing on cognitive and socio-emotional skills. The study emphasizes that while cognitive abilities such as verbal and mathematical skills are important for success, factors like social and emotional development, aspirations, and self-perceptions are also critical.

The research investigates the effects of attending higher-achieving middle schools in Mexico City, where students are assigned to schools based on their placement exam scores and family preferences, creating a clear cutoff between those who attend better schools (morning shift) and those who do not (afternoon shift). The author uses a regression discontinuity design to analyze outcomes for students near this cutoff.

Initial findings indicate that moving to better schools yields modest improvements in standardized test scores but can also negatively affect other academic measures, such as GPAs, leading to a higher likelihood of failing courses or repeating grades. Students in better schools reported lower self-perceptions of academic ability and conscientiousness, which may adversely affect their aspirations and choices for further education.

Overall, the study concludes that peer comparisons and relative rankings in schools may significantly impact students’ academic experiences and self-beliefs, potentially overshadowing the benefits of higher academic inputs associated with better schools. The author suggests the need for further research on the long-term implications of such experiences and calls for interventions aimed at providing students with objective information about their abilities to help counter negative self-perceptions.

Language:

Export as PDF

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:00.145 --> 00:02.725
Let me start by sort of explaining a little bit of,

00:02.785 --> 00:04.125
of the context of this study.

00:04.625 --> 00:08.205
So I think when we think about what factors, shape sort

00:08.205 --> 00:09.765
of like children's long-term success,

00:09.945 --> 00:12.285
and sort of like what can we affect?

00:12.745 --> 00:14.485
We think of things like, you know, verbal

00:14.545 --> 00:16.045
or mathematical ability, right?

00:16.065 --> 00:17.445
So the cognitive skills.

00:18.065 --> 00:19.605
And so these are very important,

00:20.065 --> 00:23.005
but I think there's also a range of other factors, like,

00:23.065 --> 00:25.765
you know, the social or emotional skills that you develop

00:25.865 --> 00:28.965
as you grow up, your attitudes, your aspirations,

00:29.035 --> 00:31.925
your preferences that are also gonna play a very important,

00:32.265 --> 00:33.405
role in your life.

00:34.145 --> 00:36.445
And so a question that I think is very important

00:36.545 --> 00:40.005
for social scientists, for policymakers, for parents is

00:40.005 --> 00:41.285
to try to understand to

00:41.285 --> 00:44.245
what extent schools shape these other attributes, right?

00:44.625 --> 00:47.365
And so there's already kind of this large body of work

00:47.435 --> 00:51.165
that asks, well, okay, if we send kids to a better school,

00:51.265 --> 00:53.885
and you know, we might define better schools as the schools

00:53.885 --> 00:55.525
where you have higher achieving peers,

00:55.865 --> 00:58.565
or where you have sort of like more educational inputs,

00:58.835 --> 01:01.605
what happens to, to these children in terms of learning?

01:02.305 --> 01:03.885
And something that's surprising is

01:03.885 --> 01:08.005
that actually we find very modest effects, on measures

01:08.065 --> 01:10.085
of learning from going to these better schools.

01:10.545 --> 01:13.245
But we act don't know much about sort of where it happens

01:13.305 --> 01:15.965
to these other attributes like self perceptions,

01:16.075 --> 01:17.365
like educational attainment,

01:17.665 --> 01:19.525
but like, aspirations, right?

01:19.945 --> 01:23.485
So what I'm gonna try to do in this paper is to study, okay,

01:23.875 --> 01:26.845
what is the impact of going to better, middle schools

01:27.065 --> 01:28.525
for students who are at that margin

01:28.585 --> 01:30.245
of admission in Mexico City?

01:30.625 --> 01:32.525
And I'm gonna be looking at things

01:32.525 --> 01:34.605
that other papers have already looked like,

01:34.605 --> 01:35.765
like standardized scores,

01:36.185 --> 01:38.765
but I'm also gonna be able to say things about grades,

01:38.775 --> 01:41.565
about graduation, about their socio-emotional skills

01:41.825 --> 01:43.205
and their aspirations

01:43.785 --> 01:46.565
and subsequent, educational, choices.

01:47.145 --> 01:50.565
To give you a bit of a like sense of like conceptually

01:50.745 --> 01:53.005
how we might think about different channels

01:53.005 --> 01:54.885
that might be at play in this paper.

01:55.625 --> 01:58.285
Uh, think about, a child that's deciding

01:58.285 --> 01:59.725
between two different schools, right?

02:00.105 --> 02:03.125
And they might end up placing that higher achieving school.

02:03.465 --> 02:04.885
So on the one hand, right?

02:04.985 --> 02:07.205
So you place in this better or preferred school

02:07.265 --> 02:09.645
and you might have access to more resources

02:09.825 --> 02:12.205
or your teachers might be more experienced

02:12.225 --> 02:14.005
and they encourage you more, right?

02:14.005 --> 02:15.845
And so that might make you more motivated.

02:16.305 --> 02:18.285
Uh, also you might have sort

02:18.285 --> 02:19.525
of like these higher achieving peers

02:19.705 --> 02:21.725
who are gonna be less disruptive in the classroom,

02:21.865 --> 02:23.765
or they're very motivated in school

02:23.985 --> 02:26.205
and perhaps become sort role models for you.

02:26.225 --> 02:29.605
Or they're very like, you know, engaged with schooling.

02:29.665 --> 02:32.005
Uh, and so all those factors we might think

02:32.005 --> 02:34.445
that might have like a positive impact on sort

02:34.445 --> 02:35.485
of like your aspirations

02:35.745 --> 02:37.405
and like, you know, what,

02:37.505 --> 02:39.965
how you shape all these attributes about sort

02:39.965 --> 02:42.805
of like education on the other hand, right?

02:42.825 --> 02:44.845
If you think about attending a better school,

02:45.025 --> 02:46.485
it often implies

02:46.485 --> 02:49.605
that you have like a worse position in the school

02:49.605 --> 02:50.685
ability distribution.

02:51.145 --> 02:53.605
And so that could actually negatively affect students

02:53.605 --> 02:56.445
because for instance, you might have teachers

02:56.445 --> 02:58.685
that only focus on the star students

02:59.105 --> 03:01.525
or that they're grading on relative terms, right?

03:01.545 --> 03:03.725
And so that might affect you if you have very

03:03.725 --> 03:04.925
high achieving peers.

03:05.665 --> 03:09.085
Uh, educational psychologists, I also emphasize existence

03:09.085 --> 03:11.965
of other types of effects like the big fish little pond

03:11.965 --> 03:14.005
effects, which basically predict

03:14.005 --> 03:17.365
that if you have these unfavorable, ability comparisons,

03:17.865 --> 03:20.925
that's gonna be very detrimental to your academic self image

03:21.145 --> 03:23.445
and to your effort in school, right?

03:23.545 --> 03:25.605
So there's, you know,

03:25.825 --> 03:27.885
all these different channels related to rank,

03:27.985 --> 03:30.125
and there's other literature already showing

03:30.125 --> 03:33.485
that within school, having a lower rank,

03:33.485 --> 03:35.405
that's gonna add negatively,

03:35.585 --> 03:38.205
impact your long-term outcomes as a student.

03:38.865 --> 03:41.885
So in the context of this study, we're gonna see to

03:41.885 --> 03:43.445
what extent, sort of like these type

03:43.445 --> 03:45.965
of run concerns might be sufficiently important

03:46.145 --> 03:49.765
to actually overwhelm other benefits from attending these

03:49.945 --> 03:51.045
higher achieving schools.

03:51.505 --> 03:54.405
I'm gonna look at this question by student experiences

03:54.785 --> 03:56.845
of middle school students in Mexico City.

03:57.345 --> 04:00.285
So in this setting, as in many other developing countries

04:00.285 --> 04:02.565
because of capacity constraints,

04:02.665 --> 04:04.485
school buildings have two shifts.

04:04.745 --> 04:07.805
So there's a morning shift that goes from 8:00 AM to 2:00 PM

04:07.825 --> 04:11.045
and then there's an afternoon shift that goes from 2:00 PM

04:11.105 --> 04:14.285
to 8:00 PM and these operate as independent schools.

04:14.285 --> 04:16.605
They're just, sharing a school building.

04:17.705 --> 04:21.325
Now, how the, how students are assigned

04:21.325 --> 04:23.125
to schools in Mexico City is

04:23.125 --> 04:24.525
through the centralized process.

04:25.025 --> 04:26.805
And the allocation depends on tooth lakes.

04:27.105 --> 04:29.885
The first thing is families preferences for these schools.

04:30.505 --> 04:33.965
And the second thing is the, the score that,

04:34.565 --> 04:36.725
students get in a placement is exam

04:36.725 --> 04:38.045
that they take in the sixth grade.

04:38.745 --> 04:39.845
Now, what happens is that

04:39.845 --> 04:41.725
because almost everyone prefers

04:41.725 --> 04:43.525
to attend these morning shifts,

04:43.625 --> 04:44.845
within a school building,

04:45.265 --> 04:47.685
the morning shift is usually over subscribed.

04:47.945 --> 04:51.165
And so what the algorithm does is it gives priority

04:51.195 --> 04:52.485
with those to those students

04:52.485 --> 04:54.405
that have these higher placement scores.

04:55.105 --> 04:57.405
Um, but it leads to this situation

04:57.405 --> 05:00.125
where there's gonna be sort of like this clear cutoff

05:00.185 --> 05:02.485
for getting into the morning shift, right?

05:02.625 --> 05:03.885
If you don't get to the morning shift,

05:03.945 --> 05:06.925
the algorithm immediately places you to the afternoon shift

05:06.995 --> 05:09.005
with the same school building.

05:09.585 --> 05:11.765
And so you're gonna have this clear, sort

05:11.765 --> 05:13.805
of like this continuity in,

05:13.865 --> 05:16.445
school quality at this cutoff

05:16.615 --> 05:18.965
where if you ended up in this morning shift,

05:18.965 --> 05:20.485
you're gonna have higher achieving peers,

05:20.485 --> 05:22.605
you're gonna have principals that are more educated,

05:22.745 --> 05:25.005
you're gonna have teachers that are more likely

05:25.025 --> 05:27.725
to participate in incentive programs and

05:27.725 --> 05:30.685
otherwise, sort of like you placing that afternoon shift.

05:31.145 --> 05:32.525
So this is gonna allow me

05:32.525 --> 05:34.765
to implement a regression discontinuity design

05:34.775 --> 05:36.845
where basically I'm gonna be comparing students

05:36.845 --> 05:39.965
that are very, very similar in terms of ability, in terms

05:39.965 --> 05:41.645
of school preferences, but

05:41.645 --> 05:43.925
because of very small differences on

05:43.955 --> 05:47.765
that placement exam score, they ended up placing on one side

05:47.765 --> 05:50.245
of the cutoff or another side of the cutoff, right?

05:50.265 --> 05:52.165
So they go to the morning shift, they go

05:52.165 --> 05:53.805
to these higher achieving schools and

05:53.805 --> 05:55.365
otherwise, sort of like they place

05:55.365 --> 05:56.525
to these afternoon shifts.

05:56.865 --> 05:59.485
And so I'm gonna have data on,

05:59.645 --> 06:01.685
a standardized test scores for these students.

06:01.905 --> 06:03.965
I'm gonna be able to know their GPA,

06:04.545 --> 06:08.125
and at the end of middle school they apply for high school.

06:08.225 --> 06:10.445
And at that point they fill out this questionnaire

06:10.445 --> 06:13.125
that has all these information about their social emotional

06:13.125 --> 06:14.645
skills, their aspirations,

06:15.385 --> 06:17.285
and the types of, schools

06:17.515 --> 06:20.085
that they wanna attend when they go, to high school.

06:20.265 --> 06:22.925
So I'm gonna sort of like link that data to have,

06:23.325 --> 06:26.365
outcomes in line with the existing literature of

06:26.395 --> 06:28.725
what happens when you go to these higher achieving schools.

06:29.365 --> 06:31.645
I also find very modest impacts on

06:31.675 --> 06:33.085
objective measures of learning.

06:33.145 --> 06:35.525
So this is coming from like standardized test scores

06:35.525 --> 06:36.765
that are externally graded.

06:37.185 --> 06:39.285
Uh, so I construct this index,

06:39.285 --> 06:42.045
which I call the achievement index, which is a combination

06:42.265 --> 06:44.325
of these lowest tech standardized score

06:44.325 --> 06:45.605
that they take in the seventh grade

06:45.985 --> 06:49.285
and these high stakes, high school entrance exam

06:49.285 --> 06:50.645
that they take in the ninth grade.

06:51.105 --> 06:53.085
And, you know, I find these effects

06:53.085 --> 06:56.045
of like 0.03 standard deviation gains from these

06:56.045 --> 06:57.085
higher achieving schools.

06:57.625 --> 06:59.765
Uh, all the gains are on language.

06:59.925 --> 07:01.365
I don't find any gains in math.

07:01.745 --> 07:04.205
So, you know, surprising that they're not sort

07:04.205 --> 07:06.365
of like bigger benefits in terms of these learning,

07:06.425 --> 07:08.365
but we don't find negative effects in

07:08.395 --> 07:09.765
objective measures of learning.

07:10.845 --> 07:13.305
Now, when I start looking at effects on

07:13.825 --> 07:15.025
teachers' objective rating,

07:15.665 --> 07:18.625
I actually see negative impacts from placing in these

07:18.625 --> 07:19.785
higher achieving schools.

07:20.325 --> 07:23.705
So kids have lower GPAs, so

07:23.705 --> 07:26.665
they're about point 11 standard deviations lower,

07:27.055 --> 07:29.705
they're five percentage points more likely

07:29.885 --> 07:32.625
to fail at least one course in the ninth grade.

07:32.885 --> 07:35.225
And also they're more likely to repeat grades.

07:35.725 --> 07:37.225
And so the most concerning part

07:37.225 --> 07:39.465
that I've seen in the data is that consistent

07:39.495 --> 07:42.145
with all these measures, I see that when I try

07:42.145 --> 07:43.945
to measure kids at least one year

07:43.995 --> 07:47.185
after their predicted graduation, date, I see

07:47.185 --> 07:49.425
that they're one percentage point less likely

07:49.525 --> 07:51.305
to have a middle school certificate.

07:51.645 --> 07:53.945
So basically they're less likely to graduate,

07:54.285 --> 07:55.425
from middle school.

07:56.335 --> 07:59.235
And so one interpretation of this result is

07:59.235 --> 08:00.995
that the difference between these objective

08:00.995 --> 08:04.395
and subjective measures of academic performance is that

08:04.955 --> 08:07.675
teachers are basically implicitly assessing students

08:07.775 --> 08:08.875
in relative terms, right?

08:08.875 --> 08:11.475
Even though it's sort of like there's no policy

08:11.585 --> 08:13.795
that they should be doing that in schools, right?

08:14.295 --> 08:16.955
And so this is already concerning, right?

08:16.955 --> 08:19.035
Because these, you know, these,

08:19.555 --> 08:22.035
relative grading affects sort of like students' trajectories

08:22.035 --> 08:25.315
during middle schools, but there's these other channel which

08:25.475 --> 08:28.235
is that grades also serve as a feedback mechanism

08:28.615 --> 08:29.715
for students, right?

08:29.975 --> 08:33.235
And so if students have imperfect informations about their

08:33.235 --> 08:36.435
own ability, they might be using sort of these signals to,

08:36.575 --> 08:37.715
to think about like, well,

08:37.735 --> 08:40.435
how good I am at this academic stuff, right?

08:40.735 --> 08:44.275
And so I actually can use these data from the questionnaire

08:44.575 --> 08:46.835
to try to see how they think about themselves

08:46.835 --> 08:48.235
and what their self-perceptions are.

08:48.535 --> 08:51.955
And so I, I see that students actually report that they're,

08:52.095 --> 08:54.555
less academically inclined than their peers

08:55.345 --> 08:59.005
and they are scoring lower in measures of conscientiousness,

08:59.015 --> 09:01.805
which is sort of this tendency to be hardworking,

09:01.805 --> 09:04.085
to be responsible to be organized, right?

09:04.585 --> 09:06.125
And so sort of we see sort

09:06.125 --> 09:07.485
of like these negative effects on

09:07.485 --> 09:08.605
these self-reported measures.

09:09.185 --> 09:12.765
Now, you know, if these are real effects, that's very

09:12.765 --> 09:14.765
concerning because there's a lot

09:14.765 --> 09:16.805
of other work linking conscientiousness

09:16.825 --> 09:18.405
to long-term outcomes of children.

09:19.065 --> 09:22.645
But it's possible that when these self-reported measure is

09:22.645 --> 09:25.205
just self-perceptions and it's just reference bias, right?

09:25.265 --> 09:28.245
So I go to school, maybe I see other

09:28.835 --> 09:30.565
kids are very hardworking

09:30.565 --> 09:32.965
and I think, maybe I'm not that hardworking.

09:33.065 --> 09:35.725
But you know, maybe objectively you were like as hardwork

09:35.785 --> 09:37.085
as you were gonna be no matter

09:37.085 --> 09:39.245
what school you're placed in, right?

09:39.425 --> 09:41.165
So I cannot say much to that

09:41.325 --> 09:43.045
'cause I only have the self-reported data,

09:43.185 --> 09:46.925
but what I can say is that I, I have this evidence of a link

09:46.925 --> 09:48.645
that self perceptions actually also

09:48.645 --> 09:50.005
matter for other things, right?

09:50.305 --> 09:52.645
So I can link self-perceptions,

09:52.865 --> 09:56.205
or I can think about, you know, even if I think

09:56.205 --> 09:57.565
that I'm not good enough

09:57.585 --> 10:00.005
and it's not real, that might eventually affect

10:00.065 --> 10:03.085
how I think about where to apply or what I aspire to.

10:03.505 --> 10:06.165
And so I do see sort of like these negative effects on

10:06.965 --> 10:10.445
children's, aspirations to go to university.

10:10.785 --> 10:14.005
And they're also one percentage point less likely to apply

10:14.225 --> 10:17.565
to, academic programs instead choosing

10:17.745 --> 10:19.125
for vocational programs.

10:19.755 --> 10:23.485
There's a lot of work already in education, trying

10:23.485 --> 10:26.365
to understand sort of like how schools affect

10:26.875 --> 10:28.365
objective measures of learning.

10:28.885 --> 10:32.925
I think there's a lot less work done understanding, okay,

10:33.185 --> 10:35.485
how do schools affect other channels

10:35.485 --> 10:37.685
that could eventually have important implications

10:37.745 --> 10:39.925
for students' educational trajectory?

10:40.145 --> 10:41.965
So one implication is that we need

10:41.965 --> 10:45.005
to study a lot more about those potential channels.

10:45.825 --> 10:48.725
Now in this context, I think that something

10:48.725 --> 10:50.805
that's perhaps a little bit surprising is

10:50.805 --> 10:53.125
that these rank effects

10:53.185 --> 10:56.525
or these peer comparisons seem to be sufficiently important

10:56.675 --> 10:59.165
that they overwhelm other channels

10:59.165 --> 11:00.725
through which better schools might have

11:00.725 --> 11:02.365
helped students, right?

11:02.385 --> 11:06.885
So that's sort of like one, one conclusion of this study.

11:07.665 --> 11:09.605
Um, now, you know,

11:09.765 --> 11:12.325
I think things like not graduating from middle school,

11:12.505 --> 11:14.205
that's, you know,

11:14.325 --> 11:16.045
a very negative outcome for students.

11:16.285 --> 11:17.765
I think there's other types

11:17.905 --> 11:20.925
of outcomes like choosing more vocational programs

11:20.925 --> 11:22.285
that it's unclear sort of whether

11:22.285 --> 11:24.165
that's gonna have like positive

11:24.185 --> 11:27.165
or negative effects on, on student long-term outcomes.

11:27.545 --> 11:31.045
And so studying more sort of what are the long-term outcomes

11:31.065 --> 11:33.565
of attending these schools is also gonna be important.

11:34.265 --> 11:37.245
But I think in general, you know, it's gonna be und

11:37.885 --> 11:39.405
undesirable for students to

11:40.155 --> 11:43.805
form their self-perceptions about their own ability based on

11:43.835 --> 11:46.965
sort of like their relative comparisons, right?

11:46.965 --> 11:49.045
Because that could lead to distorted beliefs.

11:49.505 --> 11:52.725
And so thinking about what might be effective interventions

11:52.755 --> 11:54.285
That provide students with sort

11:54.285 --> 11:56.325
of like objective information about their own ability

11:56.865 --> 11:59.365
or perhaps, or reaffirm their sense

11:59.365 --> 12:03.165
of adequacy in school seems to be like a, a promising,

12:03.705 --> 12:05.405
policy to, to study.

00:00 - Let me start by sort of explaining a little bit of, of the context of this study. So I think when we think about what factors, shape sort of like children's long-term success, and sort of like what can we affect? We think of things like, you know, verbal

00:14 - or mathematical ability, right? So the cognitive skills. And so these are very important, but I think there's also a range of other factors, like, you know, the social or emotional skills that you develop as you grow up, your attitudes, your aspirations, your preferences that are also gonna play a very important, role in your life. And so a question that I think is very important for social scientists, for policymakers, for parents is

00:40 - to try to understand to what extent schools shape these other attributes, right? And so there's already kind of this large body of work that asks, well, okay, if we send kids to a better school, and you know, we might define better schools as the schools where you have higher achieving peers, or where you have sort of like more educational inputs, what happens to, to these children in terms of learning? And something that's surprising is that actually we find very modest effects, on measures

01:08 - of learning from going to these better schools. But we act don't know much about sort of where it happens to these other attributes like self perceptions, like educational attainment, but like, aspirations, right? So what I'm gonna try to do in this paper is to study, okay, what is the impact of going to better, middle schools for students who are at that margin of admission in Mexico City? And I'm gonna be looking at things

01:32 - that other papers have already looked like, like standardized scores, but I'm also gonna be able to say things about grades, about graduation, about their socio-emotional skills and their aspirations and subsequent, educational, choices. To give you a bit of a like sense of like conceptually how we might think about different channels that might be at play in this paper. think about, a child that's deciding

01:58 - between two different schools, right? And they might end up placing that higher achieving school. So on the one hand, right? So you place in this better or preferred school and you might have access to more resources or your teachers might be more experienced and they encourage you more, right? And so that might make you more motivated. also you might have sort of like these higher achieving peers

02:19 - who are gonna be less disruptive in the classroom, or they're very motivated in school and perhaps become sort role models for you. Or they're very like, you know, engaged with schooling. and so all those factors we might think that might have like a positive impact on sort of like your aspirations and like, you know, what, how you shape all these attributes about sort of like education on the other hand, right?

02:42 - If you think about attending a better school, it often implies that you have like a worse position in the school ability distribution. And so that could actually negatively affect students because for instance, you might have teachers that only focus on the star students or that they're grading on relative terms, right? And so that might affect you if you have very high achieving peers.

03:05 - educational psychologists, I also emphasize existence of other types of effects like the big fish little pond effects, which basically predict that if you have these unfavorable, ability comparisons, that's gonna be very detrimental to your academic self image and to your effort in school, right? So there's, you know, all these different channels related to rank, and there's other literature already showing that within school, having a lower rank,

03:33 - that's gonna add negatively, impact your long-term outcomes as a student. So in the context of this study, we're gonna see to what extent, sort of like these type of run concerns might be sufficiently important to actually overwhelm other benefits from attending these higher achieving schools. I'm gonna look at this question by student experiences of middle school students in Mexico City. So in this setting, as in many other developing countries

04:00 - because of capacity constraints, school buildings have two shifts. So there's a morning shift that goes from 8:00 AM to 2:00 PM and then there's an afternoon shift that goes from 2:00 PM to 8:00 PM and these operate as independent schools. They're just, sharing a school building. Now, how the, how students are assigned to schools in Mexico City is through the centralized process. And the allocation depends on tooth lakes.

04:27 - The first thing is families preferences for these schools. And the second thing is the, the score that, students get in a placement is exam that they take in the sixth grade. Now, what happens is that because almost everyone prefers to attend these morning shifts, within a school building, the morning shift is usually over subscribed. And so what the algorithm does is it gives priority

04:51 - with those to those students that have these higher placement scores. but it leads to this situation where there's gonna be sort of like this clear cutoff for getting into the morning shift, right? If you don't get to the morning shift, the algorithm immediately places you to the afternoon shift with the same school building. And so you're gonna have this clear, sort of like this continuity in,

05:13 - school quality at this cutoff where if you ended up in this morning shift, you're gonna have higher achieving peers, you're gonna have principals that are more educated, you're gonna have teachers that are more likely to participate in incentive programs and otherwise, sort of like you placing that afternoon shift. So this is gonna allow me to implement a regression discontinuity design where basically I'm gonna be comparing students

05:36 - that are very, very similar in terms of ability, in terms of school preferences, but because of very small differences on that placement exam score, they ended up placing on one side of the cutoff or another side of the cutoff, right? So they go to the morning shift, they go to these higher achieving schools and otherwise, sort of like they place to these afternoon shifts. And so I'm gonna have data on,

05:59 - a standardized test scores for these students. I'm gonna be able to know their GPA, and at the end of middle school they apply for high school. And at that point they fill out this questionnaire that has all these information about their social emotional skills, their aspirations, and the types of, schools that they wanna attend when they go, to high school. So I'm gonna sort of like link that data to have, outcomes in line with the existing literature of

06:26 - what happens when you go to these higher achieving schools. I also find very modest impacts on objective measures of learning. So this is coming from like standardized test scores that are externally graded. so I construct this index, which I call the achievement index, which is a combination of these lowest tech standardized score that they take in the seventh grade and these high stakes, high school entrance exam

06:49 - that they take in the ninth grade. And, you know, I find these effects of like 0.03 standard deviation gains from these higher achieving schools. all the gains are on language. I don't find any gains in math. So, you know, surprising that they're not sort of like bigger benefits in terms of these learning, but we don't find negative effects in objective measures of learning.

07:10 - Now, when I start looking at effects on teachers' objective rating, I actually see negative impacts from placing in these higher achieving schools. So kids have lower GPAs, so they're about point 11 standard deviations lower, they're five percentage points more likely to fail at least one course in the ninth grade. And also they're more likely to repeat grades. And so the most concerning part

07:37 - that I've seen in the data is that consistent with all these measures, I see that when I try to measure kids at least one year after their predicted graduation, date, I see that they're one percentage point less likely to have a middle school certificate. So basically they're less likely to graduate, from middle school. And so one interpretation of this result is that the difference between these objective

08:00 - and subjective measures of academic performance is that teachers are basically implicitly assessing students in relative terms, right? Even though it's sort of like there's no policy that they should be doing that in schools, right? And so this is already concerning, right? Because these, you know, these, relative grading affects sort of like students' trajectories during middle schools, but there's these other channel which is that grades also serve as a feedback mechanism

08:28 - for students, right? And so if students have imperfect informations about their own ability, they might be using sort of these signals to, to think about like, well, how good I am at this academic stuff, right? And so I actually can use these data from the questionnaire to try to see how they think about themselves and what their self-perceptions are. And so I, I see that students actually report that they're, less academically inclined than their peers

08:55 - and they are scoring lower in measures of conscientiousness, which is sort of this tendency to be hardworking, to be responsible to be organized, right? And so sort of we see sort of like these negative effects on these self-reported measures. Now, you know, if these are real effects, that's very concerning because there's a lot of other work linking conscientiousness to long-term outcomes of children.

09:19 - But it's possible that when these self-reported measure is just self-perceptions and it's just reference bias, right? So I go to school, maybe I see other kids are very hardworking and I think, maybe I'm not that hardworking. But you know, maybe objectively you were like as hardwork as you were gonna be no matter what school you're placed in, right? So I cannot say much to that 'cause I only have the self-reported data,

09:43 - but what I can say is that I, I have this evidence of a link that self perceptions actually also matter for other things, right? So I can link self-perceptions, or I can think about, you know, even if I think that I'm not good enough and it's not real, that might eventually affect how I think about where to apply or what I aspire to. And so I do see sort of like these negative effects on children's, aspirations to go to university.

10:10 - And they're also one percentage point less likely to apply to, academic programs instead choosing for vocational programs. There's a lot of work already in education, trying to understand sort of like how schools affect objective measures of learning. I think there's a lot less work done understanding, okay, how do schools affect other channels that could eventually have important implications for students' educational trajectory?

10:40 - So one implication is that we need to study a lot more about those potential channels. Now in this context, I think that something that's perhaps a little bit surprising is that these rank effects or these peer comparisons seem to be sufficiently important that they overwhelm other channels through which better schools might have helped students, right? So that's sort of like one, one conclusion of this study.

11:07 - now, you know, I think things like not graduating from middle school, that's, you know, a very negative outcome for students. I think there's other types of outcomes like choosing more vocational programs that it's unclear sort of whether that's gonna have like positive or negative effects on, on student long-term outcomes. And so studying more sort of what are the long-term outcomes

11:31 - of attending these schools is also gonna be important. But I think in general, you know, it's gonna be und undesirable for students to form their self-perceptions about their own ability based on sort of like their relative comparisons, right? Because that could lead to distorted beliefs. And so thinking about what might be effective interventions That provide students with sort of like objective information about their own ability or perhaps, or reaffirm their sense

11:59 - of adequacy in school seems to be like a, a promising, policy to, to study.

Faculti AI can make mistakes