WEBVTT

1
00:00:00.001 --> 00:00:01.520
thank you very much.

2
00:00:02.820 --> 00:00:12.330
 There you go.

3
00:00:12.700 --> 00:00:30.959
My name is Valentina. I'm from the University of Bristol in the United kingdom, and I'll be the chair of the sessions on school and inequalities. We have 4 presentation, so we can do 15 min, but maybe a tiny bit longer. We're not super pressure for time, and maybe have all of the questions at the end in a nice longer

4
00:00:31.030 --> 00:00:37.519
kind of Q&A session before we start. I would like to start with the acknowledgment of country

5
00:00:37.560 --> 00:01:03.820
on behalf of everyone gathered. I pay respect to this period of country here in Nam where we're meeting. I acknowledged the past and present cooling nation war, injury, war warrant as the sovereign people. Here the cultural knowledge holders to maintain and promote the norm, aboriginal law, customs and constitution, responsibilities for country and the people. That country continues to sustain generation after generation into the future. For all of us.

6
00:01:03.830 --> 00:01:12.310
I recognize the important contribution that first nations think thinkers have made to sociology, including challenges to our dominant ways of thinking.

7
00:01:12.420 --> 00:01:19.959
We thank all first nation delegates in the room today for their contributions to knowledge in the discipline.

8
00:01:21.060 --> 00:01:22.899
and I think our first speaker

9
00:01:23.220 --> 00:01:25.240
is Anna.

10
00:01:26.140 --> 00:01:27.400
I'll click

11
00:01:28.160 --> 00:01:29.190
should work.

12
00:01:30.640 --> 00:01:32.110
Yes, perfect

13
00:01:35.730 --> 00:01:56.300
Hello, and welcome to my presentation about the role of the socio-economic and micro-specific classroom composition when it comes to competences in mathematics and science of refugee students by Levita Maclinka. And I'm working in a project where we focus on the educational integration of refugee students in Germany.

14
00:01:57.110 --> 00:02:20.280
And when we see this graph, we can see why this is important. So the numbers of refugees in Germany is constantly increasing, especially in 2,014, due to the war in Syria, and again in 2,022, due to the war in Ukraine. But yeah, globally, there, Any crisis or the number of refugees will increase even further in the future.

15
00:02:20.310 --> 00:02:30.540
By now we have over 3 million refugees living in Germany, and we know that one third of them is actually underaged, and therefore in compulsory

16
00:02:30.840 --> 00:02:32.850
school age.

17
00:02:33.050 --> 00:02:47.490
yeah. And so the whole topic of integration through education is a big deal in Germany right now, not only for politics, but also for research and recent studies show that refugee students tend to have

18
00:02:47.690 --> 00:03:01.009
lower achievements, schooling achievements than students without a migration or background, but also lowered in students with a migrational background, but no flight experience.

19
00:03:01.430 --> 00:03:23.589
So there's something about the individual effects of refugees in Germany. But we want to focus on the compositional effects. And our research question is, what is the association between the socioeconomic and micro-specific classroom composition and individual compenses in mathematics and science of refugee students

20
00:03:24.820 --> 00:03:31.089
to give you a little bit of context. I'd like to tell you something about the situation of refugee students in Germany.

21
00:03:31.310 --> 00:03:37.580
What we know about refugee students is that they tend to have Low German skills which is mainly due to their flight experience

22
00:03:37.650 --> 00:03:54.079
so they can't prepare for the language in the hosting country in the country of origin, because they don't have the means. They don't have the time, because the flight might be spontaneous, and also in many cases they don't know in which hosting country they might end up.

23
00:03:54.580 --> 00:04:01.219
and therefore the German skills are very low when they come to Germany even lower than for other migrant groups.

24
00:04:01.360 --> 00:04:19.789
Also they tend to have a very low socioeconomic status, which is, on the one hand, due to their lack of language skills which are very important on the German labor market, and also because there are institutional regulations, so they are not allowed to work in the first phase of their of them being in Germany.

25
00:04:20.040 --> 00:04:31.159
 Also, educational certificates from their country of origin might not be accepted in Germany, and that leads to a low socioeconomic status of refugees in general.

26
00:04:31.400 --> 00:04:45.269
and, moreover, they are more often placed in neighborhoods with a low, mean, socio-economic status, which is, on the one hand, because the rents are cheaper there, but also because there are institutional regulations which place them there.

27
00:04:45.610 --> 00:04:57.859
So the schools in these neighborhoods reflect this mean socioeconomic status of the neighborhoods, and so refugee students are more often enrolled in schools with low, mean, socioeconomic status.

28
00:04:58.430 --> 00:05:20.609
but not least, the educational trajectories of refugee students are somewhat special, because they have mainly interrupted educational trajectories, on the one hand, in the country of origin, because the educational infrastructure in the country of origin was disrupted through to the crisis, but also during their flight. They couldn't attend appropriate schooling.

29
00:05:20.910 --> 00:05:28.209
Also, when it comes to Germany it takes on average 7 months until they actually go to the German school system.

30
00:05:28.370 --> 00:05:57.490
 so yeah, when they are actually in the German school system, they come to the welcome classes, at least for the refugee students I'm talking about who came from 2,014 and ongoing years. and these welcome classes are special classes for migrants who just came into the German school system, and they're separated from other students. So it takes one or 2 years until they actually transition to regular classes. These welcome classes emphasize the German language acquisition.

31
00:05:57.800 --> 00:06:15.769
but they are mainly on lower school tracks, and even when the refugee students or students in a welcome classes transition to regular classes, they remain in this low school tracks just because it comes in handy. There is no relation to their actually competences.

32
00:06:16.510 --> 00:06:18.120
So what we see is that

33
00:06:18.150 --> 00:06:22.939
refugee students in Germany are in a disadvantaged position already.

34
00:06:23.640 --> 00:06:36.580
When I talk about my theoretical assumptions, I will talk about 2 main mechanisms of of compositional effects. On the one hand, we have the socioeconomic composition, and on the other the migrant-specific composition.

35
00:06:37.160 --> 00:06:50.230
about this socioeconomic composition. We take one step. Back. yes, families. They have a socioeconomic status, and this affects the financial and cultural resources of the family.

36
00:06:50.310 --> 00:06:51.950
And this, again.

37
00:06:52.480 --> 00:07:06.560
affect, how can parents support their children in schooling. And yeah, extra curriculum. And all these things. And this again has an effect on the individual competences of the children living in these families.

38
00:07:07.110 --> 00:07:10.939
When we are in the classroom, and we aggregate all the

39
00:07:11.290 --> 00:07:18.409
a socioeconomic status of the students in this classroom we get the socio-economic classroom composition.

40
00:07:18.420 --> 00:07:21.960
and the teacher of this class is aware of.

41
00:07:22.590 --> 00:07:30.470
This is also economic classroom composition, and there is a perception and also an expectation, towards the students in this classroom.

42
00:07:30.490 --> 00:07:38.900
and this again shapes the competences, because in classrooms with a low, mean, socio-economic status, the

43
00:07:40.570 --> 00:07:49.640
the the teaching might be less cognitive, engaging, and also it is slower. So this can explain why

44
00:07:49.660 --> 00:07:59.780
individuals with the same individual characteristics might have different competences. regarding in which classroom they are placed.

45
00:08:01.760 --> 00:08:13.299
When I talk about micro-specific composition, I mean 2 different things. So on the one hand, we can talk about the ethnic composition which refers to the country of origin of the students in more classroom.

46
00:08:13.420 --> 00:08:20.900
What we think is more important is the ethnolingual composition, because it refers to the language of origin, and

47
00:08:21.300 --> 00:08:36.500
the whole thing about language is very important when it comes to competences, because the German language skills are very important for all competences that are acquired in the German school system. So there are 2 ways to

48
00:08:36.820 --> 00:08:42.250
learn German. On the one hand, we have the formal language acquisition through German lessons.

49
00:08:42.280 --> 00:08:59.930
And yeah, there's this teacher like it's especially in German classes, where the teacher thoughts about grammar, also writing. So it's this formal way we know from school. But, on the other hand, in the classroom. We have this informal way of peer interaction.

50
00:09:00.200 --> 00:09:19.529
So it's more about the language learning opportunities with peers. And I will talk about this because we think it's very important. So  in a classroom, there's yeah. There are different languages of origin. But if there is one student who speaks Arabic and 50, another student speaks Arabic.

51
00:09:19.570 --> 00:09:32.609
There is no need for this person to speak German to this peers, because he can make friends. He can make conversation, and also there are less language learning opportunities for German, because not many students actually speak German as their mother tongue.

52
00:09:32.980 --> 00:09:48.380
This again affects the German skills, because yet the less you speak German, the lower are your German skills. And this, again has an effect on all competences that are taught in subjects in German, such as science and mathematics.

53
00:09:49.540 --> 00:10:00.550
To sum it up, our first hypothesis states that a high socioeconomic classroom composition is positively associated with competences in mathematics and science of refugee students.

54
00:10:00.870 --> 00:10:14.879
The second hypothesis accounts for the ethnolingual composition, and says, the higher the share of students of the same non-german origin language, the lower the competences in mathematics and science or refugee students

55
00:10:15.920 --> 00:10:31.689
really quick about my data for our data. We used the Ecobi building strength from 2,018. This is a national assessment study in Germany for ninth graders and the competences measured were mathematics and science. In that case.

56
00:10:32.110 --> 00:10:38.850
we only included refugee students who were at least one year in the German school system.

57
00:10:39.330 --> 00:10:56.340
and that leads to a number of 520 students. In our analysis. in 343 classrooms on the classroom level we all we included for the means of Sacramento, for example, we included all

58
00:10:56.550 --> 00:10:58.470
as students in the classroom.

59
00:10:58.570 --> 00:11:03.590
So we see we have a multi-level model used for our analysis.

60
00:11:03.660 --> 00:11:30.509
really quick about my variables as dependent variables. We use the competences in mathematics, biology, chemistry, and physics. We also included the language, the German language skills which are measured through C test. The C test is a writing test. So this accounts for the formal language acquisition I talked about earlier. We also included the share of students with the same non-german origin language to account for the

61
00:11:30.890 --> 00:11:37.490
ethnolingual composition of the class to test hypothesis. H. 2. We included some

62
00:11:37.510 --> 00:11:53.359
control variables, and on a classroom level we controlled for school track and mean socioeconomic status of the school class to test hypothesis. H. One, and also the share of students with their migrant background to account for the ethnic composition of school classes

63
00:11:54.450 --> 00:12:02.710
coming to my results. I only have the time to present the results for mathematic in detail, but we can discuss science results later.

64
00:12:03.950 --> 00:12:12.680
So what you see here is our analysis. We controlled in all models for the controls and for school types on the school class level.

65
00:12:13.230 --> 00:12:19.879
what we see in model one is that we included the mean socioeconomic status of the school class.

66
00:12:19.940 --> 00:12:25.630
and we see there is a significant, positive effect on the math compenses of refugee students.

67
00:12:25.670 --> 00:12:33.719
So the socioeconomic composition does indeed have a positive effect on the math competences of refugee students

68
00:12:34.600 --> 00:12:46.529
in Model 2. We included German skills as we think they are very important, and indeed they are, and the mean source of economic status becomes insignificant when we control for German skills

69
00:12:47.880 --> 00:13:06.149
in Model 3, we included the share of students with migrant background to account for the ethnic composition of school classes, and we can't find an significant effect. But that's what we expected, because, as I told you earlier, we think the ethnolingual composition is much more important than country of origin.

70
00:13:07.070 --> 00:13:21.950
So in Model 4, we included a share of students with the same non-german origin language to account for the ethnolingual composition. But we can't find and significant effect on refugees, math competences, either

71
00:13:22.370 --> 00:13:38.419
because we think that German skills and the ethnolingual composition are connected. We excluded the German skills in Model 5 to see if there is an effect of the ethnolingual composition, then. But we can't find one, as you can see in Model 5.

72
00:13:41.310 --> 00:13:51.619
To sum it up, hypothesis H. One which refers to the socioeconomic composition was indeed confirmed for mathematics and physics

73
00:13:51.920 --> 00:13:58.209
and hypothesis. Age 2 about the atmalingual composition was rejected for all subjects.

74
00:14:00.400 --> 00:14:11.190
So again, the socioeconomic composition seems to be important for competences of refugee students in mathematics and physics. But the German skills are much more relevant as these are.

75
00:14:11.230 --> 00:14:21.880
Yeah, significant throughout all models. So the former language acquisition which we measured with the German skills seems to be the most important factor in our analysis.

76
00:14:21.930 --> 00:14:27.079
The language, learning opportunities, the informal way of German language acquisition.

77
00:14:27.340 --> 00:14:52.489
It is not associated with competences. So this hypothesis was rejected. When we keep in mind what I told you before about refugee students, that they are more often in classrooms with low, mean, socioeconomic status, and also tend to have low German skills and reflect this on the results of our study. We see that refugee students are in a very disadvantaged situation in the German school system.

78
00:14:54.230 --> 00:15:23.739
Of course, our study has limitations. We could only do a cross-sectional study. also, we couldn't include German states on a level 3 in our analysis. But for further studies it would be interesting because German states differ in their educational system, and also in their support for refugees. So there might be an extra context effect for the competences of refugee students.

79
00:15:24.300 --> 00:15:42.660
Also, we had mathematics and science as dependent variables. And what I told you before, we think that there is a connection between ethnolingual composition and mathematics and science competences. But this is only an indirect effect which is mediated through

80
00:15:43.020 --> 00:15:57.149
the German skills. So yeah, that might be one reason why there wasn't and significant effect. So in further studies. It would be more interesting to test for the direct effect of ethnolingual composition and German competences.

81
00:15:58.370 --> 00:16:27.110
The topic is still relevant, as we have a high share of refugees from Ukraine now in Germany and the situation, the educational situation for them is slightly different. For example, welcome classes are not as popular anymore as they were before. In any case, we think that the additional language support for refugee students is the most important thing to pursue. And yeah, that's what I end with. Thank you for your attention.

82
00:16:38.610 --> 00:16:44.130
Thank you, Hannah. Now we go to the second presentation, which is

83
00:16:44.560 --> 00:16:45.620
remote.

84
00:16:45.690 --> 00:16:47.839
So let's see

85
00:16:48.330 --> 00:16:49.459
something like that.

86
00:16:52.850 --> 00:16:56.220
 stop sharing.

87
00:16:56.610 --> 00:16:57.710
Hi, Benjamin.

88
00:16:57.890 --> 00:16:59.000
Benjamin Rosche: I

89
00:16:59.190 --> 00:17:01.040
you can now share your screen.

90
00:17:01.180 --> 00:17:02.750
Benjamin Rosche: Okay? Awesome.

91
00:17:15.480 --> 00:17:19.689
Benjamin Rosche: And you all. Now see? Basically this, just the slides.

92
00:17:22.740 --> 00:17:24.840
Benjamin Rosche: If okay.

93
00:17:26.420 --> 00:17:32.139
Benjamin Rosche: and I hope you can give me okay. because I can't see you.

94
00:17:32.710 --> 00:17:44.679
Benjamin Rosche: Yes, we can see the slides, so you can go. Okay. Hi, everyone. My name is Benjamin Roche. I'm a Phd. Candidate at Cornell.

95
00:17:44.730 --> 00:18:01.189
Benjamin Rosche: and I'm actually on the job market this year. So if anyone has a job for me, I'm very excited to present this current work, which is on socioeconomic segregation in us high school friendship networks

96
00:18:01.670 --> 00:18:12.409
Benjamin Rosche: and I'm going to start with What we already know about segregation patterns in us high schools.

97
00:18:12.580 --> 00:18:32.539
Benjamin Rosche: So we have the most solid evidence regarding age or grade gender and racial segregation in us high schools, by contrast to evidence regarding socio-economic segregation is rather mixed. We have some studies who find a socioeconomic segregation like I,

98
00:18:32.540 --> 00:18:43.229
Benjamin Rosche: 2,017, but then others like, how at all, 2,019 do not find a socioeconomic segregation. And then in a bunch of other studies.

99
00:18:43.270 --> 00:18:53.569
Benjamin Rosche: associate economic segregation wasn't explicitly studied because it was just a control variable mostly in studies of racial segregation.

100
00:18:54.430 --> 00:19:06.819
Benjamin Rosche: And then, regarding the determinants of segregation. they have been most systematically examined regarding racial segregation to study the them and Lewis at 10

101
00:19:07.020 --> 00:19:16.749
Benjamin Rosche: And regarding other types of segregation, we are actually lacking comprehensive analyses which kind of like tease apart, how

102
00:19:16.950 --> 00:19:27.419
Benjamin Rosche: much each I can mechanism. into this segregation and regarding socioeconomic segregation. The

103
00:19:27.610 --> 00:19:32.249
Benjamin Rosche: mechanisms considered in the literature are for one

104
00:19:32.310 --> 00:19:44.119
Benjamin Rosche: systematic disparities in the ses composition between schools which leads to a socio-economic segregation within friendship networks

105
00:19:44.240 --> 00:19:56.289
Benjamin Rosche: because of exposure differences. And then within schools we have settings such as like, course, T courses or extra curricular activities

106
00:19:56.430 --> 00:19:59.560
Benjamin Rosche: which induced friendship, formation.

107
00:19:59.570 --> 00:20:14.720
Benjamin Rosche: and also socioeconomically segregated, and hence they also affect the socioeconomic segregation and friendship networks. Next, homophily. So the tendency to be friends similar and others

108
00:20:14.720 --> 00:20:33.330
Benjamin Rosche: also can induce socioeconomic segregation. But crucially, not only ses homophily induces socioeconomic segregation, but also a racial, homophily or I'm offly based on school performance and may into socioeconomic segregation.

109
00:20:33.330 --> 00:20:56.340
Benjamin Rosche: because race and school performance are correlated with ses, and then finally, network mechanisms, such as the tendency to befriend others with whom we have a friend in common. which is called try at a closure amplifies existing decoration patterns.

110
00:20:56.660 --> 00:21:02.319
Benjamin Rosche: For instance, if in a network ties among high ses students,

111
00:21:02.460 --> 00:21:18.430
Benjamin Rosche: are already more likely than If if open, try it. Among these students are, more frequently closed. Then this creates even more clustering among is us students.

112
00:21:19.070 --> 00:21:32.630
Benjamin Rosche: And basically, what I'm going to do in the next 10 min is I'm going to disentangle how much each of these mechanisms induces socioeconomic segregation and friendship.

113
00:21:33.460 --> 00:21:38.300
Benjamin Rosche: I'm using at health for this, because at health is

114
00:21:38.360 --> 00:21:58.969
Benjamin Rosche: still the only representative data set in the Us. That includes full friendship networks and rich individual level information on students, race socioeconomic background school performance, residential locations, course taking patterns and extra curriculum activities

115
00:21:59.150 --> 00:22:07.019
Benjamin Rosche: and ses in at health is measured by composite index based on parental education, occupation, and income.

116
00:22:07.630 --> 00:22:12.630
and the final 7 will includes more than 7,000 students and 15,000 friendship line.

117
00:22:14.000 --> 00:22:19.370
Benjamin Rosche: There are several different ways to measure segregation and networks there.

118
00:22:19.410 --> 00:22:21.930
Benjamin Rosche: and and I argue that

119
00:22:22.250 --> 00:22:34.359
Benjamin Rosche:  diet level segregation measures are often more easily interpreted than aggregate network segregation measures, and in the prior literature, often.

120
00:22:34.450 --> 00:22:53.099
Benjamin Rosche: researchers have used the absolute difference in ses scores between friends as a measure of socioeconomic dis a distance, and my approach is different. In that I categorize students into low medium and high ss.

121
00:22:53.260 --> 00:23:00.850
Benjamin Rosche: in the appendix. I also categorize them in 4 categories. but it doesn't really matter for the results.

122
00:23:01.160 --> 00:23:06.290
Benjamin Rosche: to then differentiate basically between same ses size.

123
00:23:06.460 --> 00:23:09.950
Benjamin Rosche: which are low, low, medium medium, and high, high ties

124
00:23:09.990 --> 00:23:18.940
Benjamin Rosche: and cross ses size, which are ties between low and medium, ses students low and high, scs students and medium. And

125
00:23:19.200 --> 00:23:28.699
Benjamin Rosche: and I do this because I suspect that the interaction patterns differ between different points in the ses. This free distribution.

126
00:23:29.690 --> 00:23:49.989
Benjamin Rosche: And now basically, I'm I'm just like measuring I'm counting the number of same and cross ses ties that I find in each school network. Essentially. And then I compare these observed number of same across ses ties to the number expected

127
00:23:49.990 --> 00:24:03.609
Benjamin Rosche: in networks that are that are random like, in in which students select friends at random. And the way I do this is, I basically just rewire the edges

128
00:24:03.640 --> 00:24:09.830
Benjamin Rosche: within a network randomly. And then this way, we kind of get a baseline of like how many

129
00:24:10.000 --> 00:24:25.950
Benjamin Rosche: same and cross ses ties we would have to expect if friendship selection would be completely random. So here in this network, for instance, which is one of one small component of a larger network, you you see evidence for

130
00:24:26.040 --> 00:24:45.389
Benjamin Rosche: socioeconomic segregation for gender segregation and also for a race segregation and basically to to understand what the baseline in this network would be is we. I basically just rewire randomly these edges, and then count again the number of same and processes.

131
00:24:45.500 --> 00:24:53.329
Benjamin Rosche: And then I can create a measure of segregation by basically just measuring how many more

132
00:24:53.670 --> 00:24:59.589
Benjamin Rosche: or less same as the same, or and cross ses ties do we have

133
00:24:59.700 --> 00:25:08.050
Benjamin Rosche: compared to what we would expect under random segregation. And I make this a percentage. That's my my segregation, Russia.

134
00:25:08.590 --> 00:25:15.589
Benjamin Rosche: Let's look at some results. So here, basically, you can see

135
00:25:15.730 --> 00:25:19.019
Benjamin Rosche: the the segregation by

136
00:25:19.420 --> 00:25:30.010
Benjamin Rosche: edge type. And here, for instance, you can see that ties among low ses students are about 20% more likely than random selection.

137
00:25:30.180 --> 00:25:58.480
Benjamin Rosche: among medium, ses 5%, and among high scs. 15. And then when we look at the cross ses size, we see that ties from low to medium or medium to low scs. Students are about like 8% less likely than we would expect under random selection. And then here you can see the ties between low and high ses. Students are particularly unlikely.

138
00:25:58.480 --> 00:26:16.180
Benjamin Rosche: Yeah, like, I'm on about 25% less likely than we would expect expect. And then interestingly, high between medium and high or high and medium, are about as likely, if not a little more likely, than we would expect under random selection.

139
00:26:16.190 --> 00:26:17.090
Benjamin Rosche: Hence.

140
00:26:17.270 --> 00:26:37.860
Benjamin Rosche: kind of to summarize a little bit what's going on in in high school networks? It is really a separation of low ses students, because, basically, both medium and high ses. Students do not want to be or like are practically not friends with low ses students, and they very much keep to themselves.

141
00:26:38.880 --> 00:26:49.420
Benjamin Rosche: And my methodological approach. Now to disentangle the determinants is that I use exponential random graph modeling to model friendship, formation.

142
00:26:49.600 --> 00:26:53.780
Benjamin Rosche: And basically, I, include the

143
00:26:54.350 --> 00:27:10.190
Benjamin Rosche: aforementioned mechanisms into my model. For one thing, I'll explain this in a second. I I check how much we between school systematic differences between schools play a role. But then within schools.

144
00:27:10.240 --> 00:27:14.230
Benjamin Rosche: basically, I, I look at ses stratified settings.

145
00:27:14.490 --> 00:27:31.720
Benjamin Rosche: and in particular, I'm see whether living in the same neighborhood. So kids that go to the same school and also live in the same neighborhood. does this have an effect on friendship information? I look at course, taking patterns and extra curriculums.

146
00:27:32.220 --> 00:27:40.909
Benjamin Rosche: and then regarding homophily. I analyze Gpa Homophily raised him off me and ses a monthly.

147
00:27:41.000 --> 00:27:50.680
Benjamin Rosche: and then regarding network mechanisms, I am examining basically a Cs differences by socialities, or whether some

148
00:27:50.970 --> 00:28:02.699
Benjamin Rosche: ses groups, more likely to send out ties. Popularity differences, which means where the some ses groups are more likely to receive times.

149
00:28:02.730 --> 00:28:24.920
Benjamin Rosche: and then also differences by ses in reciprocity and triadic closure. and rather than examining. Now all these model coefficients which just give me the effect on friendship information. I use the model to simulate synthetic networks and measure the ses segregation in these networks.

150
00:28:25.140 --> 00:28:34.700
Benjamin Rosche: and I determined the impact of each mechanism on socioeconomic segregation by basically turning on each effect one by one

151
00:28:34.800 --> 00:28:48.669
Benjamin Rosche: and then simulating networks from it and always keeping the overall number of it's just constant in that way. I can kind of measure how much each mechanism increases and socioeconomic segregation

152
00:28:49.550 --> 00:28:57.170
Benjamin Rosche: here. I'm not going to go over it. I just want to show you this is a fairly realistic friendship formation model. I include

153
00:28:57.240 --> 00:29:02.360
Benjamin Rosche: very differential homophily in in the sense that I

154
00:29:02.630 --> 00:29:06.429
Benjamin Rosche: all the kind of like interaction patterns

155
00:29:06.490 --> 00:29:23.240
Benjamin Rosche: that might differ and are theoretically sensible. I kind of include as parameters and and basically all 3 mechanisms. The Se has ratified settings. I'm often the network mechanisms are represented in this friendship formation model.

156
00:29:23.800 --> 00:29:26.039
Benjamin Rosche: And basically.

157
00:29:26.370 --> 00:29:55.489
Benjamin Rosche: what I can do now is that I look at these different types of times. Ross, ses size and same ses size. And the first thing that I do is I simulate networks in which students select friend, that's totally at random, which means not only within a school, but students are also allowed to nominate friends across schools. So that kind of eliminates all within and between school segregation

158
00:29:55.830 --> 00:30:07.960
Benjamin Rosche: and for for low high ties, basically, if everything would be random, and there would be no between school segregation and about 20% of all ties should be low, high.

159
00:30:08.010 --> 00:30:15.470
Benjamin Rosche: Now, the next model I allow students to randomly select friends, but only within their school

160
00:30:15.670 --> 00:30:21.270
Benjamin Rosche: and there I find that 15.5%

161
00:30:22.270 --> 00:30:33.130
Benjamin Rosche:  of all ties across the ses. And then, finally, what we actually observe is 11.5.

162
00:30:33.270 --> 00:30:42.400
Benjamin Rosche: And so basically, almost half of all ties between low and high ses, students are unrealized, and about 4%

163
00:30:42.430 --> 00:30:51.190
Benjamin Rosche: percentage points, are lost between schools and another, 4% are lost within schools. So kind of within in between schools

164
00:30:51.240 --> 00:31:02.700
Benjamin Rosche: as segregation affects socioeconomic segregation at at the ratio of 50 50. And now I'm going to use the organs basically to

165
00:31:03.050 --> 00:31:13.370
Benjamin Rosche: to see how each of these mechanisms. and increases segregation. It explains this part, within schools.

166
00:31:13.840 --> 00:31:18.730
Benjamin Rosche: And so how much time do I have left? About 2 min.

167
00:31:19.060 --> 00:31:38.910
Benjamin Rosche: in this graph, basically, everything that's above 0, shows an increase in segregation and everything below 0. It shows a decrease in segregation. Basically, what I'm doing is here. I'm plotting. how much each mechanism

168
00:31:39.140 --> 00:31:58.739
Benjamin Rosche: either adds to same ases times which increase the segregation or reduces cross ses size. which also adds to segregation or the other way around. If it's reduces same ses size, then it decreases segregation in. If it's

169
00:31:59.000 --> 00:32:03.960
Benjamin Rosche: increases the number of cross Sas, and it also decreases segregation.

170
00:32:04.040 --> 00:32:32.559
Benjamin Rosche: And here below, basically, I show how much each mechanism basically explains out of the total of the total socioeconomic segregation. The first thing that you can see is that basically about 50% of the total socioeconomic segregation is between schools. It's not only holds for low ties between low and high as a students. This holds for socioeconomic segregation more generally.

171
00:32:34.080 --> 00:32:45.729
Benjamin Rosche: then the other thing that we see, and this is kind of very surprising, because this is high quality data. So this is likely not a measurement error. neighborhood effects.

172
00:32:45.810 --> 00:32:54.049
Benjamin Rosche: course, structure effects and extra curricular effects only explain 6% of the total socioeconomic segregation in high schools.

173
00:32:54.220 --> 00:33:00.800
Benjamin Rosche: And this is mainly because basically within these settings, kids still show basically.

174
00:33:00.850 --> 00:33:04.049
Benjamin Rosche: preferences have preferences and date

175
00:33:04.310 --> 00:33:13.089
Benjamin Rosche: eventually determine which friends they make and which they don't. So the the settings are not decisive enough to really

176
00:33:13.200 --> 00:33:14.930
Benjamin Rosche: to determine friendships.

177
00:33:15.920 --> 00:33:30.500
Benjamin Rosche: instead, you can see that basically, both race homophily and in particular in combination with ses. So basically, when when both race and ses are crossed, basically.

178
00:33:30.820 --> 00:33:46.700
Benjamin Rosche: then, we have. This explains about together about like 15 to 16% of the total socioeconomic segregation. But even after that we still have some 25% of residual as Yes,

179
00:33:46.830 --> 00:33:50.969
Benjamin Rosche: and homophily. Which means that basically

180
00:33:51.150 --> 00:34:10.249
Benjamin Rosche: accounting for all those mechanisms that I have in the model, it's still the case that same ses ties are more likely than cross. SSI, so they is still some residual ses. I'm off lead. That's quite significant. And then the other thing that's very interesting is that both reciprocity and triadic closure

181
00:34:10.250 --> 00:34:30.490
Benjamin Rosche: have a very dividend effects on socioeconomic segregation. So they increase some types of ties, but they decrease other types of ties. And that's why, together, they actually don't impact the socioeconomic segregation. All that much because these effects offset each other.

182
00:34:30.560 --> 00:34:32.240
Benjamin Rosche: Okay, to conclude

183
00:34:32.300 --> 00:34:45.200
Benjamin Rosche: us, high school friendships are socio-economically segregated, especially at the bottom of the ses. Distribution the within and between school process affect socioeconomic segregation in a 50 50 ratio.

184
00:34:45.220 --> 00:35:01.359
Benjamin Rosche: and surprisingly within schools, ses segregation, is largely driven by racial and ses on awfully, and not so much by sas. Stratified settings and or network mechanisms, because they have an ambivalent effect.

185
00:35:02.030 --> 00:35:03.930
Benjamin Rosche: That's it. Thank you very much.

186
00:35:10.800 --> 00:35:12.260
Thank you, Benjamin.

187
00:35:12.280 --> 00:35:17.219
we'll now go to the next presenter.

188
00:35:22.390 --> 00:35:24.040
Yes, perfect.

189
00:35:26.210 --> 00:35:26.890
so good.

190
00:35:36.260 --> 00:35:36.940
Yes.

191
00:35:40.060 --> 00:35:40.740
thank you.

192
00:35:45.720 --> 00:35:46.660
Thank you.

193
00:35:47.460 --> 00:36:01.830
Good evening, everyone. My name is Isami Murray. I'm currently a post-doctoral researcher at Sophia University in Japan. Today I'll talk about the effect of entering private junior high schools in Japan.

194
00:36:01.970 --> 00:36:09.299
It's outcomes on students, academic and social outcomes, consequences.

195
00:36:10.390 --> 00:36:19.390
So here's the big theme in sociology of education that we all care about in our C. 28, especially those who are interested in schooling.

196
00:36:19.670 --> 00:36:31.319
So studies on education or inequality have often shown that socio-economically advantaged families tend to use their resources to provide their children with greater access to better learning environments.

197
00:36:31.529 --> 00:36:45.019
However, there is always always a question on the row of school versus family. And we want to know whether school reproduce or reduce inequality, because that's the central question

198
00:36:45.670 --> 00:36:57.119
and research and private school. This, vector 1980, S. When the Coleman, James Coleman, talked about public and private high school differences in the Us.

199
00:36:57.149 --> 00:37:20.980
And the subsequent studies, continued until 1980 s. To the nineteenth in Britain and the United States, and more and more recently, studies have focused on the effect of private schools in diverse countries, not just the Western countries, but the cross-national studies and studies on Africa, and so on.

200
00:37:22.270 --> 00:37:50.939
So I'm focusing on private junior high school in Japan today and talk about this context. It's a type of early tracking or school choice, because public junior high schools in Japan day there. And there's no school choice people just go to neighborhood schools. But about 8% currently of the junior high school students nationally go to private junior high schools. So they do a school choice.

201
00:37:51.060 --> 00:38:11.190
and always there's a parental support to. search for these private schools and to effort to enter, to support students, to enter the take the exam to enter these schools, and usually there's a guidance by Isuku, which is a private supplementary to the institution which costs a lot of money

202
00:38:11.300 --> 00:38:24.189
and scholars. Japanese scholars in an international literature talks about these phenomenon as commercialization of exam, Oriented exam preparation, parenting in the context of parenting research.

203
00:38:24.220 --> 00:38:29.399
and also another scholar talk about these use of

204
00:38:29.430 --> 00:38:36.010
a lot of a supplementary tutoring on Academy's subjects as a Japanese version of concerted cultivation.

205
00:38:37.550 --> 00:38:43.570
So here are some basic talks about these paragraph and public junior high school. In Japan

206
00:38:43.700 --> 00:38:57.250
public schools are public funded, private schools are private funded, and the private Schools Union High School is allowed, has larger school autonomy, so they have hiring autonomy and

207
00:38:57.300 --> 00:39:20.049
and also the school admission. The principals usually can choose in private schools on the country. In public schools there is a national curriculum, and there is a significant role of local municipality in maintaining the standardized system, and also in Japanese system. There is a teacher rotation system every few years.

208
00:39:20.880 --> 00:39:28.590
There's a lot of social attention on private junior high school in Japan. This is a very.

209
00:39:28.750 --> 00:39:41.309
you know theme that captures people's attention. Not just the academics. So there is even a comic quarter that turned into a drama about the entrance. The competitive entrance in Zanta Junior High School.

210
00:39:41.370 --> 00:39:46.340
and that's why people believe that they provide better educational opportunities.

211
00:39:46.350 --> 00:40:13.129
These are the reason you won't see. But the you know, the academic going to a better, high, better. College is not the only, you know, desire of parents. You. You will see that parents have different desires in Japan. There are very few religious religious reasons to enter a private school like in the United States. It's not the big reason, but there are a variety of reasons that the parents want their to rent, ascend their students to private schools.

212
00:40:13.190 --> 00:40:23.420
So my research question is, what what are the actual consequences of attending private schools. because many research have shown that already the background.

213
00:40:23.910 --> 00:40:30.639
a higher socioeconomic background, parental background, is related to parental

214
00:40:30.700 --> 00:40:39.339
in private school, him entering attendance. But the few studies have asked the consequences. That that's what I'm focusing today

215
00:40:39.660 --> 00:40:44.680
in the music that they record a Japanese longitudinal study of children and parents.

216
00:40:44.740 --> 00:40:58.499
there's a 4 way from 2,015 to 19. It's a multi-cohort in piano and diet data, very ambitious data that has been going on in Japan.

217
00:40:58.580 --> 00:41:08.180
and I will examine changes over time. The first year. Through the third year of Juni High School after entrance

218
00:41:08.310 --> 00:41:18.489
I first show the descriptive analysis, and because of time. I won't show the fixed effect analysis, but my descriptive results are supported by the fixed effect

219
00:41:18.550 --> 00:41:24.509
analysis. I control the header. No, the unobserved

220
00:41:24.610 --> 00:41:28.709
variants, and I also look at the visiting individual.

221
00:41:28.730 --> 00:41:30.680
the effect.

222
00:41:30.820 --> 00:41:36.600
And later I'll also talk about the mechanism using the gross curve model.

223
00:41:37.290 --> 00:41:58.140
So here is the educational expectation. The descriptive descriptive statistics from elementary school, 6 year in sixth grade, a final grade in elementary school to the first, second, and 30 year of our junior high. So private Junior High School, so left is Those students who went to our public to private

224
00:41:58.170 --> 00:42:03.730
the right one. The right 2 lines are those students who went to the public

225
00:42:03.840 --> 00:42:09.000
junior high Public elementary school to public middle school.

226
00:42:09.470 --> 00:42:35.650
and I also divided by the parent to our college status. whether parents went to college or did not go to college. So you will see that the left on the left you will see some conversion patterns. If parents didn't go to college, but they went to private school. Their expectation somewhat increases in 16 years. So it's almost the same as college.

227
00:42:35.900 --> 00:42:45.479
Yeah. On the other hand, those who go to private. No public schools is rather stable, or a gradual, very gradual increase

228
00:42:45.820 --> 00:43:04.009
and degree is, on the contrary, it think, decreases luxury. If they enter a private middle school provision in high school, I think it's a relative effect, because all other students do economically better. Their grades go down. Grace is a a relative.

229
00:43:04.030 --> 00:43:08.119
Status in their classroom or in school.

230
00:43:08.890 --> 00:43:28.570
and the motivation to study both in public and private schools. It tends to decrease as students go through elementary to junior high school. But there, I see some, you know, steeper decrease for non-college parental students who go to private schools.

231
00:43:29.200 --> 00:43:38.600
and I see some positive effect here, whether the students feel that classes school lessons are fun

232
00:43:38.620 --> 00:43:52.710
when they are the first year in junior high school. They tend to like. So like school, especially those of the college. Educated parents. Children they like to.

233
00:43:53.010 --> 00:43:59.970
you know, they tend to like classes, but you know, you see, the subsequent decrease during the second and third year.

234
00:44:01.240 --> 00:44:09.969
and this is about the minutes, the time that they spent studying outside school. And there are 3 categories. The blue

235
00:44:10.050 --> 00:44:17.399
one is from formwork, and the red one is for a self-study, and Gr, one is for juku.

236
00:44:17.620 --> 00:44:41.949
And interestingly, you see that the first year in private school they really study better, especially they do a lot of homework, but it subsequently decreases as they go through second and third year. On the contrary public school students, especially because they have the competitive entrance example high School. At the end of the third year they study a lot, especially the you see the gray area as you quiz

237
00:44:42.010 --> 00:44:43.770
rather lies in Japan.

238
00:44:43.970 --> 00:44:59.670
and I further divided these by the parental education status. This is for private school, and the left one college educated parent, the right one for the non-college parent. And you see that there are some.

239
00:44:59.710 --> 00:45:02.719
the disparities, you know differences.

240
00:45:02.980 --> 00:45:26.270
If the college parents are college educated, the private school students. They don't decrease the time that they study very much, but otherwise they decreases. And, on the contrary, you see that public school students, they don't really, you know, decrease the time that they study whether Their parents are college educated or not.

241
00:45:27.090 --> 00:45:36.940
So you you start to see some compensatory, I think you know effect of our public school, at least for the time that students invest in studying.

242
00:45:37.160 --> 00:45:48.230
and this is some non-cognitive traits where the students take time to tackle difficult problems. This is a student perception. By the way, there's not much difference, you know.

243
00:45:48.260 --> 00:45:59.670
and whether students think that they can think laws degree. No, again, not much difference, and even decrease for the non-college parent. Private school students!

244
00:46:00.270 --> 00:46:06.220
And this is where the likes go against some positive.

245
00:46:06.700 --> 00:46:14.289
a positive effect. In the first year of junior high school and subsequent decrease or convergence.

246
00:46:14.530 --> 00:46:31.009
and whether they spend time enjoy spending time with their friends. There's not really much difference which is a good thing, no differences and good finding. Whether they go to a private school public school, you know not much difference, not much difference in self-esteem as well.

247
00:46:31.360 --> 00:46:51.179
So to summarize to up to this point students in private school experience a more academically challenging environment and improve satisfaction with lessons in school. In the initial year after enrollment, however, there were no notable differences in social interaction with friends, self-esteem, and non-communist skills

248
00:46:51.180 --> 00:47:03.279
who greater public school students. And these results suggest that there may not be much difference, at least during when they are in the middle school.

249
00:47:03.450 --> 00:47:16.100
and just for the variable one, variable that I saw a big difference, the classes of one. And I further did a grosser modeling. Just to examine the mechanism in. I

250
00:47:16.120 --> 00:47:37.580
added variables, both the covariates that do not bury with time and covariance the very this time the model by model, and so whether they will explain away the coefficients of private and also the the model. You know, the you know, the variance components. I looked at

251
00:47:37.850 --> 00:47:57.980
basically what I found was the. So I saw some mediating effect of active learning, which is Japanese fur, but it means presentation and discussion type of classes. Private school school students tend to like that, and it's sort of mediated their attitudes for classes of fun.

252
00:47:58.080 --> 00:48:09.020
So one implication from here is that private school students with relatively high academic ability are more likely to increase learning satisfaction through these type of classes

253
00:48:09.200 --> 00:48:16.690
and parental emotional support for children, such as whether they encourage students to

254
00:48:16.930 --> 00:48:26.700
study, and whether they understand children who materials are feeling these sort of things that mattered both for public and private school students.

255
00:48:26.750 --> 00:48:34.150
So this suggest that even if parents so sender children to private schools, parental support are still needed.

256
00:48:35.480 --> 00:48:41.459
So why did I see the rather small trivial effects of our private schools?

257
00:48:41.750 --> 00:48:53.269
I can think of some different mechanism. And and first, you can think that scores make no difference. But I will now show you some differences in school from position.

258
00:48:53.430 --> 00:49:12.630
and of course there might be underlying low rural public schools at public schools in Japan, Junior High School in Chicago, doing rather a good job. And third is that the banished families still invest and involved in any way other than just sending them to a private school, the school choice.

259
00:49:13.110 --> 00:49:18.650
So I don't have much time. But let me explain some additional context background

260
00:49:18.760 --> 00:49:41.840
in a hurry. I I analyze to nationally representative data, Tim Santalis for the for the Japanese data. And I want to show you some contextual backgrounds. So here's the school composition between our public and private junior high school students. And you see that up to the upper right hand side you'll see the and the

261
00:49:41.870 --> 00:49:43.470
and private.

262
00:49:43.480 --> 00:49:49.209
So they have a higher Cs and high achievement. That's where they are from teams

263
00:49:49.700 --> 00:50:10.039
and the student composition. The students with a special needs. Public school have more of them in the socio-economical. These and Spanish students, public school school have more of them. The school resources, you know, the average is not a different 53. So aggregate, score 53 for public school, 60 for private school.

264
00:50:10.280 --> 00:50:20.680
and some you know, the the red circuit on and their they will have programs. But there are not 2 large differences, I would say.

265
00:50:20.990 --> 00:50:23.870
and I look at other data and

266
00:50:24.060 --> 00:50:31.369
to summarize and sufficient teacher quality, I will say, is

267
00:50:31.420 --> 00:50:39.909
realized for Japanese and middle schools, public made of schools and class sizes are actually smaller. And yeah.

268
00:50:40.220 --> 00:50:49.330
and confidence by teachers and teachers feel that they, they have a challenge in providing challenging tasks for capable students in

269
00:50:49.350 --> 00:51:05.499
public school especially, I think, because of the student composition. You know, it's it's easier to do that. If they're in private school, when students can do better, they can provide challenge in that suite. I think this is a differences rather due to a composition.

270
00:51:05.640 --> 00:51:11.370
And students. Views on teachers are not really different between public and private schools.

271
00:51:11.410 --> 00:51:22.250
And this is another figure on the differences between public and private, for the bullying type of you see not much difference. Surprisingly

272
00:51:22.710 --> 00:51:24.450
so 2.

273
00:51:25.140 --> 00:51:37.190
And what I found was that a big difference in parental attitudes on measures like this, whether, when parents were asked, I want to spend money on my children's education.

274
00:51:37.200 --> 00:51:50.939
and I want to keep up with general trend in terms of my child's education, private schools, private school parents, and it to answer this very strongly the public school. parents!

275
00:51:51.020 --> 00:52:06.970
So finally, the implication is, I want to emphasize that the educational opportunity may not always lead to subsequent gap in educational processes and outcomes. So we need to try hard to disentang what a family effect versus co-effect.

276
00:52:06.990 --> 00:52:14.710
And we also need to know that parental design and satisfaction is not always the same as our school processes.

277
00:52:14.790 --> 00:52:20.649
You have some limitation, but I want to welcome your comments. Thank you very much.

278
00:52:28.610 --> 00:52:36.780
Thank you very much. We'll now go to the last presentation, and then be ready for all of the questions and discussion.

279
00:52:38.040 --> 00:52:38.920
See?

280
00:52:48.120 --> 00:52:51.970
I went into. Oh, yes.

281
00:52:52.170 --> 00:52:55.710
they wanted to present your view. Okay.

282
00:52:59.620 --> 00:53:00.650
okay.

283
00:53:01.170 --> 00:53:14.670
good afternoon. Everyone. sorry. My name is Turkey Sanada. I'm Doria University in Japan. And today I will discuss a fix to about the school education in entering higher education. I focus on tracking in high schools.

284
00:53:17.100 --> 00:53:19.099
Okay, let's start with again.

285
00:53:19.110 --> 00:53:31.580
be the introduction. I think important of the out of school education we will be discussing today is shadow education because shared education is more popular about the out of school education.

286
00:53:31.600 --> 00:53:37.950
So the sort of education is a global phenomenon all over the world.

287
00:53:37.990 --> 00:53:46.759
In fact, it has become such a majority industry that the research hierarchy is significant, economic and social impact

288
00:53:47.110 --> 00:53:50.280
in the context of Japan, this

289
00:53:50.470 --> 00:53:58.940
in the context of Japan, this surrender is no different. Many children and special results in junior high school make use of shadow education.

290
00:53:59.010 --> 00:54:06.509
as you can see, figure one on this right. The usage or shadow location is considerable among this group.

291
00:54:08.710 --> 00:54:16.310
Yeah, this slide shows the mechanics of shadow education and its implications on educational attainment.

292
00:54:16.690 --> 00:54:25.780
Firstly, it's important to note that the use of shadow education is a significant country, influenced by the social class of students.

293
00:54:26.570 --> 00:54:33.969
Research has shown that children from higher social classes are more likely to use shadow education.

294
00:54:34.450 --> 00:54:43.900
This leads to an important consequences. so the education might be contributing to increase social inequality

295
00:54:44.670 --> 00:54:53.630
as highlighted by the previous research. The use of each other. Education enlarged the gap between different social classes.

296
00:54:55.760 --> 00:55:04.209
Okay? And let's focus on the Japanese educational system and in its relationship with shadow education. In this slide

297
00:55:04.980 --> 00:55:24.009
in Japanese educational system children are free to choice any school at every traditional stage. This means that when students move from junior high school to high school, they can select the school the best fit their educational goals and aspirations.

298
00:55:24.840 --> 00:55:39.410
Interestingly, shadow education has a significantly promoted educational achievement among lower-crossed Junior High school students. However, this research, mainly focused on the transition stage

299
00:55:40.190 --> 00:55:57.350
from junior high school to high school. So this brings us to the mainly such question of our presentations today. how does the effect use of social education during high school students affect the progression to higher education.

300
00:55:58.600 --> 00:56:04.630
So that's me. From previous studies

301
00:56:04.890 --> 00:56:08.670
broadly, there are 2 trends in shadow education research.

302
00:56:08.910 --> 00:56:19.539
The first tournament focuses on determining who used subtle education. The second torrent seeks to understand the effect of shadow education.

303
00:56:20.060 --> 00:56:23.329
This study is placed in the second torrent.

304
00:56:23.990 --> 00:56:29.480
However, opinions on the shadow allegations effect are divided. In Japan.

305
00:56:29.770 --> 00:56:38.499
Some researchers have found syringe counter-positive effects, but, on the other hand, some researchers have no so no signature effect.

306
00:56:40.500 --> 00:56:52.129
while considerable research focuses on the junior high school years studies concerning on the high school years could be more extensive. In Japanese context.

307
00:56:53.220 --> 00:56:58.030
this gap may be due to the strong high school tracking system in Japan.

308
00:56:58.860 --> 00:57:05.159
However, when we consider the relationship between shadow education and the Japanese educational system.

309
00:57:05.290 --> 00:57:16.749
it's not worthy that shuttle. Education can serve as a tool for students to recover from failure, disadvantage in high school and toance exams

310
00:57:17.570 --> 00:57:27.589
based on the above the question, arise. how does shadow education are, how how does shadow education affect higher educational and enormment?

311
00:57:28.650 --> 00:57:29.780
So

312
00:57:29.970 --> 00:57:46.109
in this study I suggest these hypothesis. the first hypothesis, H. One purpose is that children who use shadow education are more likely to achieve higher educational outcomes than those who do not.

313
00:57:46.940 --> 00:57:53.510
My second hypothesis, I to suggest that even children in this advantage high school trucks

314
00:57:53.620 --> 00:57:59.319
can achieve higher educational outcomes, more easy using shadow education.

315
00:58:00.180 --> 00:58:06.709
In H 2 hypothesis, we can check the strength of the high school truck in Japan.

316
00:58:07.210 --> 00:58:13.349
By using shoulder education, students can jump over high school trucks in Japan.

317
00:58:14.330 --> 00:58:18.980
Yeah, this is hypothesis to it. Image yeah.

318
00:58:19.030 --> 00:58:28.019
the Br are also the high school talking systems and the lateral shows hypothesis to yeah jumper about truck.

319
00:58:29.500 --> 00:58:45.309
Okay? Methods. Today, I I use the Japanese license called binary survey provides a of data about individual life course which makes it highly suitable for our study.

320
00:58:46.710 --> 00:58:52.350
and we employed 2 main steps in terms of statistical approach.

321
00:58:52.560 --> 00:59:01.510
Firstly, we calculate the in-based probability, logic. Sorry the inbound probability weight, that is, propens d square.

322
00:59:01.650 --> 00:59:11.660
This is achieved using binary logic model, where the dependent variable Y is whether or not shadow education was used in the high school

323
00:59:12.430 --> 00:59:25.530
we used Ipw this time, because if we analyze without any corrections. bias can be introduced to the research duty problems such as unobserved heterogeneity.

324
00:59:26.450 --> 00:59:30.729
In the second step we employ a partial proportional OS models.

325
00:59:30.830 --> 00:59:42.160
This model allows us to examine the order of achievement different levels our orders categorically dependent, variable with higher education. I I

326
00:59:42.580 --> 00:59:50.459
and model one, applies the basic partial proportional OS model to our data.

327
00:59:50.540 --> 01:00:07.860
then in moderative, we extend model one by adding an interaction time between shadow educational use and high school truck to examine whether the effect of shadow education varies on depend depending on the types of high school truck.

328
01:00:09.580 --> 01:00:16.010
Yeah. And this is a barriers. Our independent variable is higher education.

329
01:00:16.220 --> 01:00:28.530
For this study I have classified higher education into 4 categories. early to universities, other universities, junior colleges, and vocational schools.

330
01:00:29.050 --> 01:00:33.930
and that treatment variable is experienced in shadow education.

331
01:00:33.960 --> 01:00:43.509
This includes various forms of shadow education, such as grammar, school tutacles, private tutor, and the corresponding education.

332
01:00:44.100 --> 01:00:51.629
Finally, our dependent variable is high school truck. We have a categorized high school truck into 3 types.

333
01:00:52.240 --> 01:01:01.629
and that is they sorry, and that and that, and based on the university enrollment, late

334
01:01:01.940 --> 01:01:13.889
preparations, middle level school and the low level school. Our analysis aim to understand how these variables interpret and contribute to educational outcomes.

335
01:01:17.370 --> 01:01:18.040
You should

336
01:01:19.030 --> 01:01:29.410
okay, we begin by examining descriptive statistics. Everyone presents data on the use of shadow education during high school.

337
01:01:29.700 --> 01:01:35.569
revealing that 43% of miles and 46 of females

338
01:01:35.670 --> 01:01:38.600
takes advantage of shadow education.

339
01:01:38.870 --> 01:01:39.780
Yeah.

340
01:01:40.460 --> 01:01:50.500
And the next return to tablet, which demonstrates the impact of shadow education on higher education. That is, partial proportion of those models.

341
01:01:51.080 --> 01:01:59.570
This table, focused on salary on the lasers for primary variables. Initially, we consider the finding for males.

342
01:01:59.960 --> 01:02:13.209
The number is displaced next to the variables in the model which loses the parallelism. meat. Losing the parallelism assumption means in the proportional OS model.

343
01:02:13.940 --> 01:02:24.839
The lizard indicates that using shadow education is summoning the academic harder between universities and either junior college or vocational school

344
01:02:25.450 --> 01:02:35.600
students who take part in shadow education are twice as likely to progress from junior high junior college or vocational school to university.

345
01:02:36.530 --> 01:02:51.720
and the interaction term also confirms that students from less good academic trucks can achieve higher educational attainment through shadow education. But this is 10% of Shimmy's count effects.

346
01:02:53.700 --> 01:03:06.739
Yeah, let's say we dive into Table three's findings on V. Myers. In this model for females we maintain the parallelism. Assumption was the primary effect of shadow education.

347
01:03:06.850 --> 01:03:17.260
so there is no number. Next to the variable of the use of shadow education. Moderators suggest that federal education hosted educational advancement.

348
01:03:17.390 --> 01:03:21.609
However, the interaction term for us to be instruments.

349
01:03:23.070 --> 01:03:38.020
Yeah, and this is a summary of the results. But we have already discussed the result boundary. So we will skip this this one, this one. Okay, to summarize our findings in the context of our initial hypothesis

350
01:03:38.390 --> 01:03:43.519
based on our nations. The first hypothesis is partially accepted.

351
01:03:43.930 --> 01:03:50.709
We found evidence supporting that shadow education can facilitate higher educational attainment.

352
01:03:51.160 --> 01:04:00.119
But even with social education the gap between early to universities and other universities could not be overcome.

353
01:04:00.570 --> 01:04:11.169
Second hypothesis is partially accepted. while we observed the effect. In certain instances it was inconsistent accuracy or scenarios.

354
01:04:11.620 --> 01:04:20.440
These findings together indicate that Shewich councillors that shadow education can play in influences, educational outcomes.

355
01:04:22.040 --> 01:04:33.589
Yeah, to conclude our presentation, we found that shadow education experiences contribute to be effective even at the stage of higher education.

356
01:04:34.070 --> 01:04:40.600
For most, her most students. Shadow education enhanced their educational achievement.

357
01:04:40.650 --> 01:04:47.820
collaborating usefulness. shadow education as a tool for academic achievement.

358
01:04:48.500 --> 01:05:00.569
However, it's closer to note that she gave me counter difference was observed among males when comparing the effect between all to universities and other universities in Japan.

359
01:05:01.290 --> 01:05:08.319
This suggests that impact overshadow education varies with the division of higher education.

360
01:05:08.970 --> 01:05:22.390
Hosamoa, our study also finds that boys from level high school trucks in each country increased their chances of achieving higher educational outcomes through the shadow education.

361
01:05:23.060 --> 01:05:35.930
These findings under that potential overshadow education to level the playing field and promote educational inequality. To a promote educational equity.

362
01:05:36.380 --> 01:05:50.500
That is my research and conclusion. But what I want to tell most in this presentation is that part shuttle allegation is openly forwarded as a mechanism that expands socially, inequality on social gap.

363
01:05:50.520 --> 01:05:56.430
but it also has a potential to contribute equalization.

364
01:05:56.620 --> 01:06:07.999
that is, reference and a acknowledgment, and that supplement is pre-conceitable and binary allows it for models.

365
01:06:08.230 --> 01:06:19.169
Okay, thank you for listening. Thank you for listening. And if you have any questions I would appreciate if they could be asked in simple English. Thank you.

366
01:06:26.470 --> 01:06:27.170
Let's see.

367
01:06:27.450 --> 01:06:39.670
Thank you very much. That was a fantastic presentation. Now we can open the floor to any question to any of the presenters, or even if some of the presenters have questions for each other. let me just try and get

368
01:06:40.360 --> 01:06:42.130
stop share. Yes.

369
01:06:42.820 --> 01:06:43.640
perfect.

370
01:06:48.980 --> 01:06:50.260
It's a good

371
01:06:50.330 --> 01:06:54.490
I don't know what?

372
01:06:57.510 --> 01:06:58.539
Never seen this before.

373
01:06:59.140 --> 01:07:02.160
Still there!

374
01:07:06.340 --> 01:07:11.099
Yes, I think we're getting there. More time to think about questions.

375
01:07:14.490 --> 01:07:16.220
There we go.

376
01:07:17.380 --> 01:07:19.560
Yes. perfect.

377
01:07:20.990 --> 01:07:27.139
Any question. Then I'll start. I have a question for Anna. I was wondering if.

378
01:07:27.820 --> 01:07:32.010
when you're looking at the shared ethnolinguistic

379
01:07:32.050 --> 01:08:01.069
and say composition, if looking at the school level instead of the classroom level, might give you some more. Some evidence of that would confirm your hypothesis, at least partially, because I I guess it depends. I don't know enough about schools in Germany, so it might depend a bit on so on the size of the school. Maybe really large schools. Then it will be too much of a broader setting, but it's possible that in smaller schools, then children will still, and adolescents will still be able to find

380
01:08:01.200 --> 01:08:05.370
friends who share the same language even across classrooms.

381
01:08:05.910 --> 01:08:23.049
Thank you for your question. Research has shown that the classroom level is much more important as students tend to engage there more often than in schools in general, and for Germany in ninth grade, which was my sample. They

382
01:08:23.090 --> 01:08:40.600
are really just with the same people all the time. So in another educational system it might be different. But in Germany they really it's the class the whole time. So they're not engaging with other students in the school that much, and that's why we decided to do it for the classroom level.

383
01:08:44.660 --> 01:08:50.400
Benjamin Rosche: I'm not I. I have a comment. on just that point. I also have the feeding

384
01:08:50.510 --> 01:08:55.650
Benjamin Rosche: it. It might be nice, since you're not including school fixed effects.

385
01:08:55.860 --> 01:09:07.340
Benjamin Rosche: but you're looking at the average ses in the classroom to to at least like differentiate. You know the mean within the classroom.

386
01:09:07.670 --> 01:09:20.270
Benjamin Rosche: And there's like this within, between differentiation, you can basically how much is between schools. between school differences in the ses and then within school means and ses

387
01:09:20.300 --> 01:09:23.509
Benjamin Rosche: and that would at least confirm

388
01:09:23.890 --> 01:09:38.789
Benjamin Rosche: it would just listen. Show you that. Yes, indeed, it is really between within schools as as you claim. And I I doubt it. because I think that. Basically the ses

389
01:09:38.940 --> 01:09:47.829
Benjamin Rosche: composition will the most drastically vary between schools and not within

390
01:09:48.180 --> 01:09:49.529
Benjamin Rosche: between classrooms.

391
01:09:51.830 --> 01:10:00.270
Thank you for the comment. I think we did actually do that at 1 point, I just couldn't include it in our presentation. But thank you very much.

392
01:10:00.580 --> 01:10:09.099
Benjamin Rosche: Yeah. And I have one more comment before I forget it. So one thing I didn't really understand is, why didn't you compare basically

393
01:10:09.370 --> 01:10:20.960
Benjamin Rosche: the ses compositional effects on math compenses with on German students? Because, in a way, we already have studies that show compositional effects on

394
01:10:20.970 --> 01:10:43.060
Benjamin Rosche: on math competencies. Right? And so by setting the subgroup of refugees, you're kind of implicitly assuming, or you at least curious about differences you, it might work differently among the subgroup of refugees. And so, in order to determine this, you need the reference category of Germans.

395
01:10:44.090 --> 01:10:55.489
Yeah, also. Very good question. It was a very long process until we find our final sample, and there was a lot of negotiating, and in the end we decided for refugee students.

396
01:10:55.640 --> 01:11:23.300
To be honest, I can't really tell you why, we decided, but I think it was sort of a sample thing. Why, we in like of our data, why, we decided just to reduce a sample on refugee students. But yeah, in in further studies, we should totally compare refugee students and student without without migration background, but also students with a migration of background and also first-generation students, because that is like

397
01:11:23.380 --> 01:11:47.760
the thing you can actually compare right students who just came into the German school system, but without flight experience. I think we didn't. We couldn't do that with the data we have, because the number of students with like first-generation students without flight. Experience is really low in Germany or in our data, at least. And so that's why we couldn't do that yet. But yeah, it's planned for further studies. Thank you.

398
01:11:52.540 --> 01:11:54.080
Any other questions.

399
01:11:59.770 --> 01:12:24.840
I had a little question or comment for Benjamin. I was wondering who found it really, really interesting. Your research. And when you look at the within school, what makes a difference. It seems that it's homophily that makes most of the difference which I find. I don't know if you would agree, but a little bit boring and sad as in maybe the the kind of certification within school there is more scope for changing things.

400
01:12:24.840 --> 01:12:34.079
So I was wondering if this is because it's just that society is so stratified, and it's it doesn't necessarily make a difference. Children are already.

401
01:12:34.140 --> 01:12:46.720
let's say, hanging out with children that are similar to themselves, or if maybe because, if I remember correctly, this is about high school students, if there is already something that happens before, and by that time the kind of processes are already in place.

402
01:12:47.080 --> 01:12:47.820
Benjamin Rosche: Umhm.

403
01:12:48.010 --> 01:13:02.010
Benjamin Rosche: yeah. I mean, I also looked at the situation in Germany and in Europe, or generally, and I do find a significant differences between socioeconomic segregation and us high schools. And then

404
01:13:02.100 --> 01:13:05.960
Benjamin Rosche: in German secondary school. So

405
01:13:06.740 --> 01:13:17.989
Benjamin Rosche: these like results aren't final. So I I don't want to say too much about it. But basically, I do think that the Us context is like the

406
01:13:18.000 --> 01:13:22.170
Benjamin Rosche: socioeconomic segregation in us high schools is much bigger

407
01:13:22.320 --> 01:13:24.160
Benjamin Rosche: and

408
01:13:24.330 --> 01:13:35.370
Benjamin Rosche: and I also think that that the classroom structure works very differently or differently there, because there is not just one classroom. It's really like

409
01:13:35.390 --> 01:13:45.290
Benjamin Rosche: course, taking patterns that are decisive. and in general The rates met a much more than

410
01:13:45.370 --> 01:13:54.759
Benjamin Rosche: Then classrooms. And a another thing that I find interesting about these results that I found is that basically it shows that

411
01:13:54.820 --> 01:14:00.870
Benjamin Rosche: the classroom in in us high schools. Now, the classroom is actually not a good proxy

412
01:14:00.990 --> 01:14:06.890
Benjamin Rosche: for peer groups. at least not for the socioeconomic segregation. And so

413
01:14:06.970 --> 01:14:21.100
Benjamin Rosche:  The the peer groups are structured and are socio-economically different from from the classrooms. Essentially, that's what what this result shows. And that shows a little bit the limits of assuming

414
01:14:21.170 --> 01:14:30.829
Benjamin Rosche: that in a classroom context, we have pure effects. So if we find, let's say that the mean ses

415
01:14:31.140 --> 01:14:55.009
Benjamin Rosche: of a of of a classroom has an effect on something, and then we say, Okay, this must be pure effects, basically what the the result here shows. Well, actually, friendship networks are different from from classroom networks and And so this might be a teacher effect, or this might be a school like a classroom climate effect, but not necessarily a peer effect.

416
01:14:56.250 --> 01:14:57.119
Thank you.

417
01:14:58.450 --> 01:14:59.220
Yes.

418
01:15:03.850 --> 01:15:28.980
hi! I'm I'm Katie from Gpn, and I have a also a question to Benjamin, that this is just a confirmation question. in your results that you found that the bidding school and within school effects are almost a 50 50 and within school. In fact, it's actually the the the homophobia effect of the

419
01:15:29.440 --> 01:15:56.789
the the and but in the not part, but the in the other Friday you found that it's especially The impact due by between the low, low cohort and the high high ff and gear group effect. So does that mean that this withincraft effect is mostly captured

420
01:15:56.860 --> 01:16:04.060
by the low, low and high, high ff, and frenzy.

421
01:16:04.970 --> 01:16:11.129
Benjamin Rosche: Oh, so you're basically here in this slide here to this slide, right?

422
01:16:11.190 --> 01:16:12.440
And

423
01:16:12.540 --> 01:16:24.539
Benjamin Rosche: yeah. And so you're saying, basically, I'm I'm not entirely sure what I understood you. You're saying this effect, do you try it one more time.

424
01:16:28.060 --> 01:16:31.579
So for in the past. Fr. That

425
01:16:32.220 --> 01:16:37.779
you felt that it's the configurated

426
01:16:38.300 --> 01:16:45.050
phenomenon it's in among low, low friendship and

427
01:16:45.080 --> 01:16:58.630
a high- friendship, and in the fact that you you you found that in in a whizzing school effect. this, if

428
01:16:59.110 --> 01:17:06.130
 right? The homophily effect is particularly strong

429
01:17:07.180 --> 01:17:09.670
Benjamin Rosche: combined. Okay.

430
01:17:09.850 --> 01:17:26.809
Benjamin Rosche: I think I get it. Now, yeah. So basically. So, first of all, this here, of course, is here. I'm I'm taking out the between school differences. So this is just within schools. that's also by you. When you see here these expected numbers.

431
01:17:26.820 --> 01:17:35.000
Benjamin Rosche: If if if we would look at all schools overall. he should be more or less the same.

432
01:17:35.030 --> 01:18:02.839
Benjamin Rosche: because if you differentiate schools in 3 groups, then the expected number for each type of tie is kind of the same. But it's not the same, these expected numbers because they are systematic between school differences. And I'm I'm just looking at the differences within schools basically here. So this is the first thing. And the second thing is, the first graph is really just univariate. I'm just showing descriptively.

433
01:18:02.840 --> 01:18:21.709
Benjamin Rosche: which types of ties. most likely. And here I'm running in the background. fully conditional exponential random graph model which basically, this ses effect, is a conditional effect. It's just like conditional on all the variables that I have in the model.

434
01:18:27.230 --> 01:18:28.480
He's got it, thank you.

435
01:18:29.230 --> 01:18:32.369
Any other comments or questions.

436
01:18:35.720 --> 01:18:39.989
I had a little question for the last presenter for Turkey.

437
01:18:40.170 --> 01:19:06.969
I was wondering in the difference of result that you saw between boys and girls. I find that very interesting. And you say that shadow education can potentially be good for rebalancing some of the ses inequalities. But then it seems that maybe in terms of gender inequality. Some of your your results show that it's not necessarily very positive, and I was wondering if you had

438
01:19:07.060 --> 01:19:10.790
any more to add to that or any interpretation of why

439
01:19:10.820 --> 01:19:14.010
you're seeing this different effect. Thank you.

440
01:19:15.340 --> 01:19:26.360
Okay, thank you for your comment. And that is very, very insightful question. But inside here, in fact, I I don't.

441
01:19:40.490 --> 01:19:48.359
Yeah, but yeah, sorry I can't. I answer this in this? No worries.

442
01:19:48.430 --> 01:19:51.260
maybe something, too, for all of us to think about

443
01:19:52.300 --> 01:19:55.230
any more comment or questions

444
01:19:55.880 --> 01:19:57.090
from anyone.

445
01:20:01.040 --> 01:20:02.959
It doesn't seem like.

446
01:20:04.230 --> 01:20:06.209
there you go. Very good.

447
01:20:07.700 --> 01:20:26.720
So, if you may, I, if May amo, and I also have the last present. I saw it, Teduki, and you said that. So the shadow schools actually helps particularly for the lower less.

448
01:20:26.900 --> 01:20:29.370
 no, no.

449
01:20:29.530 --> 01:20:41.340
let's achieve the students, and and not so much or the high, already high educated students. Could you explain more on why it is the case

450
01:20:44.130 --> 01:20:45.290
you

451
01:20:46.060 --> 01:20:55.640
and just thinking about the Japanese shadow graphics. I cannot really understand the mechanism behind that.

452
01:21:00.470 --> 01:21:06.040
Yeah, thank you. A comment. Yeah, that is a very important question.

453
01:21:06.240 --> 01:21:15.889
I see, in my opinion in high schools. Teacher

454
01:21:16.400 --> 01:21:41.339
teach a student after school, but in Lower Lower Lower High School track students can't get the opportunity of after school after school. The teachers.

455
01:21:41.430 --> 01:21:45.630
 so

456
01:21:46.080 --> 01:21:47.740
so

457
01:21:47.770 --> 01:22:00.599
in the high school students there is no huge count effect, or the difference of high school drugs is is shown sure.

458
01:22:03.450 --> 01:22:04.340
Okay.

459
01:22:07.070 --> 01:22:07.939
thank you

460
01:22:09.080 --> 01:22:10.120
anymore.

461
01:22:12.220 --> 01:22:20.270
I had a little. It's not really a question, but it's the descriptive result. But in some of what you presented is you me at some point.

462
01:22:20.350 --> 01:22:27.060
There was a in the descriptor. There was a graph where you were looking at the school. It's fun change!

463
01:22:27.240 --> 01:22:29.050
And seeing that

464
01:22:29.060 --> 01:22:50.079
In private schools. It seems that it increases at the beginning, and then over the years it goes down again, and this was a lot more pronounced for children with non-college-educated parents, and I was wondering, because I've done a little bit of research on school is fun as a concept, and how children

465
01:22:50.370 --> 01:23:05.379
report it or not. If there is anything that has to do with kind of the environment of the private schools that are in. So it kind of. For for children coming from less educated families. At some point the environment is not necessarily

466
01:23:05.580 --> 01:23:16.309
enjoyable anymore because they're maybe they're not made to feel like they belong, and there is some effects of this side. I was just wondering if you had any idea about that, because I found it very interesting.

467
01:23:16.460 --> 01:23:17.340
Thank you.

468
01:23:19.960 --> 01:23:36.009
Thank you very much for the question. I was also wondering about why there's a deep, gradual decrease in of second decrease after the first. initial increase. And as you said that they are the minority in private junior high schools

469
01:23:36.550 --> 01:23:44.960
about, I think 30%. No college educated parents

470
01:23:45.060 --> 01:24:00.309
students are about 30%. So we are a minority compared to those who go to a public school. So maybe one reason, another reason is the existence of the and and and for the public middle school students

471
01:24:00.640 --> 01:24:17.999
to enter high school and private junior high school students. They are, many of them, of 6 year integrated schools, luxury so they can. There's a reason that they can relax. They don't have to pay up less pressure to study. There is another reason.

472
01:24:18.090 --> 01:24:39.500
and I was thinking of another. So the final reason I can think of is that they can be disenchanted with the environment. They can be used to the attractive environment, even if it's initially attractive to them, they can get used to it, and that's why they there's a decrease. But they are all and still hypotheses. Thank you very much. Thank you.

473
01:24:41.050 --> 01:24:44.039
If there are no more questions.

474
01:24:45.960 --> 01:24:52.850
then thank you very much to everyone that attended, and a big clap to the presenters that did a great job.