Levels
Department (Level 2) → Student (Level 1)
(department_id → student)
Glimpse of how the data is structured:
## # A tibble: 20,239 x 6
## department_id stdid grade reservation z_seg_g1 z_seg_g2
## <chr> <chr> <dbl> <int> <dbl> <dbl>
## 1 IR001CS IR001CS1001 2 0 -0.602 -1.28
## 2 IR001CS IR001CS1002 2 0 1.81 0.843
## 3 IR001CS IR001CS1003 2 1 -0.0310 -0.779
## 4 IR001CS IR001CS1004 2 1 0.158 -0.473
## 5 IR001CS IR001CS1005 2 0 -0.392 2.35
## 6 IR001CS IR001CS1006 2 0 -0.392 -0.145
## 7 IR001CS IR001CS1007 2 0 -0.392 NA
## 8 IR001CS IR001CS1008 2 1 -0.0310 -0.330
## 9 IR001CS IR001CS1009 2 0 -0.852 -1.15
## 10 IR001CS IR001CS1010 2 1 NA NA
## # ... with 20,229 more rows
Null model
Model summary
lm1.1 <-
lmer(z_seg_g2 ~ 1 + (1 |
department_id),
data = df %>% filter(reservation == 1))
lm1.2 <-
lmer(z_seg_g2 ~ 1 + (1 |
department_id),
data = df %>% filter(reservation == 0))
lm1.3 <-
lmer(z_seg_g4 ~ 1 + (1 |
department_id),
data = df %>% filter(reservation == 1))
lm1.4 <-
lmer(z_seg_g4 ~ 1 + (1 |
department_id),
data = df %>% filter(reservation == 0))
knitreg(
list(lm1.1, lm1.2, lm1.3, lm1.4),
custom.note = "%stars",
stars = c(0.01, 0.05, 0.1),
include.ci = FALSE,
include.variance = TRUE,
custom.model.names = c(
"Cohort 1 Res",
"Cohort 1 Non-res",
"Cohort 2 Res",
"Cohort 2 Non-res"
)
)
Statistical models
|
|
Cohort 1 Res
|
Cohort 1 Non-res
|
Cohort 2 Res
|
Cohort 2 Non-res
|
|
(Intercept)
|
-0.20***
|
0.17***
|
-0.15***
|
0.17***
|
|
|
(0.03)
|
(0.04)
|
(0.03)
|
(0.04)
|
|
AIC
|
12118.50
|
10441.20
|
11528.03
|
10997.02
|
|
BIC
|
12137.63
|
10459.97
|
11547.08
|
11015.86
|
|
Log Likelihood
|
-6056.25
|
-5217.60
|
-5761.01
|
-5495.51
|
|
Num. obs.
|
4348
|
3853
|
4234
|
3935
|
|
Num. groups: department_id
|
98
|
98
|
97
|
96
|
|
Var: department_id (Intercept)
|
0.08
|
0.10
|
0.05
|
0.15
|
|
Var: Residual
|
0.92
|
0.84
|
0.87
|
0.91
|
|
p < 0.01; p < 0.05; p < 0.1
|
Model ICC
bind_rows(
icc(lm1.1, by_group = T),
icc(lm1.2, by_group = T),
icc(lm1.3, by_group = T),
icc(lm1.4, by_group = T),
.id = "model"
) %>%
pivot_wider(names_from = model, values_from = ICC) %>%
select(
level = Group,
"Cohort 1 Res" = "1",
"Cohort 1 Non-res" = "2",
"Cohort 2 Res" = "3",
"Cohort 2 Non-res" = "4"
) %>%
kable()
| department_id |
0.0783756 |
0.1058881 |
0.0572255 |
0.1371008 |
Adding student level predictors
Model summary
lm2.1 <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + (1 |
department_id),
data = df %>% filter(reservation == 1)
)
lm2.2 <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + (1 |
department_id),
data = df %>% filter(reservation == 0)
)
lm2.3 <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + (1 |
department_id),
data = df %>% filter(reservation == 1)
)
lm2.4 <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + (1 |
department_id),
data = df %>% filter(reservation == 0)
)
knitreg(
list(lm2.1, lm2.2, lm2.3, lm2.4),
custom.note = "%stars",
stars = c(0.01, 0.05, 0.1),
include.ci = FALSE,
include.variance = TRUE,
custom.model.names = c(
"Cohort 1 Res",
"Cohort 1 Non-res",
"Cohort 2 Res",
"Cohort 2 Non-res"
)
)
Statistical models
|
|
Cohort 1 Res
|
Cohort 1 Non-res
|
Cohort 2 Res
|
Cohort 2 Non-res
|
|
(Intercept)
|
-0.06
|
-0.00
|
0.03
|
-0.11
|
|
|
(0.09)
|
(0.10)
|
(0.08)
|
(0.10)
|
|
z_seg_g1
|
0.25***
|
0.23***
|
|
|
|
|
(0.02)
|
(0.02)
|
|
|
|
female
|
0.02
|
-0.01
|
-0.04
|
0.05
|
|
|
(0.04)
|
(0.04)
|
(0.03)
|
(0.04)
|
|
age
|
0.01
|
-0.01
|
0.02
|
0.01
|
|
|
(0.03)
|
(0.03)
|
(0.02)
|
(0.03)
|
|
ses
|
-0.05*
|
0.06**
|
0.01
|
0.05*
|
|
|
(0.03)
|
(0.03)
|
(0.02)
|
(0.03)
|
|
b_score
|
-0.18**
|
-0.06
|
-0.16**
|
0.09
|
|
|
(0.08)
|
(0.08)
|
(0.07)
|
(0.08)
|
|
mother_ed_ind
|
-0.00
|
-0.01
|
-0.01
|
0.01
|
|
|
(0.01)
|
(0.01)
|
(0.01)
|
(0.01)
|
|
father_ed_ind
|
0.00
|
0.01
|
0.00
|
-0.00
|
|
|
(0.01)
|
(0.01)
|
(0.01)
|
(0.01)
|
|
b_i_area
|
0.03
|
0.05
|
0.01
|
0.03
|
|
|
(0.04)
|
(0.04)
|
(0.03)
|
(0.04)
|
|
b_eb1_hstype
|
0.01
|
-0.01
|
0.00
|
-0.00
|
|
|
(0.02)
|
(0.02)
|
(0.02)
|
(0.02)
|
|
b_i_school_years_english
|
-0.00
|
-0.00
|
-0.01**
|
-0.00
|
|
|
(0.00)
|
(0.00)
|
(0.00)
|
(0.00)
|
|
z_seg_g3
|
|
|
0.28***
|
0.30***
|
|
|
|
|
(0.02)
|
(0.02)
|
|
AIC
|
7335.15
|
6629.24
|
8163.92
|
7527.46
|
|
BIC
|
7411.97
|
6705.03
|
8242.48
|
7604.58
|
|
Log Likelihood
|
-3654.58
|
-3301.62
|
-4068.96
|
-3750.73
|
|
Num. obs.
|
2722
|
2514
|
3113
|
2786
|
|
Num. groups: department_id
|
98
|
98
|
97
|
95
|
|
Var: department_id (Intercept)
|
0.06
|
0.06
|
0.04
|
0.08
|
|
Var: Residual
|
0.81
|
0.76
|
0.76
|
0.81
|
|
p < 0.01; p < 0.05; p < 0.1
|
Model ICC
bind_rows(
icc(lm2.1, by_group = T),
icc(lm2.2, by_group = T),
icc(lm2.3, by_group = T),
icc(lm2.4, by_group = T),
.id = "model"
) %>%
pivot_wider(names_from = model, values_from = ICC) %>%
select(
level = Group,
"Cohort 1 Res" = "1",
"Cohort 1 Non-res" = "2",
"Cohort 2 Res" = "3",
"Cohort 2 Non-res" = "4"
) %>%
kable()
| department_id |
0.0634569 |
0.0698856 |
0.051064 |
0.0879738 |
Adding department level predictors
Student level predictors are still included (even though not shown in results below)
Model summary
lm3.1 <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diverse_dorms + mentoring_programs + integration_courses + diversity_office + (1 |
department_id),
data = df %>% filter(reservation == 1)
)
lm3.2 <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diverse_dorms + mentoring_programs + integration_courses + diversity_office + (1 |
department_id),
data = df %>% filter(reservation == 0)
)
lm3.3 <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diverse_dorms + mentoring_programs + integration_courses + diversity_office + (1 |
department_id),
data = df %>% filter(reservation == 1)
)
lm3.4 <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diverse_dorms + mentoring_programs + integration_courses + diversity_office + (1 |
department_id),
data = df %>% filter(reservation == 0)
)
knitreg(
list(lm3.1, lm3.2, lm3.3, lm3.4),
custom.note = "%stars",
stars = c(0.01, 0.05, 0.1),
include.ci = FALSE,
include.variance = TRUE,
custom.model.names = c(
"Cohort 1 Res",
"Cohort 1 Non-res",
"Cohort 2 Res",
"Cohort 2 Non-res"
),
omit.coef = "(female)|(b_score)|(ther_ed)|(area)|(^ses)|(hstype)|(age)|(official)|(school)|(seg)"
)
Statistical models
|
|
Cohort 1 Res
|
Cohort 1 Non-res
|
Cohort 2 Res
|
Cohort 2 Non-res
|
|
(Intercept)
|
-0.08
|
-0.04
|
0.06
|
0.12
|
|
|
(0.11)
|
(0.13)
|
(0.10)
|
(0.12)
|
|
diverse_dorms
|
0.12*
|
-0.09
|
-0.04
|
-0.16*
|
|
|
(0.07)
|
(0.08)
|
(0.06)
|
(0.08)
|
|
mentoring_programs
|
-0.15*
|
0.07
|
0.09
|
-0.10
|
|
|
(0.08)
|
(0.08)
|
(0.07)
|
(0.09)
|
|
integration_courses
|
0.03
|
0.06
|
-0.06
|
-0.03
|
|
|
(0.07)
|
(0.07)
|
(0.06)
|
(0.08)
|
|
diversity_office
|
0.04
|
0.04
|
-0.05
|
-0.05
|
|
|
(0.07)
|
(0.07)
|
(0.06)
|
(0.07)
|
|
AIC
|
7322.28
|
6637.32
|
8136.43
|
7517.03
|
|
BIC
|
7434.38
|
6748.00
|
8251.07
|
7629.64
|
|
Log Likelihood
|
-3642.14
|
-3299.66
|
-4049.22
|
-3739.52
|
|
Num. obs.
|
2698
|
2503
|
3083
|
2770
|
|
Num. groups: department_id
|
97
|
97
|
96
|
94
|
|
Var: department_id (Intercept)
|
0.06
|
0.06
|
0.04
|
0.08
|
|
Var: Residual
|
0.82
|
0.77
|
0.77
|
0.82
|
|
p < 0.01; p < 0.05; p < 0.1
|
Model ICC
bind_rows(
icc(lm3.1, by_group = T),
icc(lm3.2, by_group = T),
icc(lm3.3, by_group = T),
icc(lm3.4, by_group = T),
.id = "model"
) %>%
pivot_wider(names_from = model, values_from = ICC) %>%
select(
level = Group,
"Cohort 1 Res" = "1",
"Cohort 1 Non-res" = "2",
"Cohort 2 Res" = "3",
"Cohort 2 Non-res" = "4"
) %>%
kable()
| department_id |
0.0643599 |
0.0725649 |
0.0533912 |
0.0846347 |
Adding random slope
Diverse Dorms
lm4.1a <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diverse_dorms + (1 + diverse_dorms |
department_id),
data = df %>% filter(reservation == 1)
)
lm4.2a <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diverse_dorms + (1 + diverse_dorms |
department_id),
data = df %>% filter(reservation == 0)
)
lm4.3a <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diverse_dorms + (1 + diverse_dorms |
department_id),
data = df %>% filter(reservation == 1)
)
lm4.4a <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diverse_dorms + (1 + diverse_dorms |
department_id),
data = df %>% filter(reservation == 0)
)
knitreg(
list(lm4.1a, lm4.2a, lm4.3a, lm4.4a),
custom.note = "%stars",
stars = c(0.01, 0.05, 0.1),
include.ci = FALSE,
include.variance = TRUE,
custom.model.names = c(
"Cohort 1 Res",
"Cohort 1 Non-res",
"Cohort 2 Res",
"Cohort 2 Non-res"
),
omit.coef = "(female)|(b_score)|(ther_ed)|(area)|(^ses)|(hstype)|(age)|(official)|(school)|(seg)"
)
Statistical models
|
|
Cohort 1 Res
|
Cohort 1 Non-res
|
Cohort 2 Res
|
Cohort 2 Non-res
|
|
(Intercept)
|
-0.13
|
0.02
|
0.06
|
0.03
|
|
|
(0.10)
|
(0.12)
|
(0.09)
|
(0.12)
|
|
diverse_dorms
|
0.10
|
-0.04
|
-0.04
|
-0.19**
|
|
|
(0.07)
|
(0.08)
|
(0.07)
|
(0.08)
|
|
AIC
|
7355.69
|
6646.19
|
8182.67
|
7541.84
|
|
BIC
|
7462.06
|
6751.13
|
8291.45
|
7648.62
|
|
Log Likelihood
|
-3659.85
|
-3305.10
|
-4073.33
|
-3752.92
|
|
Num. obs.
|
2722
|
2514
|
3113
|
2786
|
|
Num. groups: department_id
|
98
|
98
|
97
|
95
|
|
Var: department_id (Intercept)
|
0.08
|
0.10
|
0.08
|
0.11
|
|
Var: department_id diverse_dorms
|
0.07
|
0.02
|
0.02
|
0.07
|
|
Cov: department_id (Intercept) diverse_dorms
|
-0.05
|
-0.04
|
-0.04
|
-0.06
|
|
Var: Residual
|
0.81
|
0.76
|
0.76
|
0.81
|
|
p < 0.01; p < 0.05; p < 0.1
|
Mentoring Programs
lm4.1b <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + mentoring_programs + (1 + mentoring_programs |
department_id),
data = df %>% filter(reservation == 1)
)
lm4.2b <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + mentoring_programs + (1 + mentoring_programs |
department_id),
data = df %>% filter(reservation == 0)
)
lm4.3b <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + mentoring_programs + (1 + mentoring_programs |
department_id),
data = df %>% filter(reservation == 1)
)
lm4.4b <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + mentoring_programs + (1 + mentoring_programs |
department_id),
data = df %>% filter(reservation == 0)
)
knitreg(
list(lm4.1b, lm4.2b, lm4.3b, lm4.4b),
custom.note = "%stars",
stars = c(0.01, 0.05, 0.1),
include.ci = FALSE,
include.variance = TRUE,
custom.model.names = c(
"Cohort 1 Res",
"Cohort 1 Non-res",
"Cohort 2 Res",
"Cohort 2 Non-res"
),
omit.coef = "(female)|(b_score)|(ther_ed)|(area)|(^ses)|(hstype)|(age)|(official)|(school)|(seg)"
)
Statistical models
|
|
Cohort 1 Res
|
Cohort 1 Non-res
|
Cohort 2 Res
|
Cohort 2 Non-res
|
|
(Intercept)
|
0.02
|
-0.07
|
0.01
|
0.00
|
|
|
(0.10)
|
(0.12)
|
(0.09)
|
(0.12)
|
|
mentoring_programs
|
-0.12
|
0.10
|
0.04
|
-0.17*
|
|
|
(0.08)
|
(0.08)
|
(0.06)
|
(0.10)
|
|
AIC
|
7354.89
|
6647.27
|
8187.17
|
7542.54
|
|
BIC
|
7461.26
|
6752.20
|
8295.95
|
7649.32
|
|
Log Likelihood
|
-3659.45
|
-3305.63
|
-4075.59
|
-3753.27
|
|
Num. obs.
|
2722
|
2514
|
3113
|
2786
|
|
Num. groups: department_id
|
98
|
98
|
97
|
95
|
|
Var: department_id (Intercept)
|
0.08
|
0.09
|
0.04
|
0.16
|
|
Var: department_id mentoring_programs
|
0.21
|
0.08
|
0.12
|
0.03
|
|
Cov: department_id (Intercept) mentoring_programs
|
-0.12
|
-0.06
|
-0.06
|
-0.06
|
|
Var: Residual
|
0.81
|
0.76
|
0.76
|
0.81
|
|
p < 0.01; p < 0.05; p < 0.1
|
Integration courses
lm4.1c <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + integration_courses + (1 + integration_courses |
department_id),
data = df %>% filter(reservation == 1)
)
lm4.2c <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + integration_courses + (1 + integration_courses |
department_id),
data = df %>% filter(reservation == 0)
)
lm4.3c <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + integration_courses + (1 + integration_courses |
department_id),
data = df %>% filter(reservation == 1)
)
lm4.4c <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + integration_courses + (1 + integration_courses |
department_id),
data = df %>% filter(reservation == 0)
)
knitreg(
list(lm4.1c, lm4.2c, lm4.3c, lm4.4c),
custom.note = "%stars",
stars = c(0.01, 0.05, 0.1),
include.ci = FALSE,
include.variance = TRUE,
custom.model.names = c(
"Cohort 1 Res",
"Cohort 1 Non-res",
"Cohort 2 Res",
"Cohort 2 Non-res"
),
omit.coef = "(female)|(b_score)|(ther_ed)|(area)|(^ses)|(hstype)|(age)|(official)|(school)|(seg)"
)
Statistical models
|
|
Cohort 1 Res
|
Cohort 1 Non-res
|
Cohort 2 Res
|
Cohort 2 Non-res
|
|
(Intercept)
|
-0.06
|
-0.05
|
0.06
|
-0.06
|
|
|
(0.09)
|
(0.11)
|
(0.09)
|
(0.11)
|
|
integration_courses
|
-0.00
|
0.08
|
-0.07
|
-0.10
|
|
|
(0.07)
|
(0.07)
|
(0.06)
|
(0.07)
|
|
AIC
|
7357.53
|
6647.08
|
8170.43
|
7542.99
|
|
BIC
|
7463.90
|
6752.02
|
8279.21
|
7649.78
|
|
Log Likelihood
|
-3660.77
|
-3305.54
|
-4067.22
|
-3753.50
|
|
Num. obs.
|
2722
|
2514
|
3113
|
2786
|
|
Num. groups: department_id
|
98
|
98
|
97
|
95
|
|
Var: department_id (Intercept)
|
0.08
|
0.08
|
0.09
|
0.13
|
|
Var: department_id integration_courses
|
0.03
|
0.09
|
0.14
|
0.31
|
|
Cov: department_id (Intercept) integration_courses
|
-0.03
|
-0.06
|
-0.12
|
-0.20
|
|
Var: Residual
|
0.81
|
0.76
|
0.76
|
0.81
|
|
p < 0.01; p < 0.05; p < 0.1
|
Divresity Offices
lm4.1d <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diversity_office + (1 + diversity_office |
department_id),
data = df %>% filter(reservation == 1)
)
lm4.2d <-
lmer(
z_seg_g2 ~ z_seg_g1 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diversity_office + (1 + diversity_office |
department_id),
data = df %>% filter(reservation == 0)
)
lm4.3d <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diversity_office + (1 + diversity_office |
department_id),
data = df %>% filter(reservation == 1)
)
lm4.4d <-
lmer(
z_seg_g4 ~ z_seg_g3 + female + age + ses + b_score + mother_ed_ind + father_ed_ind + b_i_area + b_eb1_hstype + b_i_school_years_english + stu_res_official + fac_res_official + diversity_office + (1 + diversity_office |
department_id),
data = df %>% filter(reservation == 0)
)
knitreg(
list(lm4.1d, lm4.2d, lm4.3d, lm4.4d),
custom.note = "%stars",
stars = c(0.01, 0.05, 0.1),
include.ci = FALSE,
include.variance = TRUE,
custom.model.names = c(
"Cohort 1 Res",
"Cohort 1 Non-res",
"Cohort 2 Res",
"Cohort 2 Non-res"
),
omit.coef = "(female)|(b_score)|(ther_ed)|(area)|(^ses)|(hstype)|(age)|(official)|(school)|(seg)"
)
Statistical models
|
|
Cohort 1 Res
|
Cohort 1 Non-res
|
Cohort 2 Res
|
Cohort 2 Non-res
|
|
(Intercept)
|
-0.06
|
-0.02
|
0.07
|
-0.07
|
|
|
(0.09)
|
(0.11)
|
(0.09)
|
(0.10)
|
|
diversity_office
|
0.02
|
0.04
|
-0.05
|
-0.06
|
|
|
(0.06)
|
(0.07)
|
(0.05)
|
(0.08)
|
|
AIC
|
7314.77
|
6627.51
|
8117.43
|
7512.69
|
|
BIC
|
7420.98
|
6732.37
|
8226.04
|
7619.37
|
|
Log Likelihood
|
-3639.39
|
-3295.76
|
-4040.72
|
-3738.35
|
|
Num. obs.
|
2698
|
2503
|
3083
|
2770
|
|
Num. groups: department_id
|
97
|
97
|
96
|
94
|
|
Var: department_id (Intercept)
|
0.07
|
0.05
|
0.07
|
0.08
|
|
Var: department_id diversity_office
|
0.01
|
0.00
|
0.04
|
0.02
|
|
Cov: department_id (Intercept) diversity_office
|
-0.02
|
0.01
|
-0.05
|
-0.01
|
|
Var: Residual
|
0.82
|
0.77
|
0.77
|
0.82
|
|
p < 0.01; p < 0.05; p < 0.1
|
Model comparison
Cohort 1 reservation
compare_performance(lm1.1, lm2.1, lm3.1, lm4.1a, lm4.1b, lm4.1c, lm4.1d)
## # Comparison of Model Performance Indices
##
## Model | Type | AIC | BIC | R2_conditional | R2_marginal | ICC | RMSE | Sigma
## -------------------------------------------------------------------------------------------
## lm1.1 | lmerMod | 12118.50 | 12137.63 | 0.08 | 0.00 | 0.08 | 0.95 | 0.96
## lm2.1 | lmerMod | 7335.15 | 7411.97 | 0.13 | 0.07 | 0.06 | 0.89 | 0.90
## lm3.1 | lmerMod | 7322.28 | 7434.38 | 0.14 | 0.08 | 0.06 | 0.89 | 0.90
## lm4.1a | lmerMod | 7355.69 | 7462.06 | 0.13 | 0.07 | 0.07 | 0.89 | 0.90
## lm4.1b | lmerMod | 7354.90 | 7461.26 | 0.14 | 0.07 | 0.07 | 0.89 | 0.90
## lm4.1c | lmerMod | 7357.53 | 7463.90 | 0.13 | 0.07 | 0.07 | 0.89 | 0.90
## lm4.1d | lmerMod | 7314.77 | 7420.98 | 0.13 | 0.07 | 0.06 | 0.89 | 0.91
Cohort 1 non-reservation
compare_performance(lm1.2, lm2.2, lm3.2, lm4.2a, lm4.2b, lm4.2c, lm4.2d)
## # Comparison of Model Performance Indices
##
## Model | Type | AIC | BIC | R2_conditional | R2_marginal | ICC | RMSE | Sigma
## -------------------------------------------------------------------------------------------
## lm1.2 | lmerMod | 10441.20 | 10459.97 | 0.11 | 0.00 | 0.11 | 0.91 | 0.92
## lm2.2 | lmerMod | 6629.24 | 6705.03 | 0.13 | 0.06 | 0.07 | 0.86 | 0.87
## lm3.2 | lmerMod | 6637.32 | 6748.00 | 0.13 | 0.07 | 0.07 | 0.86 | 0.88
## lm4.2a | lmerMod | 6646.19 | 6751.13 | 0.13 | 0.07 | 0.07 | 0.86 | 0.87
## lm4.2b | lmerMod | 6647.27 | 6752.20 | 0.13 | 0.07 | 0.07 | 0.86 | 0.87
## lm4.2c | lmerMod | 6647.08 | 6752.02 | 0.13 | 0.06 | 0.07 | 0.86 | 0.87
## lm4.2d | lmerMod | 6627.51 | 6732.37 | 0.13 | 0.07 | 0.07 | 0.86 | 0.88
Cohort 2 reservation
compare_performance(lm1.3, lm2.3, lm3.3, lm4.3a, lm4.3b, lm4.3c, lm4.3d)
## # Comparison of Model Performance Indices
##
## Model | Type | AIC | BIC | R2_conditional | R2_marginal | ICC | RMSE | Sigma
## -------------------------------------------------------------------------------------------
## lm1.3 | lmerMod | 11528.03 | 11547.08 | 0.06 | 0.00 | 0.06 | 0.92 | 0.93
## lm2.3 | lmerMod | 8163.92 | 8242.48 | 0.14 | 0.09 | 0.05 | 0.86 | 0.87
## lm3.3 | lmerMod | 8136.44 | 8251.07 | 0.14 | 0.09 | 0.05 | 0.87 | 0.88
## lm4.3a | lmerMod | 8182.67 | 8291.45 | 0.14 | 0.09 | 0.05 | 0.86 | 0.87
## lm4.3b | lmerMod | 8187.17 | 8295.95 | 0.14 | 0.09 | 0.06 | 0.86 | 0.87
## lm4.3c | lmerMod | 8170.44 | 8279.22 | 0.15 | 0.09 | 0.06 | 0.86 | 0.87
## lm4.3d | lmerMod | 8117.43 | 8226.04 | 0.14 | 0.09 | 0.06 | 0.87 | 0.88
Cohort 2 non-reservation
compare_performance(lm1.4, lm2.4, lm3.4, lm4.4a, lm4.4b, lm4.4c, lm4.4d)
## # Comparison of Model Performance Indices
##
## Model | Type | AIC | BIC | R2_conditional | R2_marginal | ICC | RMSE | Sigma
## -------------------------------------------------------------------------------------------
## lm1.4 | lmerMod | 10997.02 | 11015.85 | 0.14 | 0.00 | 0.14 | 0.95 | 0.96
## lm2.4 | lmerMod | 7527.46 | 7604.58 | 0.17 | 0.09 | 0.09 | 0.89 | 0.90
## lm3.4 | lmerMod | 7517.03 | 7629.64 | 0.18 | 0.11 | 0.08 | 0.89 | 0.90
## lm4.4a | lmerMod | 7541.84 | 7648.62 | 0.18 | 0.11 | 0.08 | 0.89 | 0.90
## lm4.4b | lmerMod | 7542.54 | 7649.32 | 0.19 | 0.10 | 0.09 | 0.89 | 0.90
## lm4.4c | lmerMod | 7542.99 | 7649.78 | 0.19 | 0.10 | 0.10 | 0.89 | 0.90
## lm4.4d | lmerMod | 7512.69 | 7619.37 | 0.18 | 0.10 | 0.09 | 0.89 | 0.90