-
Notifications
You must be signed in to change notification settings - Fork 2
/
groups_comparisons.rmd
331 lines (254 loc) · 12 KB
/
groups_comparisons.rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
---
title: "Interpretable Analysis of School Policy Decisions, Comparison to schools outside the program"
author: "Charles Saluski"
# date: "1/4/2022"
output: pdf_document
---
```{r}
library(data.table)
library(dplyr)
library(stringr)
library(purrr)
library(openxlsx)
```
```{r}
district.map.dt <- fread("Data Sources CSV/building_map_data/District MAP content area and grade all disag.csv")
district.map.dt[, State.District.ID := paste0("MO-", str_pad(COUNTY_DISTRICT, 6, "left", "0"))]
dci.buildings.dt <- fread("Data Sources/DCI Data/Active Districts/Active_DCI_buildings_2017_2022.csv")
dci.districts.dt <- unique(dci.buildings.dt[, .(State.District.ID, currentSchoolYear)])
dci.districts.dt <- dci.districts.dt[order(currentSchoolYear)]
# Account for years taken off from the program by counting along the years the
# district was actually in the program
dci.districts.dt[
,
year.in.program := which(.SD$currentSchoolYear == currentSchoolYear),
by = "State.District.ID"]
dci.years.dt <- dci.districts.dt
dci.districts.dt <- dci.districts.dt[district.map.dt, on=c("State.District.ID", "currentSchoolYear" = "YEAR")]
dci.districts.dt[, year.in.program := ifelse(is.na(year.in.program), 0, year.in.program)]
dci.districts.dt <- dci.districts.dt[
(currentSchoolYear == 2019 | currentSchoolYear == 2021)
& TYPE %in% c("Total", "Non IEP Students", "IEP Non MAPA", "IEP_student")]
dci.districts.dt <- dci.districts.dt[CONTENT_AREA == "Eng. Language Arts" & GRADE_LEVEL == "03"]
year.in.program.dt <- unique(dci.districts.dt[, .(State.District.ID, currentSchoolYear, year.in.program, COUNTY_DISTRICT)])
state.average.dt <- fread("./Data Sources CSV/MAP State Average.csv")
state.average.dt[Type == "IEP (exclude MAPA)", Type := "IEP Non MAPA"]
# state.average.dt[Type == "Total (exclude MAPA)", Type := "IEP Non MAPA"]
state.average.dt[, PROF_AND_ADV_PCT_STATE := ((Advanced + Proficient) / Reportable) * 100]
state.average.dt <- state.average.dt[
Year >= 2019 & Type %in% unique(dci.districts.dt$TYPE),
.(Year, Type, PROF_AND_ADV_PCT_STATE)]
```
```{r}
discipline.dt <- fread("Data Sources CSV/District Discipline.csv")
attendance.dt <- fread("Data Sources CSV/District Proportional Attendance Rates.csv")
discipline.attendance.dt <- discipline.dt[attendance.dt, on=c("YEAR", "COUNTY_DISTRICT_CODE", "DISTRICT_NAME")]
discipline.attendance.dt <- discipline.attendance.dt[(YEAR == 2019 | YEAR == 2021) & complete.cases(discipline.attendance.dt)]
discipline.attendance.dt <- discipline.attendance.dt[, .(YEAR, COUNTY_DISTRICT_CODE, DSCPLN_INCIDENT_RATE, PROPORTIONAL_ATTENDANCE_K_8_PCT, PROPORTIONAL_ATTENDANCE_IEP_PCT)]
for (col in colnames(discipline.attendance.dt)) {
if (grepl("PROPORTIONAL", col)) {
num.col <- as.numeric(discipline.attendance.dt[, get(col)])
num.col <- ifelse(is.na(num.col), 0, num.col)
discipline.attendance.dt[, paste(col) := num.col]
}
}
discipline.map.dt <- dci.districts.dt[discipline.attendance.dt, on=c("currentSchoolYear" = "YEAR", "COUNTY_DISTRICT" = "COUNTY_DISTRICT_CODE")]
# TODO merge discipline with DCI Districts DT
discipline.map.dt <- discipline.map.dt[complete.cases(discipline.map.dt)]
# discipline.diff.dt <- merge(discipline.map.dt, discipline.map.dt,
# by = c("State.District.ID"),
# suffixes = c("_2019", "_2021")
# )[currentSchoolYear_2019 == 2019 & currentSchoolYear_2021 == 2021,
# ]
# diff.discipline.cols <- names(discipline.map.dt)[as.vector(sapply(discipline.map.dt, class)) %in% c("numeric", "num", "int")]
# for (col in diff.discipline.cols) {
# diff.col.name <- paste(col, "diff", sep=".")
# col.x.name <- c(paste(col, "2019", sep="_"))
# col.y.name <- c(paste(col, "2021", sep="_"))
# # I can't get this to work just using the variable names, so we're using get
# discipline.diff.dt[[diff.col.name]] <- discipline.diff.dt[, get(col.y.name) - get(col.x.name)]
# }
```
```{r}
dci.districts.dt <- dci.districts.dt[state.average.dt, on = c("currentSchoolYear" = "Year", "TYPE" = "Type")]
for (col in names(dci.districts.dt)) {
if (grepl("PCT", col)) {
dci.districts.dt[, paste0(col) := as.numeric(get(col))]
}
}
dci.districts.dt[, PROF_AND_ADV_PCT := PROFICIENT_PCT + ADVANCED_PCT]
# get changes from 2019 to 2021
covid.diff.dt <- merge(dci.districts.dt, dci.districts.dt,
by = c("State.District.ID", "TYPE", "GRADE_LEVEL", "CONTENT_AREA", "SCHOOL_NAME", "SCHOOL_CODE", "DISTRICT_NAME", "COUNTY_DISTRICT", "CATEGORY", "SUMMARY_LEVEL"),
suffixes = c("_2019", "_2021")
)[currentSchoolYear_2019 == 2019 & currentSchoolYear_2021 == 2021,
]
diff.cols <- names(dci.districts.dt)[as.vector(sapply(dci.districts.dt, class)) %in% c("numeric", "num", "int")]
for (col in diff.cols) {
diff.col.name <- paste(col, "diff", sep=".")
col.x.name <- c(paste(col, "2019", sep="_"))
col.y.name <- c(paste(col, "2021", sep="_"))
# I can't get this to work just using the variable names, so we're using get
covid.diff.dt[[diff.col.name]] <- covid.diff.dt[, get(col.y.name) - get(col.x.name)]
}
covid.diff.dt[, gap.2019 := PROF_AND_ADV_PCT_2019 - PROF_AND_ADV_PCT_STATE_2019]
covid.diff.dt[, gap.2021 := PROF_AND_ADV_PCT_2021 - PROF_AND_ADV_PCT_STATE_2021]
covid.diff.dt[, gap.diff := gap.2021 - gap.2019]
```
See how much the proficient and advanced percent changed before and after covid,
and see if its change was impacted by the number of years in the program.
```{r}
year.list <- list()
box.list <- list()
diff.types <- c("Total", "Non IEP Students", "IEP_student")
for (year in unique(covid.diff.dt$year.in.program_2019)) {
for (type in unique(covid.diff.dt$TYPE))
{
year.list[[paste(type, year)]] <- data.table(
value = mean(covid.diff.dt[year.in.program_2019 == year & TYPE == type]$PROF_AND_ADV_PCT.diff),
year,
type
)
box.list[[paste(type, year)]] <- data.table(covid.diff.dt[year.in.program_2019 == year & TYPE == type, .(PROF_AND_ADV_PCT.diff, year, type, PROF_AND_ADV_PCT_STATE.diff, State.District.ID, gap.diff)])
}
}
# diff.interest.cols <- c("PROPORTIONAL_ATTENDANCE_TOTAL_PCT", "PROPORTIONAL_ATTENDANCE_IEP_PCT", "DSCPLN_INCIDENT_RATE")
# for (year in unique(covid.diff.dt$year.in.program_2019)) {
# for (col in diff.interest.cols)
# {
# year.list[[paste(col, year)]] <- data.table(
# value = mean(covid.diff.dt[year.in.program_2019 >= year, get(paste0(col, ".diff"))]),
# year,
# type = col
# )
# }
# }
covid.diff.res.dt <- data.table(do.call(rbind, year.list))
covid.diff.res.dt$year <- as.factor(covid.diff.res.dt$year)
covid.box.dt <- data.table(do.call(rbind, box.list))
covid.box.dt[, n := paste(year, ", n = ", length(unique(.SD$State.District.ID)), sep = ""), by = year]
covid.box.dt$year <- as.factor(covid.box.dt$year)
```
```{r}
library(ggplot2)
dest <- "./img_out/by_iep_groups/"
if (!dir.exists(dest)) {
dir.create(dest)
}
map.change.plot <- ggplot() +
geom_boxplot(
data = covid.box.dt,
mapping = aes(x = n, y = PROF_AND_ADV_PCT.diff, fill = type)
) +
geom_errorbar(
data = unique(covid.box.dt[, .(n, type, PROF_AND_ADV_PCT_STATE.diff)]),
mapping = aes(
x = n,
ymin = PROF_AND_ADV_PCT_STATE.diff, ymax = PROF_AND_ADV_PCT_STATE.diff,
color = type)
) +
xlab("Years in Program in 2019, n = number of districts") +
ylab("3rd Grade MAP ELA difference, 2019 to 2021") +
labs(title = "MAP ELA changes by years in DCI program", fill = "MAP Group", color = "State Average")
filename <- paste(dest, "map_ela_changes.png", sep = "")
png(filename = filename, width = 6, height = 8, unit = "in", res = 200)
plot(map.change.plot)
dev.off()
```
```{r}
gap.diff.plot <- ggplot(data = covid.box.dt) +
geom_boxplot(aes(x=n, y=gap.diff, fill=type)) +
xlab("Years in Program in 2019, n = number of districts") +
ylab("3rd Grade MAP ELA State Average Difference, 2019 to 2021") +
labs(title = "MAP ELA gap changes by years in DCI program, \n normalized against state average", fill = "MAP Group", color = "State Average")
filename <- paste(dest, "map_ela_diff_changes.png", sep = "")
png(filename = filename, width = 6, height = 8, unit = "in", res = 200)
plot(gap.diff.plot)
dev.off()
```
```{r}
dci.districts.dt[, n := paste(year.in.program, ", n = ", length(unique(.SD$State.District.ID)), sep = ""), by = c("year.in.program")]
map.years.plot <- ggplot() +
geom_boxplot(
data = dci.districts.dt,
mapping = aes(x = n, y = PROF_AND_ADV_PCT, fill = TYPE)
) +
geom_errorbar(
data = unique(dci.districts.dt[, .(n, TYPE, PROF_AND_ADV_PCT_STATE, currentSchoolYear)]),
mapping = aes(
x = n,
ymin = PROF_AND_ADV_PCT_STATE, ymax = PROF_AND_ADV_PCT_STATE,
color = TYPE)
) +
facet_grid(currentSchoolYear ~ .) +
xlab("Years in Program in year, n = number of districts in 2021") +
ylab("3rd Grade MAP ELA score") +
labs(title = "MAP ELA boxplot by years in DCI program in 2021", fill = "MAP Group", color = "State Average")
filename <- paste(dest, "map_ela_by_year.png", sep = "")
png(filename = filename, width = 6, height = 8, unit = "in", res = 200)
plot(map.years.plot)
dev.off()
```
```{r}
discipline.map.dt[, n := paste(year.in.program, ", n = ", length(unique(.SD$State.District.ID)), sep = ""), by = c("year.in.program")]
discipline.years.plot <- ggplot() +
geom_boxplot(
data = discipline.map.dt,
mapping = aes(x = n, y = DSCPLN_INCIDENT_RATE)
) +
facet_grid(currentSchoolYear ~ .) +
xlab("Years in Program in year, n = number of districts in 2021") +
ylab("") +
labs(title = "Discipline rate boxplot by years in DCI program")
filename <- paste(dest, "discipline_by_years.png", sep = "")
png(filename = filename, width = 6, height = 8, unit = "in", res = 200)
plot(discipline.years.plot)
dev.off()
targets <- c("PROPORTIONAL_ATTENDANCE_K_8_PCT","PROPORTIONAL_ATTENDANCE_IEP_PCT")
discipline.attendance.long.dt <- melt(
discipline.map.dt,
measure.vars = targets,
variable.name = "attendance_type",
value.name = "attendance_rate"
)
attendance.change.plot <- ggplot() +
geom_boxplot(
data = discipline.attendance.long.dt,
mapping = aes(x = n, y = attendance_rate, fill = attendance_type)
) +
facet_grid(currentSchoolYear ~ .) +
xlab("Years in Program in year, n = number of districts in 2021") +
ylab("") +
labs(title = "Attendance boxplot by years in DCI program")
filename <- paste(dest, "attendance_by_year.png", sep = "")
png(filename = filename, width = 10, height = 8, unit = "in", res = 200)
plot(attendance.change.plot)
dev.off()
```
As all of the previous comparisons show that the schools in the DCI programs have no significant improvement over those outside of it, and in many cases are worse, we inspect whether these schools were also below performing worse before the DCI programs were implemented.
```{r}
dci.years.dt[, total.years.2022 := max(.SD$year.in.program), by = State.District.ID]
dci.total.years <- unique(dci.years.dt[, .(State.District.ID, total.years.2022)])
comparison.map.dt <- district.map.dt[CONTENT_AREA == "Eng. Language Arts" & GRADE_LEVEL == "03"]
comparison.map.dt <- comparison.map.dt[dci.total.years, on = c("State.District.ID"), nomatch = NULL]
state.average.dt <- fread("./Data Sources CSV/MAP State Average.csv")
state.average.dt[Type == "IEP (exclude MAPA)", Type := "IEP Non MAPA"]
state.average.dt[, PROF_AND_ADV_PCT_STATE := ((Advanced + Proficient) / Reportable) * 100]
comparison.map.dt <- comparison.map.dt[state.average.dt, on = c("YEAR" = "Year", "TYPE" = "Type"), nomatch = NULL]
comparison.map.dt[, PROFICIENT_PCT := as.numeric(PROFICIENT_PCT)]
comparison.map.dt[, ADVANCED_PCT := as.numeric(ADVANCED_PCT)]
comparison.map.dt[, YEAR := as.factor(YEAR)]
comparison.map.dt[, PROF_AND_ADV_PCT := PROFICIENT_PCT + ADVANCED_PCT]
comparison.map.dt[, state.avg.diff := PROF_AND_ADV_PCT - PROF_AND_ADV_PCT_STATE]
yearly.map.diff.by.cohort.plot <- ggplot(comparison.map.dt) +
geom_boxplot(
mapping = aes(x = YEAR, y = state.avg.diff, fill = TYPE)
) +
facet_grid(
total.years.2022 ~ .
)
filename <- paste(dest, "map_by_year_and_cohort.png", sep = "")
png(filename = filename, width = 6, height = 12, unit = "in", res = 200)
plot(yearly.map.diff.by.cohort.plot)
dev.off()
```