Hawaii/Hawaii_SGPt_Baseline_2019_A_Matrix_Calculations.R at master · CenterForAssessment/Hawaii · GitHub

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
################################################################################
###                                                                          ###
###       Hawaii Time Dependent SGP Analyses -- Create Baseline Matrices     ###
###                                                                          ###
################################################################################

### Load necessary packages
require(SGP)
require(data.table)
require(cfaTools)
#debug(studentGrowthPercentiles)

###   Load the results data from the base file with time dependent data
load("Data/Base_Files/Hawaii_Data_LONG_2015_2025_TIME_DEPENDENT.Rdata")

###   Create a smaller subset of the LONG data to work with.
Hawaii_Baseline_Data_TIME_DEPENDENT <- data.table::data.table(Hawaii_Data_LONG_2015_2025_TIME_DEPENDENT[Year >= 2015 & Year <= 2019 & !is.na(TStartDt) & !is.na(TEndDt) & TEndDt > TStartDt & grade %in% c("3", "4", "5", "6", "7", "8", "11"),
	c("IDNO", "Domain", "Year", "grade", "Scale_Score", "Proficiency_Level", "Valid_Case", "TStartDt", "TEndDt"),])
setnames(Hawaii_Baseline_Data_TIME_DEPENDENT, c("IDNO", "Domain", "Year", "grade", "Scale_Score", "Proficiency_Level", "Valid_Case", "TStartDt", "TEndDt"), c("ID", "CONTENT_AREA", "YEAR", "GRADE", "SCALE_SCORE", "ACHIEVEMENT_LEVEL", "VALID_CASE", "TStartDt", "TEndDt"))

### Tidy up data
Hawaii_Baseline_Data_TIME_DEPENDENT[,ID:=as.character(ID)]
Hawaii_Baseline_Data_TIME_DEPENDENT[,YEAR:=as.character(YEAR)]
Hawaii_Baseline_Data_TIME_DEPENDENT[,GRADE:=as.character(GRADE)]
Hawaii_Baseline_Data_TIME_DEPENDENT[,SCALE_SCORE:=as.numeric(SCALE_SCORE)]
Hawaii_Baseline_Data_TIME_DEPENDENT[,ACHIEVEMENT_LEVEL:=as.character(ACHIEVEMENT_LEVEL)]
Hawaii_Baseline_Data_TIME_DEPENDENT[,VALID_CASE:=as.character(VALID_CASE)]

### Convert TStartDt and TEndDt to create date and time variables
### NOTE: Some values of TStartDt and TEndDt are missing (NA) or are set to 1900-01-01
###       for data 2019 and prior. Only create date/time variables when valid.
Hawaii_Baseline_Data_TIME_DEPENDENT[!is.na(TStartDt) & !is.na(TEndDt) & TEndDt > TStartDt, ':='(
    DATE_START = as.IDate(TStartDt),    # Test start date
    DATE_END = as.IDate(TEndDt),        # Test end date
    TIME_START = as.ITime(TStartDt),    # Test start time
    TIME_END = as.ITime(TEndDt)         # Test end time
)]

Hawaii_Baseline_Data_TIME_DEPENDENT[,DATE := DATE_END]

### Calculate time span within testing year (how long the test window was open)
Hawaii_Baseline_Data_TIME_DEPENDENT[, DATE_SPAN_WITHIN_YEAR := DATE_END - DATE_START]

### Create lagged date variables using cfaTools::getShiftedValues()
### This creates DATE_LAG_1 (previous year's date) for each student by domain
Hawaii_Baseline_Data_TIME_DEPENDENT <- getShiftedValues(
    Hawaii_Baseline_Data_TIME_DEPENDENT,
    shift_group = c("ID", "CONTENT_AREA"),   # Group by student and subject
    shift_period = "YEAR",                # Shift by year
    shift_variable = "DATE"               # Variable to shift
)

### Create DATE_LAG_3 (date from 3 years prior) for grade 11 analyses
Hawaii_Baseline_Data_TIME_DEPENDENT <- getShiftedValues(
    Hawaii_Baseline_Data_TIME_DEPENDENT,
    shift_group = c("ID", "CONTENT_AREA"),
    shift_period = "YEAR",
    shift_variable = "DATE",
    shift_amount = 3L                     # Look back 3 years
)

### Calculate DATE_SPAN_SGPt: Time elapsed between assessments
### This is THE CRITICAL VARIABLE for time-dependent SGP calculations
### - Grades 3-8: Use 1-year lag (consecutive years)
### - Grade 11: Use 3-year lag (accounts for testing schedule)
Hawaii_Baseline_Data_TIME_DEPENDENT[GRADE %in% c("3", "4", "5", "6", "7", "8"),
                            DATE_SPAN_SGPt := DATE - DATE_LAG_1]
Hawaii_Baseline_Data_TIME_DEPENDENT[GRADE == "11",
                            DATE_SPAN_SGPt := DATE - DATE_LAG_3]

### Remove temporary lag variables
Hawaii_Baseline_Data_TIME_DEPENDENT[, c("DATE_LAG_1", "DATE_LAG_3") := NULL]

### Set all VALID_CASE values to "VALID_CASE"
Hawaii_Baseline_Data_TIME_DEPENDENT[, VALID_CASE := "VALID_CASE"]

### Set data.table key for efficient operations
### Ordering: VALID_CASE, Year, Domain (subject), IDNO (student)
setkey(Hawaii_Baseline_Data_TIME_DEPENDENT, VALID_CASE, CONTENT_AREA, YEAR, GRADE, ID)

###   Read in Baseline SGP Configuration Scripts and Combine
source("SGP_CONFIG/2019/BASELINE/Matrices/SGPt/READING.R")
source("SGP_CONFIG/2019/BASELINE/Matrices/SGPt/MATHEMATICS.R")

HI_BASELINE_CONFIG <- c(
	READING_2019.config,
	MATHEMATICS_2019.config
)

###   Run SGPt analyses via abcSGP
Hawaii_SGP <- abcSGP(
        sgp_object = Hawaii_Baseline_Data_TIME_DEPENDENT,   # Input data (2015-2019)
        steps = c("prepareSGP",                             # Prepare data structure
                  "analyzeSGP",                             # Run quantile regression models
                  "combineSGP"),                             # Merge results back to data
#                  "outputSGP"),                             # Export results
        sgp.config = HI_BASELINE_CONFIG,                             # Use configurations for Reading & Math
        sgp.percentiles = TRUE,                             # Calculate SGPt percentiles
        sgp.projections = FALSE,                            # Disable projections
        sgp.projections.lagged = FALSE,                     # Disable lagged projections
        sgp.percentiles.baseline = FALSE,                   # Disable baseline percentiles
        sgp.projections.baseline = FALSE,                   # Disable baseline projections
        sgp.projections.lagged.baseline = FALSE,            # Disable baseline lagged projections
        save.intermediate.results = FALSE,                  # Don't save intermediate files
        SGPt = TRUE,                                        # ENABLE TIME-DEPENDENT SGP
#        outputSGP.output.type = "LONG_FINAL_YEAR_Data",    # Output only 2022 data
#        outputSGP.directory = "Data/2019",                  # Save results to Data/2019/
        parallel.config = list(BACKEND="PARALLEL", WORKERS=list(TAUS=4))                   # Parallel processing config
)

###   Create baseline matrices
### Utility functions

convertToBaseline <- function(baseline_matrices) {
    tmp.list <- list()
    if (is.null(baseline_matrices)) {
        return(NULL)
    } else {
        for (i in names(baseline_matrices)) {
            for (j in seq_along(baseline_matrices[[i]])) {
                baseline_matrices[[i]][[j]]@Time <- list(rep("BASELINE", length(unlist(baseline_matrices[[i]][[j]]@Time))))
            }
            names(baseline_matrices[[i]]) <- sub("[.][1234]_", "_", names(baseline_matrices[[i]]))
        }

        tmp.content_areas <- unique(sapply(strsplit(names(baseline_matrices), "[.]"), '[', 1))
        for (i in tmp.content_areas) {
            tmp.list[[paste(i, "BASELINE", sep=".")]] <- unlist(baseline_matrices[grep(i, names(baseline_matrices))], recursive=FALSE)
        }
        return(tmp.list)
    }
}

### Create list of matrices
Hawaii_SGPt_Baseline_Matrices <- convertToBaseline(Hawaii_SGP@SGP$Coefficient_Matrices)

### Save matrices
save(Hawaii_SGPt_Baseline_Matrices, file="Data/Hawaii_SGPt_Baseline_Matrices_2019.Rdata")