WIDA_CO_SGP_2024.R

################################################################################
###                                                                          ###
###    Calculate 2024 Student Growth Percentiles for Colorado WIDA/ACCESS    ###
###                                                                          ###
################################################################################

###   Load required packages
require(SGP)
require(data.table)

###   Load Data
load("Data/WIDA_CO_SGP.Rdata")
load("Data/WIDA_CO_Data_LONG_2024.Rdata")

###   Preserve basic CU/KU variables from previous analyses
setnames(WIDA_CO_SGP@Data,
         c("CATCH_UP_KEEP_UP_STATUS_3_YEAR",
           "CATCH_UP_KEEP_UP_STATUS_BASELINE_3_YEAR",
           "CATCH_UP_KEEP_UP_INTERIM_STATUS"),
         c("CUKU_ORIG", "CUKU_BASELINE_ORIG", "CUPKUP_IS")
)


###   Add baseline matrices
SGPstateData <- SGPmatrices::addBaselineMatrices("WIDA_CO", "2021")

###   Run updateSGP to produce cohort referrenced SGPs, etc.
WIDA_CO_SGP <-
    updateSGP(
        what_sgp_object = WIDA_CO_SGP,
        with_sgp_data_LONG = WIDA_CO_Data_LONG_2024,
        steps = c("prepareSGP", "analyzeSGP", "combineSGP", "summarizeSGP"),
        sgp.percentiles = TRUE,
        sgp.projections = TRUE,
        sgp.projections.lagged = TRUE,
        sgp.percentiles.baseline = TRUE,
        sgp.projections.baseline = TRUE,
        sgp.projections.lagged.baseline = TRUE,
        sgp.percentiles.equated = FALSE,
        sgp.target.scale.scores = FALSE,  #  Run below
        save.intermediate.results = FALSE,
        parallel.config = list(
            BACKEND = "PARALLEL",
            WORKERS = list(
                PERCENTILES = 10, ## & BASELINE_PERCENTILES
                PROJECTIONS = 10,
                LAGGED_PROJECTIONS = 8
            )
        )
    )

###   Move CUPKUP to "ORIG" vars
##    (only current year gets merged due to @SGP$SGProjections renaming (?))
# table(WIDA_CO_SGP@Data[, is.na(CATCH_UP_KEEP_UP_STATUS_3_YEAR), YEAR], exclude = NULL)

WIDA_CO_SGP@Data[
    !is.na(CATCH_UP_KEEP_UP_STATUS_3_YEAR),
    CUKU_ORIG := CATCH_UP_KEEP_UP_STATUS_3_YEAR
][, CATCH_UP_KEEP_UP_STATUS_3_YEAR := NULL
]

WIDA_CO_SGP@Data[
    !is.na(CATCH_UP_KEEP_UP_STATUS_BASELINE_3_YEAR),
    CUKU_BASELINE_ORIG := CATCH_UP_KEEP_UP_STATUS_BASELINE_3_YEAR
][, CATCH_UP_KEEP_UP_STATUS_BASELINE_3_YEAR := NULL
]

###  Run combineSGP for Interim Target Levels (for both 2022 and 2024)
`%w/o%` <- function(x, y) x[!x %in% y]

##    Preserve previously calculated "interim" variables
# setnames(WIDA_CO_SGP@Data, "CATCH_UP_KEEP_UP_INTERIM_STATUS", "CUPKUP_IS")
interim.targets <- c("SGP_TARGET_INTERIM", "SGP_TARGET_INTERIM_CURRENT")

###   Clean Up pre-combined data
simple.targets <- c("SGP_TARGET_3_YEAR_CURRENT", "SGP_TARGET_3_YEAR", 
  "SGP_TARGET_BASELINE_3_YEAR_CURRENT", "SGP_TARGET_BASELINE_3_YEAR")

WIDA_CO_SGP@Data[, (simple.targets) := NULL]

SGPstateData[["WIDA_CO"]][["Growth"]][["System_Type"]] <- "Cohort Referenced"
SGPstateData[["WIDA_CO"]][["SGP_Configuration"]][['current.year.lagged.target']] <- FALSE


##    L1/L2
SGPstateData[["WIDA_CO"]][["Achievement"]][["Levels"]][["Proficient"]] <-
    c("Not Proficient",
      "Proficient", "Proficient", "Proficient", "Proficient", "Proficient", NA)

WIDA_CO_SGP <-
    combineSGP(
        WIDA_CO_SGP,
        years = "2024",
        sgp.percentiles = FALSE,
        sgp.percentiles.baseline = FALSE,
        sgp.projections.baseline = FALSE,
        sgp.projections.lagged.baseline = FALSE,
        sgp.target.scale.scores = TRUE,
        sgp.target.scale.scores.merge = "all_years_lagged_current",
        max.sgp.target.years.forward = 0:1,
        parallel.config = list(
            BACKEND = "FOREACH",
            TYPE = "doParallel",
            WORKERS = list(SGP_SCALE_SCORE_TARGETS = 8)
        )
    )

tmp.target.names <-
    grep("SGP_TARGET", names(WIDA_CO_SGP@Data), value = TRUE) %w/o% interim.targets
for (tgt in tmp.target.names) {
    new.tgt <- gsub("SGP_TARGET", "SGP_LEVEL_1_TARGET", tgt)
    WIDA_CO_SGP@Data[
        !is.na(get(tgt)),
        (new.tgt) := get(tgt)
    ][, (tgt) := NULL
    ]
}
for (cuku in grep("CATCH_UP_KEEP_UP", names(WIDA_CO_SGP@Data), value = TRUE)) {
    new.cuku <- gsub("CATCH_UP_KEEP_UP", "L1L2_CUKU", cuku)
    WIDA_CO_SGP@Data[
        !is.na(get(cuku)),
        (new.cuku) := get(cuku)
    ][, (cuku) := NULL
    ]
}

target.ss.index <-
    grep("2024.*SCALE_SCORES",
         names(WIDA_CO_SGP@SGP[["SGProjections"]]), perl = TRUE)
names(WIDA_CO_SGP@SGP[["SGProjections"]])[target.ss.index] <-
    gsub("SCALE_SCORES", "SS_L1L2", 
         names(WIDA_CO_SGP@SGP[["SGProjections"]])[target.ss.index])


##    L2/L3
SGPstateData[["WIDA_CO"]][["Achievement"]][["Levels"]][["Proficient"]] <-
    c("Not Proficient", "Not Proficient",
      "Proficient", "Proficient", "Proficient", "Proficient", NA)

WIDA_CO_SGP <-
    combineSGP(
        WIDA_CO_SGP,
        years = "2024",
        sgp.percentiles.baseline=FALSE,
        sgp.projections.baseline = FALSE,
        sgp.projections.lagged.baseline = FALSE,
        sgp.target.scale.scores = TRUE,
        sgp.target.scale.scores.merge = "all_years_lagged_current",
        max.sgp.target.years.forward = 0:2,
        parallel.config = list(
            BACKEND = "FOREACH",
            TYPE = "doParallel",
            WORKERS = list(SGP_SCALE_SCORE_TARGETS = 8)
        )
    )

tmp.target.names <-
    grep("SGP_TARGET", names(WIDA_CO_SGP@Data), value = TRUE) %w/o% interim.targets
for (tgt in tmp.target.names) {
    new.tgt <- gsub("SGP_TARGET", "SGP_LEVEL_2_TARGET", tgt)
    WIDA_CO_SGP@Data[
        !is.na(get(tgt)),
        (new.tgt) := get(tgt)
    ][, (tgt) := NULL
    ]
}
for (cuku in grep("CATCH_UP_KEEP_UP", names(WIDA_CO_SGP@Data), value = TRUE)) {
    new.cuku <- gsub("CATCH_UP_KEEP_UP", "L2L3_CUKU", cuku)
    WIDA_CO_SGP@Data[
        !is.na(get(cuku)),
        (new.cuku) := get(cuku)
    ][, (cuku) := NULL
    ]
}

target.ss.index <-
    grep("2024.*SCALE_SCORES",
         names(WIDA_CO_SGP@SGP[["SGProjections"]]), perl = TRUE)
names(WIDA_CO_SGP@SGP[["SGProjections"]])[target.ss.index] <-
    gsub("SCALE_SCORES", "SS_L2L3",
         names(WIDA_CO_SGP@SGP[["SGProjections"]])[target.ss.index])


##    L4/L5 (Stay at above L5 - do before Official/L4)
SGPstateData[["WIDA_CO"]][["Achievement"]][["Levels"]][["Proficient"]] <-
    c("Not Proficient", "Not Proficient", "Not Proficient", "Not Proficient",
      "Proficient", "Proficient", NA)

WIDA_CO_SGP <-
    combineSGP(
        WIDA_CO_SGP,
        years = "2024",
        sgp.percentiles.baseline=FALSE,
        sgp.projections.baseline = FALSE,
        sgp.projections.lagged.baseline = FALSE,
        sgp.target.scale.scores = TRUE,
        sgp.target.scale.scores.merge = "all_years_lagged_current",
        max.sgp.target.years.forward = 0:1,
        parallel.config = list(
            BACKEND = "FOREACH",
            TYPE = "doParallel",
            WORKERS = list(SGP_SCALE_SCORE_TARGETS = 8)
        )
    )

tmp.target.names <-
    grep("SGP_TARGET", names(WIDA_CO_SGP@Data), value = TRUE) %w/o% interim.targets
for (tgt in tmp.target.names) {
    new.tgt <- gsub("SGP_TARGET", "SGP_LEVEL_4_TARGET", tgt)
    WIDA_CO_SGP@Data[
        !is.na(get(tgt)),
        (new.tgt) := get(tgt)
    ][, (tgt) := NULL
    ]
}
for (cuku in grep("CATCH_UP_KEEP_UP", names(WIDA_CO_SGP@Data), value = TRUE)) {
    new.cuku <- gsub("CATCH_UP_KEEP_UP", "L4L5_CUKU", cuku)
    WIDA_CO_SGP@Data[
        !is.na(get(cuku)),
        (new.cuku) := get(cuku)
    ][, (cuku) := NULL
    ]
}

target.ss.index <-
    grep("2024.*SCALE_SCORES",
         names(WIDA_CO_SGP@SGP[["SGProjections"]]), perl = TRUE)
names(WIDA_CO_SGP@SGP[["SGProjections"]])[target.ss.index] <-
    gsub("SCALE_SCORES", "SS_L4L5",
         names(WIDA_CO_SGP@SGP[["SGProjections"]])[target.ss.index])


##    L3/L4 (Official)
SGPstateData[["WIDA_CO"]][["Achievement"]][["Levels"]][["Proficient"]] <-
    c("Not Proficient", "Not Proficient", "Not Proficient",
      "Proficient", "Proficient", "Proficient", NA)

WIDA_CO_SGP <-
    combineSGP(
        WIDA_CO_SGP,
        years = "2024",
        sgp.percentiles.baseline=FALSE,
        sgp.projections.baseline = FALSE,
        sgp.projections.lagged.baseline = FALSE,
        sgp.target.scale.scores = TRUE,
        sgp.target.scale.scores.merge = "all_years_lagged_current",
        max.sgp.target.years.forward = 0:3,
        parallel.config = list(
            BACKEND = "FOREACH",
            TYPE = "doParallel",
            WORKERS = list(SGP_SCALE_SCORE_TARGETS = 8)
        )
    )

tmp.target.names <-
    grep("SGP_TARGET", names(WIDA_CO_SGP@Data), value = TRUE) %w/o% interim.targets
for (tgt in tmp.target.names) {
    new.tgt <- gsub("SGP_TARGET", "SGP_LEVEL_3_TARGET", tgt)
    WIDA_CO_SGP@Data[
        !is.na(get(tgt)),
        (new.tgt) := get(tgt)
    ][, (tgt) := NULL
    ]
}
for (cuku in grep("CATCH_UP_KEEP_UP", names(WIDA_CO_SGP@Data), value = TRUE)) {
    new.cuku <- gsub("CATCH_UP_KEEP_UP", "L3L4_CUKU", cuku)
    WIDA_CO_SGP@Data[
        !is.na(get(cuku)),
        (new.cuku) := get(cuku)
    ][, (cuku) := NULL
    ]
}

##   Rename generic "CATCH_UP_KEEP_UP_STATUS" for all students as "Official"
##   Also rename Interim Status var(s)
setnames(WIDA_CO_SGP@Data,
         c("CUKU_ORIG", "CUKU_BASELINE_ORIG", "CUPKUP_IS"),
         c("CATCH_UP_KEEP_UP_STATUS_3_YEAR",
           "CATCH_UP_KEEP_UP_STATUS_BASELINE_3_YEAR",
           "CATCH_UP_KEEP_UP_INTERIM_STATUS")
)


#+ include = FALSE, purl = FALSE, eval = FALSE
###   Add R session Info & Save results (`cfaDocs` version 0.0-1.12 or later)
source(
    system.file(
        "rmarkdown", "shared_resources", "rmd", "R_Session_Info.R",
        package = "cfaDocs"
    )
)
WIDA_CO_SGP@Version$session_platform <- list("2024" = session_platform)
WIDA_CO_SGP@Version$attached_pkgs <- list("2024" = attached_pkgs)
WIDA_CO_SGP@Version$namespace_pkgs<-  list("2024" = namespace_pkgs)

###   Output/Save Results
outputSGP(WIDA_CO_SGP)

save(WIDA_CO_SGP, file = "Data/WIDA_CO_SGP.Rdata")