PARCC Data Analysis 2016

Spring 2016 PARCC Consortium SGP Analyses

The following code outlines the code used to conduct the Spring 2016 analyses. Note that unlike the individual state analyses that this script begins with an SGP object established earlier and uses the updateSGP function to add the Spring 2016 data and conduct those analyses.

#################################################################################
###                                                                           ###
###      SGP analysis script for PARCC consortium - Spring 2016 Analyses      ###
###                                                                           ###
#################################################################################

workers <- parallel::detectCores()/2

### Load Packages

require(SGP)
require(RSQLite)


### Load Data & configurations

load("./Data/PARCC_SGP.Rdata")
parcc.db <- "./Data/PARCC_Data_LONG.sqlite"


###  Read in the Spring 2016 configuration code and combine into a single list.

source("../SGP_CONFIG/2015_2016.2/ELA.R")
source("../SGP_CONFIG/2015_2016.2/ELA_SS.R")
source("../SGP_CONFIG/2015_2016.2/MATHEMATICS.R")
source("../SGP_CONFIG/2015_2016.2/MATHEMATICS_SS.R")


PARCC_2015_2016.2.config <- c(
	ELA_2015_2016.2.config,
	ELA_SS_2015_2016.2.config,

	MATHEMATICS_2015_2016.config,
	MATHEMATICS_SS_2015_2016.config,

	ALGEBRA_I.2015_2016.config,
	ALGEBRA_I_SS.2015_2016.config,
	ALGEBRA_II.2015_2016.config,
	ALGEBRA_II_SS.2015_2016.config,
	GEOMETRY.2015_2016.config,
	GEOMETRY_SS.2015_2016.config,

	INTEGRATED_MATH_1.2015_2016.config,
	INTEGRATED_MATH_1_SS.2015_2016.config,
	INTEGRATED_MATH_2.2015_2016.config,
	INTEGRATED_MATH_2_SS.2015_2016.config,
	INTEGRATED_MATH_3.2015_2016.config,
	INTEGRATED_MATH_3_SS.2015_2016.config)


### updateSGP

PARCC_SGP <- updateSGP(
		what_sgp_object=PARCC_SGP,
		# with_sgp_data_LONG=PARCC_Data_LONG_2016,
		with_sgp_data_LONG = dbGetQuery(dbConnect(SQLite(), dbname = parcc.db), "select * from PARCC_Data_LONG_2016_2"),
		sgp.config = PARCC_2015_2016.2.config,
		steps=c("prepareSGP", "analyzeSGP"),
		sgp.percentiles = TRUE,
		sgp.projections = FALSE,
		sgp.projections.lagged = FALSE,
		sgp.percentiles.baseline=FALSE,
		sgp.projections.baseline = FALSE,
		sgp.projections.lagged.baseline = FALSE,
		sgp.percentiles.equated = FALSE,
		simulate.sgps = TRUE,
		calculate.simex= TRUE,
		save.intermediate.results=FALSE,
		outputSGP.output.type=c("LONG_Data", "LONG_FINAL_YEAR_Data"),
		parallel.config=list(BACKEND="FOREACH", TYPE="doParallel", WORKERS=list(TAUS = workers, SIMEX = workers)))


### analyzeSGP (for student growth projections)

PARCC_SGP <- analyzeSGP(
		PARCC_SGP,
		sgp.config=PARCC_2015_2016.2.config,
		sgp.percentiles=FALSE,
		sgp.projections=TRUE,
		sgp.projections.lagged=TRUE,
		sgp.percentiles.baseline=FALSE,
		sgp.projections.baseline=FALSE,
		sgp.projections.lagged.baseline=FALSE,
		parallel.config = list(BACKEND="FOREACH", TYPE="doParallel", SNOW_TEST=TRUE, WORKERS=list(PROJECTIONS = workers, LAGGED_PROJECTIONS = workers)))


### combineSGP

PARCC_SGP <- combineSGP(
		PARCC_SGP,
		sgp.target.scale.scores=TRUE,
		sgp.config=PARCC_2015_2016.2.config,
		parallel.config = ilist(BACKEND="FOREACH", TYPE="doParallel", SNOW_TEST=TRUE, WORKERS=list(SGP_SCALE_SCORE_TARGETS = workers)))

### Save results

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


### outputSGP

outputSGP(PARCC_SGP)