Program for International Student Assessment (PISA)

Build Status Build status

The authoritative source for evaluating educational achievement across nations, the Program(me) for International Student Assessment ranks the math, science, and reading skills of high school students across the developed world.

  • A large table with one row per student, a smaller table with one row per school, then multiple (optional) tables such as one row per parent or per teacher.

  • A complex sample survey designed to generalize to 15-year-old schoolchildren in more than sixty countries.

  • Released triennially since 2000.

  • Administered by the OECD.

Simplified Download and Importation

The R lodown package easily downloads and imports all available PISA microdata by simply specifying "pisa" with an output_dir = parameter in the lodown() function. Depending on your internet connection and computer processing speed, you might prefer to run this step overnight.

library(lodown)
lodown( "pisa" , output_dir = file.path( path.expand( "~" ) , "PISA" ) )

lodown also provides a catalog of available microdata extracts with the get_catalog() function. After requesting the PISA catalog, you could pass a subsetted catalog through the lodown() function in order to download and import specific extracts (rather than all available extracts).

library(lodown)
# examine all available PISA microdata files
pisa_cat <-
    get_catalog( "pisa" ,
        output_dir = file.path( path.expand( "~" ) , "PISA" ) )

# 2015 only
pisa_cat <- subset( pisa_cat , year == 2015 )
# download the microdata to your local computer
pisa_cat <- lodown( "pisa" , pisa_cat )

Analysis Examples with the survey library  

Construct a multiply-imputed, database-backed complex sample survey design:

library(DBI)
library(RSQLite)
library(survey)
library(mitools)

pisa_design <- readRDS( file.path( path.expand( "~" ) , "PISA" , "2015 cmb_stu_qqq design.rds" ) )

pisa_design <- lodown:::svyMDBdesign( pisa_design )

Variable Recoding

Add new columns to the data set:

pisa_design <- 
    update( 
        pisa_design , 
        
        gender = factor( st004d01t , labels = c( "male" , "female" ) ) ,
        
        how_many_computers_at_home = 
            factor( 
                st012q06na , 
                labels = c( "none" , "one" , "two" , "three or more" ) 
            )
 
    )

Unweighted Counts

Count the unweighted number of records in the survey sample, overall and by groups:

MIcombine( with( pisa_design , svyby( ~ one , ~ one , unwtd.count ) ) )

MIcombine( with( pisa_design , svyby( ~ one , ~ gender , unwtd.count ) ) )

Weighted Counts

Count the weighted size of the generalizable population, overall and by groups:

MIcombine( with( pisa_design , svytotal( ~ one ) ) )

MIcombine( with( pisa_design ,
    svyby( ~ one , ~ gender , svytotal )
) )

Descriptive Statistics

Calculate the mean (average) of a linear variable, overall and by groups:

MIcombine( with( pisa_design , svymean( ~ scie ) ) )

MIcombine( with( pisa_design ,
    svyby( ~ scie , ~ gender , svymean )
) )

Calculate the distribution of a categorical variable, overall and by groups:

MIcombine( with( pisa_design , svymean( ~ how_many_computers_at_home ) ) )

MIcombine( with( pisa_design ,
    svyby( ~ how_many_computers_at_home , ~ gender , svymean )
) )

Calculate the sum of a linear variable, overall and by groups:

MIcombine( with( pisa_design , svytotal( ~ scie ) ) )

MIcombine( with( pisa_design ,
    svyby( ~ scie , ~ gender , svytotal )
) )

Calculate the weighted sum of a categorical variable, overall and by groups:

MIcombine( with( pisa_design , svytotal( ~ how_many_computers_at_home ) ) )

MIcombine( with( pisa_design ,
    svyby( ~ how_many_computers_at_home , ~ gender , svytotal )
) )

Calculate the median (50th percentile) of a linear variable, overall and by groups:

MIcombine( with( pisa_design ,
    svyquantile(
        ~ scie ,
        0.5 , se = TRUE 
) ) )

MIcombine( with( pisa_design ,
    svyby(
        ~ scie , ~ gender , svyquantile ,
        0.5 , se = TRUE ,
        keep.var = TRUE , ci = TRUE 
) ) )

Estimate a ratio:

MIcombine( with( pisa_design ,
    svyratio( numerator = ~ math , denominator = ~ reading )
) )

Subsetting

Restrict the survey design to Albania:

sub_pisa_design <- subset( pisa_design , cnt == "ALB" )

Calculate the mean (average) of this subset:

MIcombine( with( sub_pisa_design , svymean( ~ scie ) ) )

Measures of Uncertainty

Extract the coefficient, standard error, confidence interval, and coefficient of variation from any descriptive statistics function result, overall and by groups:

this_result <-
    MIcombine( with( pisa_design ,
        svymean( ~ scie )
    ) )

coef( this_result )
SE( this_result )
confint( this_result )
cv( this_result )

grouped_result <-
    MIcombine( with( pisa_design ,
        svyby( ~ scie , ~ gender , svymean )
    ) )

coef( grouped_result )
SE( grouped_result )
confint( grouped_result )
cv( grouped_result )

Calculate the degrees of freedom of any survey design object:

degf( pisa_design$designs[[1]] )

Calculate the complex sample survey-adjusted variance of any statistic:

MIcombine( with( pisa_design , svyvar( ~ scie ) ) )

Include the complex sample design effect in the result for a specific statistic:

# SRS without replacement
MIcombine( with( pisa_design ,
    svymean( ~ scie , deff = TRUE )
) )

# SRS with replacement
MIcombine( with( pisa_design ,
    svymean( ~ scie , deff = "replace" )
) )

Compute confidence intervals for proportions using methods that may be more accurate near 0 and 1. See ?svyciprop for alternatives:

MIsvyciprop( ~ oecd , pisa_design ,
    method = "likelihood" )

Regression Models and Tests of Association

Perform a design-based t-test:

MIsvyttest( scie ~ oecd , pisa_design )

Perform a chi-squared test of association for survey data:

MIsvychisq( ~ oecd + how_many_computers_at_home , pisa_design )

Perform a survey-weighted generalized linear model:

glm_result <- 
    MIcombine( with( pisa_design ,
        svyglm( scie ~ oecd + how_many_computers_at_home )
    ) )
    
summary( glm_result )

Replication Example