Title: PISA Data Analysis Leveraging R: Pros and Cons

Author: Mariia Mazorchuk, Anna Bychko

Affiliation: Ukrainian Center for Educational Quality Assessment

Abstract: 
In 2018, for the first time Ukraine participated in the PISA program that aims
to evaluate the quality of education system and find a way to improve it. The
team of data analysts from the Ukrainian center of educational quality
assessment preprocessed and analyzed the PISA data and announced the
assessment results in 2019.

The team worked with a large scale assessment and that task requires specific
tools. Thus, our software team utilized the following tools such as SPSS
(http://www.spss.com.hk/software/statistics/), IDB Analyzer
(https://www.iea.nl/data-tools/tools#section-308) and PISA Data Explorer
(https://pisadataexplorer.oecd.org/ide/idepisa/).  For most operations, we
utilized R software, in particular, the intsvy package
(https://cran.r-project.org/web/packages/intsvy/intsvy.pdf) that provides for
calculations descriptive statistics, regression and correlation coefficients,
frequency tables for PISA data (including the complex sample design with
replicate weights) and rotated test forms with plausible values of achievement
scores. To visualize the data we leveraged the intsvy package and some
additional packages such as ggplot. 

During the calculation process, we encountered some issues such as the
inability to process the PISA dataset that includes all countries and to
calculate the complex estimates. To solve the issues with PISA data analysis
we leveraged R functions and since the packages of R Software Environment were
in open access, we performed the necessary calculations of statistical
characteristics. Here we present the project's results and share the
experience of processing PISA data with R functions.