Data analysis

All statistics were computed using sampling weights; standard errors based on balanced repeated replication weights were used for statistical significance and/or confidence intervals.

Analyses based on achievement test results (plausible values) were based on Rubin’s rule for multiply imputed variables.

The OECD average and PISA-D average corresponded to the arithmetic mean of the respective country estimates in each case.

Secondary analysis

The PISA data analysis manual provides SAS and SPSS codes for analyzing PISA microdata.

User-written software is also available in Stata (repest, pisatools) and R (intsvy).

Basic statistics can be computed with the PISA Data Explorer.

Types of data files


  • Student questionnaire data file
  • School questionnaire data file
  • Teacher questionnaire data file
  • Cognitive item data file



  • Individual respondent, parent/guardian, interviewer household observation questionnaire data files
  • Cognitive item data file
  • Questionnaire timing data file
Format(s) of data files
Item release policy

Most items are kept confidential for re-use in future PISA cycles in order to measure trends. A minority of items are released after each cycle to illustrate new frameworks and provide samples of tasks at different levels of proficiency.