- All statistics were computed using sampling weights.
- Estimation of standard errors for significance testing was computed by making use of replication methods (jackknife repeated replication, jackknife zone, and replicate weights) to estimate sampling variance.
- REDS data are freely available.
- REDS is an observational, non-experimental study that collected cross-sectional data. Therefore, causal inferences cannot and should not be established with REDS data.
- Information collected in REDS was self-reported by school principals, teachers, and students. Furthermore, because population features were not observed but estimated using sample data, wording such as “the estimated proportion of students with X is …” is preferable to writing “X percent of students are …”.
- Nearly all variables in REDS are categorical in nature (nominal or ordered). Analysts may therefore need to consider using categorical, nonparametric analysis methods for these types of variables.
- Techniques for continuous variables (provided that the required assumptions hold) should only be used on counts and on the derived scales obtained through data reduction or scaling methods such as factor analysis, structural equation modelling, or item response theory.
- For doing secondary analyses, data can be analyzed with the IEA IDB Analyzer. This software takes the complex data structure automatically into account by using sampling weights for accurate estimation of population features, and by applying the JRR method for accurate estimation of standard errors.
- Of note, school data from Denmark; teacher data from Burkina Faso, Denmark, Ethiopia, Kenya, and Uruguay; and student data from Burkina Faso, Denmark, Ethiopia, and Kenya may not be representative of their respective target populations—an aspect that should be noted in any analysis. Use, for example, language such as “Responding students in Denmark…” rather than “Students in Denmark…” when interpreting findings from these countries.
Types of data files
- Student questionnaire files
- Teacher questionnaire files
- School questionnaire files
- National questionnaire file
Item release policy
Responses to all items used in the study are included in the database. However, some of the items are made available for restricted use by the public. The remaining items are kept secure, thus ensuring the possibility of measuring trends over time. Access to the restricted-use items is subject to approval by IEA through its study data repository.