1. Causal inference in educational research: Causal analysis in cross-sectional observational studies.
- Author
-
Martínez-Abad, Fernando and León, Jaime
- Subjects
- *
CAUSAL inference , *REGRESSION discontinuity design , *EDUCATION research , *PANEL analysis , *CROSS-sectional method , *EARLY childhood education , *PROPENSITY score matching , *INFERENCE (Logic) - Abstract
The assumption of cause-effect relationships in ex post facto research is a widely known issue in the field of research methods in social sciences. To address this important limitation, the use of causal inference techniques has become widespread in recent years. Causal inference establishes a set of statistical procedures for drawing causal conclusions in non-experimental research. Despite its wide popularity and diffusion in the social and health sciences, its use in educational research is still marginal. Thus, this paper introduces the main causal inference techniques available to the educational researcher when observational panel data are available. After addressing the key features and potential of propensity score matching, instrumental variables, and regression discontinuity design, we present an example application of each of these techniques. We used the available databases from the PISA 2018 assessments. We included the mathematical competence as the dependent variable in all the three models implemented. Given the different characteristics of each of these techniques, the independent variable used is different in the three models applied: attendance to early childhood education in propensity score matching; student academic expectations in instrumental variables; and size of the community in which the school is located in regression discontinuity design. The article concludes by discussing the potential of this set of techniques, taking into account the needs and methodological procedures most commonly applied in educational research. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF