Select confounders using a modified disjunctive cause criterion, controlling covariates measured prior to the treatment variable¶

Strategy · All strategies
Evidence · no claims cited

Description¶

To choose covariates for causal analysis of reciprocal effects, the article follows "Vanderweele (2019), who suggested a modified disjunctive cause criterion approach". This strategy includes variables that might cause self-concept, achievement, or both, excludes possible instruments, includes proxy variables for common causes, and controls covariates measured prior to the treatment variable. The article applied it to select five sets of covariates, from demographics and socioeconomic background to motivation, well-being, and personality traits, to help satisfy the strong ignorability assumption.

Design Implications¶

Context¶

Requirements¶

• Covariates must be measured prior to the treatment variable to mitigate mediator variables and collider bias.

Constraints¶

• The article notes that even though the modified disjunctive cause criterion is very helpful to address the challenge of confounder selection, collider bias cannot ultimately be ruled out.

Target Learners¶

• lower secondary school students

Target Learning Goals¶

• causal estimation of reciprocal self-concept and achievement effects

Related Strategies¶

• Match statistical control methods to variable structure and measurement quality, preferring stratification, covariate adjustment, matching, or regression discontinuity accordingly
• Assess robustness of non-experimental composition findings with impact thresholds for confounding variables

Examples¶

-

Key Sources¶

• Hübner, N., Wagner, W., Zitzmann, S., & Nagengast, B. (2023). How Strong Is the Evidence for a Causal Reciprocal Effect? Contrasting Traditional and New Methods to Investigate the Reciprocal Effects Model of Self-Concept and Achievement. Educational Psychology Review. https://doi.org/10.1007/s10648-023-09724-6