Experiments in daily life: When causal within-person effects do (not) translate into between-person differences

Area of Science:

• Psychological science
• Methodology
• Behavioral research

Background:

• Intensive longitudinal designs capture daily life psychological dynamics.
• Observational nature limits causal inference.
• Micro-randomized interventions offer a solution for within-person (Wp) causal effects.

Purpose of the Study:

• Investigate the translation of Wp causal effects to between-person (Bp) differences.
• Identify factors that may hinder this translation.
• Provide guidance for designing and analyzing micro-randomized experiments.

Main Methods:

• Analytical derivations and simulations to model Wp and Bp effects.
• Examination of counteracting forces: contextual effects, correlated random effects, and cross-level interactions.
• Empirical illustration using a 10-day microrandomized mindfulness intervention.

Main Results:

• Wp causal effects translate to Bp differences in the absence of counteracting forces.
• Contextual effects, correlated random effects, and cross-level interactions can disrupt this translation.
• Demonstrated translation and potential disruptions using simulated and empirical data.

Conclusions:

• Micro-randomized interventions within intensive longitudinal designs are valuable for causal inference.
• Understanding cross-level translation is crucial for interpreting findings.
• Recommendations provided for experimental design and statistical analysis.