Methods paper · 2015
Micro-Randomized Trials: An Experimental Design for Developing Just-in-Time Adaptive Interventions
Klasnja P, Hekler EB, Shiffman S, Boruvka A, Almirall D, Tewari A, Murphy SA
What it found
Introduces the micro-randomized trial: at each decision point, a person is randomly given or not given a component of an intervention. That lets researchers estimate the effect of the component in the moments right after it, and how that effect changes with time and context.
Where the evidence stops
A design for running experiments, not a finding about any behaviour.
How it shapes Pursue
Pursue
The way Pursue's own suggestions should eventually be tested: moment by moment, against doing nothing.
A design input, not evidence that Pursue works.