An editoral paper about sensing special issue stating the future direction of personality in the age of multi-modal sensing. Has some really good framing points about the overall field / directions. Also some good papers referenced.
Core Observations
Need stringent standards of rigorous psychometric testing to algorithm-based trait or state scores extracted from sensed behavioural and big data.
Move away from convergent validity, focus on nomological validity
Clarity on what we are measuring or approximating and what to do with those measurement
Go beyond converging to a self-reported score. Otherwise why are we measuring those?
We should sample personality traits via other data sources too, and try to extract common variance to tap the "heart of the trait".
Disentangling description, explanation and prediction
Conceptual precision, formalization, and theory
multidisciplinarity, collaboration, and
focusing on ‘psychological’ aspects