Ask Less, Learn More: Adapting Ecological Momentary Assessment Survey Length by Modeling Question-Answer Information...

Area of Science:

• Psychological Science
• Computer Science
• Health Informatics

Background:

• Ecological momentary assessment (EMA) collects real-time, self-reported data via mobile devices in natural environments.
• High participant burden from lengthy EMA surveys can compromise data quality and compliance.
• Current EMA methods often struggle to balance data richness with participant engagement.

Purpose of the Study:

• To develop and evaluate a novel method for reducing EMA survey length while minimizing information loss.
• To enhance participant compliance and data quality in longitudinal EMA studies.
• To enable more comprehensive data collection or reduce participant burden in EMA.

Main Methods:

• Proposed a question-answer prediction model to estimate the information gain of each EMA question.
• Prioritized informative questions and skipped uninformative ones in a sequential, question-by-question approach.
• Evaluated the method using simulated question omission on four real-world EMA datasets.

Main Results:

• The proposed method reduced imputation errors by 15%-52% compared to random omission (50% skipped).
• Achieved a 34%-56% reduction in mean survey length for surveys with five answer options.
• Demonstrated real-time prediction accuracy of 72%-95% for skipped questions.

Conclusions:

• The intelligent question-skipping method significantly enhances EMA efficiency and data integrity.
• This approach offers a viable solution for reducing participant burden in longitudinal studies.
• Facilitates sustainable EMA data collection by balancing comprehensiveness and user experience.