Effects of Cognitive States on Momentary Response Biases in an Ecological Momentary Assessment Study
Journal
JMIR Preprints
Type
working paper
Date Issued
2025-10-10
Author(s)
Keller, Roman
;
Negi, Samarth
;
;
von Wangenheim, Florian
;
Chua, Xin Hui
;
Sim, Xueling
;
van Dam, Rob
;
Müller-Riemenschneider, Falk
;
Edney, Sarah Martine
;
Mair, Jaqueline Louise
Abstract
Background:
Ecological momentary assessments (EMAs) are a widely used method in health and psychological research to assess people’s thoughts, emotions, and behaviours in daily life. However, EMAs are affected by several response biases including compliance, , careless, and delayed responses, which can compromise data quality and threaten validity. Although some design and contextual factors influencing these biases have been identified, the role of momentary cognitive states (e.g., stress, fatigue, affect) remains underexplored.
Objective:
This study examined how cognitive states at the moment of responding influenced participants’ compliance, careless responding, delayed responding to EMAs.
Methods:
We conducted a secondary analysis using data from the COBRA study, a prospective observational cohort study on health behaviours in Singapore. The dataset included 29’797 EMA responses from 617 Singaporean adults who were prompted to complete six smartphone-based surveys daily for nine days. Using mixed-effects regression models, we examined how five cognitive states (stress, fatigue, hunger, positive affect, and sleep quality) predicted compliance, careless responding, and response delay, while adjusting for demographic and temporal covariates.
Results:
Overall compliance at the next prompt was high (92%), and no cognitive states were significantly associated with compliance. Careless responding was more likely at higher levels of stress and hunger, and less likely at higher levels of sleep quality, fatigue, and positive affect. Several demographic and temporal factors such as age, time of day, and number of missed prior prompts also significantly influenced response behaviors.
Conclusions:
This study demonstrates that cognitive states significantly influence careless and delayed responding in EMA studies, but they do not impact compliance. Our findings highlight the need to consider participants’ cognitive states as potential sources of systematic bias, particularly in studies investigating stress, hunger, and affective states. Future EMA research should therefore incorporate strategies to detect and mitigate response biases, such as tracking those variables and using just-in-time delivered warnings, rewards, and educational messages.
Ecological momentary assessments (EMAs) are a widely used method in health and psychological research to assess people’s thoughts, emotions, and behaviours in daily life. However, EMAs are affected by several response biases including compliance, , careless, and delayed responses, which can compromise data quality and threaten validity. Although some design and contextual factors influencing these biases have been identified, the role of momentary cognitive states (e.g., stress, fatigue, affect) remains underexplored.
Objective:
This study examined how cognitive states at the moment of responding influenced participants’ compliance, careless responding, delayed responding to EMAs.
Methods:
We conducted a secondary analysis using data from the COBRA study, a prospective observational cohort study on health behaviours in Singapore. The dataset included 29’797 EMA responses from 617 Singaporean adults who were prompted to complete six smartphone-based surveys daily for nine days. Using mixed-effects regression models, we examined how five cognitive states (stress, fatigue, hunger, positive affect, and sleep quality) predicted compliance, careless responding, and response delay, while adjusting for demographic and temporal covariates.
Results:
Overall compliance at the next prompt was high (92%), and no cognitive states were significantly associated with compliance. Careless responding was more likely at higher levels of stress and hunger, and less likely at higher levels of sleep quality, fatigue, and positive affect. Several demographic and temporal factors such as age, time of day, and number of missed prior prompts also significantly influenced response behaviors.
Conclusions:
This study demonstrates that cognitive states significantly influence careless and delayed responding in EMA studies, but they do not impact compliance. Our findings highlight the need to consider participants’ cognitive states as potential sources of systematic bias, particularly in studies investigating stress, hunger, and affective states. Future EMA research should therefore incorporate strategies to detect and mitigate response biases, such as tracking those variables and using just-in-time delivered warnings, rewards, and educational messages.
Language
English