Revising free text inputs in physical activity self-report methods: Lessons from the ACT24
Principal Investigator
Name
Robert Korycinski
Degrees
B.A.
Institution
National Cancer Institute
Position Title
Scientific Program Analyst
Email
About this CDAS Project
Study
IDATA
(Learn more about this study)
Project ID
IDATA-20
Initial CDAS Request Approval
Jan 19, 2018
Title
Revising free text inputs in physical activity self-report methods: Lessons from the ACT24
Summary
Self-report instruments can convey rich physical activity behavioral and contextual data that is typically unavailable from accelerometers and other device-based instruments. The ACT24 self-report recall provides this data by having users select from a list of activity categories, followed by specific activities unique to each category, to fill in their physical activity across the 24-hour day. Each activity is then assigned a Metabolic Equivalent of Task (MET) value from the 2000 Compendium of Physical Activities.
To allow for relatively uncommon activities, the ACT24 allows users to submit free text that describes the behavior; in some cases, users may enter free text for otherwise listed activities when they cannot locate the activity under a particular category. Researchers may recode free text entries to standardize across all reported data, but such revisions often require substantial time and effort and have an unclear impact on study results.
Thus, the purpose of this project is to investigate the change in MET minutes reported by the ACT24 after recoding and assigning free text entries to either an ACT24-listed or new activity. Given that the ACT24 platform uses the 2000 Compendium of Physical Activities, this project will also investigate the change in MET minutes after updating all activity MET values to correspond with the 2011 Compendium update. The findings may offer insights into the utility of reviewing free text entries in physical activity self-report instruments, as well as the value of including a free text option when designing instruments.
To allow for relatively uncommon activities, the ACT24 allows users to submit free text that describes the behavior; in some cases, users may enter free text for otherwise listed activities when they cannot locate the activity under a particular category. Researchers may recode free text entries to standardize across all reported data, but such revisions often require substantial time and effort and have an unclear impact on study results.
Thus, the purpose of this project is to investigate the change in MET minutes reported by the ACT24 after recoding and assigning free text entries to either an ACT24-listed or new activity. Given that the ACT24 platform uses the 2000 Compendium of Physical Activities, this project will also investigate the change in MET minutes after updating all activity MET values to correspond with the 2011 Compendium update. The findings may offer insights into the utility of reviewing free text entries in physical activity self-report instruments, as well as the value of including a free text option when designing instruments.
Aims
1) Determine the change in ACT24-reported MET minutes after a) recoding and reclassifying free text activity entries, and b) assigning MET values from the 2011 Compendium of Physical Activities to all activity entries. We will calculate changes in both total and activity-specific MET minutes.
2) Investigate whether participant free text use is associated with demographics (e.g., age, gender, BMI), self-reported activity level, or ACT24 user experience level.
Collaborators
Kelley Pettee Gabriel, Ph.D. - University of Texas Health Science Center at Houston
Heather Bowles, Ph.D. - National Cancer Institute