Dropout in a Longitudinal Survey of Amazon Mechanical Turk Workers With Low Back Pain: Observational Study
BackgroundSurveys of internet panels such as Amazon’s Mechanical Turk (MTurk) are common in health research. Nonresponse in longitudinal studies can limit inferences about change over time. ObjectiveThis study aimed to (1) describe the patterns of survey responses...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
JMIR Publications
2024-11-01
|
| Series: | Interactive Journal of Medical Research |
| Online Access: | https://www.i-jmr.org/2024/1/e58771 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1846170510465957888 |
|---|---|
| author | Nabeel Qureshi Ron D Hays Patricia M Herman |
| author_facet | Nabeel Qureshi Ron D Hays Patricia M Herman |
| author_sort | Nabeel Qureshi |
| collection | DOAJ |
| description |
BackgroundSurveys of internet panels such as Amazon’s Mechanical Turk (MTurk) are common in health research. Nonresponse in longitudinal studies can limit inferences about change over time.
ObjectiveThis study aimed to (1) describe the patterns of survey responses and nonresponse among MTurk members with back pain, (2) identify factors associated with survey response over time, (3) assess the impact of nonresponse on sample characteristics, and (4) assess how well inverse probability weighting can account for differences in sample composition.
MethodsWe surveyed adult MTurk workers who identified as having back pain. We report participation trends over 3 survey waves and use stepwise logistic regression to identify factors related to survey participation in successive waves.
ResultsA total of 1678 adults participated in wave 1. Of those, 983 (59%) participated in wave 2 and 703 (42%) in wave 3. Participants who did not drop out took less time to complete previous surveys (30 min vs 35 min in wave 1, P<.001; 24 min vs 26 min in wave 2, P=.02) and reported having fewer health conditions (5.88 vs 6.6, P<.001). In multivariate models predicting responding at wave 2, lower odds of participation were associated with more time to complete the baseline survey (odds ratio [OR] 0.98, 95% CI 0.97-0.99), being Hispanic (compared with non-Hispanic, OR 0.69, 95% CI 0.49-0.96), having a bachelor’s degree as their terminal degree (compared with all other levels of education, OR 0.58, 95% CI 0.46-0.73), having more pain interference and intensity (OR 0.75, 95% CI 0.64-0.89), and having more health conditions. In contrast, older respondents (older than 45 years age compared with 18-24 years age) were more likely to respond to the wave 2 survey (OR 2.63 and 3.79, respectively) and those whose marital status was divorced (OR 1.81) and separated (OR 1.77) were also more likely to respond to the wave 2 survey. Weighted analysis showed slight differences in sample demographics and conditions and larger differences in pain assessments, particularly for those who responded to wave 2.
ConclusionsLongitudinal studies on MTurk have large, differential dropouts between waves. This study provided information about the individuals more likely to drop out over time, which can help researchers prepare for future surveys. |
| format | Article |
| id | doaj-art-2ca6b618de354376b0b863d324784f5f |
| institution | Kabale University |
| issn | 1929-073X |
| language | English |
| publishDate | 2024-11-01 |
| publisher | JMIR Publications |
| record_format | Article |
| series | Interactive Journal of Medical Research |
| spelling | doaj-art-2ca6b618de354376b0b863d324784f5f2024-11-11T15:45:32ZengJMIR PublicationsInteractive Journal of Medical Research1929-073X2024-11-0113e5877110.2196/58771Dropout in a Longitudinal Survey of Amazon Mechanical Turk Workers With Low Back Pain: Observational StudyNabeel Qureshihttps://orcid.org/0000-0001-7782-8023Ron D Hayshttps://orcid.org/0000-0001-6697-907XPatricia M Hermanhttps://orcid.org/0000-0001-5579-5654 BackgroundSurveys of internet panels such as Amazon’s Mechanical Turk (MTurk) are common in health research. Nonresponse in longitudinal studies can limit inferences about change over time. ObjectiveThis study aimed to (1) describe the patterns of survey responses and nonresponse among MTurk members with back pain, (2) identify factors associated with survey response over time, (3) assess the impact of nonresponse on sample characteristics, and (4) assess how well inverse probability weighting can account for differences in sample composition. MethodsWe surveyed adult MTurk workers who identified as having back pain. We report participation trends over 3 survey waves and use stepwise logistic regression to identify factors related to survey participation in successive waves. ResultsA total of 1678 adults participated in wave 1. Of those, 983 (59%) participated in wave 2 and 703 (42%) in wave 3. Participants who did not drop out took less time to complete previous surveys (30 min vs 35 min in wave 1, P<.001; 24 min vs 26 min in wave 2, P=.02) and reported having fewer health conditions (5.88 vs 6.6, P<.001). In multivariate models predicting responding at wave 2, lower odds of participation were associated with more time to complete the baseline survey (odds ratio [OR] 0.98, 95% CI 0.97-0.99), being Hispanic (compared with non-Hispanic, OR 0.69, 95% CI 0.49-0.96), having a bachelor’s degree as their terminal degree (compared with all other levels of education, OR 0.58, 95% CI 0.46-0.73), having more pain interference and intensity (OR 0.75, 95% CI 0.64-0.89), and having more health conditions. In contrast, older respondents (older than 45 years age compared with 18-24 years age) were more likely to respond to the wave 2 survey (OR 2.63 and 3.79, respectively) and those whose marital status was divorced (OR 1.81) and separated (OR 1.77) were also more likely to respond to the wave 2 survey. Weighted analysis showed slight differences in sample demographics and conditions and larger differences in pain assessments, particularly for those who responded to wave 2. ConclusionsLongitudinal studies on MTurk have large, differential dropouts between waves. This study provided information about the individuals more likely to drop out over time, which can help researchers prepare for future surveys.https://www.i-jmr.org/2024/1/e58771 |
| spellingShingle | Nabeel Qureshi Ron D Hays Patricia M Herman Dropout in a Longitudinal Survey of Amazon Mechanical Turk Workers With Low Back Pain: Observational Study Interactive Journal of Medical Research |
| title | Dropout in a Longitudinal Survey of Amazon Mechanical Turk Workers With Low Back Pain: Observational Study |
| title_full | Dropout in a Longitudinal Survey of Amazon Mechanical Turk Workers With Low Back Pain: Observational Study |
| title_fullStr | Dropout in a Longitudinal Survey of Amazon Mechanical Turk Workers With Low Back Pain: Observational Study |
| title_full_unstemmed | Dropout in a Longitudinal Survey of Amazon Mechanical Turk Workers With Low Back Pain: Observational Study |
| title_short | Dropout in a Longitudinal Survey of Amazon Mechanical Turk Workers With Low Back Pain: Observational Study |
| title_sort | dropout in a longitudinal survey of amazon mechanical turk workers with low back pain observational study |
| url | https://www.i-jmr.org/2024/1/e58771 |
| work_keys_str_mv | AT nabeelqureshi dropoutinalongitudinalsurveyofamazonmechanicalturkworkerswithlowbackpainobservationalstudy AT rondhays dropoutinalongitudinalsurveyofamazonmechanicalturkworkerswithlowbackpainobservationalstudy AT patriciamherman dropoutinalongitudinalsurveyofamazonmechanicalturkworkerswithlowbackpainobservationalstudy |