Category: Technology


To Pay or Not to Pay: Using Amazon Mechanical Turk in Clinical Studies for Mood Management

Friday, November 17
12:00 PM - 1:30 PM
Location: Sapphire Ballroom O & P, Level 4, Sapphire Level

Keywords: Technology / Mobile Health | Depression | Behavioral Activation
Presentation Type: Symposium

Background: Internet interventions have become a crucial component within the healthcare system. The main goal of this study was to explore baseline and outcome differences between participants recruited through Amazon’s Mechanical Turk (AMT) and Social Media (SoM).

Method: Participants were recruited to two randomized control studies on a mood management Internet intervention. One study recruited participants through AMT and the other study from SoM. Participants were assessed for depression (PHQ-9),  anxiety (GAD-7), mood, confidence, motivation, and perceived usefulness of the intervention.

Results: The AMT sample consisted of 765 adults (Mean Age = 35.9; SD = 8.7); 69.2% female, and 464 individuals (60.65%) completed follow-up. The SoM sample consisted of 422 adults (Mean Age = 32.93; SD = 14.93); 74.9% female, and 124 (29.38%) completed follow-up. At baseline, the two samples differed on PHQ-9 scores [t(1185) = 4.02, p < .001], GAD-7 scores [t(915) = 2.95, p < .01], reported mood [t(856) = 6.57, p < .001], motivation [t(762) = 3.75, p < .001], confidence [t(793) = 4.66, p < .001], and expected usefulness of the intervention [t(833) = 2.36, p < .05].

Using a mixed effects ANOVA, a main effect of time was found for confidence [F(1,1152) = 73.71, p<0.001], motivation [F(1,1145) = 25.26, p<0.001], and usefulness[F(1,1127) = 27.29, p<0.001], such that the entire sample reported an increase from pre- to post. Furthermore, a time by sample interaction was observed for confidence [F(1,1152) = 4.38, p<0.05], motivation[F(1,1145) = 3.90, p<0.05], and usefulness[F(1,1127) = 39.34, p<0.001], suggesting that the AMT sample showed greater improvements for all three than did the SoM sample.

Conclusions: Recruitment methods through AMT were more efficient; i.e. higher recruitment rates per month, and more follow-up completers than the SoM sample. Although both sample differed in depression and anxiety, mood, confidence and motivation scores at baseline, the AMT participants showed a greater increase in confidence, motivation, and perceived usefulness. While recruitment methods through AMT are more efficient, outcomes from AMT participants should be analyzed carefully.

Haley Cook

Palo Alto University


Send Email for Haley Cook


To Pay or Not to Pay: Using Amazon Mechanical Turk in Clinical Studies for Mood Management

Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for To Pay or Not to Pay: Using Amazon Mechanical Turk in Clinical Studies for Mood Management