Using Mechanical Turk to study family processes and youth mental health: A test of feasibility.


Extensive evidence links youth mental health to family functioning, highlighting the need to document causal pathways. This will require longitudinal studies, but traditional methods for longitudinal clinical research have several limitations, including high cost and resource demands, underrepresentation of fathers, and attrition bias. We tested whether an online alternative might address limitations of-and thus provide a useful complement to-traditional methods. We used the Mechanical Turk (MTurk) survey program to obtain reports from parents (N = 177) on family functioning, the parents’ own symptoms, their children’s behavioral and emotional problems, and parenting stress, with assessments in three consecutive months. Parents provided largely high-quality data (e.g., passed consistency checks); measures showed acceptable psychometrics at each time-point; and correlations among study measures paralleled those observed in prior research. Compared to prior studies using traditional longitudinal methods, the MTurk method was (a) much lower in cost and resource requirements, (b) successful in enrolling fathers, (c) comparable in participant attrition, and (c) similar in attrition bias, participant race/ethnicity, and enrollment of single parents. Overall, findings suggest that MTurk is a viable tool with its own set of strengths and limitations, and a potentially useful complement to traditional longitudinal methods. In particular, MTurk might be a cost-effective first step in generating causal hypotheses about family processes and youth mental health, for later testing via more traditional methods.