Understanding Youths' Exposure to Neighborhood Stressors in Activity Spaces in Real Time: A Pilot Test of Ecological Momentary Assessments Triggered By Geo-Fences

Background: Youth of color are disproportionately impacted by ambient stressors in their neighborhoods which may contribute to lifelong health disparities. Several studies have found relationships between neighborhood characteristics and youth outcomes; however, most do not account for the variety of contexts that may exist within a single neighborhood. Examining youths’ exposure to different activity spaces is a more precise way of understanding youths’ exposure to stress and support in their environment. While a handful of studies have captured activity spaces relying on participant recall, using mobile technology to administer ecological momentary assessment, brief surveys answered throughout the day, based on GPS location, may provide a more accurate assessment. Before this technology can be used, however, geo fenced triggers and measures that assess features of activity spaces needed to be pilot tested.

Method: To test: 1) the feasibility and acceptability of completing brief surveys based on a participant’s location and 2) the measures developed to assess features of activity spaces, 11 youth ages 13-18 from one under resourced neighborhood in Pittsburgh were invited to participate in a pilot study in November of 2018. The scales tested assessed violence, demands, structural constraints, social support, collective efficacy and social cohesion, in activity spaces.  Prior to data collection the youth 1) completed one-on-one cognitive interviews, in which they provided feedback on the proposed survey items and 2) identified 50 activity spaces in the target neighborhood to be used as geo fenced triggers. Youth were given cellphones with the mobile application MetricWire and were asked to carry them and respond to surveys for a month. During that time 572 surveys were completed, and 13361 GPS data points were collected. The research team met with the youth after week one, two and four to discuss problems with the mobile technology, question that were unclear and general perceptions of the data collection process. The cognitive interviews and group feedback session were audio recorded and transcribed. The psychometric properties of scales were calculated, using Cronbach’s alphas, multi-level CFAs, cross classified variance component models and correlations.  Passive GPS data points were mapped to assess the youth’s general patterns of movement and exposures to spaces.   

Results: On average the youth completed 1.8 surveys a day and 21.3% were completed while in the target neighborhood but not at home. Youth reported they were being asked to complete surveys too often, there were some contexts where they could not complete surveys and that they did not mind completing many surveys in a day if they were short. Most scales demonstrated adequate psychometric properties:   physical violence (α = .94), demands (α = .77), instrumental support (α = .88), relational support (α = .77) individual efficacy (α = .93), collective efficacy (α = .77), social cohesion (α = .88).

Conclusions: Overall this method of data collection was feasible and acceptable to the youth within some parameters. The proposed measures demonstrated adequate psychometric properties, but some changes are required based on preliminary test and the youths’ feedback.  Lessons learned will be discussed.