Intergroup Avoidance: A Fly-on-the-Wall Experiment

• In Event: Advances in Survey Experiments

Abstract

This research introduces and presents the results of a ‘Fly-on-the-Wall Experiment’—a type of natural experiment where treatments are randomly observed, not artificially assigned. The data utilized exceed 8,500 hours of surveillance footage captured on “Link Light Rail” platforms between the hours of 6:00 AM and 10:00 PM over a two-week period in Seattle, Washington. RGB Color values are objectively assigned to subjects’ skin tones using an automated color meter. The dependent variable measures the distance between each pair of individuals as they successively arrive on the platform by counting the number of 1-foot (30.48 cm) tiles, including fractions, between them. The analyses assess multivariate statistical relationships among skin-tone, sex, age, ethnicity, and class with the physical proximity of study subjects. Stations’ neighborhood characteristics, the date, weather, day of the week, and time of day that observations occurred are also analyzed. Statistical models incorporate interaction effects, and subgroup-only regressions. A random sample of ethnographic interviews performed in-situ help to identify motives for avoidant behavior.

Author

• Joshua Corona, Northwestern University