Table 1.
Baseline comparability (treatment vs. control stations).
Table 2.
Victim-generated and police-generated crimes counts: Means and standard deviations (pre and post random assignment), and estimated marginal mean differences.
Fig 1.
Percent change of dosage from control conditions.
Percentage change between control and treatment stations in (A) the frequency of visits by security guards, (B) the duration of visits by security guards and (C) number of police patrols completed.
Table 3.
Delivery of proactive patrols—Duration and frequency of visits, per month.
Table 4.
Parameter estimates.
Fig 2.
Victim-generated and police-generated crimes (main effects): Estimated marginal means.
Estimated marginal means of victim and police generated of crime at differing types of locations in treatment and control stations.
Fig 3.
Sub-group analyses: Effect within crime categories (entire station complex).
Estimated marginal means of sub categories of crime in treatment and control stations.
Table 5.
Train operator self-reported incidents at participating stations (post random assignment only): Means and standard deviations per 10M passengers.
Fig 4.
Train operator’s internal reports.
Standard difference in means in categories of incidents recorded by train operators, compared between treatment and control stations.
Fig 5.
(A): Total proactive patrol dosage (duration), (B): Total proactive patrol dosage (frequency).
(A) Total number of minutes of proactive patrol, by month, in treatment and control stations. (B) Total frequency of proactive patrol, by month, in treatment and control stations.
Fig 6.
(A): Relative percent change: Victim-generated crimes (police data), (B): Relative percent change: Police-generated crimes (police data).
(A) Percentage change in victim generated crimes recorded by the police at treatment and control stations. (B) Percentage change in police generated crimes recorded by the police at treatment and control stations.