Table 1.
Baseline characteristics for premises allocated initially that remained in the intention‐to‐treat analysis and those in the per‐protocol and non‐randomized sensitivity analyses. Binary indicator variables were created and designated premises as high or low in respect of historical violence and weekend opening hours. Mean and standard deviations (SD) of raw figures are included.
| Group | Control | Intervention | ||
|---|---|---|---|---|
| n High | Mean (SD) | n High | Mean (SD) | |
| Initial allocation | (n = 300) | (n = 300) | ||
| Violence | 85 | 2.53 (3.16) | 87 | 2.78 (4.39) |
| Opening hours | 134 | 4.26 (2.94) | 132 | 4.47 (3.05) |
| Intention‐to‐treat | (n = 208) | (n = 238) | ||
| Violence | 54 | 2.41 (3.03) | 73 | 2.93 (4.75) |
| Opening hours | 97 | 4.35 (2.86) | 109 | 4.49 (2.86) |
| Per‐protocol | (n = 208) | (n = 245) | ||
| Violence | 54 | 2.41 (3.03) | 72 | 2.92 (4.75) |
| Opening hours | 97 | 4.35 (2.86) | 106 | 4.47 (2.86) |
| Non‐randomized | (n = 321) | (n = 285) | ||
| Violence | 103 | 2.64 (4.31) | 73 | 2.78 (4.46) |
| Opening hours | 172 | 4.32 (2.75) | 109 | 4.34 (2.80) |