Table 3.
Results of the analysis by site.
| Variable | Coefficients (95% CIs)a | Percentage change (95% CIs) a | pa | Change in energy purchased (kcal) | Mean number of transactions (s.d.) | Change in energy purchased (kcal) per transaction | |
|---|---|---|---|---|---|---|---|
| Availability intervention period | Site 1 | 0.007(-0.083, 0.099) | 0.68 (−8.08, 10.28) | 0.889 | 1374 | Unknown | N/A |
| Site 2 | −0.087 (−0.169, −0.005) | −8.37 (−15.64, −0.47) | 0.088 | −14,919 | 989 (87) | −15 | |
| Site 3 | −0.113 (−0.199, −0.028) | −10.70 (−18.12, −2.62) | 0.046 | −3152 | 161 (20) | −20 | |
| Site 4 | −0.057 (−0.148, 0.035) | −5.50 (−13.87, 3.70) | 0.281 | −2635 | 482 (47) | −5 | |
| Site 5 | −0.203 (−0.312, −0.093) | −18.36 (−26.90, −8.81) | 0.013 | −21,095 | 847 (76) | −25 | |
| Site 6 | −0.043 (−0.145, 0.060) | −4.21 (−13.65, 6.25) | 0.450 | −3607 | 230 (24) | −16 | |
Coefficients in bold are significant at p < 0.05.
a
As the p-values and CIs presented here have been calculated using different assumptions (it is not possible to calculate 95% CIs that correspond to the more robust Kenward-Roger adjusted p-values), the 95%CIs may cross zero while the p-values are not significant.