Table 4.
Mixed‐effect regression model, logged estimates and estimates expressed in terms of percentage change (95% CI) for rate of wine consumption a (n = 166).
| Estimate (SE) | t‐value | P‐value | 95% CI for estimate | |||
|---|---|---|---|---|---|---|
| Lower | Upper | Percentage change (95% CIs) | ||||
| Intercept | 2.15 (0.16) | 13.7 | < 0.001 | 1.85 | 2.46 | |
| Bottle size 75 cl (ref: 50 cl) | −0.06 (0.03) * | −2.1 | 0.040 | −0.12 | −0.00 | −5.8% (−10.9 to −0.4%) |
| Intervention period (ref: period 1) | 0.01 (0.03) | 0.4 | 0.691 | −0.04 | 0.07 | 1.1% (−4.3 to 6.8%) |
| Intervention order (ref: 50 cl first) | −0.02 (0.05) | −0.5 | 0.646 | −0.11 | 0.07 | −2.2% (−10.7 to 7.2%) |
| Baseline consumption (litres) | −0.41 (0.04) ** | −10.2 | < 0.001 | −0.48 | −0.33 | −33.4% (−38.3 to −28.1%) |
| Guest consumption (litres) | −0.14 (0.03) ** | −4.9 | < 0.001 | −0.19 | −0.08 | −13.0% (−17.7 to −8.0%) |
| Out‐of‐home consumption (litres) | 0.01 (0.02) | 0.5 | 0.622 | −0.03 | 0.05 | 0.9% (−2.7 to 4.7%) |
| Non‐study wine consumption at home (litres) | 0.004 (0.026) | −0.01 | 0.889 | −0.05 | 0.05 | −0.4% (−5.3 to 4.8%) |
| Price (£) per litre | 0.02 (0.01) | 1.7 | 0.089 | −0.00 | 0.05 | 2.3% (−0.3 to 4.9%) |
| ‘Usual behaviour’ duration (days) | 0.01 (0.004) * | 4.0 | < 0.001 | 0.01 | 0.02 | 1.4% (0.7 to 2.1%) |
| Number of wine drinkers in household | −0.02 (0.04) | −0.5 | 0.650 | −0.09 | 0.05 | −1.6% (−8.2 to 5.5%) |
*
Significant at the P < 0.05 level;
**
significant at the P < 0.01 level.
a
Percentage change was estimated by exponentiating the logged estimates, then subtracting these values from 1. CI = confidence interval; SE = standard error.