Skip to main content

Table 5 Regression coefficients and percentage changes for portion size intervention variables in analyses of secondary outcomes per day

From: Impact of reducing portion sizes in worksite cafeterias: a stepped wedge randomised controlled pilot trial

Analysis

Variable

Purchasing from non-intervention categories.

Coefficients (95% CIs)a

Purchasing from non-intervention categories.

Percentage change (%) (95% CIs)a

pa

Total purchasing from all categories.

Coefficients (95% CIs)a

Total purchasing from all categories.

Percentage change (%)

(95% CIs)a

pa

Overall

Intervention period

0.071 (0.010, 0.132)

7.321 (0.961, 14.083)

0.067

0.004 (− 0.051, 0.059)

0.408 (− 4.992, 6.116)

0.890

By site

Intervention period

Site 1

0.088 (−0.017,0.194)

9.242 (− 1.664, 21.358)

0.154

− 0.013 (− 0.110,0.083)

−1.338 (− 10.418, 8.663)

0.794

Site 2

0.089 (−0.009, 0.186)

9.256 (− 0.904, 20.456)

0.129

0.035 (− 0.054, 0.123)

3.513 (− 5.227, 13.060)

0.474

Site 3

0.111 (0.016, 0.205)

11.701 (1.631, 22.770)

0.064

0.060 (− 0.025, 0.146)

6.231 (− 2.465, 15.702)

0.218

Site 4

0.028 (− 0.067, 0.122)

2.816 (− 6.455, 13.005)

0.587

−0.015 (− 0.100, 0.071)

−1.479 (− 9.546, 7.307)

0.745

Site 5

0.111 (0.016, 0.205)

11.707 (1.608, 22.809)

0.065

0.008 (− 0.077, 0.094)

0.824 (− 7.451, 9.839)

0.858

Site 6

0.002 (− 0.095, 0.099)

0.226 (−9.032, 10.426)

0.965

−0.054 (− 0.142, 0.034)

−5.257 (− 13.204, 3.418)

0.275

  1. a As the p-values (the more robust Kenward-Roger corrected) and CIs (Wald) presented here have been calculated using different assumptions, there is not always an equivalence of interpretation between the 95% confidence intervals and significant p-values