Results from the ordered probit regression analysis for questionnaire 1, questionnaire 2, and merged data
| Questionnaire 1 (n=608) | Questionnaire 2 (n=688) | Total | |
| Daytime cough | –0.48 (0.04) | –0.55 (0.04) | –0.52 |
| Daytime breathlessness | –0.48 (0.04) | –0.50 (0.04) | –0.49 |
| Daytime wheeze | –0.25 (0.07) | –0.24 (0.07) | –0.25 |
| Chest tightness | –0.24 (0.07) | –0.30 (0.07) | –0.27 |
| Sleep disturbance | –0.26 (0.05) | –0.24 (0.05) | –0.25 |
The χ2 statistic (95% critical value 11.07) was used for testing the restriction that the coefficients are the same for surveys 1 and 2. Likelihood ratio (LR) = 3.54, hence the H0(no difference between weights in the two surveys) cannot be rejected. All p values ⩽0.001.