Table 3

Multivariate logistic regression analysis: ORs for residents of Liverpool and Manchester, compared to those of Glasgow, for reporting bad or very bad health, after adjustment for (1) characteristics of the samples and (2) Sense of Coherence (SOC-13) score

ModelCityPercentage of sample*OR (fully adjusted; 95% CI)Sig†, p value
1 Adjusting for age, gender, social grade, employment status, educational attainment, deprivation quintile, ethnicity, marital status, length of residence in city
Glasgow‡9.6
Liverpool8.50.74 (0.54 to 1.02)0.065
Manchester5.90.67 (0.48 to 0.94)0.022
2 Adjusting for the above, plus Sense of Coherence (SoC-13)†
Glasgow‡9.6
Liverpool8.50.67 (0.48 to 0.93)0.016
Manchester5.90.53 (0.38 to 0.76)<0.001
  • R2 values (Cox & Snell)—model 1: 0.14; model 2: 0.15.

  • Only significant (p<0.05) variables included in final models.

  • *Weighted figures.

  • †Sense of Coherence—OR 0.97 (95% CIs 0.96 to 0.98, p<0.0001).

  • ‡Reference category of variable.