Skip to main content

Table 2 Multivariate logistic regression analyses to determine which variables confound ethnicity

From: Ethnic minority women prefer strong recommendations to be screened for cancer

  Step 1: Unadjusted ethnicity only Step 2: Adjusted for socio-demographicsb Step 3: Adjusted for socio-demographics; acculturation; language Step 4: Adjusted for socio-demographics, acculturation, language; comprehension of health information
Ethnicitya
 White British 1.00 1.00 1.00 1.00
 Caribbean 2.56 [1.51–4.35]*** 2.60 [1.47–4.60]* 2.24 [1.18–4.24]* 1.80 [0.89–3.63]
 African 4.66 [2.71–8.03]*** 4.31 [2.46–7.54]** 2.90 [1.50–5.62]** 2.64 [1.34–5.23]**
 Pakistani 4.84 [2.80–8.36]*** 2.90 [1.59–5.31]** 2.06 [1.09–3.91]* 1.83 [0.92–3.64]
 Indian 7.04 [3.98–12.45]*** 4.43 [2.36–8.32]*** 3.17 [1.64–6.14]** 2.69 [1.34–5.43]**
 Bangladeshi 13.59 [7.15–25.85]*** 6.76 [3.33–13.71]*** 5.22 [2.51–10.84]*** 4.38 [2.01–9.51]***
Step: χ 2(df), p-value - 35.95 (5), <.001 11.59 (3), .009 4.20 (3), .240
Full model: χ 2(df), p-value 98.20 (5), <.001 134.12 (10), <.001 145.72 (13), <.001 149.92 (16), <.001
Nagelkerke R2 .175 .233 .251 .257
  1. *p <.05, **p <.01, ***p <.001
  2. aOdds Ratio [95% Confidence Interval]
  3. bmarital status, employment and education