www.ijbnpa.org

Table 3
Odds ratios with 95% confidence intervals (OR, 95% CI) for no recreational walking by education, mediated by neighbourhood perceptions and individual cognitions.
		Model 1 (base model): education + age + sex	Model 2: base + neighbourhood	Model 3: base + individual	Model 4: base + neighbourhood + individual
Education	% no walking	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)

1 – low (n = 281)	38.5	1.67 (1.18–2.35)	1.60 (1.13–2.27)	1.33 (0.93–1.90)	1.30 (0.91–1.87)
2 – (n = 908)	34.7	1.49 (1.17–1.90)	1.48 (1.16–1.89)	1.35 (1.04–1.75)	1.29 (0.99–1.68)
3 – (n = 366)	22.1	0.84 (0.60–1.18)	0.84 (0.60–1.19)	0.80 (0.57–1.13)	0.75 (0.53–1.06)
4 – high (n = 439)	29.2	1.00	1.00	1.00	1.00

Neighbourhood perceptions

My neighbourhood is unattractive
disagree			1.00		1.00
agree			1.32 (1.06–1.65)		1.19 (0.95–1.50)

Individual cognitions

Attitude towards regular physical activity
positive				1.00	1.00
neutral				1.12 (0.87–1.45)	1.11 (0.86–1.43)
negative				2.30 (1.59–3.32)	2.26 (1.57–3.26)

Social influences for regular physical activity
positive				1.00	1.00
neutral				1.24 (1.01–1.52)	1.24 (1.02–1.532)
negative				1.54 (1.11–2.14)	1.54 (1.11–2.14)

Perceived behaviour control to be regularly active
(very) sure				1.00	1.00
not sure/unsure				1.21 (0.95–1.55)	1.20 (0.94–1.54)
(very) unsure				1.57 (1.11–2.22)	1.56 (1.10–2.21)

Intention to be regularly active
(very) likely				1.00	1.00
maybe				1.31 (0.99–1.73)	1.30 (0.98–1.72)
(very) unlikely				2.38 (1.59–3.57)	2.38 (1.59–3.57)

Random effects ^a
Level-2 variance (SE)		0.000 (0.000)	0.000 (0.000)	0.000 (0.000)	0.000 (0.000)

^aWeighted multilevel models were estimated with the iterative generalized least squares procedure implemented in MlwiN version 2.02.
Kamphuis et al. International Journal of Behavioral Nutrition and Physical Activity 2009 6:1 doi:10.1186/1479-5868-6-1

Table 3