In the UKHLS, 43 % of participants reported they frequently walked or cycled short journeys (21 % always and 22 % very often). Among the fifth (22 %) living in rural areas, 33 % reported frequent AT compared with 46 % of urban residents (Table 1). Urban residents were 64 % more likely to frequently travel actively than rural residents [OR = 1.64 (95 % CI 1.52, 1.77) after adjusting for age]. In total 16 % did not have a car in their household; this was reported by 19 % of urban but only 7 % of rural residents.
Overall, 54 % of adults below the age of 25 reported they frequently travelled actively, whereas about 39 % of those aged between 35 and 64 years of age did so. AT was reported by 50 % of individuals not in employment and 53 % of those in the lowest household income fifth, compared to 37 % of individuals in full-time employment and 36 % in the highest income fifth (Table 1). The corresponding percentages for urban residents were slightly higher (Table 1), but were substantially lower for rural residents: only 28 % of those in full-time employment and 27 % of those in the highest income fifth frequently travelled actively. AT increased slightly with increased educational attainment for both the full and urban sample; however, AT reduced in the rural sample from 35 % for those with no qualifications to 30 % for those with degree qualifications.
Full sample: in the model which mutually adjusted for all socio-demographic factors except car ownership, all were independent predictors (Table 2, col A). The strongest positive predictors were not being in employment compared to full-time employment [OR (95 % CI) = 1.74 (1.63, 1.86)], being an urban rather than rural resident [OR (95 % CI) = 1.61 (1.49, 1.73)], and being in the lowest compared to the highest household income group [OR (95 % CI) = 1.60 (1.45, 1.77)]. After additional adjustment for having no car in the household, not having children in the household became non-significant (Table 3, col A).
Urban residents: similarly, in the adjusted model of urban residents all socio-demographic factors were independent predictors (Table 2, col B). Frequent AT decreased with increases in age but increased with decreases in household income. As in the full sample, the strongest positive predictors for urban residents were not being in employment [OR (95 % CI) = 1.77 (1.65, 1.91)] and being in the lowest household income group [OR (95 %CI) = 1.63 (1.45, 1.82)]. Working part-time, being male, not having children in the household and being educated to degree level compared with having no qualifications were also positive predictors of AT. Having no car in the household was very strongly and independently associated with AT [OR (95 %CI) = 3.67 (3.37, 3.99)]. When car ownership was added to the model, not having children in the household and some of the equivalised household income categories became non-significant (Table 3, col B).
Rural residents: fewer socio-demographic factors predicted frequent AT in the adjusted model for rural residents (Table 2, col C) than in the model for urban residents. In contrast to the urban population, neither education nor having children in the household were significantly associated with AT. However, younger age, being male, working part-time or not being in employment remained significant positive predictors, as did being in the three lowest income groups compared to the highest fifth. Associations also tended to be weaker; however, being male was a stronger predictor of AT than in the urban population. As for urban residents, not having a car was a strong independent predictor of AT [OR (95 % CI) = 3.90, CI 3.20, 4.75]. After controlling for this, fewer income categories were significantly associated with AT (Table 3, col C).
In addition, we examined whether the urban–rural differences were inflated by AT prevalence and patterns in London. When the urban analyses were rerun excluding London residents, the urban–rural differences remained (data not shown).