Table 2.
Main effects of the potential moderators, perceived park proximity, and quality
| Recreational walking | Other recreational MVPA | |||
|---|---|---|---|---|
| Logit model: OR of being non-participant a (95% CI) | Negative binomial model: min/week (95% CI) | Logit model: OR of being non-participant b (95% CI) | Negative binomial model: min/week (95% CI) | |
| Potential moderators | ||||
| Gender (ref. = male) | 0.72 (0.59, 0.88)** | 0.99 (0.90, 1.90) | 0.85 (0.70, 1.04) | 0.85 (0.74, 0.99)* |
| Education level (ref. = ≤ 10 years of education) | ||||
| ≤12 years of education/trade/certificate | 0.86 (0.68, 1.99) | 1.04 (0.93, 1.17) | 0.80 (0.64, 0.99)* | 1.00 (0.86, 1.17) |
| University degree | 0.77 (0.61, 0.96)* | 0.97 (0.97, 1.08) | 0.47 (0.38, 0.61)*** | 0.88 (0.75, 1.04) |
| Retirement status (ref. = non-retired) | 0.79 (0.63, 0.97)* | 1.31 (1.19, 1.45)*** | 0.87 (0.69, 1.08) | 1.18 (1.06, 1.32)** |
| Functional limitations | 1.14 (1.09, 1.17)*** | 0.95 (0.93, 0.97)*** | 1.16 (1.12, 1.22)*** | 0.95 (0.92, 0.98)** |
| Area of residence (ref. = urban) | ||||
| Rural | 0.82 (0.64, 1.05) | 1.04 (0.95, 1.15) | 1.09 (0.87, 1.38) | 0.90 (0.77, 1.06) |
| Fringe | 0.90 (0.43, 1.92) | 0.84 (0.63, 1.11) | 1.34 (0.70, 2.56) | 0.77 (0.49, 1.20) |
| Park characteristics | ||||
| Park proximity | 0.91 (0.85, 0.98)** | 1.03 (0.99, 1.07) | 0.98 (0.91, 1.05) | 0.98 (0.93, 1.04) |
| Park quality | 1.00 (0.98, 1.02) | 1.02 (1.01, 1.02)*** | 0.94 (0.85, 1.03) | 1.01 (0.94, 1.10) |
OR = odds ratio; CI = confidence interval; *p < 0.05; **p < 0.01; ***p < 0.001.
aOR of being non-participant in recreational walking; bOR of being non-participant in other recreational MVPA.
ZINB models evaluate two processes simultaneously. In the logit model, they analyze the relationships between the independent variables and the odds of non-participation in recreational walking or other recreational MVPA. In the negative binomial model, they analyze the relationships with weekly minutes of recreational walking or other recreational MVPA for participants who did engage in some recreational walking or other recreational MVPA. Negative binomial model parameters represent the proportional increase in minutes/week recreational walking or other recreational MVPA with a one-unit increase in the predictor.
The model for recreational walking included 2303 observations, of which 757 were zero observations, and was adjusted for age, suburb SES, and other recreational MVPA.
The model for other recreational MVPA included 2298 observations, of which 1406 were zero observations, and was adjusted for age, suburb SES, and recreational walking.