I have read with great interest the study about the mode of travel to work and health-related outcomes. Since I am not knowledgeable of demographics and travel behaviour in the UK, I have a few questions about the sample.
How representative is the sample (though significantly large) of the underlying population? For instance, 69.3% of those who walked to work and 63.9% of those who cycled to work were women, compared to 51.7% of those who commuted by public transit or by private automobile. Are these gender differences in mode of travel to work similar to what one would find in a travel behaviour study in the UK?
The sample's ethnic background suggests that 96.3% of those who cycled to work were White with almost no reasonable representation of South Asians, Blacks, Chinese or other visible minorities in the cycling cohort? In fact, the sample is largely White comprising at least 93% of the respondents. Is the sample representative of the racial composition in the UK?
I may have missed it in the draft, but I can't recall seeing a reference to the location of the respondents being in an urban (large metropolitan area) versus a rural or suburban location. The mode of travel differs by location in an urban-suburban context. Distance to work influences the mode of travel, especially non-motorized trips.
Walking to work also implies that one enjoys the privilege of living close to work. Given the exorbitant housing prices and rents in large cities like London, I'd assume the opportunity to live within the walking distance of one's work is not equally distributed among all cohorts differentiated by income, age, and household size.
Lastly, the Baseline health status suggests that those who commuted by car or public transit had "significantly" higher incidence of Diabetes history, hypertension, Cancer history, longstanding illness, CVD, and depression than those who walked or biked (Table 1). Can one actually control (in the strictest statistical terms) for these differences if the subsamples are systematically that different at baseline?
Also, without being knowledgeable of the incidence of diseases in the UK, how could one be certain that this small group of 6,751 mostly White and overwhelmingly female 'healthy' cyclists (at baseline) are indeed pointing us to something significant rather than being luckier than the rest?
Since the decline in odds is more pronounced for those whose cycling commutes are longer than the median commutes, the results are essentially driven by 3,375 respondents out of a total of 263,540. Given its better than the rest baseline health status, would this subsample of 3,375 mostly White and overwhelmingly female cohort experience an increase in the odds for illnesses if it were to change the mode of travel to car or public transit?
Rapid Response:
Representative sample
I have read with great interest the study about the mode of travel to work and health-related outcomes. Since I am not knowledgeable of demographics and travel behaviour in the UK, I have a few questions about the sample.
How representative is the sample (though significantly large) of the underlying population? For instance, 69.3% of those who walked to work and 63.9% of those who cycled to work were women, compared to 51.7% of those who commuted by public transit or by private automobile. Are these gender differences in mode of travel to work similar to what one would find in a travel behaviour study in the UK?
The sample's ethnic background suggests that 96.3% of those who cycled to work were White with almost no reasonable representation of South Asians, Blacks, Chinese or other visible minorities in the cycling cohort? In fact, the sample is largely White comprising at least 93% of the respondents. Is the sample representative of the racial composition in the UK?
I may have missed it in the draft, but I can't recall seeing a reference to the location of the respondents being in an urban (large metropolitan area) versus a rural or suburban location. The mode of travel differs by location in an urban-suburban context. Distance to work influences the mode of travel, especially non-motorized trips.
Walking to work also implies that one enjoys the privilege of living close to work. Given the exorbitant housing prices and rents in large cities like London, I'd assume the opportunity to live within the walking distance of one's work is not equally distributed among all cohorts differentiated by income, age, and household size.
Lastly, the Baseline health status suggests that those who commuted by car or public transit had "significantly" higher incidence of Diabetes history, hypertension, Cancer history, longstanding illness, CVD, and depression than those who walked or biked (Table 1). Can one actually control (in the strictest statistical terms) for these differences if the subsamples are systematically that different at baseline?
Also, without being knowledgeable of the incidence of diseases in the UK, how could one be certain that this small group of 6,751 mostly White and overwhelmingly female 'healthy' cyclists (at baseline) are indeed pointing us to something significant rather than being luckier than the rest?
Since the decline in odds is more pronounced for those whose cycling commutes are longer than the median commutes, the results are essentially driven by 3,375 respondents out of a total of 263,540. Given its better than the rest baseline health status, would this subsample of 3,375 mostly White and overwhelmingly female cohort experience an increase in the odds for illnesses if it were to change the mode of travel to car or public transit?
Competing interests: No competing interests