Extended abstract for Anxiety as a Cost of Commuting to Work:
By Bruce H. Smith
In this study there were 1067 households with at least one member commuting regularly to and from a regular workplace surveyed. They were asked a variety of questions from how many minutes they spent walking, driving, riding public transportation, or bicycling to and from work. They were also asked about their hourly wages so that they would be able to convey their time costs of traveling to and from work in dollars. Questions designed to gain money costs of travel were also asked, at that time the daily bus fare was $.60 a ride in Columbus. With the week being the unit of time studied the bus fare was $6.00 per week, for most passengers. The surveyors said that they were careful not to take samples of people that the respondents might be using the “park and ride” system, as they would not accurately reflect the purpose of the study. Respondents that drove their own cars were asked to point out expenditures per interval for interest, insurance, gasoline, and sundry other expenses. The depreciation of the cars was also taken into account and was determined by the Blue Book value of the car. In the study a total of 173 out of the 1067 people interviewed reported that they walked or rode their bicycles to work everyday. A binary logit and probit were run to analyze the impact on the log of odds and probability of variations in a set of variables. There where a set of explanatory variables which were ATC (total automobile costs) and PTC (total public transportation costs), socioeconomic variables, sex, race, age, and education. In the binary logit model, the R2 index was 0.998 and the multiple correlation coefficients for the probit was 0.893. The statistical significance of both the total money cost and the psychological cost variables support the design specification of the variables and the model. The surveyors saw a constant in the answers of the public transportation riders that the busses stop frequently and are therefore much slower than commuting by car. They suggested that they make high-speed bus lanes so that the public transportation can be “competitive” with the automobile. As far as the biking goes the survey found that the costs of bikes failed to have an impact on the odds of choosing a mode except for biking itself. Some of the socioeconomic variables for biking came through in this study, one being that males are far more likely to ride their bike than females. Also age is significant in the choice to bike or not bike to work. There are tables included along with the results of the logit and binary logit model choices.
Data Sources:
The author did his own survey based on the 1986 Modal Choice Interview Survey (MCIS) for Columbus, Ohio.
Data variables for Binary Logit and Probit Results for Modal Choice where:
Dependent variable- 1 if automobile, 0 if public transportation
ATC-Total auto costs
PTC-total bus costs
Sex-1 for male
Race-1 for black
Log of age-took the log of the surveyed respondent’s age
Education-how many years (level, did you graduate college, high school, ect.) you have
Psychological driving costs
Psychological bus costs
In the Multinominal Logit Model of Modal Choice they added a few more variables including:
BTC- Total Bicycling Costs
WTC-Total Walking Costs
Psychological Bicycling Costs
Psychological Walking Costs
Smith, Bruce. H., Journal of Urban Economics“Anxiety as a Cost of Commuting to Work.” Vol. 29, iss. 2, pgs. 260-266., Ball State University, Department of Economics, Muncie, Indiana 1991.