In major U.S. cities today, there seems to be a fundamental mismatch between poor workers and the low-skilled jobs that they are qualified to hold. In general, low-skilled jobs are spread along the urban fringe while low-skilled workers tend to be concentrated in inner-city areas. Although there are many reasons why the urban poor remain poor, one of them could be that prospective low-skilled workers are not able to commute to the places where jobs are located. The availability of adequate public transit options might determine whether a poor inner-city worker holds a job or not.

What effect does the presence or lack of viable public transit options have on urban poverty?

Metropolitan areas with poor public transit infrastructure have higher rates of poverty.

Poverty is our dependent variable. But how do we define the quality of a city’s public transit network? We chose to define a variable called “extent” that would serve as a measure of an MSA’s public transportation coverage. We obtained data from the National Transit Database on number of miles of rail and bus lines for all the MSAs in our study. We then summed the two values for each MSA, and divided this value by land area. This gave us our extent measure -- miles of transit per square mile: the closer to 1, the better an MSA’s transit coverage.

1. Identify all MSAs with populations greater than 500,000.
2. Compile numbers on extent of public transportation networks for MSAs in the sample.
3. Collect data and input into an Excel spreadsheet. Identify inconsistencies in the data.
4. Identify and control for other determinents of poverty and extent of public transit.
5 Construct an indexed variable representing "extent of public transit."
6. Run regressions with poverty as the dependent variable and observe coefficients, statistical signficance, and R-squared of our regressions before and after adding our transportation variable.
7. Write up data analysis of our results, discussing their statistical significance / insignificance. Discuss whether or not results support or nullify our hypothesis. If results are weak, propose ways to improve further research.

Regression: Poverty against Population, Payroll, Unemployment Rate, Extent of Transit Coverage

Click the images below to enlarge
 
 
Our results support our original hypothesis: our variable Extent1 is statistically significant and has the appropriate sign. It shows that the greater the extent of an MSA’s public transit network, the lower its level of poverty. However, our R-squared of .459 indicates a moderate but not strong correlation between our dependent and independent variables. Clearly, there are other variables that act on poverty which we
have not included in our analysis.