In this study I will be looking at data from the Southern Maine subset within the 2000 United States Census. This sample includes a 5% random sampling of the residents of Androscoggin, Cumberland, Kennedy, Lincoln, Sagadahoc, and York counties. The data that I will use comes from Integrated Public Use Microdata Series database for the United States (IPUMS-USA) compiled by the Minnesota Population Center. For the purposes of this report, I will consider every individual within the 5% sample as the sole representative of a distinct family, and, accordingly, limit my data set to those between the ages of 18 and 65. The focus of the study is to find first what the determinants are of total family income, and, next, if the same factors are able to explain whether a family lives below the poverty threshold. I will attempt to locate any explanatory factors such as education, race, gender, age, marital status, hours worked, etc. that may possibly account for the disparities in family income and status in regard to the poverty line.
[...] Despite any real life implications my data may have for reducing poverty in Southern Maine, no single program, subsidy, or referendum will vastly change the economic outlook. In the results for my economic variables, each level up until achievement of a professional degree reduced the probability of leading an impoverished family substantially more. Therefore, it is natural to assume that any policy promoting education will be instrumental in reducing poverty. To that end, schools need to increase efforts in higher education promotion. [...]
[...] Southern Maine has a limited minority population due to its isolation, and I expect that most minorities who do come do so with a distinct, well-paying profession in mind for the head of the family. Lastly, Female having a positive but statistically insignificant influence suggests that the labor market is relatively gender-blind. Most of the studies suggesting otherwise are outdated, so this contrary result is relatively consistent with my personal expectations. For my second model, I created the dummy dependant variable Pov2K. [...]
[...] Because the variable included only years and degrees that were successfully completed by the individual at the time, I was able to recode it into dummy variables for each year I felt would have an effect on income and poverty status. I began with Hsdeg for those observations who had acquired only a high school degree or a GED which was of my sample. Moving up, I included a Somecol dummy variable for those who had completed at least one year of college but never earned a degree: of the sample. [...]
[...] Econometric Models, Estimation Methods, and Specification Testing This section will progress through the econometric steps I took to arrive at my two final models as well as analyze what factors influence family income and whether or not a family is below the poverty threshold. Table shows the summary for the particular variables I extracted from the IPUMS dataset. It shows the results of the Proc Means test filtered for only the variables in my final income model. TABLE 1 Summary Statistics of IPUMS Data Variable Mean Std Dev Minimum Maximum My first step was to create an OLS model which could analyze explanatory power of the specific variables I designated on income. [...]
[...] In my study I will explore specific variables for each general level of degree from high school to doctorate, so I expect the relationship between each of these variables to be directly correlated with income and to present inverse marginal probabilities when set against poverty. I expect that another significant factor will be the labor market structure, specifically the number of hours worked in a typical week. Similarly to education, I do not view each additional hour worked per week as uniformly important in determining income. [...]
Online readingwith our online reader
Content validatedby our reading committee