Writing a term paper. R programming requires.
Draft Term paper/Project – min. 10 pages, min. 2000 words: One major goal of this course is to provide you with skills and knowledge of both the theory and the practical tools necessary to start your own research. The best way to achieve this goal is to write an original research paper. The paper will discuss why you chose the topic, economic model, econometrics specification, data and empirical findings. Along with your term paper, please upload your codebook (data sources, description of variables as discussed in Bailey Chapter 2) and the R script you wrote/developed for the term paper. Paper Structure (min. 10 pages, min. words 2000): I. Title page: should include the title of the paper, your name and the abstract II. Abstract: This should be less than 50 words and summarize the topic, methodology, and main findings. It should appear on your title page. III. Introduction: This section should state the nature and objectives of the project. Make sure to provide some background or motivation for why your project is interesting. IV. Literature Review: Literature review is a comprehensive summary of previous research on your chosen topic. The literature review surveys scholarly articles, books, and other sources relevant to your particular area of research. It creates a "landscape" for the reader, giving them a full understanding of the developments in the field. V. Description of the model. The model should be clearly stated and any equations carefully explained. You should write out the econometric model you plan to estimate, and discuss the expected impact of the exogenous variables in your model. VI. Data description and model estimation. You should use the techniques developed in class to analyze your data and estimate your model. Make sure to describe the dataset you are using by providing summary statistics of important variables. Your results should be reported and discussed in this section and could include: parameter estimates, standard errors, t-statistics, F-statistics, R-squared, tests for autocorrelation, heteroskedasticity, and possible multicollinearity, as appropriate. VII. Conclusion. Review the major findings as well as possible extensions for future work. Make sure to mention any limitations of your approach as well as alternative explanations of your results. Policy implications, if any, could also be included in this section. VIII. Tables and graphs. Your paper must include at least one table and one graph. The tables and graphs should be well-labeled and accessible to the reader—do not merely print out your regression output with cryptic variable names directly from R. IX. References. You should have a minimum of 4 relevant papers that you have come across in your literature review. Please follow APA format to list references. The link below will help you understand how to organize your references following APA format: Appendix If you have a lot of regression results or other details in your theoretical/statistical model that merit to be included yet, they may distract the reader, you may include them in an appendix. Microsoft Word - ECON 3161 Final Paper.docx Impact of Educational Attainment on Wages Anri Sorel and Erin Shinners ECON 3161: Econometric Analysis Dr. Shatakshee Dhongse Fall 2019 Abstract: Research in labor economics often tests the effect of educational attainment on wages. Our paper examines this relationship using individual-level data from the 2017 American Community Survey Public Use Microdata Sample for the state of Georgia. Across three regression models, we select four additional test variables: age, English fluency, race, and gender. We hypothesize that educational attainment will be the key determinant in wages. The results of our regression models did not fully support our hypothesis, with the explanatory variable gender having the most significant effect. I. Introduction The effect of educational attainment on wages and subsequent economic growth is a widely tested model in U.S. labor economics. The importance of studying education cannot be understated, as it has a strong correlation to different facets of development and financial well-being on the individual, state, and global levels. Rising costs for primary, secondary, and tertiary education, coupled with an increasingly competitive and volatile labor market, provide an environment continually and inherently curious about the relationship between educational attainment and subsequent wages upon employment. This relationship is of utmost financial importance to individuals debating the cost-benefit analysis of investing in current or future schooling. Additionally, administrators in education and policymakers must keep this relationship in mind when marketing and budgeting for educational programs. Using cross-sectional data from the 2017 Public Use Microdata Survey conducted by the American Community Survey on individuals in the State of Georgia, our study examines the impact of educational attainment on wages adjusted for inflation. It is classically accepted that a productive worker sees the highest returns in the form of wages. The human capital theory adopted in the mid-twentieth century gave a name to the categories and variables responsible for increasing worker productivity. One of the categories, economic capital, consists of traits possessed by humans that allow them to contribute to their personal economic value. Education is a key trait in economic capital believed to increase a worker's overall productivity and marketability to employers. Naturally, this is associated with an increase in wages. With the general acceptance of the human capital theory coupled with the questions of the stakeholders mentioned above (potential students, administrators, and policymakers), there has been a notable increase in the number of econometrics studies testing the correlation of educational attainment on wages. In decades past in the United States, exhibiting higher levels of education was often associated with the promise of more gainful employment. Previous economic models traditionally support this view. Today, there is an increase in educational attainment. However, the prospect of yielding a high wage, or for gaining employment at all, is societally thought to have diminished significantly with the preponderance of other determining variables. In this study, we hypothesize that despite the dismal current social convention, educational attainment continues to be the key determinant of wages as predicted by the human capital theory. In our simple regression model, we expect the independent variable educational attainment to be positively correlated to the dependent variable wage. For our multiple regression model, we expanded upon education and selected additional independent variables for testing: age and ability to speak English. While we expect the multiple regression model to also yield a positive correlation between the newly introduced independent variables, we maintain our hypothesis that educational attainment yields the greatest positive impact on worker's wages. II. Literature Review Literature about labor economics and education frequently touches on the Mincer equation. Traditionally, the relationship between educational attainment and wages is referred to as the "Mincer rate of return," after the economist who popularized it, Jacob Mincer. Patrinos (2016) proposes this model as an estimate on the individual private monetary return in the form of wages for additional years of attained education. After tabulating the Mincerian rate of return to education for 136 economies, Patrinos’s (2016) reports a result in alignment with our hypothesis that returns to education are positive. Patrinos (2016) contributes to wage-education literature, stating that when considering the return on education, one must also consider the statistical significance of years of experience as an explanatory variable due to its strong correlation to an increased human capital that yields higher wages. Failure to include experience in the model triggers the omitted variable bias and results in an underestimation of the impact of education on wage. After totaling the surveyed economies, the average return to education is 9.7%. His work takes the model a step further, adding in categorization by gender, showing that female returns to education are always higher. Psacharopoulos and Patrinos (2004) discuss a general overview of the return to education until 2004 and offer a critique of traditional sampling methodology. They conclude, interestingly, that while returns on education are positive, they are falling. From 1992 to 2004, the return on education fell 0.6 percentage points, despite an increase in the supply of educational programs and an increase in enrollment. This confirms the social convention that is held today that motivated our topic: one may be more educated in the modern market, but their wages aren't necessarily comparable to those of your equal in decades past. Psacharopoulos and Patrinos (2004) also touch on the critiques of traditional regression studies of education and wage. Traditional studies draw on data compiled using a survey of firms, which results in bias thanks to the nature of the study. Surveys of firms are often skewed towards the urban working class, which is not representative of any single country's true population. Psacharopoulos and Patrinos (2004) state that surveys of households are preferred when regression education on wages. This is encouraging for our results, given our data source, the American Community Survey Public Use Microdata Sample, draws on both household and individual data across the State of Georgia. Psacharopoulos and Patrinos (2018) provide a more concise overview of twenty-first-century data on returns to education in their decennial review. This study expands on their 2004 update, primarily reinforcing the idea that across 705 estimates from 1950 to 2014, the private returns to education remain positive yet declining. This study adds that while falling, the decline is slight, with the 2004 average rate yielding a return of 9.7% and the 2018 average rate yielding a return of 9.5%. Psacharopoulos and Patrinos (2018) conclude that this decline is insignificant, and the result of an environment where the cost of education remains the same or increases despite an increase in supply. They illustrate this idea further by focusing on a multitude of individual countries across a timespan of decades. Card (2001) explores the wage-education model using a slightly different perspective than that of other literature. He touches on literature from the late twentieth century that states returns on education are positive with respect to wage; however, the model should be evaluated on a case by case basis where education is optimized with special attention paid to the costs of attending school and not simply by paying attention to highest wage. Card (2001) also elaborates that "individuals may have different aptitudes and tastes for schooling relative to work," and as such, educational attainment should be selected with one’s future industry in mind. Throughout his work, Card (2001) reiterates the importance of considering an individual's environment when evaluating an education-wage equation, particularly one's incentivization to yield a higher wage and the marginal cost to attain additional education. In the instance of excluding both incentivization and the marginal cost of education, Card (2001) states that the OLS regression estimate of education on wages is biased. While the literature is dated, this conclusion is important to consider in our attempt to uncover the relationship between wages and education in our data while still considering the possibility of unknown variables in the error term. Our study hopes to contribute to the existing literature with similar results. Per our hypothesis, we expect our explanatory variable educational attainment to yield a higher wage and to hold as the strongest variable in a multiple regression model against other variables of interest. Additionally, we hope to expand on the literature by diversifying our additional variables in the multiple regression model to include the age of the individual and the worker's English fluency. When comparing our results to that of previous literature, it is important to remember the source of our data. While we expect similar results, our models are based on data specific to the state of Georgia. Also, our data will examine a single year in time to give more concise and modern results, as opposed to a year-by-year comparison. III. Data A. Source of Data We drew our variables from the American Community Survey Public Use Microdata Sample. The Public Use Microdata Sample draws a sample from the aggregate Census data to allow for easy customization and tabulation. For our data, we selected the individual Public Use Microdata Sample, meaning the variables report on single people and not households. Additionally, we selected the year 2017, a 1-year sample, which is representative of 1% of the United States population. For reporting purposes, we decided to focus on the state of Georgia for relevance to the expected audience. The regression was calculated using anywhere from approximately 4,000 to 6,000 observations. B. Description of the Variables Variable Source Year Observation* Type logwagp ACS PUMS 2017 47,795 Dependent schl ACS PUMS 2017 96,647 Independent agep ACS PUMS 2017 81,031 Independent eng ACS PUMS 2017 11,082 Independent Figure 1 *The difference in observations is discussed in Part C (Descriptive Statistics of the Variables) The explained variable to be tested is the natural log of the population wage (logwagp). This variable specifically records an individual’s wages or salary income in the past 12 months adjusted into constant dollars to account for inflation. A natural log was selected