Reference no: EM131918624

Assignment

1) Identify a topic that is interesting, specific, and feasible. It must be all three. An interesting question might be, "Does reading make you a better person?". This is interesting, but not specific. A viable research question might be, "Does a mandated extra hour of reading during the sixth grade raise future earnings?" It must also be feasible. There may be no data on mandated reading programs in sixth grade. You must consider all three requirements at once. One way to do this is to look at a number of available data sets and come up with a question for which you know the data are available. I have created a couple of videos that explain step-by-step how to use IPUMS. Other lists of public-use data sets are available.

2) Conduct a literature review. Find prior research that looked at your question, or any questions related to yours. You must have at least 10 prior research articles. They don't have to be exactly on your topic, but as close as you can find. The goal of the literature review is to inform the reader of the place of your research in the ongoing conversation about the topic you are researching. For example, you might study the effect of the introduction of DNA databases on crime rates. Perhaps others have looked at this questions, but used much older data than you have. Describe how these prior papers addressed the question and the benefit of using more recent data. Perhaps no one else has looked at this. In that case, discuss other papers on the effect of technology on crime solving. Don't list and describe one paper after the other. Summarize the conversation among researchers on your topic and your place in that conversation.

3) Apply economic theory to the research question in order to generate testable predictions. Use economic reasoning (it doesn't have to be mathematical) to explain what you expect to find. For example, the introduction of DNA databases raises the probability of a criminal being caught and therefore raises the cost of committing the crime. Since the cost goes up, we should expect crime to go down, holding other factors constant.

4) Find the appropriate data and variables to test the theoretical predictions. You should not settle on an idea and definitely not write a literature review before being sure you have data that can answer your question. As mentioned above, pick the question and the data at the same time.

5) Identify the appropriate econometric models to use to test the theory. This depends on the nature of your dependent variable, as well as other aspects of your data. Perhaps you dependent variable is earnings, a continuous variable. Then you would use ordinary least squares (OLS). If your dependent variable might be married, which is one if the person is married, zero if not. Then you must use a probit or logit model. Or perhaps you are using data on obesity rates for different countries over many years. This is called "panel data", and you must use methods appropriate to this type of data.

6) Understand how to interpret empirical results. Explain your results, whether they confirm or reject your hypotheses, whether the size of the effects you find make sense, whether they line up with prior research.

a. Cleaning the data. Data rarely come in a form ready for immediate analysis. Perhaps wage data are listed as 99999 when the respondent to the survey refused to answer the question. You have replace the 99999 with a missing value so it will not be used. You will certainly have to create your own variables from the ones provided. Perhaps there is a variable indicating the gender of the respondent, with 1 indicating male and 2 indicating female. You would have to create a dummy variable for female, where 1 indicates a female and 0 a male.Running the analysis. You may have to run regressions on a number of subsets of the data and with different sets of variables.

b. Creating a presentation of your results, using Excel, Stata, and PowerPoint.

c. Including all the above, you must create a detailed description of each step in your analysis. Specifically, you must create a Stata do file that specifies each step in the cleaning and analysis of your data, from the raw data to the final presentation. All research must be REPLICABLE or it is worthless.

2) Writing up the results in a paper. The writing must be grammatical, clear, and complete. Anticipate questions and answer them in one of the sections of the paper.

3) Presenting your research in front of an audience. You must be able to present your question, data, and results in a way that a non-specialist can understand. You must make PowerPoint slides that are neither too crowded nor too bare.