Research Project Overview

Introduction

As a part of this class, you will write an empirical research paper. You can work individually or with a partner. You will pick a question and a dataset and use the tools from this class to answer your question. Over the semester, you will need to meet the following milestones:

  • Feb 6: Finalize the Team
  • Feb 25: Interim Submission I
  • March 18: Research Project Feedback Meeting
  • April 8: Interim Submission II
  • May 6: Final Submission

The first interim submission is worth 5% of your total grade, while the second is worth 10%. The final submission is worth 15%. I will also meet with you mid-semester to give feedback on your project, which you should incorporate into your final paper.

The goal of Interim Submission I is for you to choose your data and research question. Interim Submission II then requires you to conduct preliminary data analysis. The Final Submission will be 7 to 10 pages and will thoroughly discuss all the analyses you performed. The final paper should demonstrate your writing ability and ability to perform and interpret statistical analyses correctly.

Below, I outline the type of question you should choose and the criteria your dataset must meet, and also include some external sources for finding data and inspiration for your project. After you’ve completed reading the content on this page, follow these links for detailed instructions for the Interim Submission I, Interim Submission II, and Final Submission.

Research Question and Data

You may choose any question and data, provided they meet the criteria outlined below.

Question

Start by identifying a topic that captivates your interest or curiosity. This could be inspired by a recent podcast you listened to or a news item that caught your attention. The key is to choose a subject that genuinely engages you, as this will make the research process more enjoyable and meaningful.

Focus your research on a cause-effect question, rather than a predictive one. In other words, rather than undertaking a project aimed at “Using weather to predict sales,” you should investigate a question such as “How does weather affect sales?” While predictive questions are important, this project is designed to enhance your understanding of causality and the challenges involved in establishing it. Therefore, your question should concern the impact of one variable on another.

Finally, ensure that your question is well-defined and feasible, meaning it should be clear and specific, and can be addressed within the limitations of time, resources, and the data available.

Variables

Once your research question is set, the next step would be to identify the variables required to investigate it. For example, if your question is “How does weather affect sales?”, you need to decide how to define ‘weather’—are you focusing on temperature, rainfall, or other weather conditions? For this project, you are required to select a primary independent variable, a dependent variable, and at least two control variables.

  • Primary Independent Variable: This is the main variable you hypothesize as having an influence or effect on another. For instance, if you were studying the impact of weather on sales, temperature or rainfall would be your primary independent variable.
  • Dependent Variable: This is the variable affected by the independent variable. In the example of weather and sales, ‘sales’ would be the dependent variable.
  • Control Variables: These are variables that affect both the dependent and the independent variable, potentially confounding our primary relationship of interest. For example, ‘month of sale’ would be a confounding variable as it affects both sales and weather conditions. For instance, people tend to buy more gifts during the holiday season when the weather is also colder. So we might observe higher sales during colder months, but this could be due to the holiday season rather than the weather. To account for such confounding effects, you will incorporate the control variables you select later in your analysis.

flowchart LR
    X(Primary Independent Variable) --> Y(Dependent Variable)
    C(Control Variables) --> X & Y

Dataset

You have the option to select an external dataset or choose one from the list of datasets I’ve compiled for this class. The compiled datasets are available in the Dropbox folder: Econ340 Datasets. Should you choose an external dataset, make sure it includes at least 100 observations and contains all the variables you plan to utilize. If you require assistance with formatting an external dataset, kindly email me at least a week in advance of your submission deadline. I would be more than happy to help you. See External Resources for some external data sources.

The Chicken and Egg Question

Deciding whether to first select your question or choose a dataset can feel like the classic chicken and egg dilemma. A question is infeasible without available data, just as data lacks purpose without a compelling question. In my opinion, the best approach is to first identify an interesting question, then consider the variables you’ll need to explore it, and finally search for datasets that may include those variables.

However, it’s not uncommon to encounter a situation where the data you need isn’t readily available. When faced with this hurdle, you’re prompted to reassess: Could you refine your original question based on the data you do have access to, or should you explore a different question altogether—one that aligns with measurable variables in the available data? This iterative approach underscores the flexible nature of research, where adjustments to the research question and the variables of interest may be necessary to accommodate the data you can actually obtain.

External Resources

Here is a list of publicly accessible datasets that may be useful:

If you need some inspiration to come up with a research question, here are some links to get you thinking:

Newspapers and other news outlets are also good sources to learn about current issues.