Final Group Project Instructions

Statistics for Social Research, Fall 2025

In the final project for this course, you will complete a data analysis project in a group of 2-3 students. This project will provide you with the opportunity to apply the methods and R skills you’ve developed over the course of the semester to analyze statistical relationships among variables in a dataset of your choice. Specifically, you will investigate the relationship between one outcome variable (Y) and one key predictor/explanatory variable (X).

You will select a dataset to work with, identify a compelling research question that the data could help to answer, and answer the question using that data. The final group project report should be 2000-3000 words.

Overview of Milestones and Deadlines

1. Oct 30, 2025 Sign up for group matching here (optional)

2. Nov 13, 2025 One-paragraph summary of project idea and names of group members due at 11:59pm

3. Dec 2, 2025 Drafts of 1) Introduction and 2) Data, sample, and key variables due at 11:59pm

4. Dec 13, 2025 Final report due at 11:59pm

5. Dec 13, 2025 Peer and self evaluation form. due at 11:59pm

Selecting a Topic and Dataset

As noted above, you will investigate the relationship between one outcome variable (Y) and one main predictor/explanatory variable (X).

Your selection of focal variables (X and Y) must be logical and well-motivated. As such, you are expected to find and cite at least two academic articles to i) develop and motivate a research question about the relationship between these focal variables and ii) pose a testable hypothesis about what the relationship might look like.

You may:

1. Use the Future of Families Dataset: for your convenience, I will upload a copy of the Future of Families Dataset to Brightspace. I will introduce this dataset in class. It contains a wide range of variables suitable for answering many sociological questions.

2. Choose your own adventure: you may also choose to locate your own dataset if you are interested in a specific topic. Here are some links to additional resources that you might find helpful:

a. Pew Research: https://www.pewresearch.org/download-datasets/ (mostly opinion polls)

b. 538: https://data.fivethirtyeight.com/ (sports, pop culture, politics).

c. Data is Plural: https://www.data-is-plural.com/archive/ (assorted datasets on all sorts of topics)

d. Directory of a lot of political science data: https://github.com/erikgahner/PolData

Please note that within your chosen dataset, your outcome variable (Y) must be an interval or ratio variable (e.g., a crime rate, hourly wages, annual income, number of protest events attended, continuous test scores, etc.) There are no restrictions on the other variables you may select.

General advice for choosing data sources (if not using Future of Families data)

● If you want to analyze the relationship between X and Y, make sure that these two variables are included in the data set.

● Ensure that the dataset has a sufficiently large sample size and that there are variables that you will actually be able to analyze with the skills you developed in this class. If you want to do more advanced data cleaning and/or merging, you’re welcome to do so, but it’s not required.

● Try to look for a ‘codebook’ or some other document that explains what the variables mean and how they are coded. Be especially careful with how different datasets record missing values.

● For some projects, preparing the data for analysis takes longer than the actual analysis itself. Try to find a dataset where you do not need to extensively recode / clean up the data before you run your analyses. This makes the final project easier.

● In a similar vein, if the data set is greater than about 50MB (this is not a hard cutoff), R commands and analyses tend to take longer.

Loading alternate data formats

You might encounter datasets that are saved in formats other than CSVs. R can load nearly all forms of data.

If you run into a data file with the extension .dta, run the following sample code (make sure you’ve successfully set your working directory first though):

install.packages(“haven”) # install if you don’t already have it

library(haven)

my_dataframe. <- read_dta(“name_of_file.dta”)

For other data file types, I recommend Googling the filename extension and “load into R” to see what function you might need to read in the data.

Group Formation

Groups can consist of 2-3 people.

Options for forming groups:

1. Choose your own group

2. Opt into group matching: fill out this form. by October 30, 2025

Please note that at the end of the semester, I will distribute a peer and self evaluation form. to better understand how your group worked together.

Requirements for Summary of Project Idea

(Due November 13, 2025)

Groups should write a one-paragraph note to describe what data set you will use and what your tentative research question is. Your research question should be one that you’re able to answer using a regression analysis. In this paragraph, you should do the following:

● State your research question.

● Formulate a hypothesis related to the research question. This hypothesis should be rooted in some sort of existing literature or theory. In other words, you need to present a plausible story why the hypothesis might be true.

● Describe your outcome variable and how it’s measured

● Describe your primary predictor/explanatory variable and how it’s measured

Please upload a document with this paragraph to Brightspace. Only one submission per group is needed.

Requirements for Final Project Report

Due December 2, 2025: Draft i) Introduction and ii) Data, sample, and key variables sections

Due December 13, 2025: Full final report

Submission Format

Please conduct your analyses and complete your write-up in a single RMarkdown File. You can start a new RMarkdown file from scratch by going to File -> New File -> R Markdown in RStudio. Please select the option to output to PDF.

The final RMarkdown file should load the data you have selected, run any preprocessing that you need to conduct (i.e., clean your data for analysis if needed), produce any summary statistics or plots of your focal variables, and conduct the main regressions of interest for the project.

The Rmd file and the resulting knitted PDF should both be uploaded to Brightspace. Only one submission is required per group.

Outline of Project Components

The following outline the key sections of your project. Please see me if you have any questions.

Introduction -- 5 points

1) Provide a description of the phenomena to be studied (i.e., the focal relationship, what’s the outcome variable, what’s the key predictor)

2) Briefly explain why we should care about the relationship under consideration

a) How does your project expand our knowledge of how the social/ political/economic world works?

b) Identify any potential policy implications of this relationship

3) Preview your data and analytic techniques (two or three sentences maximum)

a) What methods are you employing? Unless otherwise approved, everyone should be relying on ordinary least squares regression (i.e., linear regression)

4) Briefly highlight the central findings of your research. Please note that this component does not need to be included in the draft you submit on December 2nd.

Literature review, research question(s), and hypothesis -- 10 points

1) What does the existing literature/research lead us to believe about whether, why, and/or how X and Y are related? Please cite at least two academic sources.

2) Are there any alternative explanations that might explain why X and Y are associated? That is, what other statistical control variables will you need to include in your regression model if you are interested in how X impacts Y?

3) Given your brief review of the literature, clearly state your central research question(s).

4) Drawing on existing literature and/or your own ideas, state and justify one or more hypotheses about what you expect to see in your data analysis. Please make sure that your hypotheses are 1) directly related to your research question and 2) can be addressed with the data you have.

Data, sample, and key variables -- 10 points

1) Briefly discuss the data that you will be analyzing in the subsequent sections

a) What is the unit of analysis?

b) How, when, and where was the data collected? You may have to briefly review the dataset’s codebook.

2) How many observations are in your sample? If you’ve restricted the sample in some way, please indicate what you did and why.

3) What are the variables that you will be analyzing in the subsequent sections?

a) What are their levels of measurement? What kind of variables are they? What are the units that they’re measured in (e.g., miles, points, percentage)? You may summarize this in a table if you wish.

Descriptive statistics, plots, and interpretations -- 15 points

1) Calculate the appropriate descriptive/summary statistics for the focal variables used in your analysis (the outcome variable and key explanatory variable)

a) Frequency and proportion tables for ordinal/nominal variables

b) Mean, median, standard deviation, and range for interval/ratio variables

2) Report these in a professional-looking table (I will show you how to do this in R)

3) Produce a plot that summarizes the distribution of the outcome variable (e.g., a histogram)

4) Produce a scatterplot that illustrates the main relationship of interest. The outcome variable should be on the Y axis and the explanatory variable should be on the X axis.

5) Briefly interpret the tables and figures provided in this section in words. What do they tell you about the distribution of your key variables?

Regression and interpretation -- 25 points

1) Estimate an ordinary least squares regression analysis of your focal relationship (i.e., a simple linear regression model, Y ~ X)

a) Report the results in a professional table (I will show you how to do this in R)

b) Provide an interpretation of the results presented in your regression table including the model fit (R^2) and the regression coefficients (and their statistical significance)

c) Given your interpretation of results presented in the previous section, do you find evidence to support your research hypothesis?

2) What variable(s) do you need to control for to address confounding in the effect of X on Y? Provide a brief (1-2 sentence) reason for each control variable you think you should incorporate.

3) Estimate an ordinary least regression analysis of your focal relationship, but include the controls you reference above (i.e., a multiple regression model, Y ~ X + controls).

a) Report the results in a professional table (I will show you how to do this in R)

b) Provide an interpretation of the results presented in your regression table including the model fit (R^2) and the regression coefficients (and their statistical significance)

c) Given your interpretation of results presented in the previous section, do you find evidence to support your research hypothesis?

4) Briefly explain why you believe your results changed or did not change between the simple linear regression and multiple regression models.

Conclusion -- 15 points

1) Review the central purpose of your paper (to examine the relationship between X and Y)

2) Recap your central findings using clear and precise language. Be sure to connect your findings back to your primary research question. What overall answer do you arrive at? Evaluate the extent to which you find support for your hypothesis.

3) Critically assess the relationship that you examined by addressing each of the following points:

a) Evaluate whether your analysis can be interpreted as revealing a causal relationship. Discuss what additional forms of unobserved confounding might remain.

b) Identify at least one limitation of your research (all research has limitations!). You may consider any number of issues we’ve discussed in class, such as sampling, non-response, measurement, internal validity, and external validity.

c) Identify at least one way your analysis could be improved (e.g., what better data could be collected, what other research designs could better get at the question)