Homework 3

Author

Affiliation

Your Name Here

Old Dominion University

Before You Start

The section below gives you a situation and some data. Decide what is worth looking into, use what you have learned to look into it, and write it up as a short report for a reader who has not seen the data. Do not use any data that is not already provided to you. At the end, be sure to complete the Process Reflection.

What I have given you:

This week’s tools are the course methods I expect to see in your report.
An outline to use for the report. Note that the outline is only a suggestion, as you may find that the different sections bleed into one another, so you can structure your report however you’d like.
Some inspiration for the report. Again, these are merely suggestions, and the best reports will go somewhere we did not think of.

How to use AI: You may (and should!) use AI tools to help you throughout, but what you submit is your responsibility, not the AI’s. Read what it gives you, check it against the course material, and change whatever you would not stand behind. Never submit something you don’t understand, and make sure that what you submit is written in your own voice.

What to submit: Submit the rendered .html file on Canvas, along with the video described at the end. Your submission will be graded with the course rubric.

This week’s tools (Modules 1–6)

Multiple regression, with controls, dummy variables, and interactions. Interpret each coefficient in its units and say what is being held constant. Be sure to consider tools from previous modules, too.

Outline

The data: Describe the data. For example: what is one row? How many rows? Provide some summary measures of the data. What do the important distributions look like? What about some important pairwise relationships? Make note of the choices you made about the data (dropped, recoded, transformed) and why.
Analysis: Provide an analysis of the data which may include tables, figures, and/or models. Say why you include, and what you learn, from each.
Findings and discussion: In plain language, what can you conclude? What can’t you conclude? What do you think about the topic? What else would you want to know or have included in the data?

Racial Differences in the Labor Market

In 2001 and 2002, two economists sent thousands of fictitious resumes in response to real job ads in Boston and Chicago. The economists randomly assigned names to the resumes: some sounded stereotypically white and others stereotypically African-American. They also randomly varied the qualifications listed on each resume. The outcome is whether the employer called back. (data; documentation; the published paper; a non-technical memo.) These economists were interested in seeing whether there were differences in the probability an applicant was called back conditional on the race their name signaled.

To investigate how things may have changed over the next ten years, a second set of researchers collected data from the 2011 American Community Survey (ACS), a Census Bureau survey of roughly 1% of U.S. households each year. (data) Each row is an adult, aged 25 to 54, living in the city of Boston or Chicago, who is either non-Hispanic white or non-Hispanic Black. The variables (city, ethnicity, gender, college, military) are defined to match the resume data as closely as possible. The outcome is employed, which records whether the person had a job. Because the ACS is a survey, each person also has a weight, pwgtp, that says how many people in the population that row represents.

Some inspiration

Do the data contain evidence of discrimination (race, gender)? Does it differ by city? Does it differ by year?
What does each data set measure, and who is in it? Are the outcomes comparable?
What happens to the estimate(s) of discrimination as you add control variables? Does the same thing happen in both datasets? Why or why not?

Process Reflection

In your own words (Claude Code is not allowed to write this, and it knows that): how did you use Claude on this homework, where did its first attempt fall short, and what did you change or reject, and why?

Video Presentation

Record one continuous screen recording, about ten to twelve minutes for the presentation (no shorter than five, no longer than twenty) plus the two questions below, and upload it to Canvas. Your recording must show your entire screen (including the time) and include your voice. Showing your face is optional.

Part 1: Presentation (20 points). Present your report as if you were briefing a boss or client who has not seen it. Say what question you asked, what data and methods you used, what you found, and what you think it means and does not mean. You should keep your rendered report open while you talk so you can reference items like tables, figures, etc., but you should not read from the report.

Part 2: Questions (10 points). Immediately after, open a new Claude Code session (have it open in the background so you don’t waste time opening it during the recording) and paste the prompt from https://alexcardazzi.github.io/econ311/oral_check_prompt.txt, with your report’s file name filled in. Claude will ask you two questions about your report, as if it were a boss, client, or audience member. Claude may ask why you made a particular choice in your report, so be ready to explain it and to say how a different choice would have changed your results. If a question is unclear, you may ask Claude to reword or clarify it in your own words, once per question. Answer each one out loud, then type “next question” to move on. Keep your report open if you like, but your answers should show that you understand your work, not that you can read it back. Do not edit the prompt, and do not ask Claude to help you answer.

--- title: "Homework 3" author: "Your Name Here" institute: "Old Dominion University" format: html: theme: united code-tools: true code-fold: true code-summary: "Code" code-copy: hover link-external-newwindow: true tbl-cap-location: top fig-cap-location: bottom self-contained: true editor: source --- ```{r setup, include=FALSE} # INCLUDE THIS AT THE START OF EACH HOMEWORK # DO NOT EDIT THIS knitr::opts_chunk$set(fig.align = 'center') knitr::opts_chunk$set(out.width = '90%') knitr::opts_chunk$set(results = 'hold') knitr::opts_chunk$set(fig.show = 'hold') knitr::opts_chunk$set(dev='svg') knitr::opts_chunk$set(error = TRUE) par(mar = c(4.1, 4.1, 1.1, 4.1)) ``` ## Before You Start The section below gives you a situation and some data. Decide what is worth looking into, use what you have learned to look into it, and write it up as a short report for a reader who has not seen the data. Do not use any data that is not already provided to you. At the end, be sure to complete the **Process Reflection**. **What I have given you**: - **This week's tools** are the course methods I expect to see in your report. - An **outline** to use for the report. Note that the outline is only a suggestion, as you may find that the different sections bleed into one another, so you can structure your report however you'd like. - **Some inspiration** for the report. Again, these are merely suggestions, and the best reports will go somewhere we did not think of. **How to use AI**: You may (and should!) use AI tools to help you throughout, but what you submit is your responsibility, not the AI's. Read what it gives you, check it against the course material, and change whatever you would not stand behind. Never submit something you don't understand, and make sure that what you submit is written in your own voice. **What to submit**: Submit the rendered `.html` file on Canvas, along with the video described at the end. Your submission will be graded with the [course rubric](https://alexcardazzi.github.io/econ311/rubric.html). --- ### This week's tools (Modules 1–6) Multiple regression, with controls, dummy variables, and interactions. Interpret each coefficient in its units and say what is being held constant. Be sure to consider tools from previous modules, too. ### Outline - **The data**: Describe the data. For example: what is one row? How many rows? Provide some summary measures of the data. What do the important distributions look like? What about some important pairwise relationships? Make note of the choices you made about the data (dropped, recoded, transformed) and why. - **Analysis**: Provide an analysis of the data which may include tables, figures, and/or models. Say why you include, and what you learn, from each. - **Findings and discussion**: In plain language, what can you conclude? What can't you conclude? What do you think about the topic? What else would you want to know or have included in the data? ## Racial Differences in the Labor Market In 2001 and 2002, two economists sent thousands of fictitious resumes in response to real job ads in Boston and Chicago. The economists randomly assigned names to the resumes: some sounded stereotypically white and others stereotypically African-American. They also randomly varied the qualifications listed on each resume. The outcome is whether the employer called back. ([data](https://vincentarelbundock.github.io/Rdatasets/csv/AER/ResumeNames.csv); [documentation](https://vincentarelbundock.github.io/Rdatasets/doc/AER/ResumeNames.html); [the published paper](https://www.aeaweb.org/articles?id=10.1257/0002828042002561); a [non-technical memo](https://web.archive.org/web/20231107162800/https://www.ala.org/ala/hrdr/placementservice/hrpractice2.pdf).) These economists were interested in seeing whether there were differences in the probability an applicant was called back conditional on the race their name signaled. To investigate how things may have changed over the next ten years, a second set of researchers collected data from the 2011 American Community Survey (ACS), a Census Bureau survey of roughly 1% of U.S. households each year. ([data](https://alexcardazzi.github.io/econ311/data/acs_2011_boston_chicago.csv)) Each row is an adult, aged 25 to 54, living in the city of Boston or Chicago, who is either non-Hispanic white or non-Hispanic Black. The variables (`city`, `ethnicity`, `gender`, `college`, `military`) are defined to match the resume data as closely as possible. The outcome is `employed`, which records whether the person had a job. Because the ACS is a survey, each person also has a weight, `pwgtp`, that says how many people in the population that row represents. ### Some inspiration - Do the data contain evidence of discrimination (race, gender)? Does it differ by city? Does it differ by year? - What does each data set measure, and who is in it? Are the outcomes comparable? - What happens to the estimate(s) of discrimination as you add control variables? Does the same thing happen in both datasets? Why or why not? ## Process Reflection In your own words (Claude Code is not allowed to write this, and it knows that): how did you use Claude on this homework, where did its first attempt fall short, and what did you change or reject, and why? ## Video Presentation Record one continuous screen recording, about ten to twelve minutes for the presentation (no shorter than five, no longer than twenty) plus the two questions below, and upload it to Canvas. Your recording must show your entire screen (including the time) and include your voice. Showing your face is optional. **Part 1: Presentation (20 points).** Present your report as if you were briefing a boss or client who has not seen it. Say what question you asked, what data and methods you used, what you found, and what you think it means and does not mean. You should keep your rendered report open while you talk so you can reference items like tables, figures, etc., but you should not read from the report. **Part 2: Questions (10 points).** Immediately after, open a **new** Claude Code session (have it open in the background so you don't waste time opening it during the recording) and paste the prompt from [https://alexcardazzi.github.io/econ311/oral_check_prompt.txt](https://alexcardazzi.github.io/econ311/oral_check_prompt.txt), with your report's file name filled in. Claude will ask you two questions about your report, as if it were a boss, client, or audience member. Claude may ask why you made a particular choice in your report, so be ready to explain it and to say how a different choice would have changed your results. If a question is unclear, you may ask Claude to reword or clarify it in your own words, once per question. Answer each one out loud, then type "next question" to move on. Keep your report open if you like, but your answers should show that you understand your work, not that you can read it back. Do not edit the prompt, and do not ask Claude to help you answer.