Check-In 5 - Multiple Regression and Memo
Multiple regression lets you examine whether two more or more explanatory variables explain the variable you are interested in, simultaneously. Or an equivalent way of thinking about this, that has a different intuition, is that it allows us to examine whether an explanatory variable we are interested in explains the variable of interest after controlling for (or holding constant) variation in other variables. Using the simple regressions you ran in Check-In-4 as a starting point, run a multiple regression with two or more variables on the right-hand side. In this regression, you are trying to get to the heart of the questions listed in Section 1. Run and tabulate this regression. You have to decide what tells the story in the clearest and most compelling manner.
Now take a step back and write a paragraph explaining your results
Finally, pull your results together in a memo (see next section for further details).
Deliverables
The final memo should be one to two pages (including your table, and a figure if you choose to include one). Adhere strictly to the page limit --- your audience needs to be briefed briefly!
This should be a doc file or PDF. Also attach your Stata do file (clearly annotated) to the assignment submission, as a separate file.
Use a 12-point font and a maximum of double spacing. Check-in 5 is due no later than 6:00pm on December 17.
Points to Consider in Writing the Memo
Your audience
Write at a level that this audience can appreciate. You are addressing people who are familiar with basic numerical concepts (averages, percents). However, most are probably not aware of concepts like hypothesis testing, regression coefficients, and so on – after all, you didn’t know about these until you took Stat 1. Such technical terms do not belong in your memo, unless you also explain them in plain English. A big part of your assignment is to convey your findings, using words that your audience can understand.
Tips on organizing and writing a memo
Typically, memos of this sort are organized as follows:
1. Introduction. What are you going to talk about? Why are these issues important? (Note that because this is a statistics class rather than a policy course, this will be just one short paragraph.)
2. Methods: What is your data source for addressing these issues? Very briefly, how did you analyze the data? (Again, a very short paragraph, for this class.)
3. Findings: What did you find? (Normally, this section will be the bulk of your memo. For this class, it will comprise one table and perhaps one figure, and one interpretive paragraph). Here are a few tips:
Do
• Include one good table and perhaps one good figure.
• Use this section to tell the story that was revealed in your analysis.
• Highlight the most important and interesting findings.
• Tell important things first and save details for last (or omit them altogether).
• Use bolding and bullets to separate elements of the story, as needed.
• Draw your reader’s attention to important trends in the tables. For example, “As shown in Table 1, Group 1 and Group 2 differed on the following characteristics …”.
Do not
• Bury important findings in minutiae
• Use space in your “Findings” section to recount every statistic in your table(s). In other words, be selective!
4. Conclusions: How do your findings address the question that you posed in the introduction? What do you conclude? Briefly, what are the limitations of your analysis? (For us, this is one paragraph.)