QM222 Nov. 28 Presentations Some additional tips on the project For presentation schedule, see: https://docs.google.com/spreadsheets/d/1pcfaSpsS6 TISPccPsJf7ykuhTIzqsWN7OqURMO_BLM4/edit#gid= 0 QM222 Fall 2016 Section D1

Reminder: Makeup Test Makeup on the parts of the test about omitted variable bias/causality v. correlation: Part I Question 5(both parts) and Question 6(first part) plus Part III, together worth a total of 24 points) Friday December 9th during “section” Or: during the QM222 final (Sat night Dec. 17) Makeup on the entire test: Sat night Dec. 17. You will get a special version for our section. Test Grading If you got less than 84 on the original midterm, you will get the better of the two tests. If you got 84 or more on the original midterms, you will get the last grade if it is the best. If the last grade is worse, you will get the average of the two grades. I will pass out a signup on Wed December 7 for the December 9 test. (By then, you may have an idea about your project score.) I will pass out a signup on Mon Dec 12 for one of the two tests during the final. QM222 Fall 2016 Section D1

Today’s Presenters Monique Rans Sho Nihei James Stuart Kyle Parsons Ariel Lavi Clarification questions allowed during the presentation. (I will clarify things to class during presentation if I think it is needed.) 5 minutes Q/A after the presentation: Save your major comments or questions until then. I will give you additional suggestions in writing. QM222 Fall 2016 Section D1

More tips on the project: Regression table Use real understandable labels in the regression table, not the variable names from the regressions You should not include the Stata regressions OR typed out equations in the text or the Appendix – They are redundant. I was looking at some of the examples. They were old and many did also add in the Stata regressions etc. I have now posted new sample projects from last year. In the table, include SEE, AdrR2, #obs at the bottom. Include coef’s t-stats OR se’s in parentheses under each coefficient. QM222 Fall 2016 Section D1

How long should it be? The discussion of results should be the bulk of the project. You should discuss what the client can learn from the regression in detail. If you think your text is still too short, think about how you can use the results to develop scenarios or predictions that would be useful to the client. QM222 Fall 2016 Section D1

Write professionally Aim the topic at the lawyers (your clients) not me. Describe the analysis and results as intuitively as possible, without using a lot of statistics terms except when absolutely necessary (and perhaps in parentheses or footnotes.) Readers will understand: “controlling for” (or holding X2 and X3 constant) and “statistically significant” (or “we are more than 95% confident that” this variables increases salaries) or “explain only 2% of the variation in Y”.. Most will NOT understand words such as t-statistic, p-value, confidence interval, adjusted R-square. Do NOT explain these terms to them– they are not taking a statistics course. They just want the answer. Do NOT write about your process of what you did …. (first I did this, then I realized that… etc.) Most things should be written in third person such as: This report investigates whether …… As can be seen in the first columns of Table 2, …. QM222 Fall 2015 Sections C1 and F1

How important is each variable? The t-stat tells you if the impact of the variable might be zero, i.e. if it is statistically significant. How can you tell how much each variable contributes to explaining the variation in Y? In other words, is the variable important? Does it make a meaningful difference. I can suggest two ways you can do this. Using your best regression, drop that variable and see how much the adjusted R-squared changes. For each variable, multiply coef * (max X – min X), where the maximum X is the maximum in your sample (same for min). This is the largest change in Y that this variable can be responsible for. When you have a quadratic, a “variable” includes both terms of the quadratic. QM222 Fall 2016 Section D1