Overview:
In week 4 you worked with a data set related to advertising/marketing at car dealerships (209 observations with simple dummy and categorical dummy variables). Specifically, you were provided with data on what each car dealership spent on marketing as a % of its gross sales. You were also given information that described the dealership. You will work with your machine-ready data set and will the concepts of hypothesis testing to see what can be learned from this inferential statistics approach.
This project is a bit structured, in that there are specific questions to answer. You should still craft a report that is visually appealing, well organized, incorporates the technical writing tips covered, and that has an accurate and well explained analysis. The visual look and organizational structure is up to you. But be clear that it is a report for a client and not a set of homework questions. So be sure to craft it that way.
Data/Coding Expectations:
So that it is easy to follow your work, you were asked to follow some guidelines when crafting your machine-ready data set. Before starting Data Analysis #1, you need to be certain that your machine ready data set is okay. You have been provided an document that allowed you to confirm (see announcements). Please make revisions to your data set if necessary before starting on this project.
Always code YES=1 and NO=0.
Use the variable names as given to you in the spreadsheet. If you changed them, please change them back to the original labels.
For now, retain all categories in the categorical dummy variable and we will agree which to omit later, when we move on to project #2. (This makes it easier to put into subgroups.)
Audience:
You are serving as a consultant for the CAPITAL GROUP, a company in Northern Virginia that works with dealerships, sometimes acquiring them and sometimes providing them financial capital. They do have some entry-level analysts on the staff, so you are mainly speaking to them. As such, you need to be clear about what you are analyzing and how you are doing it (ex: what hypotheses are you testing, what level of confidence are you reporting for results, etc.). You may assume that the analysts can follow the statistical work if you lay it out well and are explicit about what you are doing. However, key takeaways should still be noted so that the report is accessible to others who are not statisticians.
Your Task:
1. Your client is not familiar with the data set that you are working with. They want you to start with a discussion of descriptive statistics so that they may get a better sense of the data. They are also curious about which variables might be strongly related to each other. Discuss this sufficiently.
2. Your client would like to see a box plot for the Marketing % variable to get a sense of how it is distributed, whether it is skewed in one direction or not, whether there are outliers, etc. Discuss this sufficiently.
3. Your client would like to see 2 scatter plots: a) Marketing % vs. Distance to lower tax, and b) Marketing % vs. Size (employees). These should be accompanied by a good explanation of what they show.
4. Your client would like you to test some hypotheses. They are interested in knowing if there is a statistically significant difference in the mean of Marketing % (portion of gross sales spent on marketing) when looking at certain subgroups of dealerships. The client has preconceived notions for some hypotheses but not others. You must provide the relevant hypothesis test (using appropriate notation) for each.
Is the mean of Marketing % statistically different when comparing dealerships that are locally owned with those that are not locally owned? (Your client has no prior expectation.)
Is the mean of Marketing % statistically different when comparing dealerships that have 32 or more employees with those that have 31 or fewer employees. (Your client has no prior expectation.)
The client assumes that the mean of Marketing % is statistically higher for dealerships in cities than it is for dealerships in rural locations (ignoring suburbs) Confirm if this is/is not true.
The client assumes that the mean of Marketing % is statistically higher for dealerships without road frontage than it is for dealerships with road frontage. Confirm if this is/is not true.
REMINDER: Addressing the hypothesis test questions above requires that you create the relevant subgroups. A quick way to do this is to sort by the dummy variables and then copy and paste into a clean page on your spreadsheet. Name the pages appropriately. (For example, if I have test scores for students, I can sort by a dummy variable called MALE. That means all the males would be together and all the females. I can copy and paste values to a new page.) THERE IS A VIDEO ON SORTING AND ON PASTING VALUES IF YOU NEED TO REVIEW THIS.
4. The client wants to know what the analysis above allows them to conclude about causal stories (cause and effect). In your CONCLUSION, make sure the client understand what they can and cannot conclude from descriptive statistics and hypothesis testing that has been conducted.
In this report you must:
Offer a good discussion of descriptive statistics that informs the Capital Group.
Craft effective graphs and accompanying discussion that inform the Capital Group.
Perform the desired hypothesis tests (t-test assuming equal variance and not “paired” observations bc that is something different) in Excel.
Write out your hypotheses for each test that reflects the expectations communicated. They must appear fully specified in the paper in notational form.
Create well-constructed tables of Excel results in the paper. (You can decide how to organize the tables.) Remove the stuff that is not relevant for your test (like 2-tailed results if you are running a 1-tailed test).
Be clear about level of significance you claim, if you are claiming a significant result.
Interpret your results well, with appropriate language and correct conclusions.
Offer an introduction and conclusion (with conclusion that addresses the question above).
Follow technical writing guidelines covered in lecture.
Use space efficiently, being as thorough as possible in the space permitted.
Use precise language and show mastery of the topics.
Ensure the report is visually appealing (not too much white, not too busy, nicely organized, attractive).
Organize your paper in a manner that promotes readability but also allows for easy grading (we can find your parts).
Edit carefully to ensure paper sounds professional, with minimal grammar and spelling errors.
Please do not include analytical constructs that were not covered in this course. Your content must be tied to modules from Econ 560.
Submit both the PDF and Word versions.
Guidelines:
Your report may be no more than 4 pages and coaches will not read beyond 4 pages.
Submit your report single-spaced but with a blank line before and after tables, graphs, etc. to ensure visual appeal and readability.
Use TIMES ROMAN 12 point font.
Use 1″ margins.
Have your name in the HEADER (only there) of your paper, not taking up space. Please do not include a cover page.
Use an organizational structure that makes your paper readable. Please align sections with the questions above for ease of grading, with analysis in the order indicated above. This may require headings and subheadings, but the design is up to you.
Use good technical writing techniques, especially in reference to tables, discussing takeaways, etc.
Speak to the audience identified.
Report should be well edited and attractively organized.
You do NOT need to cite any data source.
While content will be given more weight, organization and writing quality, and ability to follow instructions contribute towards your grade.
Grading Rubric
Your assignment will be graded according to the grading rubric that should be visible below the assignment. Be certain that you have read instructions and have followed them carefully.
Submission Instructions:
You will submit both PDF and Word versions in separate assignments. You must submit both and they must be identical.
Turnitin will be employed. (We do understand that certain basic parts may match others but as this is open-ended, we should not see high levels of matches.)
The due date is communicated in the assignment. This is EASTERN time zone.
If you have second thoughts and decide you need to resubmit, you will have that option. Canvas will show us the most recent version in grading mode while retaining all others. You do not need to alert us.
AFTER the deadline, please do not resubmit. Canvas will apply automated late penalties.
Individual Work/Expectations for Academic Integrity:
This is to be individual work.
You may not provide or receive assistance from others.
You may not utilize any resources such as ChatGPT or CourseHero or any other resources that is not indicated as acceptable.
You should not ask others for help. You should not accept help from others. This includes looking at past years’ papers, receiving someone else’s completed assignment, or having others review your work.
Your paper may not incorporate anything found on the Internet (although not likely there is anything helpful for this project).
The expectation is that the paper you are submitting is 100% your work and that you have not acted in a manner that gives you an advantage over others.
The penalty for violating the rules for individual work are stated in the syllabus and they are severe. Please do not put yourself in a position to receive this penalty.
Reach out for any clarifications needed.
Scientific Notation:
Excel might give some results in scientific notation using E notation. So, you need to understand how to read the results. See the information posted to the bottom of the module if you are unfamiliar with this.
However, it is best to get rid of scientific notation, meaning that some results might be reported as .0000.
Data Analysis Toolpak & HELP Videos
You will need the Excel Data Analysis Tookpak installed. If you have not done so, you should do that NOW. (You probably already have it since the preliminary module had you install it.) There are videos on what to do to install and where to find it once you have. See the preliminary module prior to Week 1.
If you do not have software with the Toolpak, contact the LU HELP DESK at 434.395.help. I believe that students can get software for free. Keep in mind that you have known about the requirement since the start of class.
Once you install the Toolpak, you have to go to your DATA tab in the ribbon and then click on it. I see my DATA ANALYSIS button on the far right. This brings you to a menu where hypothesis testing is one of the options. (I also see some data analysis tab on the main ribbon but I think this is different and students have had issues with that. Be sure you are in the right place.)
There are videos to help you with doing hypothesis testing in Excel.
Questions/Issues & Answers:
What is the convention for rounding? It is common to carry 4 digits beyond the decimal, but we do not want much more than that.
Can I express in scientific notation? The convention is that you remove scientific notation (video on how to do this is available in the modules). This does mean that some numbers will be reported as .0000 which makes clear that it is very small.
How should I handle tables? This is up to you. They must look professional and you can organize them in any manner you wish that might create efficiency. Unattractive tables will likely reduce your score for visual appeal, so keep that in mind. Technical writing skills still apply, so be sure you are paying attention to that content.
Does it matter if I use one vs. two tailed? Yes, it does. You have been given information from the company that suggests whether one or two-tailed is appropriate. Your hypotheses should reflect this.
What happens if I go over in pages? The coaches will stop reading at the end of the communicated length and you will not get credit for anything past that point. Also, there will be a deduction for not following instructions. Please stay within the length limit.
Can I do extra analysis? Please do not. We want you to do your best work for the questions assigned, and that will take up a lot of space. Please stick to what has been asked.
I am running out of room. What can I do? Use your space efficiently. There is a video about putting tables/graphs side by side, and you can wrap text as long as it is tidy and professional in appearance and not too smooshed together.
I think my data set is messed up. Remember that you still have access to the original data set in an earlier module. It will not take long to go back to that and recreate your dummy and categorical variables.
Does data need to be next to each other for hypothesis tests like it did for descriptive stats? It looks like Excel will ask you to input the range for the first data range and then the second data range separately. That should allow you to have data in whatever order you want for hypothesis tests. That will not be the case when we do regression for project #2.
