Group B Report: Facebook A/B Testing Analysis

We want you to have some experience conducting analysis on your own. Real data can be hard to find, though, and is often messy, so I've simulated a number of datasets for you to analyze.
In a folder on the Classes website there are a number of text files and matching csv files (a spreadsheet filetype). Each text file describes a (hypothetical) scenario, and the matching .csv includes the related data.
For a research report, you pick a dataset and conduct the appropriate analyses on it with the statistical software of your choice (or by hand if you really want to, I guess). We'll be covering some of this software in Week 2.
You will not be told exactly which test(s) to run. Instead, you will determine this for yourself based on the scenario in the text file. The end product should be a document in which you describe the analyses you conducted, report the test results, and draw reasonable conclusions.
The report has the following requirements:
One page long. Longer submissions won't be accepted. I want you to learn to be concise!
Twelve point font, 1.5-line spacing.
Report your statistical analyses in (approximately) APA format. We need to be able to understand it, but formatting isn't the point of the assignment. Please round your numbers to 2 or 3 digits as appropriate. We won't take points off if you don't, but it makes the reports easier to read.
Some sort of graph or plot is not mandatory, but is strongly encouraged.
Discuss, at least briefly, your conclusions: say what the test results mean for the (hypothetical) study. Also mention if you notice anything strange or have any concerns about the results.
There is also a rough rubric, which you can find on the NYU Classes site. When you are done with a report, email it to Graham as a Word document or PDF to receive a grade.


You are required to do two of these research reports at some point during the class. You can do these two reports at any point during the class, but since you won't be taught how to do most analyses until about halfway through, most of you will probably want to do these assignments during the second half of the course.
The research reports are divided up into two groups, Group A and Group B, and these groups are in separate folders on NYU Classes. You must do one research report on a Group A scenario and one research report on a Group B scenario. This is to help make sure you get the chance to practice your skills on different kinds of problems.
If you are not satisfied with your grade on one or both of your research reports, you may do a replacement research report on a new dataset. You can only do a replacement report once for each research report. We want you to be able to learn from your mistakes but also I don't want you each doing 10 of these. If you do a replacement for your research report on a Group A scenario, the replacement should also be (a different) Group A scenario, and a replacement for a research report on a Group B scenario should be on another Group B scenario.
Essentially this means you can do up to four of these (two from each Group) and your grade will be based on your best from each Group. Obviously if you're happy with your first two, you don't need to do four. Please don't send us more than four, the extras won't be graded.