Some material we are covering in this class cannot be found in the book.
I will do my best here to put copies of the material in this section.
Matched Pairs - Conducting
hypothesis tests between two samples where the same individuals
appear in each sample. Similar to Chapter 11, but described in a
better way.
Proportions - Conducting
hypothesis tests with proportions rather than means. Related to chapter
19, but more comprehensive.
Estimation - Confidence intervals
and the concept of estimating parameters such as the mean of a population.
A compressed version of the notes (because the original scanned version
is 9MB) is here.
Project Information
The structure of the project is as follows: try to find questions outside of mathematics that
you are interested in answering with statistics. Anything is fair game, and I'd like you to
present ideas to me in my office by Friday April 1. Try to set up an appointment so I
can definitely be around for long enough to chat. Bring about 10 related questions that you'd
like to answer, and we'll try to cut that down to 3 or so.
You should expect to spend about 10 hours working on this project, depending on how long it
takes you to write things up. That time might break down as
2-3 hours: Preparation, designing questions, finding data sets
2-3 hours: Analysis of your data, answering questions
4-6 hours: Writing up your results and presentation
You can find previous projects listed here. The writeups here are maybe a little more involved
than I necessarily expect, but these were solid.
I have also created a sample project writeup, linked below. This took me exactly 1 hour, and
is indicative of the amount of work I expect. The quality of course would be higher with more
time to edit, but it's still a good example.
Potential sources of data can be found at the
bottom of this page, although you certainly shouldn't limit yourself to this.
Data Set Links
For those of you who are having trouble coming up with a project idea,
feel free to check out available data sets and see if something
interests you. I would of course consider it better for you to work on
a project related to your research, but if you're not working on
anything right now, this might be a useful source of material.
Once again, I cannot encourage you enough to start looking for a data
set (or compiling your own) early. Many great ideas for a statistics
project have been foiled by lousy data. Try coming up with 10 questions
you want to answer, and the looking for the data sets to do so. Once
you've found potential answers to 5 of those 10 questions, you're
probably good to go.