I am going to work with Amy Bruckman on the MOOSE crossing tutorial mini-project.
Most of the work done on this project have been fairly simple statistics and frequency analysis. There are important problems that require difficult statistics to solve.
Probably the question most people want the answer to is "What makes someone more likely to spend money on the web?" and "What demographics would be most profitable to target?" To answer these questions we plan do a regression analysis. This will suggest certain factors that predict buying behavior on the web, which will be acomposite variable made up of the answers to several questions. With demographic data I can do ANOVA and other mean testing to predict which demographic groups spend most on the web. The findings of these tests will suggest other questions to explore. This will probably take us over 20 hours. But just in case, there are a few other questions we would be interested in looking at.
We would like to find out what predicts web satisfaction. Does the operating system have any effect? How about age? Are those who use commercial services more satisfied? How important is connection speed? Which groups are more likely to use a catalog as opposed to the web?
deliverables: An HTML paper for the WWW survey website of publishable quality describing our statistical tests, hypotheses, and conclusions.