Categorical data and X vs Y plots

Administrivia:
- If you haven't emailed your proposal, please do so soon. The next deadlines aren't as forgiving (lateness will be penalized--since this is for zero points--no penalty will be extracted!)
- Feel free to turn in your survey form on Thursday. We'll let you know of any problems earlier if you do so. You'll be able to start collecting data earlier.

Holocaust question survey: 65% are certain it happened, vs 91% are certain it happened. Which is correct? (see overhead)
Be sure to fill out your own questionares. You can catch some of these difficulties.
What about sensitive questions?
Randomized response, a justification for obtuse questions.
Have you used drugs (i.e. committed a felony) or were you born in a month which doesn't have the letter "r" in it?
Hint: May, June, July
Suppose 1/2 of the people say yes. Then: 1/3 said yes because of birthday. 1/6 said yes because of felony. 1/2 said no because of wrong birthday AND no felony.
We are only interested in those NOT born in May, June or July
Conclusions: 25% have committed a felony.
Better still would be to have them roll a die

Dad's nutrition study of Bagbanna village
weight vs. age for boys and girls (ages 8 - 12)
two regression lines? One?
In other words, which grows faster
Need to do a combined regression to decide. So MUST do one multiple regression first.
No colinearity: Ages and sex are unrelated
Do example

Last modified: Tue Mar 21 12:37:48 2000