Collecting data and project introduction

• Administrivia:
  • reading necessary for exam is no listed in the schedule
  • Homework 2 description is on the web (link from schedule)

• Collect data for sample survey
  • Simple things like title might matter: "Penn. statisfaction survey" vs. Stress test
  • Order of questions
  • How you ask the questions
  • Forced choice vs free choice (more homework vs. less homework don't allow abstain)

• Landon (selection bias/non-response bias)
• TV marti (overhead)
• Self selection reader surveys (or wharton professor surveys)
• Simple random sampling
  • random digit dialing (bummer about all those wrong numbers)
  • random from phone book (bummer about unlisted number or two listings)
  • If you call only once (bummer about people who work)
  • Call back 3 times is genaral rule
  • Ideal, visit house if you don't get through on the phone

• Stratified sampling
  • Match population in sample
  • Get 1/4 of your sample from each year (freshman, sophmore, junior, senior)
  • If you match too much, you can't randomize anymore

Last modified: Thu Feb 24 11:26:50 2000