Radon Lab

Our goal is to provide estimates for the proportion of houses in the state that exceed the EPA recommended action levels. Begin to answer the following two sets of questions using the data provided in the spreadsheet data/Radon.gnumeric

1. State-wide estimate.

• Use the number of houses in the sample with radon levels that exceed 4 pCi/l to estimate the proportion of houses in the state that exceed this level.

  Keep in mind the sampling scheme when creating the estimate. That is, in some counties one sampled house represents 500 houses and in other counties one house represents 10,000 houses.

• Attach standard errors to your estimate, or provide an interval estimate for the proportion of houses in Minnesota that exceed 4 pCi/l.

2. County-level estimates.
Next consider making county estimates for the percentage of houses that exceed a certain level of radon concentration.

• Many counties have as few as four or five observations. For a county with only four observations, the estimated proportion of houses in the county to exceed 4 pCi/l can only be either 0, 25, 50, 75, or 100%, according to whether 0, 1, 2, 3, or 4 houses exceed the level. Can you use the information for the entire state or for neighboring counties to help provide a more reasonable estimate for a county with a small sample?

  Pick one county with few observations, and modify its estimate according to the procedure you described.

• How would you use the sample survey results for each county to estimate the proportion of households in that county with radon levels that exceed 4 pCi/l? (You may choose a different level other than 4pCi/L if you wish. Be sure to explain why you chose that level.)

  Using estimates for each county, color or shade a county map of Minnesota, indicating at least three levels of indoor radon concentration.

  Use the Gui colorMap() in R to do color the map. To use the colorMap function do the following in R:

  • library(Radon)
  • mystuff <- scan("myData.csv")
  • colorMap(mstuff)
  • Refer to the handout for how to use the GUI.

Wrap-up What do you conclude about the geographical distribution of radon concentration in Minnesota? What factors limit the validity of your conclusions?

You work for the State Department of Public Health in Minnesota, and your job is to summarize your findings in a report that will be reviewed by the head of the Department (your boss's boss) who will use your report (and other information) to determine what actions to take in informing the public of the potential radon problem (or not).

The report should be at most 2-pages in length. It may include a technical appendix as part of the 2-pages where you justify the methods that you used in determining your estimates and coloring your map. In addition you are to turn in:

• County map
• Table of your estimates
• Gnumeric worksheet that shows your calculations (on the 87 county values -- not all 1003 values).

Read Appendix A for instructions on how to prepare your written report. Pay particular attention to the suggestions for organizing your report (intro, methodology, conclusions, etc). Be sure to provide references if you use material from the book or other sources.