Class Project

Project 1

Select a sample of 100 voters (from a population of 480) to determine support for a mayoral candidate. You may choose either a random sample or a stratified random sample. Possible stratifying variables include race, gender and age. It is anticipated that at least 75% of African-Americans, 30% or less of Caucasians and approximately 50% of Hispanics support the mayoral candidate. Across all ethnic groups, support among older voters (>50 years) will be much greater than support among younger voters and support may be slightly stronger for females than for males. You have a Minitab Portable Worksheet of all registered voters including their ethnicity, age and gender.

Project 2

A city needs to determine the total actual income (this can only be determined through an audit) of its 500 taxpayers. It hires an auditor for $1000 to estimate total actual income and provides the auditor with a Minitab Portable Worksheet of reported incomes of its taxpayers. The cost of sampling records for different income classes appears in the table below. You should also anticipate that the variability in the reporting errors should also increase with reported income.

Actual Income Cost of Audit
$0-29,999 $10
$30,000-39,999 $20
$40,000-49,999 $30
$50,000-99,999 $50
$100,000-149,999 $75
$150,000 or more $125

Project 3

You want to project Lexinton County's average 1996 traffic intensity from the 1994 traffic intensity study using a sample of 30 intersections. Data from the 1994 traffic intensity study is available in a Minitab Portable Worksheet . The worksheet also contains the roadway associated with each sampling site since you may want to stratify along certain high-growth corridors (i.e., US 1, SC 6, US 378, I-20 and I-26). The variance in each stratum should be proportional to the average traffic volume in each stratum.

Part I

Choose any one of the three projects and develop a sampling plan to obtain the most efficient estimator possible. By April 2, you need to communicate to me the size of your sample (when relevant), your stratification scheme (if used) and the allocation of sample points among your strata. I will conduct the sampling and send you your data, which you will then analyze. A detailed explanation of your decision-making in the design phase can be postponed until the final report.

Part II

The major focus of your written project should be an explanation of your sampling plan rather than a discussion of the analysis. I anticipate that your project should include a brief Introduction, a discussion of your Sampling Plan, Data Analysis and Conclusions. Your project grade will be based on the quality of your sampling plan, your written report and the precision of your estimator. Here are some questions to keep in mind while preparing the elements of your written report:

Sampling Plan

Is the decision to stratify or not stratify explained? Is the selection of the strata and strata sample sizes arbitrary or is the selection process explained carefully? Are decisions supported by data analysis? Are any important ideas omitted or misused in the selection of strata and strata sample sizes?

Data Analysis

Are point estimates and error bounds produced? How do the error bounds compare to bounds estimated when constructing the sampling plan (if relevant)? How do actual costs compare to predicted costs (if relevant)? Are the results presented in isolation or are they related back to the original goals of the experiment?

Format

Are there serious mechanical errors in the written report? Are graphs and tables easy to refer to? Are there any major stylistic problems (poor organization, lack of transition elements, excessive use of passive voice)?

University of South Carolina Department of Statistics Return to Department of Statistics Home Page