You've probably heard the advertisements or seen the commercials stressing the importance of completing the Census 2000 questionnaire: "This is your future.  Don't leave it blank."

The census is how America knows what America needs (U.S. Census Bureau, the Official Statistics).  Every ten years the Census Bureau mails questionnaires to each household in the country.  The completed questionnaires yield valuable information regarding the population.  This information helps lawmakers, legislatures, and other decision makers understand the needs of the communities in which we live.  Without this information, the amount of money the government spends on services, such as education, transportation, and the elderly, may not be enough to support the need. 
 

In this activity you will select two quantitative variables on which you will gather information from each of the fifty states.  You will use the World Wide Web to research all relevant information; it is not necessary to use census data.  You will observe the relationship between these variables in terms of the strength, form, and direction of the scatterplot.  You will model this relationship using least-squares regression techniques discussed in class.  Finally, you will summarize your findings in a well written report, which you will present to the class upon completion.
 

Your report must fulfill two goals:
(1)     Produce a professional report of research findings using technology.
(2)     Demonstrate your understanding of the topics related to least-squares regression as discussed in class.
 

• Choose the variables.

You must choose two quantitative variables that are of interest to you with respect to the fifty states.  The data should not have an expected or exact relationship which could easily be calculated beforehand.  However, if you do not expect some moderate to strong association between the variables, you will not be able to determine a suitable model.

Click here to view examples.

• Collect data.

Once you have chosen your variables, you must use the internet to find data on each state.  Because Census 2000 is not complete, you may want to view other online sources, such as almanacs and directories, to gather your data.

Click here to view possible web sites.

• Generate a spreadsheet.

Once you find the data you need, copy and save it to an Excel spreadsheet or within the Graphical Analysis program.  When pasting your data in any spreadsheet, be sure that each value is linked to the appropriate state.  For instance, one variable may display data in ascending order while the other variable may be displayed in alphabetical order by state. 

• Construct a scatterplot.

Construct a scatterplot of your data, being certain to include appropriate labels and scales on your graph, as well as a title.  Observe the relationship between the two variables, and describe this association in terms of its strength, direction, and form.  Identify any influential observations or possible outliers. 

• Determine a model.

Calculate an equation for the least-squares regression line that models the data in the scatterplot.  Interpret the slope and y-intercept of this model with respect to the variables.  How appropriate is this model?  Would it be reasonable to use this model to make predictions?  Does there appear to be an explanatory-response relationship between the variables?  Use the correlation coefficient and r² values to answer these questions.
 
• Summarize your results.

Summarize the results of your analysis in a well written report, which will be presented to the class.  The report should have each of the following sections in the order given:
 
• Description of variables.

You should include an explanation of your choice of variables and documentation of internet sites as well as other references, if necessary. 
• Data presentation.

Computer generated tables and graphical representations of the raw data are expected without interpretation.  Clear labeling is expected. 
• Linear model.

The linear regression equation used to model the relationship between your variables should also be included in your report.  Be sure to define the variable labels.  You must interpret the slope and y-intercept of your model with respect to the variables.  You should also discuss the appropriateness of the model in terms of the observed association and the correlation coefficient and r² values.
• Conclusion.

Summarize the results of your research.  Discuss how your findings may help lawmakers, legislatures, and other decision makers understand the needs of the communities in which we live.   You may want to point out how Maryland ranks with respect to the other forty-nine states.  What recommendations or suggestions do you have, if any, to improve or maintain the state's rank?

Click here to view a sample summary.

You will be evaluated on your final report and presentation using the rubric below.
 

This activity was designed to enable you to use technology to search for information and to present research results in a professional manner.  In this activity you also furthered your investigation of least-squares regression by computing an equation to model your collected data.  For many statisticians, this type of research and analysis is routine.  Hopefully, you will find this to be a valuable lesson.

Most of the introduction came from the U.S. Census Bureau.

I'd like to thank Kim Champagne for saving this file from mass corruption.

Return to the top of the page