Steps (all statistical analysis to be done in Excel and/or StatCrunch):
1. Watch the TED talk by Hans Roling titled The best stats youve ever seen. You will need to include comments on this in your paper. Here is a link: http://www.ted.com/talks/hans_rosling_shows_the_best_stats_you_ve_ever_seen?language=en
2. Download the Excel data from IvyLearn
3. Create histograms of each of the variables (one histogram for fertility, one for life expectancy). Use the histograms to identify the shapes of the distribution. StatCrunch will be the easier tool to use for this particular task. If you use Excel, you are required to use a histogram plug-in or bin the data yourself (not an easy task).
4. Calculate some descriptive statistics for each of the variables, including but not limited to the mean, median and standard deviation. Organize these numbers nicely in a table.
5. Using fertility as the predictor variable and life expectancy as the response variable, create a scatter diagram, come up with the least-squares regression line (you need to state the actual equation for the least-squares regression line) and calculate the linear correlation coefficient as well as the coefficient of determination. Make sure that you understand all interpretations and include them in your paper. Please carefully review the rubric below to see the full list of required interpretations.
6. Use the regression line to predict life expectancy for the United States given fertility and then compare this to the actual value in the United States. Hint: you will have to plug the actual fertility value for the US into your equation.
7. Name some possible lurking variables that may be at work here.
8. Explain the difference between correlation and causation and why we cannot say that there is a cause and effect relationship in this situation.
9. Explain why we cannot use our regression model to predict the life expectancy of one particular individual.
10. Take a look at the website where this data was pulled from and comment on how the model might have been different if we used the data from 20, 40 or 60 years ago. Navigate to http://gapminder.org and click on Tools. You should see a bubble chart appear. Use the x-axis and y-axis dropdown menus to ensure that life expectancy (years) is selected on the y-axis and children per woman (total fertility) is selected on the x-axis.
11. Put everything together into an organized paper (in a Word document or PDF) and submit on IvyLearn.
Graded Item Points Possible Points Earned
Organization/Formatting
Paper is well organized with clear section headings 5
Paper contains data in tables and displayed in graphics when appropriate 5
Graphical Representations/Descriptive Statistics
Histograms for each variable 10
Comments on shape of distribution for each variable 5
Descriptive statistics for each variable, well organized 10
Regression
Scatterplot 5
Least-squares regression line 10
Linear correlation coefficient 5
Coefficient of determination and interpretation 5
Interpretation of slope and y-intercept 10
US prediction and comparison to actual value 5
Additional Discussion/Thoughts
Lurking variables 5
Correlation/causation 5
Why we cannot use for individual predictions 5
Commentary on how model would have been different in the past 5
Tie-in of Hans Rosling TED talk 5
BOOK:
Title: Fundamentals of Statistics
Author: Sullivan
Edition: 5th
Publisher: Pearson
