Popcorn Correlation

Go into Minitab.

Open the data and select just our class

  1. File / Open Worksheet
    1. Go to R:
    2. Choose the \pub folder
    3. Open the popcorn.mtw worksheet
  2. Go to Manip / Subset Worksheet
    1. Name the worksheet "Section 2"
    2. Click on Condition and enter "sect=2"
    3. Click OK

Descriptive Statistics

Find the means for the two variables you've been assigned. We will use these numbers later.

Finding the Correlation

  1. Go to Stats / Basic Statistics / Correlation
  2. Enter the two variables you have been assigned.
  3. Click OK

The correlation is _________.

Now, go back to Correlation and enter the variables in the other order. The correlation when the variables are switched is _________. How does the correlation coefficient change when the variables are switched?

Changing the Scale

Go to an empty column and label it as newx and the next one as newy. The first variable we will convert into minutes by dividing the time by 60 and the second variable we will subtract the mean so that it is centered about zero.

  1. Go to Calc / Calculator
    1. Store the results into newx
    2. Double click on your first variable and then put "/60" (without the quotes). Example: Act/60
    3. Click OK
  2. Go to Calc / Calculator
    1. Store the results int newy
    2. Double click on your second variable and then put "-###" where ### is the mean you found for that variable earlier. For example, if the mean of the pop column was 97.3 seconds, it would read "pop-97.3"
    3. Click OK

Find the correlation between newx and newy. The correlation is __________. How does the correlation coefficient change when the scale on the variables is changed?

Drawing Plots

  1. Go to Stat / Regression / Fitted Line Plot and enter your Y and X variables.
  2. Go into Storage and choose Residuals and Fits
  3. Hit OK

Do there appear to be any outliers that seriously affect the results? If so, look at the data and determine who gave those results.

Follow the next steps very carefully!

  1. Go to Graph / Plot
  2. Enter the Y and X variables for Graph 1.
  3. Enter FITS1 and the X variable for Graph 2
  4. Click on Symbol in Item 1 and then click Edit Attributes.
    1. Graph 1 should be a solid blue circle of size 3
    2. Graph 2 should have a type of None
  5. Add Item 2 to be Connect for each Graph
  6. Click on Item 2 and choose Edit Attributes
    1. Graph 1 should have a type of None
    2. Graph 2 should be a solid black line of size 2
  7. Click on Frame and choose Multiple Graphs and set them to overlay
  8. Click on Frame and Reference
    1. Add a graph in the Y direction that is a dashed red line with position equal to the mean of your Y variable.
    2. Add a graph in the X direction that is a dashed red line with position equal to the mean of your X variable
  9. Click OK

The intersection of the red reference lines represent the centroid (x-bar, y-bar) and your regression equation should pass through the centroid.

Leave your graph open and save your project as POP# (your group #) into the R:\01\ folder.