Stat 2000, Section 002, Homework Assignment 7 (30 Points)

(10/23/2015 - Due Friday 10/30/2015 by 10:35am)

• 0) Reading: Sections 4.1, 4.2 & 4.3
  
  
• 1) Please work on the following textbook exercises in Moore/McCabe/Craig:
  
  
  • Exercise 3.102 (2 points), 3.111 (2), 3.112 (2), 3.116 (2), 3.126 (2), 4.3 [using Table B] (2), 4.7 (1), 4.19 (2), 4.21 (1), 4.22 (1), 4.25 (1), 4.26 (1), 4.29 (2), 4.31 (2), 4.34 (2)
    
    
  Note: Always use Table B from the textbook at the specified line (even if the question states that you could use software or Table B)!
  
  
• 2) Regression with StatCrunch (5 points):
  The file http://www.math.usu.edu/~symanzik/teaching/2015_stat2000_fall/hw07_examdata.csv contains exam data for 47 students from a previous class.
  
  For this exercise, we are primarily interested in the question how the "Total" exam score is affected by three other variables, in particular HWScore, CanvasTime, and Mystery.
  
  You should do this as follows:
  • Calculate the correlations between Total and the three other variables of interest. Which are strongest and which are weaker? What are the directions?
  • Create the scatterplots for Total (the response variable) and the three other variables of interest. Based on these three plots, can we fit a regression line in each case? Justify your answer.
  • For those plots where a regression line is meaningful, fit the regression line and construct at least three meaningful residual plots for each of the lines you have fitted.
  • Carefully assess your residual plots. Do these plots suggest that your fitted regression line is suitable to describe the relationship? Be specific and indicate which of the residual plots suggest that your fitted regression line is meaningful and which residual plots are not very informative.
  • In those cases where your fitted regression line is meaningful, use your regression line to make the following predictions of Total, using 20, 50, and 80 for HWScore; 100, 300, and 1000 for CanvasTime; and 100, 300, and 700 for Mystery. Are all of these meaningful predictions? Be specific and use the proper statistical terms.
  • Provide an overall assessment how the three variables affect the "Total" exam score. Which variables seem to have a strong effect on the "Total" exam score and which variables seem to have a minor (possibly neglectable) effect? Think carefully about the relationship between Total and Mystery!
  • Write a computer-based conclusive report of your results (a total of three to four pages, including all figures). Look at the solutions of HW 5, Exercise 2, how to structure such a report. Just the computer output will result in at most 50% of the possible points for this exercise.
  
  Use statistical software of your choice (StatCrunch, CrunchIt, Excel, etc.) for this question!