Lab Exercise 2

Module 7

Question 1.
The following is a table showing the grades earned by a number of students along with the
amount of time they spent studying, the number of classes they missed, their GPA, and their
age. You are to run a multiple regression on this data.

Studen Minutes GP Ag Test

t Studied Classes Missed A e Score
3.7
1 75 2 8 21 93
2.8
2 70 9 3 18 77
2.1
3 0 14 7 17 60
3.6
4 90 3 4 20 89
3.8
5 120 1 9 22 100
3.9
6 105 3 6 20 95
2.7
7 125 8 3 18 77
2.4
8 60 12 9 19 71
3.0
9 45 10 4 18 77
3.1
10 120 2 1 19 96

a) Which variable(s) are the independent and the dependent variables, and why to you
believe this is the case?

b) Run the multiple regression in Excel. You will need to use the Data Analysis Add-in.
When the Add-In Window appears, choose Regression. Part of the Regression Window is
shown below to help you set the correct options.
 When entering the data into the Input X and Input Y boxes, be sure to include the
labels in the row immediately above the data for student 1 when you set your
data for x and y.
  Set the regression parameters as shown in the following diagram. Note that the
data part of the window is not shown, but you have to enter the correct
information into the areas for follows (there is a diagram below:

Display the results of your regression (they will be in the spreadsheet which has your
name on it).
 Do not include the residual data at the bottom of the output or the plots which
result. You will use some of these later. Show only the part of the spreadsheet
from “Summary Output” to above the Residual Information.
 It would be helpful if you made the columns wide enough to show the full data.

c) Using the results above, show the formula for the regression line. Use two significant
decimal places.

d) Interpret the regression line. What does each of the parameters tell you?

e) If a studied 80 minutes, missed 5 classes, had a GPA of 3.13, and was 27 years of age,
what would be their predicted score on the test.

f) Copy and paste your Normal Probability Plot here and interpret the plot (what does it
tell you?).
Question 2.

Here is a data set of grades earned by students in three classes of the same course.

Class
Grade 1 2 3
A 8 6 12
B 20 24 18
C 17 19 15
D 5 8 12
F 2 1 5

You are interested in determining if the grades in the classes are independent from each other.
In other words, are the grades allocated to the various classes consistent or are their differences
in the grade distributions? In order to do this, you will need to use Pearson’s chi-square, and
you will use Excel to help you solve this question.

First, copy the above data into Excel. Then add totals for the row and column data, and for the
overall data. Your work should look something like the following:

Class
Grade 1 2 3 Total
A 8 6 12
B 20 24 18
C 17 19 15
D 5 8 12
F 2 1 5
Total

a) Copy and paste your completed work here:

Your next step is to find the PROBABLE numbers for each grade for each class. This sounds
complicated, but it can be done quite easily in Excel. Make a copy of your above work and paste
it below the existing work. Then delete out the data shown above. You should have something
like the following:

Class
Grade 1 2 3 Total
A 8 6 12 26
B 20 24 18 XX
C 17 19 15 XX
D 5 8 12 XX
F 2 1 5 XX
Total 52 XX XX 172

Probables
Class
Grade 1 2 3 Total
A 26
B
C
D
F
Total 52 172

To calculate the probables, multiply the row total by the column total for each value. Then
divide by the overall total. For the first row and first column – the A’s for Class 1 - (which has the
value of 8), the calculation would be (26 * 52)/172 = 7.860465. Hence, we expect there to be
7.860465 A’s. Do the same for all the data (all grades for all classes). If you calculate the totals of
the probables, they will be the same as for the original data, so it is a good way to check your
work.

b) Copy and paste your completed probables work here:

c) The formula for the test statistic for Chi-Square is:

 which can appear quite scary. However, here is what you need to do:
1. Subtract each observed value from its expected value and square it.
For the first row and column, this would be (8 - 7.860465)2.
2. Divide the result by the expected value for that row/column
combination - in this case, by 7.860465.
You should now have (8 – 7.860465)2 / 7.860465
3. Add up all the values for all the row/column pairs.
The following might help. You can enter the values from your earlier
work, and then perform the calculations in Excel.

Observe Expecte Observed Minus Squared Divided by

d d Expected Squared Expected
8
20
17
5
2
6
24
19
8
1
12
18
15
12
5
Total 172

d) The value of the “Squares Divided By Expected” is the Chi-square ( χ 2) test statistic.

e) Now you need to obtain the critical value for χ 2. First you need to determine the
degrees of freedom. The degrees of freedom are (r-1)(c-1) where r = number of rows
and c = number of columns FROM YOUR ORIGINAL DATA TABLE AT THE BEGINNING OF
THIS ASSIGNMENT. What are your degrees of freedom?

f) Assume you want to be 95% confident in your result. Find the critical value of χ 2 from
the tables in the back of the book. Remember that a χ 2 is always one-tailed.

g) What is your conclusion?