STATISTICAL METHODS
EXERCISES: Correlation and regression (Chapter 4)
Question 1
The following data were recorded from a study that was intended to demonstrate that productivity declines
significantly if workers are required to work for longer than two hours after the lunch break. Worker
productivity was expressed as an index, where the value of the index was directly proportional to productivity.
The other variable, time, indicates the number of minutes between the end of the lunch break and the time at
which the productivity reading was made. The data were obtained from 8 randomly selected workers.
Worker 1 2 3 4 5 6 7 8
Productivity (P) 75 69 78 65 64 78 66 61
Time (T) 101 109 95 116 122 99 113 125
The estimated least squares regression equation of productivity on time is:
a) Pˆ = 134.951 − 0.595T
b) Pˆ = −0.595 + 134.951T
c) Tˆ = 222.321 − 1.616P
d) Tˆ = −1.616 + 222.321P
e) Pˆ = −0.595 + 222.321T
Question 2
With reference to Question 1:
Based on your calculation of the coefficient of determination, which of the following statements is true?
a) There is a strong positive linear relationship between productivity and time
b) There is a strong negative linear relationship between productivity and time
c) Approximately 1.94% of the variation in the time is accounted for by the productivity
d) Approximately 3.84% of the variation in the productivity is accounted for by time
e) None of the above statements are true
1
Question 3
An analyst believes that the number of businesses filing for bankruptcy each month increases as the interest
rate increases. The analyst collected the following data from 8 randomly selected months with the aim of
regressing the number of filings for bankruptcy (F) on the monthly interest rate (R).
Number of filings (F) 8 17 10 5 36 22 28 15
Interest rate (R) 13.5 14.0 13.0 13.5 17.5 14.5 16.0 15.5
The estimated number of businesses filing for bankruptcy when the monthly interest rate is 13.5% is
approximately
a) 5
b) 6.5
c) 8
d) 10.27
e) Cannot be estimated
Question 4
A production manager has compared the dexterity test scores (D), scores of how quick one is with one’s hands,
of five assembly line employees with their hourly productivity (P). The summary statistics are as follows:
D = 70 D 2
= 1006 P = 306 P 2
= 18984 DP = 4362
Which one of the following is correct?
a) The strength of the linear relationship between dexterity and productivity is 0.714
b) The coefficient of determination is equal to 95.46%
c) The least squares regression line of P on D is Pˆ = 3 + 19.2D
d) The predicted productivity of an employee with a dexterity score of 5 is equal to 34.2
Question 5
Vending Machines for Africa sells vending machines to major corporates to sell snacks on the various floors
in their organisation. As part of its standard purchase contract, Vending Machines for Africa agree to perform
regular services on the vending machines. To better schedule its service calls, Vending Machines for Africa
wishes to obtain information concerning the time it takes to perform the required service. The company
collected data for 11 service calls. The data consisted of the number of vending machines (V) to be serviced
on a call, and the time in minutes (T), on the service call. This data are summarised by the following statistics:
V = 43 V 2
= 201 T = 1189 T 2
= 149037 VT = 5466
2
The least squares regression line, which may be used to predict the time taken on a service call for the above
data is (rounding off intermediate and final calculations to 3 decimal places):
a) Tˆ = 24.859 + 10.915V
b) Tˆ = 10.915 + 24.859V
c) Vˆ = 0.04 + 24.859T
d) Vˆ = −0.415 + 0.040T
Question 6
A chain store advertised on television over a period of ten weekends. The number of advertisements, ranging
from 2 to 6, and the sales generated (in thousands of Rand) were recorded. The store would like to predict
sales based on the number of advertisements that were aired. The following output shows the results of a
regression analysis.
Regression Statistics
Multiple R 0.93
R Square 0.87
Adjusted R Square 0.85
Standard Error 3.08
Observations 10
ANOVA df SS MS F Significance F
Regression 1 490.05 490.05 51.62 0.00
Residual 8 75.95 9.49
Total 9 566
Standard P- Lower Upper Lower Upper
Coefficients t Stat
Error value 95% 95% 95.0% 95.0%
Intercept 41.20 2.92 14.09 0.00 34.46 47.94 34.46 47.94
X 4.95 0.69 7.18 0.00 3.36 6.54 3.36 6.54
The correlation coefficient is:
a) 0.83
b) 0.85
c) 0.87
d) 0.93
e) 0.96
3
Question 7
With reference to Question 6:
Which one of the following statements is correct?
a) The coefficient of determination indicates that there is a strong positive relationship between sales and
the number of advertisements
b) The sales account for 93% of the variation in the number of advertisements
c) Approximately 7% of the variation in the sales is due to other factors
d) Approximately 13% of the variation in the number of advertisements is accounted for by the variability
in the sales
e) Approximately 87% of the variation in the sales is accounted for by the variability in the number of
advertisements
Question 8
With reference to Question 6:
The predicted sales generated by 4 television adverts is:
a) R61
b) R184.60
c) R4950
d) R61000
e) Invalid prediction (extrapolation)
4
SOLUTIONS
Question 1
A
Question 2
E
Question 3
D
Question 4
D
Question 5
B
Question 6
D
Question 7
E
Question 8
D
