In: Statistics and Probability
Enterprise Industries produces Fresh, a brand of liquid laundry detergent. In order to manage its inventory more effectively and make revenue projections, the company would like to better predict demand for Fresh. To develop a prediction model, the company has gathered data concerning demand for Fresh over the last 30 sales periods (each sales period is defined to be a four-week period). The demand data are presented in table concerning y (demand for Fresh liquid laundry detergent), x1x1 (the price of Fresh), x2x2 (the average industry price of competitors' similar detergents), and x3x3 (Enterprise Industries’ advertising expenditure for Fresh). To ultimately increase the demand for Fresh, Enterprise Industries’ marketing department is comparing the effectiveness of three different advertising campaigns. These campaigns are denoted as campaigns A, B, and C. Campaign A consists entirely of television commercials, campaign B consists of a balanced mixture of television and radio commercials, and campaign C consists of a balanced mixture of television, radio, newspaper, and magazine ads. To conduct the study, Enterprise Industries has randomly selected one advertising campaign to be used in each of the 30 sales periods in table below. Although logic would indicate that each of campaigns A, B, and C should be used in 10 of the 30 sales periods, Enterprise Industries has made previous commitments to the advertising media involved in the study. As a result, campaigns A, B, and Cwere randomly assigned to, respectively, 9, 11, and 10 sales periods. Furthermore, advertising was done in only the first three weeks of each sales period, so that the carryover effect of the campaign used in a sales period to the next sales period would be minimized. Table below lists the campaigns used in the sales periods.
To compare the effectiveness of advertising campaigns A,
B, and C, we define two dummy variables.
Specifically, we define the dummy variable DB
to equal 1 if campaign B is used in a sales period and 0
otherwise. Furthermore, we define the dummy variable DC to equal 1
if campaign C is used in a sales period and 0 otherwise. The table
presents the Excel and Excel add-in (MegaStat) output of a
regression analysis of the Fresh demand data by using the model
Historical Data Concerning Demand for Fresh Detergent | ||||
Sales Period |
Price for Fresh, x1 |
Average Industry Price, x2 |
Advertising Expenditure for Fresh, x3 |
Demand for Fresh, y |
1 | 3.89 | 3.86 | 5.56 | 7.36 |
2 | 3.74 | 4.05 | 6.72 | 8.52 |
3 | 3.75 | 4.35 | 7.21 | 9.20 |
4 | 3.78 | 3.78 | 5.52 | 7.55 |
5 | 3.61 | 3.85 | 7.03 | 9.33 |
6 | 3.66 | 3.86 | 6.54 | 8.25 |
7 | 3.65 | 3.78 | 6.71 | 8.77 |
8 | 3.81 | 3.81 | 5.23 | 7.86 |
9 | 3.80 | 3.66 | 5.27 | 7.14 |
10 | 3.80 | 4.04 | 6.06 | 8.06 |
11 | 3.90 | 4.11 | 6.53 | 7.80 |
12 | 3.96 | 4.05 | 6.25 | 8.13 |
13 | 3.75 | 4.17 | 7.02 | 9.15 |
14 | 3.71 | 4.27 | 6.96 | 8.89 |
15 | 3.74 | 4.16 | 6.85 | 8.98 |
16 | 3.80 | 4.17 | 6.87 | 8.86 |
17 | 3.70 | 4.22 | 7.10 | 9.26 |
18 | 3.82 | 4.35 | 7.04 | 9.03 |
19 | 3.79 | 4.14 | 6.88 | 8.78 |
20 | 3.85 | 3.76 | 6.58 | 7.90 |
21 | 3.81 | 3.77 | 6.23 | 7.65 |
22 | 3.73 | 3.69 | 6.09 | 7.28 |
23 | 3.73 | 3.92 | 6.51 | 8.04 |
24 | 3.51 | 3.66 | 7.03 | 8.57 |
25 | 3.60 | 4.15 | 6.89 | 8.76 |
26 | 3.64 | 4.25 | 6.89 | 9.26 |
27 | 3.70 | 3.61 | 6.52 | 8.22 |
28 | 3.70 | 3.74 | 5.76 | 7.60 |
29 | 3.81 | 3.80 | 5.81 | 7.90 |
30 | 3.74 | 4.25 | 6.86 | 9.23 |
Advertising Campaigns Used by Enter prise Industries |
|
Sales Period |
Advertising Campaign |
1 | B |
2 | B |
3 | B |
4 | A |
5 | C |
6 | A |
7 | C |
8 | C |
9 | B |
10 | C |
11 | A |
12 | C |
13 | C |
14 | A |
15 | B |
16 | B |
17 | B |
18 | A |
19 | B |
20 | B |
21 | C |
22 | A |
23 | A |
24 | A |
25 | A |
26 | B |
27 | C |
28 | B |
29 | C |
30 | C |
Regression Statistics | |
Multiple R | .9698 |
R Square | .9405 |
Adjusted R Square | .9281 |
Standard Error | .1811 |
Observations | 30 |
ANOVA | df | SS | MS | F | Significance F |
Regression | 5 | 12.4458 | 2.4892 | 75.8909 | .0000 |
Residual | 24 | .7872 | .0328 | ||
Total | 29 | 13.2329 | |||
Coefficients | Standard Error | t Stat | p-value | Lower 95% | Upper 95% | |
Intercept | 8.7835 | 1.8061 | 4.863 | .0001 | 5.0559 | 12.5111 |
Price X1 | -2.6123 | .4714 | -5.542 | .0000 | -3.5851 | -1.63947 |
Ind Price X2 | 1.5397 | .2240 | 6.874 | .0000 | 1.0774 | 2.0020 |
AdvExp X3 | .5034 | .0963 | 5.226 | .0000 | .3046 | .7023 |
DB | -.2654 | .0835 | -3.178 | .0041 | -.4378 | -.0930 |
DC | .2466 | .0814 | 3.029 | .0058 | .0786 | .4146 |
Predicted values for: Demand using an Excel add-in (MegaStat) | ||||||
95% Confidence Interval | 95% Prediction Interval | |||||
Predicted | lower | upper | lower | upper | Leverage | |
8.64161 | 8.51365 | 8.76957 | 8.24653 | 9.03669 | .117 | |
y = β0 + β1x1x1 + β2 x2x2+ β3x3x3 + β4DB + β5DC + ε
(a) In this model the parameter
β4 represents the effect on mean demand of
advertising campaign B compared to advertising campaign
A, and the parameter β5 represents the
effect on mean demand of advertising campaign C compared
to advertising campaign A. Use the regression output to
find and report a point estimate of each of the above effects and
to test the significance of each of the above effects. Also, find
and report a 95 percent confidence interval for each of the above
effects. (Round your answers to 4 decimal
places.)
The point estimate of the effect on the mean of campaign B compared to campaign A is b4 = . |
The 95% confidence interval = [, ]. |
The point estimate of the effect on mean of campaign C compared to campaign A is b5 = . |
The 95% confidence interval = [, ]. |
(b) The prediction results at the bottom of the
output correspond to a future period when Fresh’s price will be
x1x1 = 3.70, the average price of similar detergents will be x2x2 =
3.90, Fresh’s advertising expenditure will be x3x3= 6.50, and
advertising campaign C will be used. Show how yˆy^=
8.64161 is calculated. Then find, report, and interpret a 95
percent confidence interval for mean demand and a 95 percent
prediction interval for an individual demand when x1x1 = 3.70, x2x2
= 3.90, x3x3 = 6.50, and campaign C is used.
(Round your answers to 5 decimal places.)
yˆy^ = |
Confidence interval = [, ] |
Prediction interval = [, ] |
(c) Consider the alternative model
y = β0 + β1x1x1 + β2x2x2 +
β3x3x3 + β4DA +
β5DC + ε
Here DA equals 1 if advertising campaign
A is used and equals 0 otherwise. Describe the effect
represented by the regression parameter β5.
β5 = effect on mean of Campaign (Click to select)ABC compared to Campaign B.
(d) The Excel output of the least squares point
estimates of the parameters of the model of part c is as
follows.(Round your answer to 4 decimal
places.)
Coefficients | Standard Error | t Stat | p-value | Lower 95% | Upper 95% | |
Intercept | 8.7835 | 1.8061 | 4.863 | .0001 | 5.0559 | 12.5111 |
Price X1 | -2.6123 | .4714 | -5.542 | .0000 | -3.5851 | -1.6395 |
Ind Price X2 | 1.5397 | .2240 | 6.874 | .0000 | 1.0774 | 2.0020 |
AdvExp X3 | .5034 | .0963 | 5.226 | .0000 | .3046 | .7023 |
DA | -.2654 | .0835 | -3.178 | .0041 | -.4378 | -.0930 |
DC | .2466 | .0814 | 3.029 | .0058 | .0786 | .4146 |
Use the Excel output to test the significance of the effect represented by β5 and find a 95 percent confidence interval for β5. Interpret your results.
95 percent confidence interval for β5 [ , ] |
β5 is significant at alpha = 0.1 and alpha = 0.05 because p-value = . |
Thus there is strong evidence that β5 (Click to select)is greater thanis less thanis equal to 0. |
(a)
From the output,
The point estimate of the effect on the mean of campaign B compared
to campaign A is b4 = -.2654
The 95% confidence interval = [-.4378, -.0930].
The point estimate of the effect on mean of campaign C compared to
campaign A is b5 = .2466
The 95% confidence interval = [.0786, .4146].
(b)
The estimated regression equation is,
yˆ= 8.7835 -2.6123x1 + 1.5397x2+ 0.5034x3 - 0.2654DB +
0.2466DC
For x1 = 3.70, x2 = 3.90, x3 = 6.50, DB = 0 and DC = 1
yˆ= 8.7835 -2.6123 * 3.70 + 1.5397 * 3.90 + 0.5034 * 6.50 - 0.2654
* 0 + 0.2466 * 1 = 8.6415
Confidence interval = [8.5137, 8.7696]
Prediction interval = [8.2465, 9.0367]
(c)
Since DC = 1, Campaign C is used, the answer is,
β5 = effect on mean of Campaign C compared to Campaign B.
(d)
95 percent confidence interval for β5 [0.0786 , 0.4146 ]
β5 is significant at alpha = 0.1 and alpha = 0.05 because p-value =
0.0058 (less than 0.1 and 0.05).
Since the confidence interval contains all positive values,
Thus there is strong evidence that β5 is greater than equal to
0.