MTB > execute 'aa-467.mtb'
MTB > Note. Bill Qualls - Multiple Regression Example
MTB > Note. Source: Amir Aczel, Complete Business Statistics,
MTB > Note. 2ed, Irwin Publishing, Page 467, Example (a)
MTB > Note. This is file AA-467.MTB
MTB >  
MTB >  
MTB > Note. "Alka-Seltzer recently embarked on an in-store
MTB > Note. promotional campaign, with displays of its antacid
MTB > Note. featured prominently in supermarkets. The company
MTB > Note. also ran its usual radio and television commercials.
MTB > Note. Over a period of 10 weeks, the company kept track of
MTB > Note. its expenditure on radio and television advertising,
MTB > Note. variable X1, and well as its spending on in-store
MTB > Note. displays, variable X2. The resulting sales for each
MTB > Note. week in the area studied were recorded as the
MTB > Note. dependent variable, Y."
MTB >  
MTB >  
MTB > read c1-c3
DATA> 12  5 72
DATA> 11  8 76
DATA> 15  6 78
DATA> 10  5 70
DATA> 11  3 68
DATA> 16  9 80
DATA> 14 12 82
DATA>  8  4 65
DATA>  8  3 62
DATA> 18 10 90
DATA> end
     10 ROWS READ
MTB >  
MTB >  
MTB > name c1 'RadioTV' c2 'InStore' c3 'Sales'
MTB >  
MTB >  
MTB > Note. First, let's try using Radio/TV (only) as a predictor...
MTB > regr c3 1 c1

The regression equation is
Sales = 45.0 + 2.38 RadioTV

Predictor       Coef       Stdev    t-ratio        p
Constant      45.014       3.909      11.52    0.000
RadioTV       2.3810      0.3076       7.74    0.000

s = 3.108       R-sq = 88.2%     R-sq(adj) = 86.7%

Analysis of Variance

SOURCE       DF          SS          MS         F        p
Regression    1      578.82      578.82     59.92    0.000
Error         8       77.28        9.66
Total         9      656.10



 
MTB >  
MTB >  
MTB > Note. Next, let's try using In-store displays (only)...
MTB > regr c3 1 c2

The regression equation is
Sales = 58.5 + 2.43 InStore

Predictor       Coef       Stdev    t-ratio        p
Constant      58.482       3.253      17.98    0.000
InStore       2.4335      0.4559       5.34    0.000

s = 4.240       R-sq = 78.1%     R-sq(adj) = 75.3%

Analysis of Variance

SOURCE       DF          SS          MS         F        p
Regression    1      512.26      512.26     28.49    0.000
Error         8      143.84       17.98
Total         9      656.10



MTB >  
MTB >  
MTB > Note. Finally, let's use both (combined) as a predictor...
MTB > regr c3 2 c1 c2

The regression equation is
Sales = 47.2 + 1.60 RadioTV + 1.15 InStore

Predictor       Coef       Stdev    t-ratio        p
Constant      47.165       2.470      19.09    0.000
RadioTV       1.5990      0.2810       5.69    0.000
InStore       1.1487      0.3052       3.76    0.007

s = 1.911       R-sq = 96.1%     R-sq(adj) = 95.0%

Analysis of Variance

SOURCE       DF          SS          MS         F        p
Regression    2      630.54      315.27     86.34    0.000
Error         7       25.56        3.65
Total         9      656.10

SOURCE       DF      SEQ SS
RadioTV       1      578.82
InStore       1       51.72



MTB > stop
*** Minitab Release 8.21 *** Minitab, Inc. ***
Storage available 16174