Sunteți pe pagina 1din 5

Q.2.

An insurance company provides auto insurance and is analyzing the data obtained from
fatal crashes. A sample of the motor vehicle deaths is randomly selected for a two-year
period. The number of fatalities is listed below for the different days of the week. At the 0.05
significance level, test the claim that accidents occur on different days with equal frequency.
The analysis is provided below. (5 marks)

Q.3. A razor manufacturer is interested in getting information on the possible commercial


acceptance of a new razor. He wants to know the characteristics that differentiate between
prospective buyers/non-buyers of the product. He is interested primarily in ascertaining the
relative importance of the following razor characteristics:

 Quality
 Brand Name
 Free Blades
 Packaging

The above stated points affect a potential buyer’s overall evaluation of the razor’s
desirability. Each of the 18 respondents rates the product according to each of the four
characteristics and then indicates whether he would be a prospective buyer of the product or
not. The rating is done on an 11 point scale (where 0 represents very poor and 10 excellent).
Function
I. Do the variables have a strong relationship with the discriminant
1
function? Is the model significant? (2+2 marks)
Quality 1.219

Brand Name .104

Free Blades -.338


II. Do all the variables discriminate the grouping variable? Which variable Packaging -.268

most significantly discriminates the grouping variable and why? (4


marks)

We have four independent variable quality, Brand name, free blade and packaging out
of which Quality is the best predictor with coefficient 1.219, followed by free blade
(-.338) , then packaging(-.268) and then the brand name with value -1.04.

III. Does the analysis classify the cases appropriately? (4 marks)

The model was able to classify 83% of the original group cases correctly. And after
removing one of the count and repeating the test again the percentage fell down to
72.2% which is illustrated by the cross validated grouped number. Theis can be
determined using classification result matrix

The centroid of the two groups buyer and non buyer are same and hence the classif

Group Statistics

Buyer / Non-Buyer Mean Std. Deviation Valid N (listwise)


Wilks' Lambda

Standardized Canonical
Test of Wilks'
Unweighted Weighted Discriminant Function
Function(s) Lambda Chi-square df Sig.
Buyer Quality 7.44 1.944 9 9.000 Coefficients
1 .492 9.936 4 .042
Brand Name 5.78 1.986 9 9.000 Function

1
Free Blades 5.22 2.108 9 9.000
Quality 1.219
Packaging 3.44
Structure Matrix 1.424 9 9.000
Brand Name -.104
Non-Buyer Quality 4.00 2.000 9 9.000
Function Free Blades -.338
Brand Name 4.33 1.803 9 9.000
Packaging -.268
Free Blades 4.33
1 1.414 9 9.000

Packaging
Quality 3.67 1.936 9 9.000
.911
Total Quality 5.72 2.608 18 18.000
Brand Name .397
Brand Name 5.06 1.984 18 18.000
Free Blades .258 1.801
Free Blades 4.78 18 18.000

Packaging
Packaging 3.56 -.068 1.653 18 18.000

Pooled within-groups correlations


between discriminating variables
Pooled Within-Groups Matrices
and standardized canonical
Eigenvalues
discriminant functions
Quality Brand Name Free Blades Packaging
Variables ordered by absolute size % of Cumulative Canonical
Correlation Quality 1.000 .633 .549 .209
of correlation within function. Function Eigenvalue Variance % Correlation
Brand Name .633 1.000 .541 .327
1 1.033a 100.0 100.0 .713
Free Blades .549 .541 1.000 .064
a. First 1 canonical discriminant functions were used in the
Packaging .209
Canonical Discriminant .327 .064 1.000 analysis.
Function Coefficients

Function

Quality .618

Brand Name -.055

Free Blades -.188

Packaging -.157

(Constant) -1.800

Unstandardized coefficients

Tests of Equality of Group Means

Wilks'
  Lambda F df1 df2 Sig.
Quality .538 13.729 1 16 .002
Brand Name .860 2.610 1 16 .126
Free Blades .935 1.103 1 16 .309
Packaging .995 .077 1 16 .785
Functions at Group Centroids

Function

Buyer / Non-Buyer 1

Buyer .958

Non-Buyer -.958

Unstandardized canonical discriminant


functions evaluated at group means

Prior Probabilities for Groups

Cases Used in Analysis


Buyer /
Non-Buyer Prior Unweighted Weighted

Buyer .500 9 9.000

Non-Buyer .500 9 9.000

Total 1.000 18 18.000

Classification Resultsb,c

Predicted Group Membership


Buyer / Non-
Buyer Buyer Non-Buyer Total

Original Count Buyer 8 1 9

Non-Buyer 2 7 9

% Buyer 88.9 11.1 100.0

Non-Buyer 22.2 77.8 100.0

Cross-validateda Count Buyer 7 2 9

Non-Buyer 3 6 9

% Buyer 77.8 22.2 100.0

Non-Buyer 33.3 66.7 100.0

a. Cross validation is done only for those cases in the analysis. In cross validation, each case is
classified by the functions derived from all cases other than that case.

b. 83.3% of original grouped cases correctly classified.


a. 72.2% of cross-validated grouped cases correctly classified.

S-ar putea să vă placă și