This action might not be possible to undo. Are you sure you want to continue?

Discriminant Analysis

18-2

Chapter Outline

1) Overview 2) Basic Concept 3) Relation to Regression and ANOVA 4) Discriminant Analysis Model 5) Statistics Associated with Discriminant Analysis 6) Conducting Discriminant Analysis i. Formulation ii. Estimation iii. Determination of Significance

iv. Interpretation

v. Validation

18-3

Chapter Outline

7) i. ii. Multiple Discriminant Analysis Formulation Estimation

iii. Determination of Significance iv. Interpretation v. 8) 9) Validation Stepwise Discriminant Analysis Internet and Computer Applications

10) Focus on Burke 11) Summary 12) Key Terms and Concepts

1 ANOVA REGRESSION One DISCRIMINANT ANALYSIS One 18-4 Similarities Number of dependent variables Number of independent variables Differences Nature of the dependent variables Nature of the independent variables One Multiple Multiple Multiple Metric Categorical Metric Metric Categorical Metric . and Discriminant Analysis Table 18.Similarities and Differences between ANOVA. Regression.

Examination of whether significant differences exist among the groups. . The objectives of discriminant analysis are as follows: Development of discriminant functions. Determination of which predictor variables contribute to most of the intergroup differences. in terms of the predictor variables. Classification of cases to one of the groups based on the values of the predictor variables.18-5 Discriminant Analysis Discriminant analysis is a technique for analyzing data when the criterion or dependent variable is categorical and the predictor or independent variables are interval in nature. Evaluation of the accuracy of classification. which will best discriminate between the categories of the criterion or dependent variable (groups). or linear combinations of the predictor or independent variables.

However. with G groups and k predictors. When three or more categories are involved. The second function. In general. not all the functions may be statistically significant. discriminant functions. the technique is known as two-group discriminant analysis. it is possible to derive only one discriminant function. it is possible to estimate up to the smaller of G . The first function has the highest ratio of between-groups to within-groups sum of squares. and so on.1. uncorrelated with the first. has the second highest ratio. .18-6 Discriminant Analysis When the criterion variable has two categories. In multiple discriminant analysis. the technique is referred to as multiple discriminant analysis. The main distinction is that. or k. more than one function may be computed. in the two-group case.

. are estimated so that the groups differ as much as possible on the values of the discriminant function. . or weights (b). .18-7 Discriminant Analysis Model The discriminant analysis model involves linear combinations of the following form: D = b0 + b1X1 + b2X2 + b3X3 + . + bkXk where D b 's X 's = = = discriminant score discriminant coefficient or weight predictor or independent variable The coefficients. This occurs when the ratio of between-group sum of squares to within-group sum of squares for the discriminant scores is at a maximum.

The centroid is the mean values for the discriminant scores for a particular group. The means for a group on all the functions are the group centroids. . Classification matrix. Sometimes also called confusion or prediction matrix. Centroid. the classification matrix contains the number of correctly classified and misclassified cases. There are as many centroids as there are groups. It is a measure of association between the single discriminant function and the set of dummy variables that define the group membership. Canonical correlation measures the extent of association between the discriminant scores and the groups. as there is one for each group.18-8 Statistics Associated with Discriminant Analysis Canonical correlation.

Discriminant scores.18-9 Statistics Associated with Discriminant Analysis Discriminant function coefficients. Large Eigenvalues imply superior functions. The unstandardized coefficients are multiplied by the values of the variables. For each discriminant function. . the Eigenvalue is the ratio of between-group to withingroup sums of squares. Eigenvalue. when the variables are in the original units of measurement. The discriminant function coefficients (unstandardized) are the multipliers of variables. These products are summed and added to the constant term to obtain the discriminant scores.

. Group means and group standard deviations. serves as the metric dependent variable in the ANOVA. in turn. These are computed for each predictor for each group. The pooled within-group correlation matrix is computed by averaging the separate covariance matrices for all the groups. Pooled within-group correlation matrix. with the grouping variable serving as the categorical independent variable. Each predictor. These are calculated from a one-way ANOVA.18-10 Statistics Associated with Discriminant Analysis F values and their significance.

Large values of (near 1) indicate that group means do not seem to be different. Also referred to as discriminant loadings. Total correlation matrix. Wilks' for each predictor is the ratio of the within-group sum of squares to the total sum of squares. The standardized discriminant function coefficients are the discriminant function coefficients and are used as the multipliers when the variables have been standardized to a mean of 0 and a variance of 1. Wilks' . Sometimes also called the U statistic. Structure correlations. Small values of (near 0) indicate that the group means seem to be different. . Its value varies between 0 and 1. the structure correlations represent the simple correlations between the predictors and the discriminant function.18-11 Statistics Associated with Discriminant Analysis Standardized discriminant function coefficients. If the cases are treated as if they were from a single sample and the correlations computed. a total correlation matrix is obtained.

18.18-12 Conducting Discriminant Analysis Fig.1 Formulate the Problem Estimate the Discriminant Function Coefficients Determine the Significance of the Discriminant Function Interpret the Results Assess Validity of Discriminant Analysis .

The other part. or the experience of the researcher. and the independent variables. The criterion variable must consist of two or more mutually exclusive and collectively exhaustive categories.Conducting Discriminant Analysis Formulate the Problem 18-13 Identify the objectives. Often the distribution of the number of cases in the analysis and validation samples follows the distribution in the total sample. called the estimation or analysis sample. is used for estimation of the discriminant function. called the holdout or validation sample. The predictor variables should be selected based on a theoretical model or previous research. the criterion variable. is reserved for validating the discriminant function. One part of the sample. .

Of these. 30 households were included in the analysis sample and the remaining 12 were part of the validation sample.18-14 A Sample Problem Suppose we want to determine the salient characteristics of families that have already visited a vacation resort during the last two years. Data were obtained from a pretest sample of 42 households. .

18-15 Variables in Data Set DEPENDENT or CRITERION variable INDEPENDENT or PREDICTOR variables The households that visited a resort during the last two years are coded as 1. as 2 (VISIT variable as criterion with two categories). those that did not. Annual family income (INCOME) Attitude toward travel (TRAVEL) Importance attached to family vacation (VACATION) Household size (HSIZE) Age of the head of household (AGE) .

7 75.2 Resort Visit Annual Family Income ($000) Attitude Toward Travel Importance Household Age of Attached Size Head of to Family Household Vacation Amount Spent on Family Vacation No.2 70.18-16 Information on Resort Visits: Analysis Sample Table 18. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 50.2 57.0 64.4 71.0 5 6 7 7 6 8 5 2 7 7 6 5 1 4 5 8 7 5 5 6 7 3 4 5 6 7 8 8 2 6 3 4 6 5 4 5 3 6 4 5 5 4 6 3 2 43 61 52 36 55 68 62 51 57 45 44 64 54 56 58 M (2) H (3) H (3) L (1) H (3) H (3) M (2) M (2) H (3) H (3) H (3) H (3) M (2) H (3) H (3) .2 49.9 48.5 52.1 73.3 62.0 46.3 62.9 56.1 68.

4 37.3 55.0 46.4 44.5 41.3 41.0 33.1 38.2 50.0 37.2 43.1 36.3 Attitude Toward Travel 5 4 2 5 6 6 1 3 6 2 5 8 6 3 3 Importance Household Age of Attached Size Head of to Family Household Vacation 4 3 5 2 6 6 2 5 4 7 1 3 8 2 3 3 2 2 4 3 2 2 3 5 4 3 2 2 3 2 58 55 57 37 42 45 57 51 64 54 56 36 50 48 42 Amount Spent on Family Vacation L L M M M L M L L L M M L L L (1) (1) (2) (2) (2) (1) (2) (1) (1) (1) (2) (2) (1) (1) (1) No. Resort Visit 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Annual Family Income ($000) 32.8 57.2 cont.18-17 Information on Resort Visits: Analysis Sample Table 18.1 35. 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 .

5 38.Information on Resort Visits: Holdout Sample Table 18.1 35.0 49.0 54. 1 2 3 4 5 6 7 8 9 10 11 12 1 1 1 1 1 1 2 2 2 2 2 2 50.2 4 7 6 5 6 5 4 5 6 2 7 2 7 4 7 4 6 6 3 3 5 6 3 2 3 7 4 3 6 3 4 5 3 5 3 3 45 55 58 60 46 56 54 39 44 51 37 49 M(2) H (3) M(2) M(2) H (3) H (3) L (1) L (1) H (3) L (1) M(2) L (1) .0 68.6 54.6 39.3 Resort Visit Annual Family Income ($000) Attitude Toward Travel Importance Household Age of Attached Size Head of to Family Household Vacation Amount Spent on Family Vacation 18-18 No.0 45.8 63.4 37.0 62.

Conducting Discriminant Analysis 18-19 Estimate the Discriminant Function Coefficients The direct method involves estimating the discriminant function so that all the predictors are included simultaneously. . the predictor variables are entered sequentially. In stepwise discriminant analysis. based on their ability to discriminate among groups.

091 .000 4.197 vacation .2344 15 15.000 age 53.070 1.091 .089 -.014 -.453 .520 9.000 51.9198 15 15. .067 2. . Deviation V alid N (lis tw is e) Mean Std. Deviation V alid N (lis tw is e) Mean Std.217 12.000 travel 5.017 .000 . Deviation V alid N (lis tw is e) income 60.4) Group Statis tics visit 1 Mean Std.867 1.043 1.000 2.000 Tes ts of Equality of Group M eans Wilks ' Lambda .197 1.400 1.9518 15 15.017 hs iz e .9780 30 30.000 -.084 1.2710 15 15.567 1.Results of Two-Group Discriminant Analysis (Table 18.800 1.277 5.800 .000 4.5740 30 30.954 F 33.001 .333 1.014 travel .257 income travel vacation hs iz e age Contd.133 8.913 7.197 .733 8.0517 15 15.3309 30 30.000 vacation 5.333 1.5511 15 15.043 age -.021 .000 50.000 .000 4.933 8.9411 15 15.338 df 1 1 1 1 1 1 df 2 28 28 28 28 28 Sig.070 .000 18-20 Unw eighted Weighted 2 Unw eighted Weighted Total Unw eighted Weighted Pooled Within-Groups M atrice s Correlation income travel vacation hs iz e age income 1.017 -.000 .017 -.000 3.000 41.990 14.7706 15 15.8307 15 15.925 .197 .636 1.089 -.000 4.000 51.000 .084 -.796 2.8205 15 15.933 2.657 .143 .7952 30 30.0998 30 30.000 hs iz e 4.824 .

096 .346 .164 income travel vacation hs iz e age income hs iz e vacation travel age Pooled w ithin-groups c orrelations betw een discriminating variables and standardized canonic al disc riminant f unctions Variables ordered by absolute s iz e of correlation w ithin func tion.213 .) Eigenvalues Canonical Function Eigenvalue % of V arianc e Cumulativ e % Correlation 1 1.743 . First 1 canonical discriminant f unctions w ere us ed in the analysis. .469 .233 . .0 .801 a.130 df 5 Sig.0 100. 18-21 Wilks ' Lam bda Test of Function(s) 1 Wilks ' Lambda .4 cont’d.000 Standar dize d Canonical Dis crim inant Function Coe fficie nts Function 1 .541 .Results of Two-Group Discriminant Analysis (Table 18. Contd.786a 100.822 .209 Structure Matrix Func tion 1 .359 Chi-s quare 26.

025 -7.120 .427 .050 .975 18-22 income travel vacation hs iz e age (Cons tant) Unstandardiz ed coef f icients Functions at Group Ce ntroids Func tion 1 1.291 visit 1 2 Unstandardized c anonical disc riminant f unctions evaluated at group means Contd.Results of Two-Group Discriminant Analysis (Table 18. .291 -1.085 .4 cont’d.) Canonical Dis crim inant Function Coe fficie nts Func tion 1 .

000 18-23 visit 1 2 Total Prior .0 100.768 -36.500 .459 1.4) Prior Pr obabilitie s for Groups Cases Us ed in Analysis Unw eighted Weighted 15 15.678 1.381 .500 1.0% of original grouped cas es c orrec tly clas sified.832 -57.000 a Clas sification Res ults Original Clas sification Function Coe fficie nts Count % visit income travel vacation hs iz e age (Cons tant) 1 .000 30 30. Fisher's linear dis criminant f unctions .0 100.938 3.0 20. 90.0 a.532 2 .000 15 15.0 Total 15 15 100.Results of Two-Group Discriminant Analysis (Table 18.936 visit 1 2 1 2 Predicted Group Membership 1 2 12 3 0 15 80.0 .509 .218 .322 .628 2.

the Wilks' statistic is the product of the univariate for each function. . If the null hypothesis is rejected.Conducting Discriminant Analysis 18-24 Determine the Significance of Discriminant Function The null hypothesis that. The significance level is estimated based on a chisquare transformation of the statistic. If several functions are tested simultaneously (as in the case of multiple discriminant analysis). indicating significant discrimination. In SPSS this test is based on Wilks' . the means of all discriminant functions in all groups are equal can be statistically tested. in the population. one can proceed to interpret the results.

we can obtain some idea of the relative importance of the variables by examining the absolute magnitude of the standardized discriminant function coefficients. Another aid to interpreting discriminant analysis results is to develop a characteristic profile for each group by describing each group in terms of the group means for the predictor variables. .Conducting Discriminant Analysis Interpret the Results 18-25 The interpretation of the discriminant weights. With this caveat in mind. also called canonical loadings or discriminant loadings. or coefficients. Some idea of the relative importance of the predictors can also be obtained by examining the structure correlations. These simple correlations between each predictor and the discriminant function represent the variance that the predictor shares with the function. is similar to that in multiple regression analysis. there is no unambiguous measure of the relative importance of the predictors in discriminating between the groups. Given the multicollinearity in the predictor variables.

can then be determined by summing the diagonal elements and dividing by the total number of cases. or the percentage of cases correctly classified. estimated by using the analysis sample. offer a leaveone-out cross-validation option. The hit ratio. It is helpful to compare the percentage of cases correctly classified by discriminant analysis to the percentage that would be obtained by chance. .Conducting Discriminant Analysis 18-26 Access Validity of Discriminant Analysis Many computer programs. Classification accuracy achieved by discriminant analysis should be at least 25% greater than that obtained by chance. such as SPSS. The discriminant weights. The cases are then assigned to groups based on their discriminant scores and an appropriate decision rule. are multiplied by the values of the predictor variables in the holdout sample to generate discriminant scores for the cases in the holdout sample.

66333 2.20000 3.93333 Group Standard Deviations 1 2 3 Total 5.71594 2.20000 5.00000 0.97804 1.19722 1.09732 9.00000 51.40000 4.21667 4.00000 0.57000 50.00231 8.70000 4.19722 1.02512 1.00000 0.35702 1.57395 AGE Pooled Within-Groups Correlation Matrix INCOME TRAVEL VACATION INCOME TRAVEL VACATION HSIZE AGE 1.50000 56.18-27 Results of Three-Group Discriminant Analysis Table 18.00474 -0.38050 -0.09981 1.50555 1.03588 0.13529 1.34022 1.22080 -0.50000 4.00000 Contd.00000 -0.86667 4.56667 50.29718 6.97000 51.61434 12.00000 6.11000 64.20939 1.10000 4.5 Group Means AMOUNT INCOME TRAVEL VACATION HSIZE AGE 1 2 3 Total 38.88856 2.60117 8.90000 4.79523 1.30000 49.33089 HSIZE 8.10000 3.01326 1.25263 7. .30681 0.48551 1.05120 0.93333 3.

944 1.78790 0. Wilks' (U-statistic) and univariate F ratio with 2 and 27 degrees of freedom.00 3.8190 0.42076 0.4450 : * marks the two canonical discriminant functions remaining in the analysis. .33991 -0.14198 -0.00 0.53354 0.5 cont.87411 0.8020 5.517 4 0.88214 F 38.26215 0.00 93.76851 0.1797 0.49474 FUNC 2 -0. Variable INCOME TRAVEL VACATION HSIZE AGE Wilks' Lambda 0.93 0.18-28 Results of Three-Group Discriminant Analysis Table 18.24 100.16317 0.634 1.830 1.804 Significance 0.1626 0.0400 0.1664 44.1840 CANONICAL DISCRIMINANT FUNCTIONS Function 1* 2* Eigenvalue 3.88060 0.93 6.12932 0.831 10 0.8902 : 1 0.0000 0. Standardized Canonical Discriminant Function Coefficients INCOME TRAVEL VACATION HSIZE AGE FUNC 1 1.07 Cum Canonical After Wilks' % Correlation Function Chi-square df Significance : 0 0.04740 0.2469 % of Variance 93.52447 Contd.

27833 0.6952264E-01 0.45362* 0.44578 0.16576 FUNC 2 -0.5928055E-01 0.85556* 0.18-29 Results of Three-Group Discriminant Analysis Table 18.19319* 0.6284206E-01 (constant) -11.1002796 AGE 0.1542658 -0.58829* 0. Structure Matrix: Pooled within-groups correlations between discriminating variables and canonical discriminant functions (variables ordered by size of correlation within function) INCOME HSIZE VACATION TRAVEL AGE FUNC 1 0.65867 3 2.6197148E-01 TRAVEL 0.791600 Canonical discriminant functions evaluated at group means (group centroids) Group FUNC 1 FUNC 2 1 -2.09442 -3.41847 2 -0.04100 0.1265334 0.24020 Contd.34079* Unstandardized canonical discriminant function coefficients FUNC 1 FUNC 2 INCOME 0.2612652 HSIZE -0.4223430 VACATION -0.21935 0.07749 0.14899 0.1867977 0.40479 -0.5 cont. .

0% Percent of grouped cases correctly classified: 86.0% 1 10. Classification Results: Actual Group Group Group Group 1 2 3 No.0% 0 0.0% 0 2 0. of Cases 10 10 10 Predicted Group Membership 1 2 3 9 90. of Cases 1 2 3 Group 1 4 3 75.18-30 Results of Three-Group Discriminant Analysis Table 18.00% .5 cont.0% 0 0.0% 3 75.0% Group Group 2 3 4 4 0 0.0% 1 25.0% Percent of grouped cases correctly classified: 75.0% 1 10.0% 8 80.0% 0.0% 1 25.0% 20.0% 3 75.0% 1 0 25.67% Classification results for cases not selected for use in the analysis Predicted Group Membership Actual Group No.0% 0 0.0% 9 90.

0 -4.0 * indicates a group centroid -6.0 6.0 -2.0 4.0 0.0 1 1 1 1 *1 1 1 1 0.18-31 All-Groups Scattergram Fig.0 23 12 * 2 2 1 2 2 2 3 3* 3 3 3 3 3 -4.0 . 18.2 Across: Function 1 Down: Function 2 4.0 2.

0 2.0 -4.0 -8.3 13 13 Across: Function 1 13 Down: Function 2 13 13 * Indicates a 13 group centroid 13 113 112 3 112233 *1 1 1 2 2 2 2 3 3 * 1 1 2 2* 223 233 1122 2233 1122 223 11122 233 11222 223 1122 233 11122 2233 11122 223 1122 11122 233 8.0 0.0 -6.0 8.0 .0 6.0 -2. 18.0 4.18-32 Territorial Map Fig.0 0.0 4.0 -8.0 -4.

An F ratio is calculated for each predictor by conducting a univariate analysis of variance in which the groups are treated as the categorical variable and the predictor as the criterion variable. if it meets certain significance and tolerance criteria.18-33 Stepwise Discriminant Analysis Stepwise discriminant analysis is analogous to stepwise multiple regression (see Chapter 17) in that the predictors are entered sequentially based on their ability to discriminate between the groups. A second predictor is added based on the highest adjusted or partial F ratio. taking into account the predictor already selected. The predictor with the highest F ratio is the first to be selected for inclusion in the discriminant function. .

The Mahalanobis procedure is based on maximizing a generalized measure of the distance between the two closest groups.18-34 Stepwise Discriminant Analysis Each predictor selected is tested for retention based on its association with other predictors selected. The selection of the stepwise procedure is based on the optimizing criterion adopted. . The process of selection and retention is continued until all predictors meeting the significance criteria for inclusion and retention have been entered in the discriminant function. The order in which the variables were selected also indicates their importance in discriminating between the groups.

18-35 SPSS Windows The DISCRIMINANT program performs both twogroup and multiple discriminant analysis. To select this procedure using SPSS for Windows click: Analyze>Classify>Discriminant … .

5.65 5.5.: .:0945.5...: ::::.. 6510.&50.979.5.4..5:0945.31.#96/34 :.6:0945.5.506.:0945..50.3:: .%:3.3:: 6943.65605.50.945.5.

..6510.3:: 6943.5:0945.5.5.#96/34 .

65:.656../3::631/:30.:.0.69.5.3::..1:0945.473:6336:.51.:. 15.6.473 :9:9169.:.9.655 ..9:.09..065::.69: '7910.6316.909 "57.5106330.65:.473::169:.97./34:.965.65 '6.5. 50.965..331.9.4.31.31.5..5.9069.50.1:0945.331.51.0.9.473 .65 ".4.54/960.6569. 517515.6/0.1:.3461369796:9:.3303:.69.6. .656.7950 6. .9/./3.9/.3:. 69.6.:165.1/.31.9.5..::5.3:: :.3...4730.5...9.0./3: '09.669469 4.:.9.1:.

.

0:6.51964.9..31.: 6:631:9503115.9: ..5.1:..659:69.. 195..55 97.94.9..9:. :.6. &. :.. .0.473#96/34 &776:.:.5.473 ..51...79.0.473 .:.61. .39.945.1.65:.3.96/...43:.35.6.47366:631: ".0.3:::.

.

90611.9. .&./3:5. 3.3..965.195......6...6..0.1.9.:09.435064 !" .50.1..11 56.6 0.9./3 '6:631:..:..3 '% 4769.9:69.01..9.6:.9.9:..6.:...91./3: 55.65 '"! 6:631: & 6..43..69: !#!!'69#%'"% . #!!'69%'%"!./3.:.166:631 .: &'.

.

14769.65 .. %:69. .506:6316465.16&75....91.3 .:5. 5694.0.01& .43 '6.:.. 5064 '9...0.43 6:631.43 .3 .6.65 !6 :.6565%:69.65 .473 './3 55..3::&.

.

.3::&.3 .65 .16&75.0...3 . 5064 '9. 55.65 !6 :.:5.../3 065.01& .. 5694.6565%:69.0...:.43 6:631.506:6316465.14769.91.65 .43 '6.43 . .6. %:69.473 '.

473 '..: 6316./3 .&.5694.:.6565%:69.

3 .43 .91..0..01& .3 .6. 5064 '9. 55.506:6316465..43 '6.65 .0..14769.65 !6 :.. %:69..43 6:631.16&75. ..65 .

: .65605.5.4.5:0945.5.:0945.5.3:: :.6510..50..

5./3:.3::.4.465967: . 1:0945...9 50311:43.9 .7910.50..9..69:.69 ...5.33..61:0945...95.7910.56:3 5:.7:1:0945. '190.5.65:6.91:85.33/.5./3.4..:165.61 563::.

:6'6.%:3.

.. 08 $ 1/ 1/ ..3 83..05:47149...' 349.0 3.5.' 349.0/9$ 3.0 0898 /.0/9$ 3..94% /090 /0903& 349.85:47 39/044! /090 /0903& 0898 /. 08 349...0 3../-2.. 8 0.0/9$ 349.989..967:0945.05:47149.3 0.3 /090 /090 /0903& 98... 08 08 349. 0.9$5:47 8.9$5:47 .3 024. 0.989..0 .94% .1 0... 0.79 0..3 024... 349.. 8. 349.0 3.79 024.5.0/9$ 349...3 349.:6148980% 83.79 024.0 0898 /..79 024.. 0.79.0/9$ 3.. 0. 0.' 349../3 65.79 024. 0..0774 80.3:: '..:6148980% 08 08 349.79 0.

.

' 349. 08 08 349.3: 0.8/.3:193.0774.85:47 39/044! 39..3: 349.0774..7.../3..7.80-.343.327.79 0.140809:48-.0774.3: 349.3 %:3. 0.79 024..79 349.7.8/. 0.80-. 8349.3: 0..' 8349../0/7.8/30090-8349.3:139349../0/7..3:193.327. 39..98/3.140809:48-. 0.:6'6. 08 024.7..327.98/3.85:47 39/044! 349.349.3 349. 0../3.343.-/070/7480-.0774.8/30090-8349.-/070/7480-.327.3:139349...

3:193./-2..343. 8 8 8 349.3.5./3 065.967:0945.30 80:. 8 349./3.0774 ../-2.8.30 .:2: 0.7..1 79.3:14980% .3:14980% 8 349. .3: 0.:68 07.9.07:9. 349.30 80:..:79$ .:68 .343. 88.. ./-2.3..3:: '.8/.3:93./0/7.327.1 65.327.'14 . 0:.5.987 . 093/08:0708349.9$ $ $ 1/ 1/ 07..343.1104349.

.

:6'6.%:3.

/0/7..:.3:1 93.343./0/7.8/../3..025:479.79 024.3 024.983& 93.3:: '..1104349.3: 349.../0/7.5.983& 349.8349. 349.343.1104.3: 93./3. 0.327.8/.3: 8930.1 65. 0.. .1104349./3.967:0945./3..9834 0.343.08349..79 0.8.3: 8/479305:479..3:1 83./09..025:479.:. 8/479305:479./09.3 0...1 83.3: 98.8.9834 93.327.327.08349.5.327. 08 349.8349./3 065./0/7.3:93..983& 349.343.. 98..1104.3:93.983& 8930.

.

98.94% 5870-20 5870-20 5:47/09.188./05:47.3 /0188. 747! . 93:4 93:4 .9.038 708 93. 08 349./07! 5:47/09.. 349.8/7. 8 8 .. 9:80#349. 9:80#349.3:349.-47!747! %:3. 98.-47!747! 85:47741809-. ..808.37414 . 0..37 ..79 0.94% .188.188./07! 98. 98. .8349..33/08&808.37 . 85:47741809-. /090 /0903& 88...3:193. 8930...9834 0.:6'6.0774.1104349.327.79 024..94% 98...

5.3:: '.967:0945./3 .5.

.

6510.5.3:: .5.5:0945.

65 ':50.:..5.731:0945.9 8..506:0945..1 2 5&#&&.7673.1/.50..:.5.:6 43.65 '53376.:.0.:165.:5.050.1:43..69.79610..33.3 50.:16532: :9. 4.945..32: :..50.5:6.0.0 2 :.6.3:: ..65:5.::.65:.5. .5..4.0.:.331:0945.:.:/..9.65.9.33967:.5.5/:.30.5.56:3 .:.&50.5033::.

.5:694.979...579601.1:0945.5 :50.9.65650.9.0 .:.:::90.53376. :8.1510.656...: ..6 5.9:3.:.5.

6510.5..3:: 5.979.5:0945.: .5.%:3.

656.7399::65..94:6. 50.43..979.: &641..:0.:: :43.967: .455.50.65979:5..5.06335...51:0945.. 4..65 56..65 0605.5:69..69.93..979.4769./3: .5.455./:63..506.9.65.07910.1.5. .69:0.6..5650.9:.336.3:6 /6/.93.5/.07963 69.5.1:0945.0. 93.0.4769.:96.:.5.9111:0945.7910.51.9.6 1367.65:.506.50.9.5..4/6:4.9.5..5.906993.69 .69.:690605./3:./3:/.9..3:60. '5.69:51:0945.1:0945.54510.7910.51.65:/.6.9:.331 0.6.9674..3::9:3..4769..506..5.5..7910. 7910...::.5:641.90.36.09675..5..69:.5.9.16..50.15: '::473 06993.9.56/.3:: 5.15: 691:0945.:..0967/1:09/5 .9.51/.9:56 5..543.7910.

6510.5.31.6:0945.3:: .5.3:: 00::.5:0945.5.5.

97969.50647.:&#&&69..3.4::0. .

65.

6.096::.

631/6/..51 115/. 9.9.5.4.65 '1:0945.9.669 .51/0.6.::.1:0945..50 .473 '0.473.91:0945...5.7910.0.51.7905..009.:373..3:::631/. :069:69.:165.3 03.9 ..779679..6316.01/1:0945.3.31.5..7905.:.731/.::06990.::.659.5.0 .:.60..5..5.6/.51/0.5..5.9451/:445.::06990..303.::0.5..:: ..473.3:6.50 3.6567..3::.5/ 1.::10.1/:5.5.9.6.::51.3345../3:5.5.69 .60647.:..5.10:6593 '..6316..6967:/.7905.3:: :. :069:.1.354/960..65.943.:...9..::1/1:0945.::5.60.65....

.

:6'9. %:3.

3 #6631..967:0945..3:: '.3 '%'"!& 967&.5..5.51.91.65: '6./3 967 .5: "!'!" '6.5..

9 !" '% '"!& !" '% '"! & .967:6993..65 ..

.

.

.

65.1 .

.

:6'9. %:3.

5. 32: .5./3 065.3:: '.967:0945.

9.515.9.9..5650.3. &50.6..0 .50 6993....50 !"!&% !!'!'"!& 50.4/1.932: 5.5119:69164 .3..:./3 !" '% '"! & 32: .:.6550.9.65 64.652 .

60..3:0945.65:94.5650.5.91&50.51.5650.50 4.31:0945.50.: !" '% '"! & ! .50.:8.5..911.65605.3:: &..555.5.92:.

.

! .

65.1 .

.

%:3.:6'9.

967:0945.9 .90./3 065.5.5.3:: '.5.9 #6631.. &.

5650.967:06993..65 ! ! .51:0945.5.510.65:/./3:69191/:606993.9. 50.../3:.65: .9.550.65.31:0945.5.

5650.31:0945..51.5.650605.: ! ! !" .50. !" & '"! '% 5:.9110.

.

'% '"! .

.

& .

.

.

5. 065:.. .

.

5: 96705. ..65:.5.961: 967 ! ! .5650.50.9674.1.31:0945.3..

.

.

1 . 65.

.

:6'9. %:3.

./3 065.3:: #910.1 967 4/9:7 0..::1 .::56.::0.967:0945.::06990.659:3...:: #910.::06990.5.:30.3967 967 967 967 !6 6.::0.:: 967 967 967 #905.3967 !6 6.3:: '.1 967 4/9:7 #905.:690.5.: 0..169:5. 3.696710.696710.303.303.65%:3.5.::1 3.

.

33.

967:&0.65 ...4 096::50.65 6550.99.

510..961 .:.967 05.

.

.

.

.

.7 096::50. '99.3 .961 . 96705.65 510.65 6550.:.69.

.

.

.

.

.

.

.

9.366:.69.967: 5 9. 7910.967:.33/./3.5.9.7910.9..69:.65.7:1:0945.:165.50.9.111/.:..3.3 9.0065.9.3::6.7399::65 :.5050 .69:.39.5.9.:.7910.61:0945.1697..169./3.255.6..3:::. 4.5.169503:655..0.6:0..:165.9 5..1. ..1:0945.69/ 06510.:09.7:43.1:30.1:.51 ....6:.5.7::0945.51.95.5:50.5..639..5.6/ :30.:.5.07910.50. :06517910..965.91:85..6.5.7. 9...9..9 .:.69.5.303..690.3:: &.09.9:../3 '7910.69.5.1 .9./. &.6 :.5009. .

.

3:: .5.5.69:30.337910.965.:.50/.519. &.:165 ../3:9:30.915. 1:0945.1:.:..:1 65.07910.1 ' ./55.5.1 '7960::6:30.5.7:796019:/.59.51 5.6.3:6 510..65/.56/: 796019 :/.5009.67.519.3.50.:1654.65 ':30.1699.1:.5.3...::60.:.5 .5.69 503:65.31 4.5.94769.1.9.97910..65.7::0945.6036::.4509. 967: '691950.:96..5.69:4.69::30.967: .656.5/.:50.:..5051:0945.65.45.167.65:065.9.65.

.

&#&&516: '&% !!'7969..479694:/6.6.

.5.:::0945.3:: '6:30.5143.Ã .5.731:0945. .:796019:5&#&&69516:0302 5.33. 967.5.

- Defining the Marketing Research Problem and Developing an Approach.finaL
- Malhotra10 Tif (1)
- Malhotra Mr05 Ppt 22
- malhotra05_tif.doc
- Defining the Objectives
- Marketing Research Module 8 Sampling Design
- Marketing Research Module 5 Qualitative Research
- Marketing Research Module 4 Experimentation
- Malhotra Mr05 Ppt 21
- Malhotra Mr05 Ppt 20
- discriminant analysis
- C2
- Sampling Techniques
- TN04 - Segmentation Technical Note
- Naresh K. Malhotra Review of Marketing Research
- malhotra05.ppt
- defining the marketing research problem and developing an approach
- Malhotra Selected
- Research Design
- Malhotra 05
- Session 5 - Sampling and Data Collection (BB)
- MRAO6ETBM_chapter_7.doc
- Sample Assignment eGM07.doc
- data collection
- 04 malhotra04
- malhotra13.ppt
- malhotra12_tif.doc
- Marketing Research Module 2 Secondary Data
- malhotra16_tif.doc
- malhotra_mr05_ppt_06
- Malhotra18-Discriminant Analysis-With SPSS Output Inserts-2003 Format

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd