# Ch apter E ightee n

Discriminant and Logit Analysis

18-1

Cha pter Ou tl ine
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

Cha pter Ou tl ine
1) i. ii. Multiple Discriminant Analysis Formulation Estimation

iii. Determination of Significance iv. Interpretation v. 2) Validation

Stepwise Discriminant Analysis

18-3

Ch apter O utl ine
9) The Logit Model i. ii. Estimation Model Fit

iii. Significance Testing iv. Interpretation of Coefficients v. An Illustrative Application

10) Summary

Sim il aritie s and Diffe renc es b etwe en ANO VA , R eg re ss ion , and D isc rim ina nt An aly sis
Table 18.1

ANOVA
Sim il ari ti es Number of de pen den t varia bl es Number of indep en den t varia bl es On e

REGR ESS ION D ISCR IMIN ANT/LOG IT
On e One

Mu lt ipl e

Mu lt ipl e

Multi pl e

Differ en ces Nature of the de pen den t Me tr ic Me tr ic varia bl es Nature of the indep en den t Ca te goric al Me tr ic varia bl es

Categor ic al Metr ic
18-5

Discr imina nt Ana lysis
Discr imi nan t an al ys is is a technique for analyzing data when the criterion or dependent variable is categorical and the predictor or independent variables are interval in nature. The objectives of discriminant analysis are as follows:  Development of discri min ant fun ct io ns, or linear combinations of the predictor or independent variables, which will best discriminate between the categories of the criterion or dependent variable (groups).  Examination of whether significant differences exist among the groups, in terms of the predictor variables.  Determination of which predictor variables contribute to most of the intergroup differences.  Classification of cases to one of the groups based on the values of the predictor variables.  Evaluation of the accuracy of classification.

18-6

Discr imina nt Ana lysis

When the criterion variable has two categories, the technique is known as two -gr oup d iscr imi na nt analysis. When three or more categories are involved, the technique is referred to as mu ltip le di scr imi nan t an al ys is . The main distinction is that, in the two-group case, it is possible to derive only one discriminant function. In multiple discriminant analysis, more than one function may be computed. In general, with G groups and k predictors, it is possible to estimate up to the smaller of G - 1, or k, discriminant functions. The first function has the highest ratio of between-groups to within-groups sum of squares. The second function, uncorrelated with the first, has the second highest ratio, and so on. However, not all the functions may be 18-7

Geometric Interpretation
Fig. 18.1
X2 G1
1 1 1 1 1 1 1 1 1
2 11 1 1 2 22 21

2 2 2 2 2 2 22 22

G2

G1 G2 X1 D
18-8

Discr iminant An alysis Mo del
The dis cr imi nant an alysis 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, or weights (b), are estimated so that the groups differ as much as possible on the values of the discriminant function. This occurs when the ratio of between-group sum of squares to withingroup sum of squares for the discriminant scores is at a maximum.
18-9

St ati stics A sso ciated w ith Discr iminant An alysis

Ca nonical co rre lat io n. Canonical correlation measures the extent of association between the discriminant scores and the groups. It is a measure of association between the single discriminant function and the set of dummy variables that define the group membership. Ce ntr oid. The centroid is the mean values for the discriminant scores for a particular group. There are as many centroids as there are groups, as there is one for each group. The means for a group on all the functions are the group centroids.

Cl as sif icat ion mat rix . Sometimes also called confusion or prediction matrix, the classification matrix © 2007 Prentice Hall contains the number of correctly classified and

18-10

Stati sti cs Asso ciated w ith Discr imina nt Ana lysis

Discr imi nan t function co ef fici en ts. The discriminant function coefficients (unstandardized) are the multipliers of variables, when the variables are in the original units of measurement. Discr imi nan t sco res. The unstandardized coefficients are multiplied by the values of the variables. These products are summed and added to the constant term to obtain the discriminant scores. Eig en val ue . For each discriminant function, the Eigenvalue is the ratio of between-group to within-group sums of squares. Large Eigenvalues imply superior functions.

18-11

St ati stics A sso ciated w ith Discr iminant An alysis

F va lu es an d the ir si gni fican ce . These are calculated from a one-way ANOVA, with the grouping variable serving as the categorical independent variable. Each predictor, in turn, serves as the metric dependent variable in the ANOVA.
Group mean s and gro up st an dar d de vi at ions. These are computed for each predictor for each group. Po ole d wi thin -gro up co rrel at io n ma tr ix. The pooled within-group correlation matrix is computed by averaging the separate covariance matrices for all the groups.
18-12

Stati st ics A sso ciated with Di sc ri mi nant Ana ly si s

St anda rdiz ed d is crimin an t fu nct ion coefficient s. 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. St ruc tu re corr elat ions. Also referred to as discriminant loadings, the structure correlations represent the simple correlations between the predictors and the discriminant function. Tot al c orre la tion ma trix . If the cases are treated as if they were from a single sample and the correlations computed, a total correlation matrix is obtained. λ for each Wilks 'λ . Sometimes also called the U statistic, Wilks' predictor is the ratio of the within-group sum of squares to the total sum of squares. Its value varies between 0 and 1. Large values of λ (near 1) indicate that group means do not seem to be different. Small values ofλ (near 0) indicate that the group means seem to be different.
18-13

Co nduc ti ng Di scrim inant An al ysi s
Fig. 18.2

Formulate the Problem Estimate the Discriminant Function Coefficients Determine the Significance of the Discriminant Function Interpret the Results Assess Validity of Discriminant Analysis

Co nduc ti ng Di scrim inant An alysi s Formulate th e Pro bl em
 

Identify the objectives, the criterion variable, and the independent variables. The criterion variable must consist of two or more mutually exclusive and collectively exhaustive categories. The predictor variables should be selected based on a theoretical model or previous research, or the experience of the researcher. One part of the sample, called the estimation or an al ysis samp le, is used for estimation of the discriminant function. The other part, called the holdout or va li dat ion samp le , is reserved for validating the discriminant function. Often the distribution of the number of cases in the analysis and validation samples follows the distribution in the total sample. 18-15

Info rm at ion on Resort Vis it s: Analys is Sample
Table 18.2
Amount Annual Res or t (\$ 00 0) Attitud e Famil y to Fa mily Importance Toward Ho use ho ld Ag e of Hea d of Visi t Inc ome Ho use hold Famil y Si ze Attached N o. Vaca tion

Sp ent on Travel Vacatio n

1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 1 10 1 11 1 12 1 13 1 14 1 © 2007 Prentice Hall

50.2 70.3 62.9 48.5 52.7 75.0 46.2 57.0 64.1 68.1 73.4 71.9 56.2 49.3

5 6 7 7 6 8 5 2 7 7 6 5 1 4

8 7 5 5 6 7 3 4 5 6 7 8 8 2

3 4 6 5 4 5 3 6 4 5 5 4 6 3

43 61 52 36 55 68 62 51 57 45 44 64 54 56

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)

18-16

Info rm at ion on Resort Vis it s: Ana lys is Sam pl e
Table 18.2, cont.Annual
Amou nt Res ort Sp ent on Income Travel Vacati on 16 17 18 19 20 21 22 23 24 25 26 27 28 29 © 2007 Prentice Hall 2 2 2 2 2 2 2 2 2 2 2 2 2 2 (\$ 000) Attitud e Famil y to Fam ily Importance Tow ar d Ho use ho ld Ag e of Si ze Hea d o f Visi t Ho use hol d Famil y A ttached N o. V ac ation

32.1 36.2 43.2 50.4 44.1 38.3 55.0 46.1 35.0 37.3 41.8 57.0 33.4 37.5

5 4 2 5 6 6 1 3 6 2 5 8 6 3

4 3 5 2 6 6 2 5 4 7 1 3 8 2

3 2 2 4 3 2 2 3 5 4 3 2 2 3

58 55 57 37 42 45 57 51 64 54 56 36 50 48

L L M M M L M L L L M M L L

(1) (1) (2) (2) (2) (1) (2) (1) (1) (1) (2) (2) (1) (1)

18-17

Info rm ati on on Re sort Vis it s: Hol dout Sam ple
Table 18.3
Amount Res or t Sp ent o n Income Travel Famil y Vacati on (\$ 00 0) Annual Attitud e Famil y to F am ily Importance Tow ard Ho use ho ld Ag e of Si ze Hea d of Visi t Ho use hol d

Attached N o.

Va cation

1 1 2 1 3 1 4 1 5 1 6 1 7 2 8 2 9 2 © 2007 10 Prentice Hall 2

50.8 63.6 54.0 45.0 68.0 62.1 35.0 49.6 39.4 37.0

4 7 6 5 6 5 4 5 6 2

7 4 7 4 6 6 3 3 5 6

3 7 4 3 6 3 4 5 3 5

45 55 58 60 46 56 54 39 44 51

M(2) H (3) M(2) M(2) H (3) H (3) L (1) L (1) H (3) L (1)

18-18

Con duc ting D isc rimi na nt A na lysi s Esti ma te t he Di sc ri mi nant Func tion C oef fic ient s
The dire ct me tho d involves estimating the discriminant function so that all the predictors are included simultaneously. In st ep wi se dis crimin an t an aly sis , the predictor variables are entered sequentially, based on their ability to discriminate among groups.

18-19

Ana lys is
Table 18.4      GROUP MEANS VISIT
1 2 Total INCOME 60.52000 41.91333 51.21667 TRAVEL VACATION 5.40000 4.33333 4.86667 5.80000 4.06667 4.9333 HSIZE 4.33333 2.80000 3.56667 AGE 53.73333 50.13333 51.93333

Group Standard Deviations 1 2 Total 9.83065 7.55115 12.79523 1.91982 1.95180 1.97804 1.82052 2.05171 2.09981 1.23443 .94112 1.33089 HSIZE 8.77062 8.27101 8.57395 AGE

Pooled Within-Groups Correlation Matrix INCOME TRAVEL VACATION INCOME TRAVEL VACATION HSIZE AGE 1.00000 0.19745 0.09148 0.08887 - 0.01431 1.00000 0.08434 -0.01681 -0.19709

1.00000 0.07046 0.01742

1.00000 -0.04301

1.00000

Wilks' (U-statistic) and univariate F ratio with 1 and 28 degrees of freedom Variable INCOME TRAVEL VACATION HSIZE AGE

Wilks' 0.45310 0.92479 0.82377 0.65672 0.95441

F 33.800 2.277 5.990 14.640 1.338

Significance 0.0000 0.1425 0.0209 0.0007 0.2572

Co nt.
18-20

Resu lts of Two -G ro up Di sc rim ina nt Ana lysi s
Table 18.4, cont.
CAN ON ICAL DISCR IM IN ANT F UN CT IONS
Fun ction 1* Eig env alue 1.78 62 % of Var ia nce 10 0.00 Cum C anon ica l Aft er Wilks ' % Co rrelatio n Funct ion λ Ch i-s quare : 0 0 .35 89 26 .130 1 00 .00 0 .800 7 : in the analy sis . d f Si gnifi cance 5 0.00 01

* marks t he 1 canon ical d isc riminan t fun ctio ns remaining

Stand ard Cano nical Dis criminan t Funct ion Co eff icients
FUN C INCO ME TR AV EL VACA TI ON HSI ZE AG E 0.74 0.09 0.23 0.46 0.20 30 1 61 1 32 9 91 1 92 2 1

Structure

Po oled within -gr oup s c or relations be twee n dis cr iminating vari ables & canon ical d isc riminan t fu nctio ns (variables or dered by siz e o f co rr elation within fu nctio n) FUN C INCO ME HSI ZE VACA TI ON TR AV EL AG E

Matri x:

1

0.82 0.54 0.34 0.21 0.16

20 2 09 6 60 7 33 7 35 4

Co nt.
18-21

Res ults of Tw o-G ro up Di sc rim ina nt Analys is
Table 18.4, cont.
Un st anda rdized Ca no nica l Discrim ina nt Fun ct ion Coefficient s FUNC 1 INCO ME 0. 84 76 710 E- 01 TRAV EL 0. 49 64 455 E- 01 VA CA TI ON 0. 12 02 813 HSI ZE 0. 42 73 893 AG E 0. 24 54 380 E- 01 (co nst ant ) -7.97 54 76 Ca non ical discrim ina nt fun ct ions evalua ted at grou p means (grou p cent ro ids ) Gro up FUNC 1 1 1.2 911 8 2 -1.29 11 8 Cla ssifica tion result s fo r cases sel ect ed for use in ana ly sis Pr edict ed Gro up Ac tua l Grou p No . of Ca ses 1 Gro up 1 15 12 80 .0 % 0 0. 0%

Memb ersh ip 2 3 20 .0 % 15 10 0. 0%

Gro up

2

15

Per cent of gro uped ca ses co rrect ly cla ssifi ed : 90. 00 %

Co nt.
18-22

Res ults of Tw o-G ro up Di sc rim ina nt Analys is
Table 18.4, cont.

Clas si fication Re sult s for cases n ot selec te d for use in the anal ysi s (ho ldou t sa mple )
Predicted Actual Group No. of Cases Group Group 1 2 6 6 Group 1 4 66.7% 0 0.0% Membership 2 2 33.3% 6 100.0%

Percent of grouped cases correctly classified: 83.33%.

18-23

Con duc ting Di scr imi na nt A na lysi s De term ine t he S ign if icanc e o f Di sc ri mi nant Func tio n

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

Int erpre t t he R esul ts
 

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

Conduc ting Di sc rim inant Ana lys is Asses s Vali dit y of Di scr im ina nt A nal ys is

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

18-26

Re sul ts o f Three -Gro up D isc rimi na nt Analy sis
Table 18.5 Grou p Me an s
AM OUN T 1 2 3 Total INCO ME 38.57 000 50.11 000 64.97 000 51.21 667 TR AV EL 4. 500 00 4. 000 00 6. 100 00 4. 866 67 VACA TI ON 4.70 000 4.20 000 5.90 000 4.93 333 HSI ZE 3 .100 00 3.40 00 0 4 .200 00 3 .566 67 AG E 50 .3000 0 49 .5000 0 56 .0000 0 51 .9333 3

Grou p Stand ard D evi ation s
1 2 3 Total 5.297 18 6.002 31 8.614 34 12.79 523 1. 715 94 2. 357 02 1. 197 22 1. 978 04 1.88 856 2.48 551 1.66 333 2.09 981 1.19 72 2 1.50 55 5 1.13 52 9 1.33 08 9 8.09 732 9.25 263 7.60 117 8.57 395

Pooled With in -G rou ps Correlati on Matri x
INCO ME INCO ME TRA VEL VACAT ION HSI ZE AG E

TR AV EL 1. 000 00 0. 035 88 0. 004 74 -0.34 02 2

VACA TI ON

HSI ZE

AG E

1.000 00 0.051 20 0.306 81 0.380 50 -0.209 39 1.00 000 0.22 080 -0 .01 32 6 1.00 00 0 -0. 025 12 1.00 000 Co nt.
18-27

Re sul ts o f Three -Gro up D isc rimi na nt Analy sis
Table 18.5, cont.
Wi lk s' (U-st at ist ic) and un iv ari ate F rat io wit h 2 and 2 7 deg rees of freed om .
Var iable INCO ME TRA VEL VACAT ION HSI ZE AG E Wilk s' L ambda 0.26 0.78 0.88 0.87 0.88 21 5 79 0 06 0 41 1 21 4

F
38.00 3 .634 1 .830 1 .944 1 .804

Sig nifi cance 0.00 00 0.04 00 0.162 6 0.18 40

0.17 97

CANONI CAL DISC RI MINA NT FUNCT IONS
Fun ction Eig env alue Sig nifi cance 1* 2* 4 3.81 90 0.24 0.24 69 % of Vari ance 9 3.93 6.07 Cum C anoni ca l After % Co rrelatio n Funct ion 93 .93 1 00.0 0 : 0 0 .890 2 0 .445 0 Wi lks ' λ Chi -s quare 0.16 64 : 1 : 44 .831 0.80 20 df 10 0.00 5 .51 7

* marks t he two canon ical disc rimin ant fu nctio ns remainin g in the analy si s.

St and ard ize d Ca no ni cal Disc rimi na nt Funct ion Coeff ici ent s
INCO ME TRA VEL VACAT ION HSI ZE AG E

FUN C 1 1.04 740 0.33 991 -0 .141 98 -0 .163 17 0.49 474

-0 .420 76 0.76 851 0.53 354 0.12 932 0.52 447

FUN C

2

Co nt.
18-28

Resul ts of T hree -Group Discr imi na nt Ana lysi s
Table 18.5, cont.
Structure Matr ix: Pooled wi th in- gr oup s co rr el atio ns betwe en d iscr iminating variab les and canon ical dis crimin ant functio ns (v aria ble s o rde re d b y si ze o f correl atio n w ithin f unctio n)
INCO ME HSI ZE VACATI ON TR AV EL AG E FUN C 1 0.85 55 6* 0.19 31 9* 0.21 93 5 0.14 89 9 0.16 57 6 FUN C 1 0.15 42 658 0.18 67 977 -0 .695 226 4E- 01 -0 .126 533 4 0 .592 805 5E- 01 -1 1.09 442 FUN C 2 -0 .278 33 0.07 74 9 0.58 82 9* 0.45 36 2* 0.34 07 9* FUN C 2

Un stand ard ize d cano nical d is cri mina nt functio n co eff icients
INCO ME TR AV EL VACATI ON HSI ZE AG E (co nstant) Gro up 1 2 3 FUN C 1 -2 .041 00 -0 .404 79 2 .445 78 0.42 23 430 0.26 12 652 0.10 02 796 0.62 84 206 E-0 1 -3 .791 600 -0 .619 714 8E- 01

Cano nica l d iscri minant f unctio ns eval uated a t gro up mea ns (g ro up centro ids)
FUN C 2 0.41 84 7 -0 .658 67 0.24 02 0

Co nt.
18-29

Re sul ts o f Three -Gro up D isc rimi na nt Analy sis
Table 18.5, cont.
Classi fic at ion Re sul ts:
Actual Grou p Gr oup Gr oup Gr oup 1 2 3 Pr edi cted No. o f Cas es 10 10 10 9 90 .0% 1 10 .0% Gro up M embers hip 1 2 1 10 .0% 9 90 .0% 0 0.0% 0 0.0% 8 80 .0% 3

0 2 0.0% 20 .0% Perc ent of gr ouped cases c orr ectly clas sif ied: 8 6.67 %

Classi fic at ion resul ts for cases n ot sel ect ed for use in the ana lysi s
Actual G ro up Gro up Gro up Gro up 1 2 3 No. o f C ases 4 4 4 Pr edi cted 1 3 75 .0% 0 0.0%

Gro up M embers hip 2 3 1 25 .0% 3 75 .0% 0 0.0% 1 25 .0% 3 75 .0%

1 0 25 .0% 0.0% Perc ent of gr ouped cases c orr ectly clas sif ied: 7 5.00 % © 2007 Prentice Hall

18-30

Al l-Groups S ca tter gram
Fig. 18.3
Acros s: Functi on 1 Down : Fu nct ion 2 4. 0 1 1 1 1 1 *1 1 1

0. 0 -4.0

23 12 * 2 2 1 2 2 2

3 3* 3 3 3 3

3

* in dic ates a grou p ce ntroid -6.0

-4.0

-2.0

0.0

2.0

4.0

6.0
18-31

Te rrit orial Ma p
Fig. 18.4
13 13 Acr oss: Function 1 13 Down : Fu ncti on 2 13 13 * Ind icates a 13 group centro id 13 113 1 1 2 1 3 1 2 2 3 3 * *1 1 1 2 2 2 2 3 3 1 1 2 * 2 2 3 2 2 3 3 1 1 2 2 2 3 3 1 12 2 2 2 3 2 1 1 1 2 2 3 3 12 2 2 1 2 2 3 2 1 2 1 2 3 3 1 12 2 1 2 2 3 3 1 12 1 2 2 2 2 3 1 1 2 1 1 12 2 2 2 3 3

8.0

4.0

0.0 -4.0

-8.0 -8.0

-6.0

-4.0

-2.0

0.0

2.0

4.0

6.0

8.0

18-32

St epw ise Discr imi na nt Ana lys is

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. 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. The predictor with the highest F ratio is the first to be selected for inclusion in the discriminant function, if it meets certain significance and tolerance criteria. A second predictor is added based on the highest adjusted or partial F ratio, taking into account the predictor already selected.

18-33

Stepw ise D iscr imi na nt Ana lys is

Each predictor selected is tested for retention based on its association with other predictors selected. 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 selection of the stepwise procedure is based on the optimizing criterion adopted. The Mah ala no bi s pr oce du re is based on maximizing a generalized measure of the distance between the two closest groups. The order in which the variables were selected also indicates their importance in discriminating between the groups.
18-34

Th e Logi t Mo del

The dependent variable is binary and there are several independent variables that are metric The binary logit model commonly deals with the issue of how likely is an observation to belong to each group It estimates the probability of an observation belonging to a particular group

18-35

Bi na ry Log it Mo del Formul ati on
The probability of success may be modeled using the logit model as:

Or

 P  log   = a +a 1 − P 
e 0

1

X

1

+a

2

X

2

+... +a

k

X

k

Or

 P =  ∑a X log  1 − P 
n e i= 0 i i

18-36

Mo del F ormu latio n
k

exp( ∑ P = 1 + exp(
i =0

a X
i k i =0 i

i

)
i

∑ a X

)

Where:

P Xi ai

= Probability of success = Independent variable i = parameter to be estimated.
18-37

Prop erti es of the Lo git Model

Although Xi may vary from − ∞ to + ∞ , P is constrained to lie between 0 and 1. When Xi approaches − ∞ , P approaches 0. When Xi approaches + ∞, P approaches 1. When OLS regression is used, P is not constrained to lie between 0 and 1.
18-38

Es ti mati on a nd Mod el F it

The estimation procedure is called the maximum likelihood method. Fit: Cox & Snell R Square and Nagelkerke R Square. Both these measures are similar to R2 in multiple regression. The Cox & Snell R Square can not equal 1.0, even if the fit is perfect This limitation is overcome by the Nagelkerke R Square. Compare predicted and actual values of Y to determine the percentage of correct predictions.
18-39

 

 

Signi fi can ce Tes ting

18-40

Inte rpre tati on o f Co eff ici ents

If Xi is increased by one unit, the log odds will change by ai units, when the effect of other independent variables is held constant. The sign of ai will determine whether the probability increases (if the sign is positive) or decreases (if the sign is negative) by this amount.

18-41

Ex plai ning Bran d Lo yal ty
Table 18.6
No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Loyalty 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Brand 4 6 5 7 6 3 5 5 7 7 6 5 7 5 7 3 4 2 5 4 3 3 3 4 6 3 4 3 5 1

Product 3 4 2 5 3 4 5 4 5 6 7 6 3 1 5 1 6 5 2 1 3 4 6 4 3 6 3 5 5 3

Shopping 5 4 4 5 4 5 5 2 4 4 2 4 3 4 5 3 2 2 4 3 4 5 3 2 6 3 2 2 3 2

18-42

Resu lts of Log isti c Re gres si on
Table 18.7

Dependent Vari able Enc oding Original Value Not Loyal Loyal Internal Value 0 1

M odel Su m m ary -2 Log likelihood 23.471(a) Cox & Snell R Square .453 Nagelkerke R Square .604

Step 1

a Estimation terminated at iteration number 6 because parameter estimates changed by less than .001.
18-43

Resu lts of Log isti c Regres si on
Table 18.7, cont.

Cl ass if icati on Ta ble a
Predicted Loyalty to the Brand Not Loyal Loyal 12 3 3 12 Percentage Correct 80.0 80.0 80.0

Observed Step 1 Loyalty to the Brand Overall Percentage

Not Loyal Loyal

a. The cut value is .500 Var iab les in t he Equation a
Step 1 Brand Product Shopping Constant B 1.274 .186 .590 -8.642 S.E. .479 .322 .491 3.346 Wald 7.075 .335 1.442 6.672 df 1 1 1 1 Sig. .008 .563 .230 .010 Exp(B) 3.575 1.205 1.804 .000
18-44

a.Variable(s) entered on step 1: Brand, Product, Shopping.

SP SS Wi ndow s
The DISCRIMINANT program performs both twogroup and multiple discriminant analysis. To select this procedure using SPSS for Windows click: An aly ze> Cl assi fy>D iscri min ant … The run logit analysis or logistic regression using SPSS for Windows, click:

An aly ze > Reg ress io n>Bin ary Logisti c …

18-45

SPSS Wi ndo ws: Two- gro up Dis cri mi nant
1. 2. 3. 4.

5.

6. 7.

8.

9.

Select ANALYZE from the SPSS menu bar. Click CLASSIFY and then DISCRIMINANT. Move “visit” in to the GROUPING VARIABLE box. Click DEFINE RANGE. Enter 1 for MINIMUM and 2 for MAXIMUM. Click CONTINUE. Move “income,” “travel,” “vacation,” “hsize,” and “age” in to the INDEPENDENTS box. Select ENTER INDEPENDENTS TOGETHER (default option) Click on STATISTICS. In the pop-up window, in the DESCRIPTIVES box check MEANS and UNIVARIATE ANOVAS. In the MATRICES box check WITHIN-GROUP CORRELATIONS. Click CONTINUE. Click CLASSIFY.... In the pop-up window in the PRIOR PROBABILITIES box check ALL GROUPS EQUAL (default). In the DISPLAY box check SUMMARY TABLE and LEAVE-ONE-OUT CLASSIFICATION. In the USE COVARIANCE MATRIX box check WITHIN-GROUPS. Click CONTINUE. Click OK.
18-46

SP SS Wi ndow s: Logi t An alysi s
1. 2. 3.

Select ANALYZE from the SPSS menu bar. Click REGRESSION and then BINARY LOGISTIC. Move “Loyalty to the Brand [Loyalty]” in to the DEPENDENT VARIABLE box. Move “Attitude toward the Brand [Brand},” “Attitude toward the Product category [Product},” and “Attitude toward Shopping [Shopping],” in to the COVARIATES(S box.) Select ENTER for METHOD (default option) Click OK.
18-47

4.

5. 6.