You are on page 1of 12

Types of Factor Analysis:

Exploratory
Confirmatory

Assumptions of FA:

High Correlation between (Use Bartley tests)


Sample size has to be very large

If KMO value is less than 0.5 , it states that the sample size is not sufficient

Types / techniques under Varimax

Oblique & Ortho

.5 is the criteria cut off value to decide the variable whih belongs to a factor.

Total Variance Explained

Component Initial Eigenvalues Extraction Sums of Squared Loadings

Total % of Variance Cumulative % Total % of Variance Cumulative %

1 6.390 13.891 13.891 6.390 13.891 13.891


2 4.352 9.460 23.351 4.352 9.460 23.351
3 2.880 6.262 29.612 2.880 6.262 29.612
4 2.715 5.902 35.515 2.715 5.902 35.515
5 2.110 4.587 40.101 2.110 4.587 40.101
6 1.806 3.925 44.027 1.806 3.925 44.027
7 1.757 3.820 47.846 1.757 3.820 47.846
8 1.647 3.581 51.427 1.647 3.581 51.427
9 1.515 3.294 54.721 1.515 3.294 54.721
10 1.298 2.823 57.544 1.298 2.823 57.544
11 1.290 2.803 60.347 1.290 2.803 60.347
12 1.205 2.620 62.968 1.205 2.620 62.968
13 1.129 2.454 65.422 1.129 2.454 65.422
14 1.005 2.184 67.606 1.005 2.184 67.606
15 1.001 2.175 69.781 1.001 2.175 69.781
16 .933 2.028 71.809
17 .888 1.930 73.739
18 .825 1.793 75.532
19 .806 1.753 77.285
20 .761 1.655 78.940
21 .721 1.567 80.507
22 .673 1.463 81.970
23 .649 1.410 83.380
24 .609 1.324 84.704
25 .580 1.260 85.964
26 .569 1.237 87.201
27 .538 1.170 88.371
28 .505 1.098 89.469
29 .459 .999 90.468
30 .448 .975 91.443
31 .417 .907 92.350
32 .396 .862 93.211
33 .367 .799 94.010
34 .332 .723 94.733
35 .302 .657 95.390
36 .287 .625 96.015
37 .278 .605 96.620
38 .236 .514 97.134
39 .231 .502 97.636
40 .214 .466 98.102
41 .180 .392 98.494
42 .163 .354 98.848
43 .158 .343 99.191
44 .148 .321 99.512
45 .114 .247 99.759
46 .111 .241 100.000

Extraction Method: Principal Component Analysis.


Total Variance Explained

Component Initial Eigenvalues Extraction Sums of Squared Loadings

Total % of Variance Cumulative % Total % of Variance Cumulative %

1 6.390 13.891 13.891 6.390 13.891 13.891


2 4.352 9.460 23.351 4.352 9.460 23.351
3 2.880 6.262 29.612 2.880 6.262 29.612
4 2.715 5.902 35.515 2.715 5.902 35.515
5 2.110 4.587 40.101 2.110 4.587 40.101
6 1.806 3.925 44.027 1.806 3.925 44.027
7 1.757 3.820 47.846 1.757 3.820 47.846
8 1.647 3.581 51.427 1.647 3.581 51.427
9 1.515 3.294 54.721 1.515 3.294 54.721
10 1.298 2.823 57.544 1.298 2.823 57.544
11 1.290 2.803 60.347 1.290 2.803 60.347
12 1.205 2.620 62.968 1.205 2.620 62.968
13 1.129 2.454 65.422 1.129 2.454 65.422
14 1.005 2.184 67.606 1.005 2.184 67.606
15 1.001 2.175 69.781 1.001 2.175 69.781
16 .933 2.028 71.809
17 .888 1.930 73.739
18 .825 1.793 75.532
19 .806 1.753 77.285
20 .761 1.655 78.940
21 .721 1.567 80.507
22 .673 1.463 81.970
23 .649 1.410 83.380
24 .609 1.324 84.704
25 .580 1.260 85.964
26 .569 1.237 87.201
27 .538 1.170 88.371
28 .505 1.098 89.469
29 .459 .999 90.468
30 .448 .975 91.443
31 .417 .907 92.350
32 .396 .862 93.211
33 .367 .799 94.010
34 .332 .723 94.733
35 .302 .657 95.390
36 .287 .625 96.015
37 .278 .605 96.620
38 .236 .514 97.134
39 .231 .502 97.636
40 .214 .466 98.102
41 .180 .392 98.494
42 .163 .354 98.848
43 .158 .343 99.191
44 .148 .321 99.512
45 .114 .247 99.759
46 .111 .241 100.000

Extraction Method: Principal Component Analysis.

Rotated Component Matrixa

Component

1 2 3 4

Role Overload -.079 -.087 .620 .218


Role Ambiguity -.013 .160 .309 .639
Role Conflict -.033 .164 .496 .491
Group/Political Pressure .068 -.066 -.003 .821
Persons' Responsibility .036 -.728 .242 -.151
Under Participation .527 .579 .240 -.205
Powerlessness .198 .749 .145 .051
Poor Peer Relations .792 .014 -.142 -.162
Instrinsic Impoversihment .749 -.023 .114 .319
Status -.700 -.346 .052 -.021
Strenuous Working
.338 -.325 .490 .218
Conditions
Unprofitability -.009 .083 .759 -.022

Extraction Method: Principal Component Analysis.


Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 7 iterations.

Logistic Regression:

When we have a binary categorical variable (DV) (i.e. Yes or No, Good OR Bad) 0 & 1
Whereas, IVs (i.e. all the variables) need to be either metric / categorical

Differences:

Multiple Regression Logistic Regression / Logit


Technique used - LSM Maximum Likelihood method
Graphical Rep - Linear Curve line (in the form of 0 & 1)
Widely used Not so often

Points to remember in log regression:

Let the expected outcome value be always = 1

BLOCK 1: METHOD
Omnibus Tests of Model Coefficients

Chi-square df Sig.

Step 11.577 4 .021

Step 1 Block 11.577 4 .021

Model 11.577 4 .021

Model Summary

Step -2 Log likelihood Cox & Snell R Nagelkerke R


Square Square

1 13.376a .474 .633

a. Estimation terminated at iteration number 6 because


parameter estimates changed by less than .001.

Both Cox & Nagelkerke needs to be considered in this test.


Both the values are calculated based on the log likelihood.
General Interpretation: Closer to 1, better the result.
Hypothesis remains the same as
Classification Tablea

Observed Predicted

Buyer / Non-Buyer Percentage

Non-Buyer Buyer Correct

Non-Buyer 7 2 77.8
Buyer / Non-Buyer
Step 1 Buyer 1 8 88.9

Overall Percentage 83.3

a. The cut value is .500

Variables in the Equation

B S.E. Wald df Sig. Exp(B)

durab 1.145 .604 3.593 1 .058 3.143

light_Weight .187 .859 .047 1 .828 1.206

Step 1a low_investment -.691 .828 .696 1 .404 .501

rot_resistance -.292 .581 .253 1 .615 .747

Constant -3.229 3.123 1.069 1 .301 .040

a. Variable(s) entered on step 1: durab, light_Weight, low_investment, rot_resistance.

Interpretation: Thus Durability has the positive and highest impact on the buying behavior of the XYZ
product.
Omnibus Tests of Model Coefficients

Chi-square df Sig.

Step 41.459 5 .000

Step 1 Block 41.459 5 .000

Model 41.459 5 .000

Model Summary

Step -2 Log likelihood Cox & Snell R Nagelkerke R


Square Square

1 458.517a .098 .138

a. Estimation terminated at iteration number 4 because


parameter estimates changed by less than .001.

Classification Tablea

Observed Predicted

admit Percentage

Reject Admit Correct

Reject 254 19 93.0


admit
Step 1 Admit 97 30 23.6

Overall Percentage 71.0

a. The cut value is .500

Variables in the Equation

B S.E. Wald df Sig. Exp(B)

gre .002 .001 4.284 1 .038 1.002

gpa .804 .332 5.872 1 .015 2.235

rank 20.895 3 .000

Step 1a rank(1) 1.551 .418 13.787 1 .000 4.718

rank(2) .876 .367 5.706 1 .017 2.401

rank(3) .211 .393 .289 1 .591 1.235

Constant -5.541 1.138 23.709 1 .000 .004

a. Variable(s) entered on step 1: gre, gpa, rank.


Tests of Equality of Group Means

Wilks' Lambda F df1 df2 Sig.

Durability .538 13.729 1 16 .002


Light Weight .860 2.610 1 16 .126
Low Investment .935 1.103 1 16 .309
Rot Resistance .995 .077 1 16 .785

Pooled Within-Groups Matrices

Durability Light Weight Low Investment Rot Resistance

Durability 1.000 .633 .549 .209

Light Weight .633 1.000 .541 .327


Correlation
Low Investment .549 .541 1.000 .064

Rot Resistance .209 .327 .064 1.000


Box's Test of Equality of Covariance Matrices

Log Determinants

Buyer / Non-Buyer Rank Log


Determinant

Non-Buyer 4 2.327
Buyer 4 2.985
Pooled within-groups 4 3.768

The ranks and natural logarithms of determinants


printed are those of the group covariance matrices.

Test Results (Similar to levens


test

Box's M 17.796
Approx. 1.287

df1 10
F
df2 1223.904

Sig. .233

Tests null hypothesis of equal


population covariance matrices.
Summary of Canonical Discriminant Functions

Eigenvalues

Function Eigenvalue % of Variance Cumulative % Canonical


Correlation

1 1.033a 100.0 100.0 .713

a. First 1 canonical discriminant functions were used in the analysis.

Wilks' Lambda

Test of Function(s) Wilks' Lambda Chi-square df Sig.

1 .492 9.936 4 .042

Standardized Canonical
Discriminant Function
Coefficients

Function

Durability 1.219
Light Weight -.104
Low Investment -.338
Rot Resistance -.268

Structure Matrix

Function

Durability .911
Light Weight .397
Low Investment .258
Rot Resistance -.068

Pooled within-groups
correlations between
discriminating variables and
standardized canonical
discriminant functions
Variables ordered by absolute
size of correlation within
function.
Standardized Canonical
Discriminant Function
Coefficients
Structure Matrix
Function
Function
1
1
Durability 1.219
Durability .911
Light Weight -.104
Light Weight .397
Low Investment -.338
Low Investment .258
Rot Resistance -.268
Rot Resistance -.068

Pooled within-groups
correlations between
discriminating variables and
standardized canonical
discriminant functions
Variables ordered by absolute
size of correlation within
function.

Canonical Discriminant Function Coefficients

Function

Durability .618
Light Weight -.055
Low Investment -.188
Rot Resistance -.157
(Constant) -1.800

Unstandardized coefficients

Functions at Group Centroids

Buyer / Non-Buyer Function

Non-Buyer -.958
Buyer .958

Unstandardized canonical
discriminant functions evaluated
at group means
Classification Resultsa,c

Buyer / Non-Buyer Predicted Group Membership Total

Non-Buyer Buyer

Non-Buyer 7 2 9
Count
Buyer 1 8 9
Original
Non-Buyer 77.8 22.2 100.0
%
Buyer 11.1 88.9 100.0
Non-Buyer 6 3 9
Count
Buyer 2 7 9
Cross-validatedb
Non-Buyer 66.7 33.3 100.0
%
Buyer 22.2 77.8 100.0
a. 83.3% of original grouped cases correctly classified.
b. 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.
c. 72.2% of cross-validated grouped cases correctly classified.

You might also like