First Model:

Table1a.Case Processing Summary

N Marginal Percentage

triptype1 Type-1 25 4.3%

Type-2 45 7.7%

Type-3 182 31.0%

Type-4 37 6.3%

Type-5 298 50.8%

Age1 <30 333 56.7%
>30 254 43.3%
income 20000-40000 66 11.2%
40000-70000 152 25.9%
70000-100000 86 14.7%
>100000 283 48.2%
Vehicles.Owner Yes 465 79.2%
No 122 20.8%
Valid 587 100.0%
Missing 0
Total 587
Subpopulation 54a

a. The dependent variable has only one value observed in 26 (48.1%)

subpopulations.

The heading of the output is “Nominal regression”; this assumes that there is no “ranking
ordering” in the categorical outcome. In table1a, observe that we have 54 subpopulations. If
there are no zero frequencies in each of the subpopulation, no warning-message will be
displayed.

Table1b.Model Fitting Information

Model Fitting
Criteria Likelihood Ratio Tests

Model -2 Log Likelihood Chi-Square df Sig.

Intercept Only 1039.794

Final 316.546 723.248 32 .000

In table1b, the "Final" row presents information on whether all the coefficients of the model
are zero (i.e., whether any of the coefficients are statistically significant). We can see from
the sig.column that p=.000, which means the full model statistically significantly predicts the
dependent variable better than the intercept-only model alone.

Table1c.Pseudo R-Square

Cox and Snell .708

Nagelkerke .777
McFadden .508

The table1c is showing the pseudo R-square that is similar to the R-square in ordinary least
square regression that indicates the proportion of variation being explained by the model.
Only about 7% is being explained by the age1 + income+ Vehicles.Owner1.

Table1d.Likelihood Ratio Tests

Model Fitting
Criteria Likelihood Ratio Tests

-2 Log Likelihood
Effect of Reduced Model Chi-Square df Sig.

Intercept 316.546a .000 0 .

WorkingFamilymembers 322.082 5.536 4 .237
Familymembers 329.664 13.118 4 .011
Children 407.526 90.980 4 .000
Age1 708.853 392.307 4 .000
income 397.522 80.977 12 .000
Vehicles.Owner 333.580 17.034 4 .002

The chi-square statistic is the difference in -2 log-likelihoods between the final model
and a reduced model. The reduced model is formed by omitting an effect from the
final model. The null hypothesis is that all parameters of that effect are 0.
a. This reduced model is equivalent to the final model because omitting the effect
does not increase the degrees of freedom.

The table1d is representing the Likelihood ratio test shows that which of independent
variables are statistically significant. We can see that family members, children, age1, income
and vehicles Owner are statistically significant because p-value = .000. The p-value of
working family members is not statistically significant because p-value=0.237.
Table1e.Parameter Estimates
95% Confidence Interval for Exp(B)
a
triptype1 B Std. Error Wald df Sig. Exp(B) Lower Bound Upper Bound
Type-1 Intercept .446 1.215 .135 1 .714
WorkingFamilymembers -.876 .515 2.892 1 .089 .417 .152 1.143
Familymembers -.016 .197 .007 1 .935 .984 .669 1.447
Children -.087 .200 .187 1 .666 .917 .619 1.358
[Age1=1.00] -.573 .674 .721 1 .396 .564 .150 2.116
[Age1=2.00] 0b . . 0 . . . .
[income=2] .053 .726 .005 1 .942 1.054 .254 4.373
[income=3] -.791 .546 2.098 1 .148 .453 .155 1.322
[income=4] -1.383 1.136 1.480 1 .224 .251 .027 2.327
[income=5] 0b . . 0 . . . .
[Vehicles.Owner=1.00] -.926 .524 3.127 1 .077 .396 .142 1.106
[Vehicles.Owner=2.00] 0b . . 0 . . . .
Type-2 Intercept -1.014 1.040 .951 1 .330
WorkingFamilymembers -.188 .273 .475 1 .490 .828 .485 1.414
Familymembers .020 .137 .020 1 .886 1.020 .779 1.335
Children .085 .169 .253 1 .615 1.089 .782 1.515
[Age1=1.00] .816 .409 3.977 1 .046 2.262 1.014 5.044
[Age1=2.00] 0b . . 0 . . . .
[income=2] -19.667 .000 . 1 . 2.876E-9 2.876E-9 2.876E-9
[income=3] -.641 .490 1.710 1 .191 .527 .202 1.377
[income=4] 1.105 .577 3.666 1 .056 3.020 .974 9.365
[income=5] 0b . . 0 . . . .
[Vehicles.Owner=1.00] -1.703 .481 12.549 1 .000 .182 .071 .467
[Vehicles.Owner=2.00] 0b . . 0 . . . .
Type-3 Intercept -22.161 2275.326 .000 1 .992
WorkingFamilymembers .443 .612 .524 1 .469 1.557 .469 5.163
Familymembers -1.323 .374 12.478 1 .000 .266 .128 .555
Children 2.248 .427 27.723 1 .000 9.473 4.102 21.877
[Age1=1.00] 24.978 2275.325 .000 1 .991 7.044E10 .000 .c
[Age1=2.00] 0b . . 0 . . . .
[income=2] -.477 1.596 .089 1 .765 .620 .027 14.156
[income=3] .180 1.099 .027 1 .870 1.197 .139 10.315
[income=4] -21.953 3795.359 .000 1 .995 2.924E- .000 .c
10
[income=5] 0b . . 0 . . . .
[Vehicles.Owner=1.00] .932 1.199 .604 1 .437 2.540 .242 26.656
[Vehicles.Owner=2.00] 0b . . 0 . . . .
Type-4 Intercept -1.554 1.020 2.320 1 .128
WorkingFamilymembers -.341 .293 1.355 1 .244 .711 .401 1.262
Familymembers .062 .135 .208 1 .648 1.064 .816 1.387
Children .005 .170 .001 1 .978 1.005 .720 1.403
[Age1=1.00] .188 .441 .181 1 .671 1.207 .508 2.866
[Age1=2.00] 0b . . 0 . . . .
[income=2] -1.141 1.081 1.113 1 .291 .320 .038 2.660
[income=3] -.575 .539 1.138 1 .286 .563 .196 1.618
[income=4] 1.119 .578 3.744 1 .053 3.063 .986 9.517
[income=5] 0b . . 0 . . . .
[Vehicles.Owner=1.00] -.624 .532 1.374 1 .241 .536 .189 1.521
[Vehicles.Owner=2.00] 0b . . 0 . . . .
a. The reference category is: Type-5.
b. This parameter is set to zero because it is redundant.
c. Floating point overflow occurred while computing this statistic. Its value is therefore set to system missing.

The table1e is representing the parameter estimates. We can see that working family
member, family members and children are not statistically significant because p-value is
greater than 0.05.
In the first set of coefficients, age1 (<30) =1 and age1 (>30) =2 is used as reference category,
since the p-value of age1=1 is 0.396 that is not statistically significant. We can say that
people who have greater than 30 age group go home to work and work to home with
additional stop and home other home trip. For income, income=5 is used as a reference
category and the p-value of income=2, income=3 and income=4 are insignificant because p-
value=0.942, 0.148, 0.224, so we can say that people whose income are in range 20000-
40000,70000-100000 and 40000-70000 do not go home to work and work to home with
additional stop and home other home trip.
The working family member, family members and children are not statistically significant
because p-value is greater than 0.05. In the second set of coefficients, age1 (<30) =1 and age1
(>30) =2 is used as reference category, since the p-value of age1=1 is 0.046 that is
statistically significant. We can say that people who have less than 30 age group go home to
recreational and back home. For income, income=5 is used as a reference category and the p-
value of income=3 is insignificant because p-value=0.191, so we can say that people whose
income are in range 40000-70000 do not go home to work and work to home with additional
stop and home other home trip. The p-value of income=4 is statistically significant because
p-value=0.000. We can say that people whose income is in range 70000-100000 do home to
recreational and back home.
The family members and children are statistically significant because p-value is 0.05 and
working family member is statistically insignificant because p-value is greater than 0.05. In
the third set of coefficients, age1 (<30) =1 and age1 (>30) =2 is used as reference category,
since the p-value of age1=1 is 0.991 that is not statistically significant. We can say that
people who have greater than 30 age group go home school home. For income, income=5 is
used as a reference category and the p-value of income=2, income=3 and income=4 are
insignificant because p-value=0.765,0.870,0.995 ,so we can say that people whose income
are in range 20000-40000, 40000-70000 and 70000-100000 do not go home school home.
The working family member, family members and children are not statistically significant
because p-value is greater than 0.05.
In the fourth set of coefficients, age1 (<30) =1 and age1 (>30) =2 is used as reference
category, since the p-value of age1=1 is 0.671 that is not statistically significant. We can say
that people who have greater than 30 age group do home shopping home and home shopping
with additional stop. For income, income=5 is used as a reference category and the p-value of
income=2 and income=3 is insignificant because p-value=0.291 and 0.286, so we can say
that people whose income are in range 20000-40000, 40000-70000 do not do home shopping
home and home shopping with additional stop. The p-value of income=4 is statistically
significant because p-value=0.000. We can say that people whose income is in range 70000-
100000 do home shopping home and home shopping with additional stop.
Second Model:

Table1.Case Processing Summary

N Marginal Percentage

Vehicles Car 190 32.4%

Bus/Van 342 58.3%

2W 55 9.4%
Age1 <30 333 56.7%
>30 254 43.3%
Vehicles.Owner1 Yes 465 79.2%
No 122 20.8%
income 20000-40000 66 11.2%
40000-70000 152 25.9%
70000-100000 86 14.7%
>100000 283 48.2%
Valid 587 100.0%
Missing 0
Total 587
Subpopulation 50a

The heading of the output is “Nominal regression”; this assumes that there is no “ranking
ordering” in the categorical outcome. In table1, observe that we have 50 subpopulations. If
there are no zero frequencies in each of the subpopulation, no warning-message will be
displayed.

Table2.Model Fitting Information

Model Fitting
Criteria Likelihood Ratio Tests

-2 Log
Model Likelihood Chi-Square df Sig.

Intercept Only 694.587

Final 313.106 381.481 12 .000

In table2, the "Final" row presents information on whether all the coefficients of the model
are zero (i.e., whether any of the coefficients are statistically significant). We can see from
the sig.column that p=.000, which means the full model statistically significantly predicts the
dependent variable better than the intercept-only model alone.

Table3.Pseudo R-Square

Cox and Snell .478

Nagelkerke .572
McFadden .360
The table3 is showing the pseudo R-square that is similar to the R-square in ordinary least
square regression that indicates the proportion of variation being explained by the model.
Only about 5% is being explained by the age1 + Vehicles.Owner1+income.

Table4.Likelihood Ratio Tests

Model Fitting Criteria Likelihood Ratio Tests

-2 Log Likelihood of
Effect Reduced Model Chi-Square df Sig.
a
Intercept 313.106 .000 0 .
time 371.401 58.295 2 .000
Age1 547.972 234.866 2 .000
Vehicles.Owner1 398.883 85.776 2 .000
income 351.164 38.057 6 .000

The chi-square statistic is the difference in -2 log-likelihoods between the final

model and a reduced model. The reduced model is formed by omitting an
effect from the final model. The null hypothesis is that all parameters of that
effect are 0.
a. This reduced model is equivalent to the final model because omitting the
effect does not increase the degrees of freedom.

The table4 is representing the Likelihood ratio test shows that which of independent
variables are statistically significant. We can see that time, age1, vehicles.owner1 and income
are statistically significant because p-value = .000 for all independent variable.
Table5. Parameter Estimates

95% Confidence Interval for Exp(B)

a
Vehicles B Std. Error Wald df Sig. Exp(B) Lower Bound Upper Bound

Car Intercept 4.926 1.264 15.201 1 .000

time -.100 .135 .551 1 .458 .905 .695 1.178

[Age1=1.00] -.574 .388 2.185 1 .139 .564 .263 1.206

b
[Age1=2.00] 0 . . 0 . . . .

[Vehicles.Owner1=1.00] -2.632 1.095 5.772 1 .016 .072 .008 .616

b
[Vehicles.Owner1=2.00] 0 . . 0 . . . .

[income=2] -1.020 .566 3.253 1 .071 .361 .119 1.092

[income=3] -1.841 .420 19.187 1 .000 .159 .070 .362

[income=4] -.109 .548 .039 1 .843 .897 .306 2.627

b
[income=5] 0 . . 0 . . . .
Bus/Van Intercept 8.367 1.322 40.069 1 .000

time -.981 .179 30.206 1 .000 .375 .264 .532

[Age1=1.00] 3.090 .395 61.133 1 .000 21.977 10.129 47.682

[Age1=2.00] 0b . . 0 . . . .

[Vehicles.Owner1=1.00] -5.287 1.123 22.159 1 .000 .005 .001 .046

[Vehicles.Owner1=2.00] 0b . . 0 . . . .

[income=2] -1.108 .586 3.567 1 .059 .330 .105 1.043

[income=3] -2.133 .454 22.098 1 .000 .118 .049 .288

[income=4] -1.412 .594 5.661 1 .017 .244 .076 .780

[income=5] 0b . . 0 . . . .

a. The reference category is: 2W.

b. This parameter is set to zero because it is redundant.

The table5 is representing the parameter estimates. We can see that there are two sets of
logistic regression coefficients. The first set of coefficients is found in the car row
representing the comparison of car category to the reference category 2W (wheeler
motorcycle). The second set of coefficients is found in the bus/van row representing the
comparison of bus/van category to the reference category 2W (wheeler motorcycle). We can
see that time is not statistically significant (p-value= .458) for first set of coefficients and
statistically significant (p-value=.000) for the second set of coefficients.
In the first set of coefficients, age1 (<30) =1 and age1 (>30) =2 is used as reference category,
since the p-value of age1=1 is 0.139 that is not statistically significant. We can say that
people who has greater than 30 age group travel by car. For income, income=5 is used as a
reference category and only income= 3 is statistically significant because p-value=.000. We
can say that people whose income is 40000-70000 use car. The p-value of income=2 and
income=3 are insignificant because p-value=0.71, 0.843, so we can say that people whose
income are 20000-40000 and 70000-100000 do not use car.
In the second set of coefficients, age1 (<30) =1 and age1 (>30) =2 is used as reference
category, since the p-value of age1=1 is 0.000 that is statistically significant. We can say that
people who has less than 30 age group travel by bus/van. For income, income=5 is used as a
reference category and income= 3 and income=4 is statistically significant because p-
value=.000. We can say that people whose income is 40000-70000 and 70000-100000 use
bus/van. The p-value of income=2 is insignificant because p-value=.059, so we can say that
people whose income are 20000-40000 do not use car.