0% found this document useful (0 votes)

33 views

Multinomial Logit: Notes

1. The document provides information on a multinomial logit regression model with occupational status as the dependent variable and race, education, and experience as independent variables. 2. Table 3 shows the logit coefficients for contrasts between different occupational categories and the base category of menial work. 3. The interactive Excel tool allows the user to calculate predicted probabilities and plot graphs by varying the independent variables and observing the changes in predicted probabilities for different occupational categories.

Uploaded by

Seokho Seo

Available Formats

Download as XLS, PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

33 views

Multinomial Logit: Notes

Uploaded by

Seokho Seo

Available Formats

Download as XLS, PDF, TXT or read online on Scribd

You are on page 1/ 38

Multinomial Logit

Do not add or delete any rows or columns.

Notes
1. Enter all of your information in the light green cells, such as this:
2. Press <Ctrl+Shift+p> to see the XPost pull-down menu (Optional).
3. For statistical interpretations of multinomial logit regression, see Long (1996: 148-88).

Table 1. Descriptive Statistics for Y - optional Name

Mean
Std
Min
Max
occ
3.398
1.368
1.00
5.00

Base Category =

Table 2. Descriptive Statistics, Variable Types, and Base Values

Indep
Std.
Variable
Variable
Mean
Dev.
Min
Max
Type
------------WHITE
0.917
0.276
0.00
1.00
B
ED
13.095
2.946
3.00
20.00
C
EXPER
20.501
13.959
2.00
66.00
C

Menial

Table 3. Estimates of Logit C

Base
Value
--1
M
M

Contrast Group
Constant
WHITE
ED
EXPER
-------------------------------------------------

-----------------------------------------------

Table 3. Estimates of Logit Coefficients

Logit coefficients b/w the contrast groups and your base category:
BlueCol

Craft

WhiteCol

prof

0.741
1.237
-0.099
0.005

-1.091
0.472
0.094
0.028

-6.239
1.571
0.353
0.035

-11.518
1.774
0.779
0.036

Calculator

Do not add or delete any rows or columns.

Notes
1. Enter all of your information in the light green cells, such as this:
2. Predicted probabilities are given for base values shown in Table 3. Note base values can only be changed in the Inpu
3. In the green cells you can choose four contrasts, three variables and specify up to five values for that variable. Predic
probabilities are computed at these values with other values equal to the base values in Table 3.

Table 1. Logit Coefficients and Factor Change of Other Contrasts

Contrast 1
Contrast 2
Contrast 3
BlueCol
Prof
craft
WhiteCol
Craft
vs.
vs.
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47

Variable
WHITE
ED
EXPER

b
-0.538
-0.878
-0.031

e^b
0.584
0.415
0.970

e^(b*std)
0.862
0.075
0.649

b
-1.099
-0.259
-0.007

e^b
0.333
0.772
0.993

e^(b*std)
0.738
0.466
0.908

b
-1.302
-0.685
-0.008

vs.

e^b
0.272
0.504
0.992

48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

or columns.

nly be changed in the Input sheet.

s for that variable. Predicted

Table 2.
Prof
e^(b*std)
0.698
0.133
0.895

Contrast 4
prof
b
0.203
0.426
0.001

vs.

whitecol

e^b
1.225
1.531
1.001

e^(b*std)
1.058
3.505
1.015

Predicted Probability
All Variables
Held at
Menial

BlueCol

Base #VALUE! #VALUE!

Mean #VALUE! #VALUE!
Minimum #VALUE! #VALUE!
Maximum #VALUE! #VALUE!
With other variables held at their base values,
vary the following variables and calculate the predicted probabilities
Menial

0 #VALUE!
1 #VALUE!
1.

BlueCol
#VALUE!
#VALUE!

white

exper

3
8
13
18
22
0
5
10
20
30

#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!

Craft
#VALUE!
#VALUE!
#VALUE!
#VALUE!

Basic Predicted Probability

WhiteCol
prof
#VALUE!
#VALUE!
#VALUE!
#VALUE!

#VALUE!
#VALUE!
#VALUE!
#VALUE!

culate the predicted probabilities at the specified values.

Craft
#VALUE!
#VALUE!

WhiteCol
#VALUE!
#VALUE!

prof
#VALUE!
#VALUE!

#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!

Table 3. Base Values

Variable
WHITE
ED
EXPER

Values
1
13.09
20.50

Plotter

Do not add or delete any rows or columns.

Notes
1. Enter all of your information in the light green cells, such as this:
2. Enter information in Table 1 about your graph. Note you can edit the graph as with any Excel chart.

Table 1. Specification of x-variable.

X-variable to vary
ed
With other variable at the base, vary this variable.
Start Value
3
Enter the start value for the x-axis.
End Value
22
Enter end value for the x-axis.
# of Points
10
2 to 20 points allowed

Figure 1. Predicted Probabilities

Menial

BlueCol

Craft

WhiteCol

prof

1.00
0.90

Predicted Probability

0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
3

Education

e any rows or columns.

as with any Excel chart.

Table 2. Base Values

Variable

prof

WHITE
ED
EXPER

Values
1.00
13.09
20.50

Table 3. Sources of Plot

x-value
3.0
5.1
7.2
9.3
11.4
13.6
15.7
17.8
19.9
22.0

Menial
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!

BlueCol
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!

Craft
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!

WhiteCol
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!

prof
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!

Discrete Change
Notes
1.
2.

Do not add or delete any rows or columns.

This page computes factor change and discrete change using the base values specified on the Inpu
are shown on Table 3 below, but can only be changed on the Input sheet. There is nothing you can
Do not add or delete any rows or columns.

Table 1. Factor Change -- Select U if Unit Factor Change; S if Std Factor Change
Type of
Variable Change
BlueCol
Craft
WhiteCol
prof
WHITE
ED
EXPER
-----------------------------------------------------------

U
S
S

3.444
0.746
1.068

1.604
1.318
1.472

4.813
2.831
1.621

5.896
9.923
1.645

Base Category =

-------------------------------------

or columns.