![Show Menu](styles/mobile-menu.png)
![Page Background](./../common/page-substrates/page0087.png)
МОНГОЛЫН ХҮН АМЫН СЭТГҮҮЛ Дугаар (367) 20, 2011
86
For the
education
variable; if a head of
household has college degree, his household
probability of living in Apartment is 49.5 per
cent greater than a head of household whose
education level is less than primary, holding all
other variables at their means. Summing up all
the results relative to the education variables,
it shows the following pattern. As education
level increases, the average absolute discrete
change increases. It tells us that as education
level of head of household increases, it is
likely to move from
Ger
to Apartment or from
House to Apartment or from
Ger
to House
(Figure 1).
Figure 1:
Change in Predicted Probability for type of dwelling (
)
-.32
-.12
.09
.29
.5
G
H
A
G H
A
GH
A
educ4-0/1
educ3-0/1
educ2-0/1
Effect directions of the variables for the
probabilities of living in
Ger
and House are
exactly same. The opposite directions of the
above are shown for the probabilities of living
in Apartment. Although the effect directions of
the variables for the probabilities of living in
Ger
and House are exactly same but they are
differed in size of effect. The size of effects
of the variables for the probability of living in
House are greater than the size of effects of the
variables for the probability of living in
Ger
(refer to the Figure 2 and Table 12).
Table 12:
Effect directions of the Probability for MNLM of Type of dwelling
Variable
Ger
House
Apartment
totperson
+
+
−
mig_percent
+
+
−
age
−
−
+
sex
+
+
−
educ
−
−
+
childnum
+
+
−
totrev
−
−
+
food
+
+
−
Figure 2:
Change in Predicted Probability for type of dwelling
( )
-.1
-.06
-.02
.03
.07
G H
A
G H A
G H A
G H
A
G
H
A
GHA
G H A
totper
mig_percent
age
sex-0/1
childnum
totrev
food