90 / 274
多元迴歸的自變數比較與多元共線性之影響：效果量、優勢性與相對權數指標的估計與應用
90
表
8
薪資對數值各
IV
相對重要性分析（
DA
與
RWA
）結果摘述
投入
IV
1
性別
2
年齡
3
已婚
4
小孩數
5
在職年資
6
每週工時
7
教育年數
優勢分析
1
性別
.046(13.8%)
[.031,.063]
?
C-
?
?
?
C+
2
年齡
.009(2.7%)
[.007,.013]
?
?
C+
?
C+
3
已婚
.009(2.7%)
[.005,.014]
?
C+
C+
C+
4
小孩數
.011(3.3%)
[.007,.018]
C+
?
C+
5
在職年資
.050(15.0%)
[.033,.068]
?
C+
6
每週工時
.037(11.1%)
[.021,.053]
C+
7
教育年數
.171(51.4%)
[.143,.201]
條件優勢
K = 0
.056
[.038,.076]
.000
[.000,.005]
.001
[.000,.004]
.007
[.002,.018]
.031
[.018,.050]
.033
[.016,.051]
.137
[.110,.168]
K = 1
.054
[.036,.072]
.011
[.009,.015]
.007
[.005,.012]
.015
[.010,.024]
.049
[.032,.070]
.033
[.017,.050]
.160
[.132,.191]
K = 2
.051
[.034,.068]
.015
[.012,.020]
.009
[.006,.015]
.016
[.012,.024]
.057
[.040,.078]
.035
[.019,.051]
.174
[.144,.205]
K = 3
.047
[.032,.064]
.014
[.011,.019]
.01
[.006,.017]
.014
[.010,.021]
.058
[.040,.079]
.037
[.021,.053]
.180
[.150,.212]
K = 4
.043
[.029,.059]
.011
[.008,.017]
.011
[.005,.017]
.011
[.006,.018]
.056
[.037,.076]
.039
[.023,.055]
.182
[.153,.213]
K = 5
.039
[.026,.054]
.007
[.004,.014]
.011
[.006,.018]
.008
[.004,.015]
.051
[.034,.070]
.040
[.024,.057]
.182
[.152,.213]
K = 6
.035
[.022,.050]
.004
[.000,.010]
.012
[.006,.019]
.006
[.002,.013]
.045
[.028,.064]
.040
[.025,.058]
.180
[.150,.211]
RIW
.047
[.032,.063]
.015
[.011,.018]
.009
[.006,.014]
.015
[.011,.021]
.047
[.032,.064]
.036
[.021,.055]
.164
[.136,.191]
分割比
14.0%
[9.9,18.7]
4.4%
[3.7,5.6]
2.8%
[1.7,4.3]
4.6%
[3.3,6.5]
14.1%
[9.9,18.7]
10.9%
[6.5,15.9]
49.3%
[42.9,54.8]
註：對角線上的數值為優勢分析的一般優勢指數與分割比（括弧內）。上三角區域內為完全優勢比較結果，
C+
表示完全優勢，
C-
表示完全劣勢，
?
表示無法確立。
[ ]
當中的數值為
1,000
次拔靴抽樣得出的
95%
區間估計。