82 / 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
82
表
4
模擬資料的相關矩陣與效果量及相對重要性指標摘要表
Case1
自變數獨立
（無共線性）
Case2
自變數相依
（壓抑／重疊效果）
Case3
自變數相依
（增強效果）
DV IV1 IV2 IV3 IV4
DV IV1 IV2 IV3 IV4
DV IV1 IV2 IV3 IV4
母體相關係數
DV
1.00
1.00
1.00
IV1
.600
1.00
.600
1.00
.600
1.00
IV2
.400
.000 1.00
.400
.900 1.00
.400
.000 1.00
IV3
.200
.000 .000 1.00
.200
.400 .400 1.00
.200
.800 .000 1.00
IV4
.000
.000 .000 .000 1.00
.000
.000 .000 .000 1.00
.000
.000 -.500 .000 1.00
相關指數
VIF r
2
r
p
2
r
sp
2
r
s
2
VIF r
2
r
p
2
r
sp
2
r
s
2
VIF r
2
r
p
2
r
sp
2
r
s
2
IV1
1.00 .360 .450 .360 .643 5.32 .360 .360 .301 .778 2.78 .360 .720 .538 .456
IV2
1.00 .160 .267 .160 .286 5.32 .160 .159 .101 .346 1.33 .160 .505 .213 .203
IV3
1.00 .040 .083 .040 .071 1.20 .040 .000 .000 .086 2.78 .040 .510 .218 .051
IV4
1.00 .000 .000 .000 .000 1.00 .000 .000 .000 .000 1.33 .000 .203 .053 .000
Total
.560 .800 .560 1.000
.560 .519 .403 1.210
.560 1.939 1.022 .709
迴歸與相對重要指數
β
β
2
rβ Dg RIW β
β
2
rβ Dg RIW β
β
2
rβ Dg RIW
IV1
.600 .360 .360 .360 .360 1.266 1.602 .759 .325 .319 1.222 1.494 .733 .449 .449
LL
.562
.320 .319 .319 1.178
.689 .292 .287 1.163
.675 .416 .416
UL
.635
.401 .401 .401 1.357
.835 .358 .352 1.278
.792 .478 .479
IV2
.400 .160 .160 .160 .160 -.734 .539 -.294 .125 .125 .533 .284 .213 .187 .187
LL
.359
.126 .126 .126 -.832
-.333 .112 .112 .494
.178 .156 .155
UL
.441
.196 .196 .196 -.632
-.257 .139 .139 .573
.248 .218 .219