Trong F test df là mấy số k; (n-k); m …
Chap 2: Single Regression: Y = B0 + B1.X + u
Test the significance of model (R )2
H0: R = 02
H1: R > 02
F-test = ESS/(k-1)
RSS/(n-k)
Test the significance of a coefficient (B1)
H0: B1 = 0
H1: B1 != 0
T-test = B1^ - (B1)
SE
Chap 3: Multiple regression: Y = B0 + B1.X1 + B2.X2
+B3.X3+…U
Test the significance of model (R )2
Test the significance of a coefficient (Bx) như trên
Test the equality of 2 coefficients
H0: B2=B3
H1: B2 != B3
Method 1: t- test : B2^ - B3^ - (B2-B3)
SE
+ +
Method 2
F_test: F = t (df: 1, n-k)2
Method 3: make a conjunction
If B2 is different from B3 assume that B2 = B3 + a
Y = B0 + B1.X1 + B2.X2 + B3.X3
 Y = B0 + B1.X1 + X2 + B3.X3(B3 + a).
 Y = B0 + B1.X1 + .X2 + .(X2 + X3)a B3
H0 : a =0
H1 : a != 0
(test bằng hệ số mới của biến X2)
T-test tương tự phần 2.Test the significance of a coefficient (Bx)
Restricted and unrestricted
Test xem khi bỏ bớt biến thì R có thay đổi hay ko (test the difference between 2 models)2
 xem biến đó có effect lên Y hay ko
H0: B2 = B3 =…=0
H1: at least one is different to 0
F-test = (RSS – RSS )/m = (R R UR 2 – R UR 2
)/m m: số biến bị bỏ điR
RSS / (n-k) (1 – R )/ (n-k)UR 2UR
Chap 4: Dummy
N categories add (N-1) dummies , the remaining one is the constant (base-group)
Test a dummy’s coefficient
Ex : wage = B0 + B1 educ + B2 female
Male: wage = B0 + B1 educ
Female : wage = B0 + B1 educ + B2
H0 : B2 = 0
H1 : B2 != 0
T-test tương tự phần 2.Test the significance of a coefficient (Bx) dung coefficient của female
Dummy Interaction
Add an interaction: educ*female
wage = B0 + B1 educ + B2 female + a1.educ*female
Male: wage = B0 + B1 educ
Female : wage = B0 + B2 + (B1+a1) educ
H0: a1 = 0
H1: a1 != 0
T-test tương tự phần 2.Test the significance of a coefficient (Bx) dung coefficient của educ*female
Chow test
Run the model separately on 2 sub-sample (ex: run on male and female)
H0: B0m = B0f; B1m = B1f ; B2m=B2f;... lừi
H1: At least 1 pair is different.
F- test: (RSS – (RSS1 + RSS2))/ k
(RSS1 + RSS2)/ (n-2k)

Preview text:

Trong F test df là mấy số k; (n-k); m …
Chap 2: Single Regression: Y = B0 + B1.X + u
Test the significance of model (R2) H0: R2 = 0 H1: R2 > 0 F-test = ESS/(k-1) RSS/(n-k)
Test the significance of a coefficient (B1) H0: B1 = 0 H1: B1 != 0 T-test = B1^ - (B1) SE
Chap 3: Multiple regression: Y = B0 + B1.X1 + B2.X2 +B3.X3+…U
Test the significance of model (R2)
Test the significance of a coefficient (Bx) như trên
Test the equality of 2 coefficients H0: B2=B3 H1: B2 != B3
Method 1: t- test : B2^ - B3^ - (B2-B3) SE + + Method 2 F_test: F = t2 (df: 1, n-k) Method 3: make a conjunction
If B2 is different from B3 assume that B2 = B3 + a Y = B0 + B1.X1 + B2.X2 + B3.X3
Y = B0 + B1.X1 + (B3 + a).X2 + B3.X3
Y = B0 + B1.X1 + a.X2 + B3.(X2 + X3) H0 : a =0 H1 : a != 0
(test bằng hệ số mới của biến X2)
T-test tương tự phần 2.Test the significance of a coefficient (Bx) Restricted and unrestricted
Test xem khi bỏ bớt biến thì R2 có thay đổi hay ko (test the difference between 2 models)
xem biến đó có effect lên Y hay ko H0: B2 = B3 =…=0
H1: at least one is different to 0
F-test = (RSSR – RSSU )/m = (R R U R2 – R
2R)/m m: số biến bị bỏ đi
RSSUR / (n-k) (1 – R2U )/ (n-k) R Chap 4: Dummy
N categories add (N-1) dummies , the remaining one is the constant (base-group) Test a dummy’s coefficient
Ex : wage = B0 + B1 educ + B2 female Male: wage = B0 + B1 educ
Female : wage = B0 + B1 educ + B2 H0 : B2 = 0 H1 : B2 != 0
T-test tương tự phần 2.Test the significance of a coefficient (Bx) dung coefficient của female Dummy Interaction
Add an interaction: educ*female
wage = B0 + B1 educ + B2 female + a1.educ*female Male: wage = B0 + B1 educ
Female : wage = B0 + B2 + (B1+a1) educ H0: a1 = 0 H1: a1 != 0
T-test tương tự phần 2.Test the significance of a coefficient (Bx) dung coefficient của educ*female Chow test
Run the model separately on 2 sub-sample (ex: run on male and female)
H0: B0m = B0f; B1m = B1f ; B2m=B2f;... lừi
H1: At least 1 pair is different.
F- test: (RSS – (RSS1 + RSS2))/ k (RSS1 + RSS2)/ (n-2k)