Академический Документы
Профессиональный Документы
Культура Документы
Camila Galindo
Mnica Reyes
Monitores:
Sergio Camelo
Juliana Ramrez
Ly Lx1 Lx2
Source |
SS
df
MS
-------------+-----------------------------Model | 3.20912138
2 1.60456069
Residual |
1.1092547
51 .021750092
-------------+-----------------------------Total | 4.31837608
53 .081478794
Number of obs
F( 2,
51)
Prob > F
R-squared
Adj R-squared
Root MSE
=
=
=
=
=
=
54
73.77
0.0000
0.7431
0.7331
.14748
-----------------------------------------------------------------------------Ly |
Coef.
Std. Err.
t
P>|t|
[95% Conf. Interval]
-------------+---------------------------------------------------------------Lx1 |
.2099216
.02888
7.27
0.000
.1519426
.2679005
Lx2 |
.278065
.0280479
9.91
0.000
.2217565
.3343736
_cons |
2.189535
.2048467
10.69
0.000
1.778288
2.600782
------------------------------------------------------------------------------
Lyx Lx1x2
Source |
SS
df
MS
-------------+-----------------------------Model |
12.345668
1
12.345668
Residual | 4.54001899
52 .087308057
-------------+-----------------------------Total |
16.885687
53 .318597868
Number of obs
F( 1,
52)
Prob > F
R-squared
Adj R-squared
Root MSE
=
=
=
=
=
=
54
141.40
0.0000
0.7311
0.7260
.29548
-----------------------------------------------------------------------------Lyx |
Coef.
Std. Err.
t
P>|t|
[95% Conf. Interval]
-------------+---------------------------------------------------------------Lx1x2 |
.4732243
.0397958
11.89
0.000
.3933682
.5530803
_cons | -.3708129
.0402139
-9.22
0.000
-.451508
-.2901178
------------------------------------------------------------------------------
Lx1x2
Lx1x2 = 0
F(
1,
52) =
Prob > F =
141.40
0.0000
Por otra parte, el estadstico F puede ser calculado por medio de la razn entre las variables
explicativas, cada una con distribucin ji cuadrado y dividida entre sus respectivos grados de
libertad. Por lo tanto:
Prueba F: [(SSRr - SSRnr)/q ] / [SSRnr/n-k-1) ~ Fq,n-k-1
(SSRr - SSRnr)/q =3.4307429
SSRnr/n-k-1 =0.021750092
Prueba F =157.7356209 ~ F 1,52
A continuacin, se presenta el procedimiento general y directo usando el paquete de computacin
Stata para abordar el problema de Modelos Restringidos:
reg
Ly Lx1 Lx2
Source |
SS
df
MS
-------------+-----------------------------Model | 3.20912138
2 1.60456069
Residual |
1.1092547
51 .021750092
-------------+-----------------------------Total | 4.31837608
53 .081478794
Number of obs
F( 2,
51)
Prob > F
R-squared
Adj R-squared
Root MSE
=
=
=
=
=
=
54
73.77
0.0000
0.7431
0.7331
.14748
-----------------------------------------------------------------------------Ly |
Coef.
Std. Err.
t
P>|t|
[95% Conf. Interval]
-------------+---------------------------------------------------------------Lx1 |
.2099216
.02888
7.27
0.000
.1519426
.2679005
Lx2 |
.278065
.0280479
9.91
0.000
.2217565
.3343736
_cons |
2.189535
.2048467
10.69
0.000
1.778288
2.600782
. test Lx1+ Lx2 =1
( 1)
Lx1 + Lx2 = 1
F(
1,
51) =
Prob > F =
157.74