class 8

---------------------------------------------------------------------------------------------------------

name: <unnamed>

log: C:\Users\Michael\Documents\newer web pages\soc_meth_proj3\fall_2016_logs\class8.log

log type: text

opened on: 19 Oct 2016, 10:17:53

. use "C:\Users\Michael\Documents\current class files\intro soc methods\cps_mar_2000_new with additional vars.dta",

> clear

. regress incwage age age_sq yrsed married male i.race if age >=25 & age <=64 [aweight= perwt_rounded]

(sum of wgt is 1.4261e+08)

Source | SS df MS Number of obs = 69305

-------------+------------------------------ F( 8, 69296) = 2008.75

Model | 1.3727e+13 8 1.7159e+12 Prob > F = 0.0000

Residual | 5.9192e+13 69296 854192824 R-squared = 0.1882

-------------+------------------------------ Adj R-squared = 0.1882

Total | 7.2919e+13 69304 1.0522e+09 Root MSE = 29227

-----------------------------------------------------------------------------------------------

incwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]

------------------------------+----------------------------------------------------------------

age | 2737.602 87.40829 31.32 0.000 2566.282 2908.922

age_sq | -30.90895 .9930601 -31.12 0.000 -32.85534 -28.96255

yrsed | 3491.357 38.51498 90.65 0.000 3415.867 3566.846

married | 3182.541 240.5636 13.23 0.000 2711.037 3654.046

male | 16597.64 222.2257 74.69 0.000 16162.08 17033.2

race |

Black/Negro | -3530.491 347.2994 -10.17 0.000 -4211.198 -2849.785

American Indian/Aleut/Eskimo | -5822.139 1181.959 -4.93 0.000 -8138.777 -3505.502

Asian or Pacific Islander | -1348.464 561.7893 -2.40 0.016 -2449.57 -247.358

_cons | -86376.09 1890.877 -45.68 0.000 -90082.2 -82669.97

-----------------------------------------------------------------------------------------------

. matrix VC=e(V)

* e(V) is a matrix that stata produces after every regression, you just have to call it up and rename it.

. matrix list VC

symmetric VC[10,10]

100b. 200. 300.

age age_sq yrsed married male race race race

age 7640.2094

age_sq -86.157016 .98616843

yrsed -134.03946 1.8818845 1483.4035

married -2133.9449 20.533356 -423.14483 57870.827

male -77.007314 1.2186348 -34.320357 -502.37335 49384.263

100b.race 0 0 0 0 0 0

200.race -483.59055 6.6633063 634.97375 13857.211 2129.1696 0 120616.88

300.race 272.13585 -1.7595027 1485.8529 4869.0456 570.44689 0 16997.268 1397026.8

650.race 563.21372 -4.0685094 -934.51661 -1567.6385 1318.4951 0 14405.105 14028.879

_cons -156200.51 1730.2182 -17685.649 17681.931 -22737.229 0 -26294.339 -46835.248

650.

race _cons

650.race 315607.2

_cons -18016.485 3575417.6

* Variance covariance matrix is symmetric because Cov (X1, X2)= Cov (X2, X1), so Stata only shows you the main diagonal, containing the variances of the coefficients, and the lower half of the symmetric matrix.

. display 7650.5407^0.5

87.467369

* And note that the variance of the coefficient (above, age) is the SE squared.

. *what about the contrast between Asians and blacks?

. lincom 650.race-200.race

( 1) - 200.race + 650.race = 0

------------------------------------------------------------------------------

incwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]

-------------+----------------------------------------------------------------

(1) | 2182.027 638.2898 3.42 0.001 930.9806 3433.074

------------------------------------------------------------------------------

* Asians earn significantly more than blacks, now let’s recreate that SE (638) from the Variance Covariance matrix.

. *OK, and now let's use the Variance Covariance matrix to generate that same SE, which is going to be sqrt(varx1 +varx2-2cov(x1,x2)). Go back to my PDF on means and variances for the formula.