class 10 log

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

log: C:\AAA Miker Files\newer web pages\soc_388_notes\soc_388_2007\class_ten_log.log

log type: text

opened on: 25 Oct 2007, 11:04:36

. *Today we will talk a little about stepwise regression.

. *demonstrating some stepwise regressions on a small and familiar dataset, LA racial intermarriage

. table husb wife, contents (sum count) row col

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

| wife

husb | black Mexican Oth Hisp All Others white Total

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

black | 4074 63 32 42 215 4426

Mexican | 25 3947 143 95 1009 5219

Oth Hisp | 16 132 239 18 304 709

All Others | 19 78 18 1022 360 1497

white | 103 1156 373 492 28453 30577

Total | 4237 5376 805 1669 30341 42428

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

. *Fitting models to the data...

. set linesize 79

. table husb wife, contents(mean intermar mean intermar_full)

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

| wife

husb | black Mexican Oth Hisp All Others white

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

black | 1 0 0 0 0

| 1 0 0 0 0

Mexican | 0 1 0 0 0

| 0 2 0 0 0

Oth Hisp | 0 0 1 0 0

| 0 0 3 0 0

All Others | 0 0 0 1 0

| 0 0 0 4 0

white | 0 0 0 0 1

| 0 0 0 0 5

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

. desmat: poisson count husb wife intermar

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

Poisson regression

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

Dependent variable count

Optimization: ml

Number of observations: 25

Initial log likelihood: -80138.505

Log likelihood: -1405.264

LR chi square: 157466.482

Model degrees of freedom: 9

Pseudo R-squared: 0.982

Prob: 0.000

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

nr Effect Coeff s.e.

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

count

husb

1 Mexican -0.026 0.031

2 Oth Hisp -1.288** 0.048

3 All Others -0.822** 0.039

4 white 1.026** 0.025

wife

5 Mexican 0.258** 0.031

6 Oth Hisp -0.880** 0.046

7 All Others -0.388** 0.038

8 white 1.119** 0.026

intermar

9 1 2.850** 0.017

10 _cons 5.271** 0.021

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

* p < .05

** p < .01

. poisgof

Goodness-of-fit chi2 = 2632.715

Prob > chi2(15) = 0.0000

. *Needs more...

. desmat: poisson count husb wife intermar_full

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

Poisson regression

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

Dependent variable count

Optimization: ml

Number of observations: 25

Initial log likelihood: -80138.505

Log likelihood: -132.836

LR chi square: 160011.338

Model degrees of freedom: 13

Pseudo R-squared: 0.998

Prob: 0.000

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

nr Effect Coeff s.e.

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

count

husb

1 Mexican 1.485** 0.061

2 Oth Hisp 0.340** 0.071

3 All Others 0.361** 0.070

4 white 2.791** 0.065

wife

5 Mexican 2.328** 0.083

6 Oth Hisp 1.262** 0.089

7 All Others 1.397** 0.088

8 white 3.498** 0.087

intermar_full

9 1 6.378** 0.099

10 2 2.533** 0.055

11 3 1.940** 0.094

12 4 3.237** 0.073

13 5 2.032** 0.051

14 _cons 1.934** 0.098

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

* p < .05

** p < .01

. poisgof

Goodness-of-fit chi2 = 87.85888

Prob > chi2(11) = 0.0000

. *This is much better, but still leaves something to be desired. But what?

. *We are going to look at a set of off-diagonal relationships, and see which o

> nes Stata would put in...

. table husb wife, contents (mean QS)

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

| wife

husb | black Mexican Oth Hisp All Others white

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

black | 0 21 31 41 51

Mexican | 21 0 32 42 52

Oth Hisp | 31 32 0 43 53

All Others | 41 42 43 0 54

white | 51 52 53 54 0

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

. *10 interaction terms for the 10 symmetric pairs of off diagonal cells.

. *It turns out that this is the greatest number of symmetric interactions you can have in a square table.

. *An alternative version of the Quasi-symmetry terms might have main diagonal plus several off-diagonal symmetries, such as QS2 below

. table husb wife, contents (mean QS2)

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

| wife

husb | black Mexican Oth Hisp All Others white

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

black | 1 0 0 41 51

Mexican | 0 2 0 42 52

Oth Hisp | 0 0 3 0 53

All Others | 41 42 0 4 0

white | 51 52 53 0 5

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

. desmat: poisson count husb wife QS

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

Poisson regression

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

Dependent variable count

Optimization: ml

Number of observations: 25

Initial log likelihood: -80138.505

Log likelihood: -89.596

LR chi square: 160097.818

Model degrees of freedom: 18

Pseudo R-squared: 0.999

Prob: 0.000

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

nr Effect Coeff s.e.

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

count

husb

1 Mexican -0.431** 0.051

2 Oth Hisp -1.866** 0.065

3 All Others -1.190** 0.057

4 white 0.625** 0.049

wife

5 Mexican 0.399** 0.051

6 Oth Hisp -0.970** 0.065

7 All Others -0.193** 0.057

8 white 1.319** 0.049

9 21 -4.596** 0.109

10 31 -3.814** 0.150

11 32 -1.956** 0.069

12 41 -4.323** 0.132

13 42 -3.148** 0.078

14 43 -3.314** 0.171

15 51 -4.274** 0.059

16 52 -2.284** 0.023

17 53 -2.047** 0.050

18 54 -2.550** 0.038

19 _cons 8.312** 0.016

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

* p < .05

** p < .01

. poisgof

Goodness-of-fit chi2 = 1.379208

Prob > chi2(6) = 0.9671

. *Quasi-Symmetry fits very well....

. desmat: poisson count husb wife QS2

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

Poisson regression

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

Dependent variable count

Optimization: ml

Number of observations: 25

Initial log likelihood: -80138.505

Log likelihood: -89.596

LR chi square: 160097.818

Model degrees of freedom: 18

Pseudo R-squared: 0.999

Prob: 0.000

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

nr Effect Coeff s.e.

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

count

husb

1 Mexican 1.428** 0.160

2 Oth Hisp 0.774** 0.130

3 All Others -0.690** 0.225

4 white 4.029** 0.212

wife

5 Mexican 2.257** 0.171

6 Oth Hisp 1.671** 0.142

7 All Others 0.307 0.232

8 white 4.724** 0.220

QS2

9 1 6.454** 0.194

10 2 2.738** 0.190

11 3 1.173** 0.201

12 4 5.454** 0.385

13 5 -0.355 0.390

14 41 1.632** 0.244

15 42 0.948** 0.257

16 51 -1.225** 0.231

17 52 -1.092** 0.182

18 53 -1.638** 0.258

19 _cons 1.858** 0.194

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

* p < .05

** p < .01

. poisgof

Goodness-of-fit chi2 = 1.379208

Prob > chi2(6) = 0.9671

. *It's the same model, you can tell bc goodness of fit is exactly the same.

. *What is left after putting all the symmetrical terms into the model, is asymmetrical terms.

. table husb wife, contents(mean Asym)

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

| wife

husb | black Mexican Oth Hisp All Others white

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

black | 0 0 0 0 0

Mexican | 21 0 0 0 0

Oth Hisp | 31 32 0 0 0

All Others | 41 42 43 0 0

white | 51 52 53 54 0

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

. *To do stepwise, we have to use desmat to generate the dummies first.

. desmat husb wife QS2 Asym

Desmat generated the following design matrix:

nr Variables Term Parameterization

First Last

1 _x_1 _x_4 husb ind(1)

2 _x_5 _x_8 wife ind(1)

3 _x_9 _x_17 QS2 ind(0)

4 _x_18 _x_24 Asym ind(0)

. *Notice, that several of the Asym terms were dropped from desmat, because they would have been colinear. You can't put more than 25 terms (in this case, 1 constant plus 24 terms) into a model to predict a 5x5 table.

. sw poisson count (_x_1-_x_8) _x_9-_x_24, forward pe(.001) pr(.01)

begin with empty model

p = 0.0000 < 0.0010 adding _x_1 _x_2 _x_3 _x_4 _x_5 _x_6 _x_7 _x_8

p = 0.0000 < 0.0010 adding _x_9

p = 0.0000 < 0.0010 adding _x_10

p = 0.0000 < 0.0010 adding _x_12

p = 0.0000 < 0.0010 adding _x_17

p = 0.0000 < 0.0010 adding _x_14

p = 0.0000 < 0.0010 adding _x_13

p = 0.0000 < 0.0010 adding _x_15

p = 0.0000 < 0.0010 adding _x_16

p = 0.0000 < 0.0010 adding _x_11

Poisson regression Number of obs = 25

LR chi2(17) = 160096.96

Prob > chi2 = 0.0000

Log likelihood = -90.022867 Pseudo R2 = 0.9989

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

count | Coef. Std. Err. z P>|z| [95% Conf. Interval]

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

_x_1 | 1.471544 .1558986 9.44 0.000 1.165988 1.7771

_x_2 | .73637 .1218848 6.04 0.000 .4974801 .9752598

_x_3 | -.5103945 .1108274 -4.61 0.000 -.7276122 -.2931768

_x_4 | 3.859923 .0999757 38.61 0.000 3.663975 4.055872

_x_5 | 2.300677 .1676349 13.72 0.000 1.972118 2.629235

_x_6 | 1.635904 .1357068 12.05 0.000 1.369923 1.901884

_x_7 | .4869843 .1235324 3.94 0.000 .2448652 .7291033

_x_8 | 4.554325 .1164357 39.11 0.000 4.326115 4.782535

_x_9 | 6.470534 .195383 33.12 0.000 6.087591 6.853478

_x_10 | 2.666644 .1756067 15.19 0.000 2.322461 3.010827

_x_12 | 5.111081 .0751948 67.97 0.000 4.963702 5.25846

_x_17 | -1.415834 .0908697 -15.58 0.000 -1.593935 -1.237733

_x_14 | .7403587 .1228662 6.03 0.000 .4995454 .9811719

_x_13 | 1.468076 .1632134 8.99 0.000 1.148184 1.787968

_x_15 | -1.039167 .110891 -9.37 0.000 -1.256509 -.8218247

_x_16 | -.9503197 .0900525 -10.55 0.000 -1.126819 -.77382

_x_11 | 1.262343 .1769777 7.13 0.000 .9154734 1.609213

_cons | 1.841846 .1947538 9.46 0.000 1.460136 2.223557

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

. desrep

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

Poisson regression

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

Dependent variable count

Optimization: ml

Number of observations: 25

Initial log likelihood: -80138.505

Log likelihood: -90.023

LR chi square: 160096.964

Model degrees of freedom: 17

Pseudo R-squared: 0.999

Prob: 0.000

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

nr Effect Coeff s.e.

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

count

husb

1 Mexican 1.472** 0.156

2 Oth Hisp 0.736** 0.122

3 All Others -0.510** 0.111

4 white 3.860** 0.100

wife

5 Mexican 2.301** 0.168

6 Oth Hisp 1.636** 0.136

7 All Others 0.487** 0.124

8 white 4.554** 0.116

QS2

9 1 6.471** 0.195

10 2 2.667** 0.176

11 4 5.111** 0.075

12 53 -1.416** 0.091

13 42 0.740** 0.123

14 41 1.468** 0.163

15 51 -1.039** 0.111

16 52 -0.950** 0.090

17 3 1.262** 0.177

18 _cons 1.842** 0.195

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

* p < .05

** p < .01

. poisgof

Goodness-of-fit chi2 = 2.233234

Prob > chi2(7) = 0.9458

. *What we got here, is only symmetrical terms from QS2, including 4 of the 5 diagonal terms and 5 of the symmetrical off-diagonal terms. None of the asymetrical terms got added.

. *Which tells us something: looking at the data this way, there do not appear to be important gender asymetries.

. *Now let's do a backward stepwise

. sw poisson count (_x_1-_x_8) _x_9-_x_24, pe(.001) pr(.01)

begin with full model

p = 0.9720 >= 0.0100 removing _x_22

p = 0.7010 >= 0.0100 removing _x_21

p = 0.5047 >= 0.0100 removing _x_19

p = 0.4576 >= 0.0100 removing _x_20

p = 0.5321 >= 0.0100 removing _x_23

p = 0.5793 >= 0.0100 removing _x_18

p = 0.5194 >= 0.0100 removing _x_24

Poisson regression Number of obs = 25

LR chi2(17) = 160096.96

Prob > chi2 = 0.0000

Log likelihood = -90.022867 Pseudo R2 = 0.9989

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

count | Coef. Std. Err. z P>|z| [95% Conf. Interval]

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

_x_1 | 1.471544 .1558986 9.44 0.000 1.165988 1.7771

_x_2 | .73637 .1218848 6.04 0.000 .4974801 .9752598

_x_3 | -.5103945 .1108274 -4.61 0.000 -.7276122 -.2931768

_x_4 | 3.859923 .0999757 38.61 0.000 3.663975 4.055872

_x_5 | 2.300677 .1676349 13.72 0.000 1.972118 2.629235

_x_6 | 1.635904 .1357068 12.05 0.000 1.369923 1.901884

_x_7 | .4869843 .1235324 3.94 0.000 .2448652 .7291033

_x_8 | 4.554325 .1164357 39.11 0.000 4.326115 4.782535

_x_9 | 6.470534 .195383 33.12 0.000 6.087591 6.853478

_x_10 | 2.666644 .1756067 15.19 0.000 2.322461 3.010827

_x_11 | 1.262343 .1769777 7.13 0.000 .9154734 1.609213

_x_12 | 5.111081 .0751948 67.97 0.000 4.963702 5.25846

_x_13 | 1.468076 .1632134 8.99 0.000 1.148184 1.787968

_x_14 | .7403587 .1228662 6.03 0.000 .4995454 .9811719

_x_15 | -1.039167 .110891 -9.37 0.000 -1.256509 -.8218247

_x_16 | -.9503197 .0900525 -10.55 0.000 -1.126819 -.77382

_x_17 | -1.415834 .0908697 -15.58 0.000 -1.593935 -1.237733

_cons | 1.841846 .1947538 9.46 0.000 1.460136 2.223557

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

. poisgof

Goodness-of-fit chi2 = 2.233234

Prob > chi2(7) = 0.9458

. desrep

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

Poisson regression

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

Dependent variable count

Optimization: ml

Number of observations: 25

Initial log likelihood: -80138.505

Log likelihood: -90.023

LR chi square: 160096.964

Model degrees of freedom: 17

Pseudo R-squared: 0.999

Prob: 0.000

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

nr Effect Coeff s.e.

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

count

husb

1 Mexican 1.472** 0.156

2 Oth Hisp 0.736** 0.122

3 All Others -0.510** 0.111

4 white 3.860** 0.100

wife

5 Mexican 2.301** 0.168

6 Oth Hisp 1.636** 0.136

7 All Others 0.487** 0.124

8 white 4.554** 0.116

QS2

9 1 6.471** 0.195

10 2 2.667** 0.176

11 3 1.262** 0.177

12 4 5.111** 0.075

13 41 1.468** 0.163

14 42 0.740** 0.123

15 51 -1.039** 0.111

16 52 -0.950** 0.090

17 53 -1.416** 0.091

18 _cons 1.842** 0.195

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

* p < .05

** p < .01

. *In this case, both forward and backward yielded the same model, but it doesn't have to be so.

. sw poisson count (_x_1-_x_8) _x_9-_x_24, forward pe(.05) pr(.1)

begin with empty model

p = 0.0000 < 0.0500 adding _x_1 _x_2 _x_3 _x_4 _x_5 _x_6 _x_7 _x_8

p = 0.0000 < 0.0500 adding _x_9

p = 0.0000 < 0.0500 adding _x_10

p = 0.0000 < 0.0500 adding _x_12

p = 0.0000 < 0.0500 adding _x_17

p = 0.0000 < 0.0500 adding _x_14

p = 0.0000 < 0.0500 adding _x_13

p = 0.0000 < 0.0500 adding _x_15

p = 0.0000 < 0.0500 adding _x_16

p = 0.0000 < 0.0500 adding _x_11

Poisson regression Number of obs = 25

LR chi2(17) = 160096.96

Prob > chi2 = 0.0000

Log likelihood = -90.022867 Pseudo R2 = 0.9989

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

count | Coef. Std. Err. z P>|z| [95% Conf. Interval]

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

_x_1 | 1.471544 .1558986 9.44 0.000 1.165988 1.7771

_x_2 | .73637 .1218848 6.04 0.000 .4974801 .9752598

_x_3 | -.5103945 .1108274 -4.61 0.000 -.7276122 -.2931768

_x_4 | 3.859923 .0999757 38.61 0.000 3.663975 4.055872

_x_5 | 2.300677 .1676349 13.72 0.000 1.972118 2.629235

_x_6 | 1.635904 .1357068 12.05 0.000 1.369923 1.901884

_x_7 | .4869843 .1235324 3.94 0.000 .2448652 .7291033

_x_8 | 4.554325 .1164357 39.11 0.000 4.326115 4.782535

_x_9 | 6.470534 .195383 33.12 0.000 6.087591 6.853478

_x_10 | 2.666644 .1756067 15.19 0.000 2.322461 3.010827

_x_12 | 5.111081 .0751948 67.97 0.000 4.963702 5.25846

_x_17 | -1.415834 .0908697 -15.58 0.000 -1.593935 -1.237733

_x_14 | .7403587 .1228662 6.03 0.000 .4995454 .9811719

_x_13 | 1.468076 .1632134 8.99 0.000 1.148184 1.787968

_x_15 | -1.039167 .110891 -9.37 0.000 -1.256509 -.8218247

_x_16 | -.9503197 .0900525 -10.55 0.000 -1.126819 -.77382

_x_11 | 1.262343 .1769777 7.13 0.000 .9154734 1.609213

_cons | 1.841846 .1947538 9.46 0.000 1.460136 2.223557

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

. poisgof

Goodness-of-fit chi2 = 2.233234

Prob > chi2(7) = 0.9458

. desrep

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

Poisson regression

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

Dependent variable count

Optimization: ml

Number of observations: 25

Initial log likelihood: -80138.505

Log likelihood: -90.023

LR chi square: 160096.964

Model degrees of freedom: 17

Pseudo R-squared: 0.999

Prob: 0.000

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

nr Effect Coeff s.e.

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

count

husb

1 Mexican 1.472** 0.156

2 Oth Hisp 0.736** 0.122

3 All Others -0.510** 0.111

4 white 3.860** 0.100

wife

5 Mexican 2.301** 0.168

6 Oth Hisp 1.636** 0.136

7 All Others 0.487** 0.124

8 white 4.554** 0.116

QS2

9 1 6.471** 0.195

10 2 2.667** 0.176

11 4 5.111** 0.075

12 53 -1.416** 0.091

13 42 0.740** 0.123

14 41 1.468** 0.163

15 51 -1.039** 0.111

16 52 -0.950** 0.090

17 3 1.262** 0.177

18 _cons 1.842** 0.195

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

* p < .05

** p < .01

. *We ended up with the same QS only model, no asymetries added, even though we put the bar for entry at an easier .05

. sw poisson count (_x_1-_x_8) _x_9-_x_24, pe(.05) pr(.1)

begin with full model

p = 0.9720 >= 0.1000 removing _x_22

p = 0.7010 >= 0.1000 removing _x_21

p = 0.5047 >= 0.1000 removing _x_19

p = 0.4576 >= 0.1000 removing _x_20

p = 0.5321 >= 0.1000 removing _x_23

p = 0.5793 >= 0.1000 removing _x_18

p = 0.5194 >= 0.1000 removing _x_24

Poisson regression Number of obs = 25

LR chi2(17) = 160096.96

Prob > chi2 = 0.0000

Log likelihood = -90.022867 Pseudo R2 = 0.9989

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

count | Coef. Std. Err. z P>|z| [95% Conf. Interval]

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

_x_1 | 1.471544 .1558986 9.44 0.000 1.165988 1.7771

_x_2 | .73637 .1218848 6.04 0.000 .4974801 .9752598

_x_3 | -.5103945 .1108274 -4.61 0.000 -.7276122 -.2931768

_x_4 | 3.859923 .0999757 38.61 0.000 3.663975 4.055872

_x_5 | 2.300677 .1676349 13.72 0.000 1.972118 2.629235

_x_6 | 1.635904 .1357068 12.05 0.000 1.369923 1.901884

_x_7 | .4869843 .1235324 3.94 0.000 .2448652 .7291033

_x_8 | 4.554325 .1164357 39.11 0.000 4.326115 4.782535

_x_9 | 6.470534 .195383 33.12 0.000 6.087591 6.853478

_x_10 | 2.666644 .1756067 15.19 0.000 2.322461 3.010827

_x_11 | 1.262343 .1769777 7.13 0.000 .9154734 1.609213

_x_12 | 5.111081 .0751948 67.97 0.000 4.963702 5.25846

_x_13 | 1.468076 .1632134 8.99 0.000 1.148184 1.787968

_x_14 | .7403587 .1228662 6.03 0.000 .4995454 .9811719

_x_15 | -1.039167 .110891 -9.37 0.000 -1.256509 -.8218247

_x_16 | -.9503197 .0900525 -10.55 0.000 -1.126819 -.77382

_x_17 | -1.415834 .0908697 -15.58 0.000 -1.593935 -1.237733

_cons | 1.841846 .1947538 9.46 0.000 1.460136 2.223557

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

. poisgof

Goodness-of-fit chi2 = 2.233234

Prob > chi2(7) = 0.9458

. desrep

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

Poisson regression

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

Dependent variable count

Optimization: ml

Number of observations: 25

Initial log likelihood: -80138.505

Log likelihood: -90.023

LR chi square: 160096.964

Model degrees of freedom: 17

Pseudo R-squared: 0.999

Prob: 0.000

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

nr Effect Coeff s.e.

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

count

husb

1 Mexican 1.472** 0.156

2 Oth Hisp 0.736** 0.122

3 All Others -0.510** 0.111

4 white 3.860** 0.100

wife

5 Mexican 2.301** 0.168

6 Oth Hisp 1.636** 0.136

7 All Others 0.487** 0.124

8 white 4.554** 0.116

QS2

9 1 6.471** 0.195

10 2 2.667** 0.176

11 3 1.262** 0.177

12 4 5.111** 0.075

13 41 1.468** 0.163

14 42 0.740** 0.123

15 51 -1.039** 0.111

16 52 -0.950** 0.090

17 53 -1.416** 0.091

18 _cons 1.842** 0.195

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

* p < .05

** p < .01

. *This particular trial with stepwise seems to be giving us very consistent results regardless of forward or backward, and regardless of the entry and removal criteria, but it does not always work out so consistently.

. *Okay, that's enough of stepwise for now.

. clear

. * A quick note about about how to create QS:

. *I am using a numeric version of the data

. use "C:\AAA Miker Files\newer web pages\soc_388_notes\LA_interar_numeric.dta", clear

. describe

Contains data from C:\AAA Miker Files\newer web pages\soc_388_notes\LA_interar_

> numeric.dta

obs: 25

vars: 8 24 Oct 2007 15:54

size: 325 (99.9% of memory free)

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

storage display value

variable name type format label variable label

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

count int %8.0g

wife byte %10.0g race_label

husb byte %10.0g race_label

intermar byte %8.0g

intermar_full byte %8.0g

QS byte %8.0g

Asym byte %8.0g

QS2 byte %8.0g

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

Sorted by:

. tabulate husb wife [fweight=count]

| wife

husb | black Mexican Oth Hisp All Other white | Total

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

black | 4,074 63 32 42 215 | 4,426

Mexican | 25 3,947 143 95 1,009 | 5,219

Oth Hisp | 16 132 239 18 304 | 709

All Others | 19 78 18 1,022 360 | 1,497

white | 103 1,156 373 492 28,453 | 30,577

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

Total | 4,237 5,376 805 1,669 30,341 | 42,428

. tabulate husb wife [fweight=count], nolab

| wife

husb | 1 2 3 4 5 | Total

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

1 | 4,074 63 32 42 215 | 4,426

2 | 25 3,947 143 95 1,009 | 5,219

3 | 16 132 239 18 304 | 709

4 | 19 78 18 1,022 360 | 1,497

5 | 103 1,156 373 492 28,453 | 30,577

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

Total | 4,237 5,376 805 1,669 30,341 | 42,428

. gen new_qs=0

. replace new_qs=(10*husb)+wife if husb>wife

(10 real changes made)

. replace new_qs=(10*wife)+husb if husb<wife

(10 real changes made)

. *I am just using the row and column numbers here to create a separate variable for each off diagonal cell, it doesn't matter what the values are as long as they are all different.

. table husb wife, contents(mean new_qs)

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

| wife

husb | black Mexican Oth Hisp All Others white

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

black | 0 21 31 41 51

Mexican | 21 0 32 42 52

Oth Hisp | 31 32 0 43 53

All Others | 41 42 43 0 54

white | 51 52 53 54 0

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

. *This is quasi-symmetry, the largest number of symmetric interactions on a square table, which in this case has r+c=5+5=10 distinct values.

. exit, clear