---------------------------------------------------------------------------------------------------------
log: C:\AAA Miker Files\newer web pages\soc_388_notes\soc_388_2007\class_eight_log.log
log type: text
opened on: 18 Oct 2007, 11:02:02
. use "C:\AAA Miker Files\newer web pages\soc_meth_proj3\cps_y2k_numeric.dta", clear
. *This is the CPS data from last class, the big dataset with 133K observations that I am about to contract
. contract maritl sex race if age>19 & age<60, zero
no room to add more variables due to width
An attempt was made to add a variable that would have increased the memory required to store an
observation beyond what is currently possible. You have the following alternatives:
1. Store existing variables more efficiently; see help compress.
2. Drop some variables or observations; see help drop. (Think of Stata's data area as the area of
a rectangle; Stata can trade off width and length.)
3. Increase the amount of memory allocated to the data area using the set memory command; see help
memory.
r(902);
. clear
. set mem 200m
Current memory allocation
current memory usage
settable value description (1M = 1024k)
--------------------------------------------------------------------
set maxvar 5000 max. variables allowed 1.733M
set memory 200M max. data space 200.000M
set matsize 400 max. RHS vars in models 1.254M
-----------
202.987M
. use "C:\AAA Miker Files\newer web pages\soc_meth_proj3\cps_y2k_numeric.dta", clear
. contract maritl sex race if age>19 & age<60, zero
. rename _freq count
. tabulate race
p25 | Freq. Percent Cum.
------------+-----------------------------------
White | 14 25.00 25.00
Black | 14 25.00 50.00
Amer Indian | 14 25.00 75.00
Asian | 14 25.00 100.00
------------+-----------------------------------
Total | 56 100.00
. tabulate race [fweight=count]
p25 | Freq. Percent Cum.
------------+-----------------------------------
White | 61,729 85.18 85.18
Black | 7,045 9.72 94.90
Amer Indian | 957 1.32 96.22
Asian | 2,737 3.78 100.00
------------+-----------------------------------
Total | 72,468 100.00
. tabulate race [fweight=count], nolab
p25 | Freq. Percent Cum.
------------+-----------------------------------
1 | 61,729 85.18 85.18
2 | 7,045 9.72 94.90
3 | 957 1.32 96.22
4 | 2,737 3.78 100.00
------------+-----------------------------------
Total | 72,468 100.00
. save "C:\AAA Miker Files\newer web pages\soc_388_notes\collapsed age race marital from cps.dta"
file C:\AAA Miker Files\newer web pages\soc_388_notes\collapsed age race marital from cps.dta saved
. *build a contrast
. * build a contrast between blacks and whites...
. gen black_v_white=0 if race==1
(42 missing values generated)
. replace black_v_white=1 if race==2
(14 real changes made)
. label define race_contrast 0 "white" 1 "black"
. label val black_v_white race_contrast
. set linesize 79
. desmat: poisson count sex maritl race black_v_white*maritl
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 28
Initial log likelihood: -82911.173
Log likelihood: -714.325
LR chi square: 164393.696
Model degrees of freedom: 14
Pseudo R-squared: 0.991
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.054** 0.008
maritl
2 married, AF spouse present -4.935** 0.061
3 married, spouse absent -3.889** 0.036
4 widowed -3.857** 0.036
5 divorced -1.747** 0.013
6 separated -3.394** 0.028
7 never married -1.030** 0.010
race
8 Black -2.660** 0.020
black_v_white.maritl
9 black.married, AF spouse present 0.780** 0.167
10 black.married, spouse absent 0.800** 0.099
11 black.widowed 1.034** 0.089
12 black.divorced 0.602** 0.042
13 black.separated 1.542** 0.060
14 black.never married 1.065** 0.029
15 _cons 9.832** 0.006
-------------------------------------------------------------------------------
* p < .05
** p < .01
. tabulate race [fweight=count]
p25 | Freq. Percent Cum.
------------+-----------------------------------
White | 61,729 85.18 85.18
Black | 7,045 9.72 94.90
Amer Indian | 957 1.32 96.22
Asian | 2,737 3.78 100.00
------------+-----------------------------------
Total | 72,468 100.00
. desmat: poisson count sex maritl race
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -1809.960
LR chi square: 240625.359
Model degrees of freedom: 10
Pseudo R-squared: 0.985
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.057** 0.007
maritl
2 married, AF spouse present -4.844** 0.055
3 married, spouse absent -3.740** 0.032
4 widowed -3.749** 0.032
5 divorced -1.709** 0.012
6 separated -3.185** 0.024
7 never married -0.900** 0.009
race
8 Black -2.170** 0.013
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
11 _cons 9.787** 0.006
-------------------------------------------------------------------------------
* p < .05
** p < .01
. replace black_v_white=0 if black_v_white==.
(28 real changes made)
. save "C:\AAA Miker Files\newer web pages\soc_388_notes\collapsed age race marital from cps.dta", replace
file C:\AAA Miker Files\newer web pages\soc_388_notes\collapsed age race marital from cps.dta saved
. table race , contents (mean black_v_white)
----------------------------
p25 | mean(black_~e)
------------+---------------
White | 0
Black | 1
Amer Indian | 0
Asian | 0
----------------------------
. *Make a note from class, when I first created the black_v_white variable it had missing values for the other two racial categories, which were then dropped from the regressions. I fixed this by setting the other categories back to zero.
. rename black_v_white black_v_others
. desmat: poisson count sex maritl race black_v_white*maritl
variable black_v_white not found
r(111);
. desmat: poisson count sex maritl race black_v_others*maritl
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -975.476
LR chi square: 242294.327
Model degrees of freedom: 16
Pseudo R-squared: 0.992
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.057** 0.007
maritl
2 married, AF spouse present -4.912** 0.058
3 married, spouse absent -3.803** 0.034
4 widowed -3.856** 0.035
5 divorced -1.761** 0.013
6 separated -3.390** 0.028
7 never married -1.009** 0.010
race
8 Black -2.655** 0.020
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
black_v_others.maritl
11 black.married, AF spouse present 0.757** 0.166
12 black.married, spouse absent 0.715** 0.099
13 black.widowed 1.032** 0.089
14 black.divorced 0.616** 0.041
15 black.separated 1.537** 0.059
16 black.never married 1.044** 0.029
17 _cons 9.825** 0.006
-------------------------------------------------------------------------------
* p < .05
** p < .01
. *Compared to married, which is category one, the category which is the default comparison category, blacks are over-represented in every other marital status group.
. poisgof
Goodness-of-fit chi2 = 1579.077
Prob > chi2(39) = 0.0000
. desmat: poisson count sex maritl race black_v_others*maritl=ind(1)*ind(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -975.476
LR chi square: 242294.327
Model degrees of freedom: 16
Pseudo R-squared: 0.992
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.057** 0.007
maritl
2 married, AF spouse present -4.912** 0.058
3 married, spouse absent -3.803** 0.034
4 widowed -3.856** 0.035
5 divorced -1.761** 0.013
6 separated -3.390** 0.028
7 never married -1.009** 0.010
race
8 Black -2.655** 0.020
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
black_v_others.maritl
11 black.married, AF spouse present 0.757** 0.166
12 black.married, spouse absent 0.715** 0.099
13 black.widowed 1.032** 0.089
14 black.divorced 0.616** 0.041
15 black.separated 1.537** 0.059
16 black.never married 1.044** 0.029
17 _cons 9.825** 0.006
-------------------------------------------------------------------------------
* p < .05
** p < .01
. poisgof
Goodness-of-fit chi2 = 1579.077
Prob > chi2(39) = 0.0000
. *exactly the same coefficients...
. *because ind(1) is the default
. desmat: poisson count sex maritl race black_v_others*maritl=ind(1)*ind(5)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -975.476
LR chi square: 242294.327
Model degrees of freedom: 16
Pseudo R-squared: 0.992
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.057** 0.007
maritl
2 married, AF spouse present -4.912** 0.058
3 married, spouse absent -3.803** 0.034
4 widowed -3.856** 0.035
5 divorced -1.761** 0.013
6 separated -3.390** 0.028
7 never married -1.009** 0.010
race
8 Black -2.039** 0.036
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
black_v_others.maritl
11 black.married, spouse present -0.616** 0.041
12 black.married, AF spouse present 0.141 0.169
13 black.married, spouse absent 0.099 0.103
14 black.widowed 0.416** 0.094
15 black.separated 0.922** 0.067
16 black.never married 0.428** 0.042
17 _cons 9.825** 0.006
-------------------------------------------------------------------------------
* p < .05
** p < .01
. *Here the coefficients for the interactions are all different, because the comparison is to divorce.
. poisgof
Goodness-of-fit chi2 = 1579.077
Prob > chi2(39) = 0.0000
. *if we add a third dimension- is there a gender difference for blacks in terms of marital status, the problem with the coefficients depending on the excluded category gets even more troublesome.
. desmat: poisson count sex maritl race black_v_others*maritl*sex=ind(1)*ind(5)=ind(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.268** 0.024
maritl
2 married, AF spouse present -6.796** 0.213
3 married, spouse absent -3.695** 0.046
4 widowed -4.837** 0.080
5 divorced -1.878** 0.020
6 separated -3.645** 0.045
7 never married -0.866** 0.013
race
8 Black -2.167** 0.058
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
black_v_others.maritl
11 black.married, spouse present -0.449** 0.065
12 black.married, AF spouse present 1.436** 0.386
13 black.married, spouse absent 0.003 0.156
14 black.widowed 0.641** 0.203
15 black.separated 0.931** 0.112
16 black.never married 0.259** 0.066
black_v_others.sex
17 black.female 0.215** 0.075
maritl.sex
18 married, spouse present.female -0.218** 0.026
19 married, AF spouse present.female 2.258** 0.223
20 married, spouse absent.female -0.442** 0.071
21 widowed.female 1.230** 0.092
22 separated.female 0.231** 0.061
23 never married.female -0.518** 0.029
black_v_others.maritl.sex
24 black.married, spouse present.female -0.293** 0.085
25 black.married, AF spouse present.female -1.578** 0.431
26 black.married, spouse absent.female 0.222 0.208
27 black.widowed.female -0.335 0.229
28 black.separated.female -0.032 0.139
29 black.never married.female 0.367** 0.086
30 _cons 9.829** 0.007
-------------------------------------------------------------------------------
* p < .05
** p < .01
. poisgof
Goodness-of-fit chi2 = 298.9646
Prob > chi2(26) = 0.0000
. desmat: poisson count sex maritl race black_v_others*maritl*sex=ind(0)*ind(5)=ind(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.268** 0.024
maritl
2 married, AF spouse present -6.796** 0.213
3 married, spouse absent -3.695** 0.046
4 widowed -4.837** 0.080
5 divorced -1.878** 0.020
6 separated -3.645** 0.045
7 never married -0.866** 0.013
race
8 Black -2.167** 0.058
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
black_v_others.maritl
11 black.married, spouse present -0.449** 0.065
12 black.married, AF spouse present 1.436** 0.386
13 black.married, spouse absent 0.003 0.156
14 black.widowed 0.641** 0.203
15 black.separated 0.931** 0.112
16 black.never married 0.259** 0.066
black_v_others.sex
17 black.female 0.215** 0.075
maritl.sex
18 married, spouse present.female -0.218** 0.026
19 married, AF spouse present.female 2.258** 0.223
20 married, spouse absent.female -0.442** 0.071
21 widowed.female 1.230** 0.092
22 separated.female 0.231** 0.061
23 never married.female -0.518** 0.029
black_v_others.maritl.sex
24 black.married, spouse present.female -0.293** 0.085
25 black.married, AF spouse present.female -1.578** 0.431
26 black.married, spouse absent.female 0.222 0.208
27 black.widowed.female -0.335 0.229
28 black.separated.female -0.032 0.139
29 black.never married.female 0.367** 0.086
30 _cons 9.829** 0.007
-------------------------------------------------------------------------------
* p < .05
** p < .01
. desmat: poisson count sex maritl race black_v_others*maritl*sex=ind(2)*ind(5)=ind(1)
*Later Note: I meant ind(2)*ind(5)*ind(1)- I am not sure why it worked with= but see desmat help.
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.483** 0.071
maritl
2 married, AF spouse present -4.910** 0.317
3 married, spouse absent -3.243** 0.140
4 widowed -3.747** 0.179
5 divorced -1.429** 0.062
6 separated -2.264** 0.088
7 never married -0.159** 0.040
race
8 Black -2.167** 0.058
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
black_v_others.maritl
11 white.married, spouse present 0.449** 0.065
12 white.married, AF spouse present -1.436** 0.386
13 white.married, spouse absent -0.003 0.156
14 white.widowed -0.641** 0.203
15 white.separated -0.931** 0.112
16 white.never married -0.259** 0.066
black_v_others.sex
17 white.female -0.215** 0.075
maritl.sex
18 married, spouse present.female -0.511** 0.080
19 married, AF spouse present.female 0.680 0.369
20 married, spouse absent.female -0.220 0.196
21 widowed.female 0.895** 0.210
22 separated.female 0.199 0.125
23 never married.female -0.151 0.080
black_v_others.maritl.sex
24 white.married, spouse present.female 0.293** 0.085
25 white.married, AF spouse present.female 1.578** 0.431
26 white.married, spouse absent.female -0.222 0.208
27 white.widowed.female 0.335 0.229
28 white.separated.female 0.032 0.139
29 white.never married.female -0.367** 0.086
30 _cons 9.380** 0.064
-------------------------------------------------------------------------------
* p < .05
** p < .01
. *If you have only two categories, switching the indicator comparison category just reverses the coefficients. But if you have many categories, it becomes a thornier problem.
. poisgof
Goodness-of-fit chi2 = 298.9646
Prob > chi2(26) = 0.0000
. *for desmat, the key to dealing with multiple categories and several dimensions, without having your coefficients all depend on the arbitrary excluded categories, is sums-to-zero coding, or what desmat calls deviation coding, which is the default in SAS, but indicator coding is the default in desmat and in the built in xi function in stata.
. desmat: poisson count sex maritl race black_v_others*maritl*sex=dev(1)*dev(2)*dev(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
race
1 Amer Indian -4.167** 0.033
2 Asian -3.116** 0.020
black_v_others
3 black -0.887** 0.019
maritl
4 married, spouse present 2.371** 0.021
5 married, spouse absent -1.080** 0.045
6 widowed -1.207** 0.050
7 divorced 0.900** 0.025
8 separated -0.293** 0.031
9 never married 1.874** 0.021
black_v_others.maritl
10 black.married, spouse present -0.440** 0.021
11 black.married, spouse absent -0.086 0.045
12 black.widowed 0.094 0.050
13 black.divorced -0.143** 0.025
14 black.separated 0.315** 0.031
15 black.never married 0.079** 0.021
sex
16 female 0.311** 0.019
black_v_others.sex
17 black.female -0.005 0.019
maritl.sex
18 married, spouse present.female -0.305** 0.021
19 married, spouse absent.female -0.288** 0.045
20 widowed.female 0.409** 0.050
21 divorced.female -0.123** 0.025
22 separated.female -0.015 0.031
23 never married.female -0.290** 0.021
black_v_others.maritl.sex
24 black.married, spouse present.female -0.014 0.021
25 black.married, spouse absent.female 0.114* 0.045
26 black.widowed.female -0.025 0.050
27 black.divorced.female 0.059* 0.025
28 black.separated.female 0.051 0.031
29 black.never married.female 0.151** 0.021
30 _cons 6.155** 0.019
-------------------------------------------------------------------------------
* p < .05
** p < .01
. poisgof
Goodness-of-fit chi2 = 298.9646
Prob > chi2(26) = 0.0000
. *the same goodness of fit, but all the coeffients in the interactions are built on deviation coding, so they are not dependent on excluded category.
. desmat: poisson count sex maritl race black_v_others*maritl*sex=dev(1)*dev(2)*dev(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
race
1 Amer Indian -4.167** 0.033
2 Asian -3.116** 0.020
black_v_others
3 black -0.887** 0.019
maritl
4 married, spouse present 2.371** 0.021
5 married, spouse absent -1.080** 0.045
6 widowed -1.207** 0.050
7 divorced 0.900** 0.025
8 separated -0.293** 0.031
9 never married 1.874** 0.021
black_v_others.maritl
10 black.married, spouse present -0.440** 0.021
11 black.married, spouse absent -0.086 0.045
12 black.widowed 0.094 0.050
13 black.divorced -0.143** 0.025
14 black.separated 0.315** 0.031
15 black.never married 0.079** 0.021
sex
16 female 0.311** 0.019
black_v_others.sex
17 black.female -0.005 0.019
maritl.sex
18 married, spouse present.female -0.305** 0.021
19 married, spouse absent.female -0.288** 0.045
20 widowed.female 0.409** 0.050
21 divorced.female -0.123** 0.025
22 separated.female -0.015 0.031
23 never married.female -0.290** 0.021
black_v_others.maritl.sex
24 black.married, spouse present.female -0.014 0.021
25 black.married, spouse absent.female 0.114* 0.045
26 black.widowed.female -0.025 0.050
27 black.divorced.female 0.059* 0.025
28 black.separated.female 0.051 0.031
29 black.never married.female 0.151** 0.021
30 _cons 6.155** 0.019
-------------------------------------------------------------------------------
* p < .05
** p < .01
. poisgof
Goodness-of-fit chi2 = 298.9646
Prob > chi2(26) = 0.0000
. desmat: poisson count sex maritl race black_v_others*maritl*sex=ind(2)*ind(5)*ind(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.483** 0.071
maritl
2 married, AF spouse present -4.910** 0.317
3 married, spouse absent -3.243** 0.140
4 widowed -3.747** 0.179
5 divorced -1.429** 0.062
6 separated -2.264** 0.088
7 never married -0.159** 0.040
race
8 Black -2.167** 0.058
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
black_v_others.maritl
11 white.married, spouse present 0.449** 0.065
12 white.married, AF spouse present -1.436** 0.386
13 white.married, spouse absent -0.003 0.156
14 white.widowed -0.641** 0.203
15 white.separated -0.931** 0.112
16 white.never married -0.259** 0.066
black_v_others.sex
17 white.female -0.215** 0.075
maritl.sex
18 married, spouse present.female -0.511** 0.080
19 married, AF spouse present.female 0.680 0.369
20 married, spouse absent.female -0.220 0.196
21 widowed.female 0.895** 0.210
22 separated.female 0.199 0.125
23 never married.female -0.151 0.080
black_v_others.maritl.sex
24 white.married, spouse present.female 0.293** 0.085
25 white.married, AF spouse present.female 1.578** 0.431
26 white.married, spouse absent.female -0.222 0.208
27 white.widowed.female 0.335 0.229
28 white.separated.female 0.032 0.139
29 white.never married.female -0.367** 0.086
30 _cons 9.380** 0.064
-------------------------------------------------------------------------------
* p < .05
** p < .01
. poisgof
Goodness-of-fit chi2 = 298.9646
Prob > chi2(26) = 0.0000
. desmat: poisson count sex maritl race black_v_others*maritl*sex=ind(1)*ind(5)*ind(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
sex
1 female 0.268** 0.024
maritl
2 married, AF spouse present -6.796** 0.213
3 married, spouse absent -3.695** 0.046
4 widowed -4.837** 0.080
5 divorced -1.878** 0.020
6 separated -3.645** 0.045
7 never married -0.866** 0.013
race
8 Black -2.167** 0.058
9 Amer Indian -4.167** 0.033
10 Asian -3.116** 0.020
black_v_others.maritl
11 black.married, spouse present -0.449** 0.065
12 black.married, AF spouse present 1.436** 0.386
13 black.married, spouse absent 0.003 0.156
14 black.widowed 0.641** 0.203
15 black.separated 0.931** 0.112
16 black.never married 0.259** 0.066
black_v_others.sex
17 black.female 0.215** 0.075
maritl.sex
18 married, spouse present.female -0.218** 0.026
19 married, AF spouse present.female 2.258** 0.223
20 married, spouse absent.female -0.442** 0.071
21 widowed.female 1.230** 0.092
22 separated.female 0.231** 0.061
23 never married.female -0.518** 0.029
black_v_others.maritl.sex
24 black.married, spouse present.female -0.293** 0.085
25 black.married, AF spouse present.female -1.578** 0.431
26 black.married, spouse absent.female 0.222 0.208
27 black.widowed.female -0.335 0.229
28 black.separated.female -0.032 0.139
29 black.never married.female 0.367** 0.086
30 _cons 9.829** 0.007
-------------------------------------------------------------------------------
* p < .05
** p < .01
. save "C:\AAA Miker Files\newer web pages\soc_388_notes\collapsed age race marital from cps.dta", replace
file C:\AAA Miker Files\newer web pages\soc_388_notes\collapsed age race marital from cps.dta saved
. *Women are less likely to be married than men, in part because women live longer (and are more likely to be widowed). For blacks, the gender difference is even stronger, that is black women are even less likely to be married than black men.
. desmat: poisson count sex maritl race black_v_others*maritl*sex=dev(1)*dev(2)*dev(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
race
1 Amer Indian -4.167** 0.033
2 Asian -3.116** 0.020
black_v_others
3 black -0.887** 0.019
maritl
4 married, spouse present 2.371** 0.021
5 married, spouse absent -1.080** 0.045
6 widowed -1.207** 0.050
7 divorced 0.900** 0.025
8 separated -0.293** 0.031
9 never married 1.874** 0.021
black_v_others.maritl
10 black.married, spouse present -0.440** 0.021
11 black.married, spouse absent -0.086 0.045
12 black.widowed 0.094 0.050
13 black.divorced -0.143** 0.025
14 black.separated 0.315** 0.031
15 black.never married 0.079** 0.021
sex
16 female 0.311** 0.019
black_v_others.sex
17 black.female -0.005 0.019
maritl.sex
18 married, spouse present.female -0.305** 0.021
19 married, spouse absent.female -0.288** 0.045
20 widowed.female 0.409** 0.050
21 divorced.female -0.123** 0.025
22 separated.female -0.015 0.031
23 never married.female -0.290** 0.021
black_v_others.maritl.sex
24 black.married, spouse present.female -0.014 0.021
25 black.married, spouse absent.female 0.114* 0.045
26 black.widowed.female -0.025 0.050
27 black.divorced.female 0.059* 0.025
28 black.separated.female 0.051 0.031
29 black.never married.female 0.151** 0.021
30 _cons 6.155** 0.019
-------------------------------------------------------------------------------
* p < .05
** p < .01
. *Here, after using what I say is the better deviation coding, there doesn't a
> ppear to be as much of a black women marriage deficit, though they are more l
> ikely than black men to be never married.
. desmat: poisson count sex maritl race black_v_others*maritl*sex=dev(1)*dev(5
> )*dev(1)
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 56
Initial log likelihood: -122122.640
Log likelihood: -335.420
LR chi square: 243574.440
Model degrees of freedom: 29
Pseudo R-squared: 0.997
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
race
1 Amer Indian -4.167** 0.033
2 Asian -3.116** 0.020
black_v_others
3 black -0.887** 0.019
maritl
4 married, spouse present 2.371** 0.021
5 married, AF spouse present -2.565** 0.092
6 married, spouse absent -1.080** 0.045
7 widowed -1.207** 0.050
8 separated -0.293** 0.031
9 never married 1.874** 0.021
black_v_others.maritl
10 black.married, spouse present -0.440** 0.021
11 black.married, AF spouse present 0.181* 0.092
12 black.married, spouse absent -0.086 0.045
13 black.widowed 0.094 0.050
14 black.separated 0.315** 0.031
15 black.never married 0.079** 0.021
sex
16 female 0.311** 0.019
black_v_others.sex
17 black.female -0.005 0.019
maritl.sex
18 married, spouse present.female -0.305** 0.021
19 married, AF spouse present.female 0.612** 0.092
20 married, spouse absent.female -0.288** 0.045
21 widowed.female 0.409** 0.050
22 separated.female -0.015 0.031
23 never married.female -0.290** 0.021
black_v_others.maritl.sex
24 black.married, spouse present.female -0.014 0.021
25 black.married, AF spouse present.female -0.336** 0.092
26 black.married, spouse absent.female 0.114* 0.045
27 black.widowed.female -0.025 0.050
28 black.separated.female 0.051 0.031
29 black.never married.female 0.151** 0.021
30 _cons 6.155** 0.019
-------------------------------------------------------------------------------
* p < .05
** p < .01
. *after changing the excluded category, using deviation coding here we have the same coefficients for the other stuff.
. *blacks remain much less likely to be married and more likely to be never married, the gender differences within the black community are not as strong.
. poisgof
Goodness-of-fit chi2 = 298.9646
Prob > chi2(26) = 0.0000
. exit, clear