------------------------------------------------------------------------------------------------------------------------------------------------------
name: STATA_Chapter10a
log: C:\Dropbox\PilesOfVariance\Chapter10a\STATA\STATA_Chapter10a_Output.smcl
log type: smcl
opened on: 27 Jan 2015, 17:03:07
.
. display as result "Chapter 10a: Descriptive Statistics for Time-Invariant Variables"
Chapter 10a: Descriptive Statistics for Time-Invariant Variables
. preserve
. collapse aget0 ytdeatht0, by(personid)
. summarize aget0 ytdeatht0
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
aget0 | 207 83.33409 2.966657 79.41918 97.77808
ytdeatht0 | 207 -7.174669 3.978486 -15.90959 -.0491803
. corr aget0 ytdeatht0
(obs=207)
| aget0 ytdeat~0
-------------+------------------
aget0 | 1.0000
ytdeatht0 | 0.1536 1.0000
. restore
.
. display as result "Chapter 10a: Descriptive Statistics for Time-Varying Variables"
Chapter 10a: Descriptive Statistics for Time-Varying Variables
. summarize time tvage tvytdeath recall
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
time | 557 2.674216 2.603101 0 8.502732
tvage | 557 85.64534 3.55728 79.41918 99.89863
tvytdeath | 557 -5.869738 3.66448 -15.90959 -.0491803
recall | 557 10.1939 3.826512 0 16
.
. display as result "Ch 10a: Empty Means, Random Intercept Model for Prose Recall"
Ch 10a: Empty Means, Random Intercept Model for Prose Recall
. mixed recall , ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1428.6775
Iteration 1: log likelihood = -1428.6775
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(0) = .
Log likelihood = -1428.6775 Prob > chi2 = .
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_cons | 9.734908 .2505783 38.85 0.000 9.243784 10.22603
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 10.45792 1.309559 8.181943 13.36701
-----------------------------+------------------------------------------------
var(Residual) | 5.164586 .3930054 4.448999 5.99527
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 217.28 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1428.678 3 2863.355 2873.353
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estat icc,
Intraclass correlation
------------------------------------------------------------------------------
Level | ICC Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid | .6694138 .0347124 .598231 .7336009
------------------------------------------------------------------------------
. estat wcorrelation, covariance,
Covariances for personid = 2:
obs | 1 2
-------------+----------------
1 | 15.623
2 | 10.458 15.623
. estat wcorrelation,
Standard deviations and correlations for personid = 2:
Standard deviations:
obs | 1 2
-------------+----------------
sd | 3.953 3.953
Correlations:
obs | 1 2
-------------+----------------
1 | 1.000
2 | 0.669 1.000
.
. display as result "Ch 10a: Saturated Means by Rounded Years in Study, Random Intercept Model for Prose Recall"
Ch 10a: Saturated Means by Rounded Years in Study, Random Intercept Model for Prose Recall
. mixed recall i.occasion, ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1420.7044
Iteration 1: log likelihood = -1420.7044
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(4) = 16.36
Log likelihood = -1420.7044 Prob > chi2 = 0.0026
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
occasion |
2 | .3199198 .2590023 1.24 0.217 -.1877153 .8275549
4 | .0794866 .2934922 0.27 0.787 -.4957475 .6547208
6 | .2801501 .3456074 0.81 0.418 -.397228 .9575282
8 | -1.214967 .3952252 -3.07 0.002 -1.989594 -.4403395
|
_cons | 9.676717 .2776392 34.85 0.000 9.132554 10.22088
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 10.64352 1.320273 8.346378 13.57289
-----------------------------+------------------------------------------------
var(Residual) | 4.920005 .3750727 4.23716 5.712895
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 225.30 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1420.704 7 2855.409 2878.738
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. contrast i.occasion,
Contrasts of marginal linear predictions
Margins : asbalanced
------------------------------------------------
| df chi2 P>chi2
-------------+----------------------------------
recall |
occasion | 4 16.36 0.0026
------------------------------------------------
. margins i.occasion,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
occasion |
0 | 9.676717 .2776392 34.85 0.000 9.132554 10.22088
2 | 9.996637 .3029749 32.99 0.000 9.402817 10.59046
4 | 9.756204 .3346483 29.15 0.000 9.100305 10.4121
6 | 9.956867 .3818312 26.08 0.000 9.208492 10.70524
8 | 8.46175 .4275118 19.79 0.000 7.623843 9.299658
------------------------------------------------------------------------------
.
. display as result "Ch 10a: Saturated Means by Rounded Age, Random Intercept Model for Prose Recall"
Ch 10a: Saturated Means by Rounded Age, Random Intercept Model for Prose Recall
. mixed recall i.roundage, ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1415.083
Iteration 1: log likelihood = -1415.083
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(16) = 28.09
Log likelihood = -1415.083 Prob > chi2 = 0.0309
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
roundage |
80 | -3.040891 1.808104 -1.68 0.093 -6.584709 .502927
81 | -3.746704 1.789519 -2.09 0.036 -7.254097 -.2393119
82 | -3.357461 1.749538 -1.92 0.055 -6.786494 .0715706
83 | -3.609251 1.785102 -2.02 0.043 -7.107986 -.1105156
84 | -3.641355 1.739709 -2.09 0.036 -7.051123 -.2315878
85 | -3.881963 1.786291 -2.17 0.030 -7.38303 -.3808968
86 | -3.511053 1.743324 -2.01 0.044 -6.927906 -.0942007
87 | -3.618165 1.7844 -2.03 0.043 -7.115525 -.1208057
88 | -3.91318 1.75241 -2.23 0.026 -7.34784 -.4785203
89 | -4.522045 1.79076 -2.53 0.012 -8.031869 -1.012221
90 | -3.863591 1.80446 -2.14 0.032 -7.400267 -.3269145
91 | -5.799055 1.842761 -3.15 0.002 -9.4108 -2.187309
92 | -3.972397 1.848589 -2.15 0.032 -7.595565 -.349228
93 | -5.703432 1.996853 -2.86 0.004 -9.617192 -1.789673
94 | -2.996408 2.059523 -1.45 0.146 -7.032998 1.040182
95 | -2.751423 2.12049 -1.30 0.194 -6.907507 1.404661
|
_cons | 13.49127 1.743314 7.74 0.000 10.07444 16.90811
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 10.47406 1.297342 8.216429 13.35202
-----------------------------+------------------------------------------------
var(Residual) | 4.813043 .3668586 4.145145 5.588559
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 228.54 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1415.083 19 2868.166 2931.488
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. contrast i.roundage,
Contrasts of marginal linear predictions
Margins : asbalanced
------------------------------------------------
| df chi2 P>chi2
-------------+----------------------------------
recall |
roundage | 16 28.09 0.0309
------------------------------------------------
. margins i.roundage,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
roundage |
79 | 13.49127 1.743314 7.74 0.000 10.07444 16.90811
80 | 10.45038 .6118924 17.08 0.000 9.251096 11.64967
81 | 9.74457 .4497905 21.66 0.000 8.862997 10.62614
82 | 10.13381 .4841158 20.93 0.000 9.184963 11.08266
83 | 9.882024 .4283558 23.07 0.000 9.042462 10.72159
84 | 9.849919 .4242857 23.22 0.000 9.018334 10.6815
85 | 9.609311 .4344014 22.12 0.000 8.7579 10.46072
86 | 9.980221 .4556477 21.90 0.000 9.087168 10.87327
87 | 9.873109 .4343894 22.73 0.000 9.021721 10.7245
88 | 9.578094 .4983522 19.22 0.000 8.601341 10.55485
89 | 8.969229 .4584126 19.57 0.000 8.070757 9.867701
90 | 9.627684 .5907544 16.30 0.000 8.469826 10.78554
91 | 7.69222 .6317473 12.18 0.000 6.454018 8.930422
92 | 9.518877 .7213114 13.20 0.000 8.105133 10.93262
93 | 7.787842 .9904464 7.86 0.000 5.846603 9.729081
94 | 10.49487 1.159317 9.05 0.000 8.222647 12.76709
95 | 10.73985 1.219186 8.81 0.000 8.350291 13.12941
------------------------------------------------------------------------------
.
. display as result "Ch 10a: Saturated Means by Rounded Years to Death, Random Intercept Model for Prose Recall"
Ch 10a: Saturated Means by Rounded Years to Death, Random Intercept Model for Prose Recall
. mixed recall i.roundytdeath, ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1417.5724
Iteration 1: log likelihood = -1417.5724
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(15) = 22.91
Log likelihood = -1417.5724 Prob > chi2 = 0.0860
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
roundytdeath |
86 | .5416656 1.536966 0.35 0.725 -2.470733 3.554064
87 | .7375234 1.396824 0.53 0.597 -2.000201 3.475247
88 | .9518062 1.392787 0.68 0.494 -1.778005 3.681618
89 | .6697124 1.353754 0.49 0.621 -1.983597 3.323021
90 | .8457886 1.363671 0.62 0.535 -1.826958 3.518535
91 | .1362158 1.327377 0.10 0.918 -2.465396 2.737828
92 | .6194919 1.359368 0.46 0.649 -2.044821 3.283805
93 | .6646477 1.328232 0.50 0.617 -1.938639 3.267934
94 | 1.135974 1.356522 0.84 0.402 -1.522759 3.794707
95 | .0859927 1.332033 0.06 0.949 -2.524745 2.69673
96 | .0091726 1.352369 0.01 0.995 -2.641423 2.659768
97 | -.1580038 1.340433 -0.12 0.906 -2.785205 2.469197
98 | .4278425 1.361838 0.31 0.753 -2.241311 3.096996
99 | -1.108634 1.356085 -0.82 0.414 -3.766511 1.549243
100 | -.5725507 1.456078 -0.39 0.694 -3.426412 2.28131
|
_cons | 9.556733 1.304465 7.33 0.000 7.000029 12.11344
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 10.60648 1.315434 8.317708 13.52504
-----------------------------+------------------------------------------------
var(Residual) | 4.848608 .3699914 4.175062 5.630815
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 226.58 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1417.572 18 2871.145 2931.134
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. contrast i.roundytdeath,
Contrasts of marginal linear predictions
Margins : asbalanced
------------------------------------------------
| df chi2 P>chi2
-------------+----------------------------------
recall |
roundytdeath | 15 22.91 0.0860
------------------------------------------------
. margins i.roundytdeath,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
roundytdeath |
85 | 9.556733 1.304465 7.33 0.000 7.000029 12.11344
86 | 10.0984 .9505272 10.62 0.000 8.235399 11.9614
87 | 10.29426 .699022 14.73 0.000 8.924198 11.66431
88 | 10.50854 .7227761 14.54 0.000 9.091924 11.92515
89 | 10.22645 .5684507 17.99 0.000 9.112302 11.34059
90 | 10.40252 .546394 19.04 0.000 9.331609 11.47343
91 | 9.692949 .4607899 21.04 0.000 8.789817 10.59608
92 | 10.17622 .5151612 19.75 0.000 9.166527 11.18592
93 | 10.22138 .4296754 23.79 0.000 9.379232 11.06353
94 | 10.69271 .4705051 22.73 0.000 9.770534 11.61488
95 | 9.642725 .4168003 23.14 0.000 8.825812 10.45964
96 | 9.565905 .4473465 21.38 0.000 8.689122 10.44269
97 | 9.398729 .4191907 22.42 0.000 8.57713 10.22033
98 | 9.984575 .4639753 21.52 0.000 9.0752 10.89395
99 | 8.448099 .4492567 18.80 0.000 7.567572 9.328626
100 | 8.984182 .6850347 13.11 0.000 7.641539 10.32683
------------------------------------------------------------------------------
.
. display as result "Ch 10a: Empty Means, Random Intercept Model for Years since Birth"
Ch 10a: Empty Means, Random Intercept Model for Years since Birth
. mixed tvage , ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1460.2136
Iteration 1: log likelihood = -1460.2133
Iteration 2: log likelihood = -1460.2133
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(0) = .
Log likelihood = -1460.2133 Prob > chi2 = .
------------------------------------------------------------------------------
tvage | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_cons | 85.40236 .2040709 418.49 0.000 85.00239 85.80233
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 5.091145 .8861275 3.619608 7.160929
-----------------------------+------------------------------------------------
var(Residual) | 7.77283 .5869496 6.703513 9.012721
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 72.93 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1460.213 3 2926.427 2936.425
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estat icc,
Intraclass correlation
------------------------------------------------------------------------------
Level | ICC Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid | .3957676 .0500664 .3029102 .4968037
------------------------------------------------------------------------------
. estat wcorrelation, covariance,
Covariances for personid = 2:
obs | 1 2
-------------+----------------
1 | 12.864
2 | 5.091 12.864
. estat wcorrelation,
Standard deviations and correlations for personid = 2:
Standard deviations:
obs | 1 2
-------------+----------------
sd | 3.587 3.587
Correlations:
obs | 1 2
-------------+----------------
1 | 1.000
2 | 0.396 1.000
.
. display as result "Ch 10a: Empty Means, Random Intercept Model for Years to Death"
Ch 10a: Empty Means, Random Intercept Model for Years to Death
. mixed tvytdeath , ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1469.4621
Iteration 1: log likelihood = -1469.462
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(0) = .
Log likelihood = -1469.462 Prob > chi2 = .
------------------------------------------------------------------------------
tvytdeath | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_cons | -5.486775 .2153242 -25.48 0.000 -5.908803 -5.064748
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 6.04281 .9770495 4.401593 8.295985
-----------------------------+------------------------------------------------
var(Residual) | 7.716211 .5823066 6.655304 8.946236
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 87.51 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1469.462 3 2944.924 2954.922
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estat icc,
Intraclass correlation
------------------------------------------------------------------------------
Level | ICC Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid | .4391889 .0482869 .3478052 .5348911
------------------------------------------------------------------------------
. estat wcorrelation, covariance,
Covariances for personid = 2:
obs | 1 2
-------------+----------------
1 | 13.759
2 | 6.043 13.759
. estat wcorrelation,
Standard deviations and correlations for personid = 2:
Standard deviations:
obs | 1 2
-------------+----------------
sd | 3.709 3.709
Correlations:
obs | 1 2
-------------+----------------
1 | 1.000
2 | 0.439 1.000
.
. display as result "Ch 10a: Empty Means, Random Intercept Model for Years in Study"
Ch 10a: Empty Means, Random Intercept Model for Years in Study
. mixed time , ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1326.4655
Iteration 1: log likelihood = -1322.7519
Iteration 2: log likelihood = -1322.7321
Iteration 3: log likelihood = -1322.7321
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(0) = .
Log likelihood = -1322.7321 Prob > chi2 = .
------------------------------------------------------------------------------
time | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_cons | 2.674216 .1101979 24.27 0.000 2.458232 2.8902
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 3.74e-20 1.03e-19 1.70e-22 8.23e-18
-----------------------------+------------------------------------------------
var(Residual) | 6.763969 .4053182 6.014436 7.606909
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 0.00 Prob >= chibar2 = 1.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1322.732 3 2651.464 2661.462
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estat icc,
Intraclass correlation
------------------------------------------------------------------------------
Level | ICC Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid | 5.53e-21 0 5.53e-21 5.53e-21
------------------------------------------------------------------------------
. estat wcorrelation, covariance,
Covariances for personid = 2:
obs | 1 2
-------------+----------------
1 | 6.764
2 | 0.000 6.764
. estat wcorrelation,
Standard deviations and correlations for personid = 2:
Standard deviations:
obs | 1 2
-------------+----------------
sd | 2.601 2.601
Correlations:
obs | 1 2
-------------+----------------
1 | 1.000
2 | 0.000 1.000
.
. display as result "Ch 10a: Fixed Quadratic, Random Intercept Model using Years since Birth"
Ch 10a: Fixed Quadratic, Random Intercept Model using Years since Birth
. mixed recall c.tvage84 c.tvage84#c.tvage84, ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1425.3132
Iteration 1: log likelihood = -1425.3132
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(2) = 6.79
Log likelihood = -1425.3132 Prob > chi2 = 0.0336
-------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
tvage84 | -.1189889 .0516498 -2.30 0.021 -.2202208 -.0177571
|
c.tvage84#c.tvage84 | .0047917 .0075791 0.63 0.527 -.010063 .0196464
|
_cons | 9.81965 .2634294 37.28 0.000 9.303338 10.33596
-------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 10.48043 1.307503 8.207034 13.38358
-----------------------------+------------------------------------------------
var(Residual) | 5.07161 .3860011 4.368786 5.8875
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 221.24 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1425.313 5 2860.626 2877.29
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estimates store FitFQRIAge,
.
. display as result "Eq 10a.1: Random Linear Model using Years since Birth"
Eq 10a.1: Random Linear Model using Years since Birth
. mixed recall c.tvage84, ///
> || personid: tvage84, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1420.1517
Iteration 1: log likelihood = -1419.8406
Iteration 2: log likelihood = -1419.8402
Iteration 3: log likelihood = -1419.8402
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(1) = 4.06
Log likelihood = -1419.8402 Prob > chi2 = 0.0439
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
tvage84 | -.0901116 .0447097 -2.02 0.044 -.177741 -.0024821
_cons | 9.811871 .2642278 37.13 0.000 9.293994 10.32975
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(tvage84) | .0899987 .0342662 .0426726 .1898119
var(_cons) | 11.26518 1.472973 8.71846 14.55582
cov(tvage84,_cons) | -.3274672 .1776429 -.6756409 .0207065
-----------------------------+------------------------------------------------
var(Residual) | 4.275193 .3840855 3.584952 5.098332
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 232.74 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1419.84 6 2851.68 2871.677
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
.
. display as result "Eq 10a.1: Fixed Quadratic, Random Linear Model using Years since Birth"
Eq 10a.1: Fixed Quadratic, Random Linear Model using Years since Birth
. mixed recall c.tvage84 c.tvage84#c.tvage84, ///
> || personid: tvage84, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1420.1353
Iteration 1: log likelihood = -1419.809
Iteration 2: log likelihood = -1419.8086
Iteration 3: log likelihood = -1419.8086
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(2) = 4.11
Log likelihood = -1419.8086 Prob > chi2 = 0.1284
-------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
tvage84 | -.0807891 .0575452 -1.40 0.160 -.1935755 .0319974
|
c.tvage84#c.tvage84 | -.0022047 .0084807 -0.26 0.795 -.0188266 .0144172
|
_cons | 9.824619 .2689636 36.53 0.000 9.29746 10.35178
-------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(tvage84) | .0922261 .0356656 .0432194 .196802
var(_cons) | 11.26274 1.47299 8.716058 14.55351
cov(tvage84,_cons) | -.3314963 .1792005 -.6827228 .0197302
-----------------------------+------------------------------------------------
var(Residual) | 4.26269 .3858777 3.569679 5.090241
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 232.25 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1419.809 7 2853.617 2876.946
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estimates store FitFQRLAge,
. lrtest FitFQRLAge FitFQRIAge,
Likelihood-ratio test LR chi2(2) = 11.01
(Assumption: FitFQRIAge nested in FitFQRLAge) Prob > chi2 = 0.0041
Note: The reported degrees of freedom assumes the null hypothesis is not on the boundary of the parameter space. If this is not true, then the
reported test is conservative.
. predict PredFQRLAge, xb,
. margins, at (c.tvage84=(-2(2)10)) vsquish,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
1._at : tvage84 = -2
2._at : tvage84 = 0
3._at : tvage84 = 2
4._at : tvage84 = 4
5._at : tvage84 = 6
6._at : tvage84 = 8
7._at : tvage84 = 10
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 9.977379 .3079918 32.39 0.000 9.373726 10.58103
2 | 9.824619 .2689636 36.53 0.000 9.29746 10.35178
3 | 9.654222 .2679669 36.03 0.000 9.129017 10.17943
4 | 9.466188 .2787929 33.95 0.000 8.919764 10.01261
5 | 9.260516 .3196708 28.97 0.000 8.633973 9.887059
6 | 9.037207 .4300079 21.02 0.000 8.194407 9.880007
7 | 8.79626 .6270603 14.03 0.000 7.567244 10.02528
------------------------------------------------------------------------------
. corr recall PredFQRLAge
(obs=557)
| recall PredFQ~e
-------------+------------------
recall | 1.0000
PredFQRLAge | 0.0593 1.0000
.
. display as result "Ch 10a: Fixed Quadratic, Random Intercept Model using Years to Death"
Ch 10a: Fixed Quadratic, Random Intercept Model using Years to Death
. mixed recall c.tvytdeath7 c.tvytdeath7#c.tvytdeath7, ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1422.6951
Iteration 1: log likelihood = -1422.6951
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(2) = 12.13
Log likelihood = -1422.6951 Prob > chi2 = 0.0023
-------------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------------+----------------------------------------------------------------
tvytdeath7 | -.0992235 .0384329 -2.58 0.010 -.1745506 -.0238964
|
c.tvytdeath7#c.tvytdeath7 | -.0148659 .0078739 -1.89 0.059 -.0302984 .0005666
|
_cons | 10.13279 .2824991 35.87 0.000 9.579097 10.68647
-------------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 10.44043 1.300531 8.178748 13.32753
-----------------------------+------------------------------------------------
var(Residual) | 5.012144 .3814878 4.31754 5.818497
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 221.74 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1422.695 5 2855.39 2872.054
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estimates store FitFQRIYTD,
.
. display as result "Eq 10a.1: Fixed Quadratic, Random Linear Model using Years to Death"
Eq 10a.1: Fixed Quadratic, Random Linear Model using Years to Death
. mixed recall c.tvytdeath7 c.tvytdeath7#c.tvytdeath7, ///
> || personid: tvytdeath7, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1414.578
Iteration 1: log likelihood = -1414.429
Iteration 2: log likelihood = -1414.4288
Iteration 3: log likelihood = -1414.4288
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(2) = 11.47
Log likelihood = -1414.4288 Prob > chi2 = 0.0032
-------------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------------+----------------------------------------------------------------
tvytdeath7 | -.086684 .0475226 -1.82 0.068 -.1798265 .0064586
|
c.tvytdeath7#c.tvytdeath7 | -.0202656 .0086927 -2.33 0.020 -.0373029 -.0032283
|
_cons | 10.14707 .2750136 36.90 0.000 9.608051 10.68608
-------------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(tvytde~7) | .1202498 .0396283 .0630334 .2294024
var(_cons) | 9.819069 1.387091 7.444324 12.95136
cov(tvytde~7,_cons) | -.0714948 .1593093 -.3837352 .2407457
-----------------------------+------------------------------------------------
var(Residual) | 4.066471 .3655795 3.409525 4.849998
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 238.28 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1414.429 7 2842.858 2866.187
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estimates store FitFQRLYTD,
. lrtest FitFQRLYTD FitFQRIYTD,
Likelihood-ratio test LR chi2(2) = 16.53
(Assumption: FitFQRIYTD nested in FitFQRLYTD) Prob > chi2 = 0.0003
Note: The reported degrees of freedom assumes the null hypothesis is not on the boundary of the parameter space. If this is not true, then the
reported test is conservative.
. predict PredFQRLYTD, xb,
. margins, at (c.tvytdeath7=(-4(2)6)) vsquish,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
1._at : tvytdeath7 = -4
2._at : tvytdeath7 = -2
3._at : tvytdeath7 = 0
4._at : tvytdeath7 = 2
5._at : tvytdeath7 = 4
6._at : tvytdeath7 = 6
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 10.16955 .3701552 27.47 0.000 9.444064 10.89505
2 | 10.23937 .3007028 34.05 0.000 9.650007 10.82874
3 | 10.14707 .2750136 36.90 0.000 9.608051 10.68608
4 | 9.892638 .261921 37.77 0.000 9.379282 10.40599
5 | 9.476082 .2725901 34.76 0.000 8.941816 10.01035
6 | 8.897402 .3559909 24.99 0.000 8.199673 9.595132
------------------------------------------------------------------------------
. corr recall PredFQRLYTD
(obs=557)
| recall PredFQ~D
-------------+------------------
recall | 1.0000
PredFQRLYTD | 0.1101 1.0000
.
. display as result "Ch 10a: Fixed Quadratic, Random Intercept Model using Years in Study"
Ch 10a: Fixed Quadratic, Random Intercept Model using Years in Study
. mixed recall c.time c.time#c.time, ///
> || personid: , variance mle covariance(unstructured),
Note: single-variable random-effects specification in personid equation; covariance structure set to identity
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1423.0044
Iteration 1: log likelihood = -1423.0044
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(2) = 11.57
Log likelihood = -1423.0044 Prob > chi2 = 0.0031
-------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
time | .2613309 .1192431 2.19 0.028 .0276188 .495043
|
c.time#c.time | -.0469071 .0158256 -2.96 0.003 -.0779248 -.0158895
|
_cons | 9.660987 .2749864 35.13 0.000 9.122023 10.19995
-------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Identity |
var(_cons) | 10.62162 1.320374 8.324879 13.55201
-----------------------------+------------------------------------------------
var(Residual) | 4.983023 .3797372 4.291667 5.785751
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) = 223.63 Prob >= chibar2 = 0.0000
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1423.004 5 2856.009 2872.672
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estimates store FitFQRIYIS,
.
. display as result "Eq 10a.1: Fixed Quadratic, Random Linear Model using Years in Study"
Eq 10a.1: Fixed Quadratic, Random Linear Model using Years in Study
. mixed recall c.time c.time#c.time, ///
> || personid: time, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1414.1476
Iteration 1: log likelihood = -1413.9187
Iteration 2: log likelihood = -1413.9179
Iteration 3: log likelihood = -1413.9179
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(2) = 10.84
Log likelihood = -1413.9179 Prob > chi2 = 0.0044
-------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
time | .2807025 .1114632 2.52 0.012 .0622386 .4991663
|
c.time#c.time | -.0479232 .0149497 -3.21 0.001 -.0772241 -.0186223
|
_cons | 9.638345 .2876754 33.50 0.000 9.074511 10.20218
-------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(time) | .1335909 .0444206 .0696212 .2563375
var(_cons) | 13.06751 1.653647 10.19709 16.74594
cov(time,_cons) | -.6552902 .2342868 -1.114484 -.1960964
-----------------------------+------------------------------------------------
var(Residual) | 3.957359 .3610391 3.309393 4.732194
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 241.80 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1413.918 7 2841.836 2865.165
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. estimates store FitFQRLYIS,
. lrtest FitFQRLYIS FitFQRIYIS,
Likelihood-ratio test LR chi2(2) = 18.17
(Assumption: FitFQRIYIS nested in FitFQRLYIS) Prob > chi2 = 0.0001
Note: The reported degrees of freedom assumes the null hypothesis is not on the boundary of the parameter space. If this is not true, then the
reported test is conservative.
. predict PredFQRLYIS, xb,
. margins, at (c.time=(0(2)8)) vsquish,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
1._at : time = 0
2._at : time = 2
3._at : time = 4
4._at : time = 6
5._at : time = 8
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 9.638345 .2876754 33.50 0.000 9.074511 10.20218
2 | 10.00806 .2632276 38.02 0.000 9.49214 10.52397
3 | 9.994383 .2781401 35.93 0.000 9.449238 10.53953
4 | 9.597323 .3051352 31.45 0.000 8.999269 10.19538
5 | 8.816878 .4341406 20.31 0.000 7.965978 9.667778
------------------------------------------------------------------------------
. corr recall PredFQRLYIS
(obs=557)
| recall PredFQ~S
-------------+------------------
recall | 1.0000
PredFQRLYIS | 0.0592 1.0000
.
. display as result "Eq 10a.2: Fixed Quadratic, Random Linear Model using Years since Birth"
Eq 10a.2: Fixed Quadratic, Random Linear Model using Years since Birth
. display as result "Controlling for Birth Cohort"
Controlling for Birth Cohort
. mixed recall c.tvage84 c.tvage84#c.tvage84 ///
> c.aget084 c.aget084#c.aget084 c.tvage84#c.aget084, ///
> || personid: tvage84, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1412.2404
Iteration 1: log likelihood = -1411.9232
Iteration 2: log likelihood = -1411.9227
Iteration 3: log likelihood = -1411.9227
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(5) = 20.17
Log likelihood = -1411.9227 Prob > chi2 = 0.0012
-------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
tvage84 | .295902 .1134579 2.61 0.009 .0735286 .5182755
|
c.tvage84#c.tvage84 | -.0453879 .0150827 -3.01 0.003 -.0749495 -.0158263
|
aget084 | -.5793354 .1543666 -3.75 0.000 -.8818883 -.2767825
|
c.aget084#c.aget084 | -.0774895 .0306508 -2.53 0.011 -.1375639 -.0174151
|
c.tvage84#c.aget084 | .125601 .0344975 3.64 0.000 .0579872 .1932149
|
_cons | 9.414132 .3509857 26.82 0.000 8.726213 10.10205
-------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(tvage84) | .0907217 .0343522 .0431916 .1905562
var(_cons) | 11.15461 1.448107 8.64869 14.38661
cov(tvage84,_cons) | -.3424301 .1769739 -.6892926 .0044325
-----------------------------+------------------------------------------------
var(Residual) | 4.110768 .3745462 3.438484 4.914495
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 232.97 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1411.923 10 2843.845 2877.173
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. * Multivariate Test of Birth Cohort Contextual Effects
. test (c.aget084=0) (c.aget084#c.aget084=0) (c.tvage84#c.aget084=0)
( 1) [recall]aget084 = 0
( 2) [recall]c.aget084#c.aget084 = 0
( 3) [recall]c.tvage84#c.aget084 = 0
chi2( 3) = 16.03
Prob > chi2 = 0.0011
. * Contextual Linear Birth Cohort on Intercept
. lincom c.aget084*1
( 1) [recall]aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.5793354 .1543666 -3.75 0.000 -.8818883 -.2767825
------------------------------------------------------------------------------
. * Contextual Quadratic Birth Cohort on Intercept
. lincom c.aget084#c.aget084*1
( 1) [recall]c.aget084#c.aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.0774895 .0306508 -2.53 0.011 -.1375639 -.0174151
------------------------------------------------------------------------------
. * Contextual Linear Birth Cohort on Linear Slope
. lincom c.tvage84#c.aget084*1
( 1) [recall]c.tvage84#c.aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | .125601 .0344975 3.64 0.000 .0579872 .1932149
------------------------------------------------------------------------------
. * Total Linear Birth Cohort on Intercept
. lincom c.aget084*1 + c.tvage84*1
( 1) [recall]tvage84 + [recall]aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.2834334 .1043786 -2.72 0.007 -.4880117 -.078855
------------------------------------------------------------------------------
. * Total Quadratic Birth Cohort on Intercept
. lincom c.aget084#c.aget084*1 + c.tvage84#c.aget084*1 + c.tvage84#c.tvage84*1
( 1) [recall]c.tvage84#c.tvage84 + [recall]c.aget084#c.aget084 + [recall]c.tvage84#c.aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | .0027236 .0215584 0.13 0.899 -.0395301 .0449774
------------------------------------------------------------------------------
. * Total Linear Birth Cohort on Linear Slope
. lincom c.tvage84#c.aget084*1 + c.tvage84#c.tvage84*2
( 1) 2*[recall]c.tvage84#c.tvage84 + [recall]c.tvage84#c.aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | .0348252 .0201822 1.73 0.084 -.0047312 .0743816
------------------------------------------------------------------------------
. estimates store FitCohAge,
. lrtest FitCohAge FitFQRLAge,
Likelihood-ratio test LR chi2(3) = 15.77
(Assumption: FitFQRLAge nested in FitCohAge) Prob > chi2 = 0.0013
. predict PredCohAge, xb,
. margins, at (c.tvage84=(-2(2)10) c.aget084=(-4(4)4)) vsquish,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
1._at : tvage84 = -2
aget084 = -4
2._at : tvage84 = -2
aget084 = 0
3._at : tvage84 = -2
aget084 = 4
4._at : tvage84 = 0
aget084 = -4
5._at : tvage84 = 0
aget084 = 0
6._at : tvage84 = 0
aget084 = 4
7._at : tvage84 = 2
aget084 = -4
8._at : tvage84 = 2
aget084 = 0
9._at : tvage84 = 2
aget084 = 4
10._at : tvage84 = 4
aget084 = -4
11._at : tvage84 = 4
aget084 = 0
12._at : tvage84 = 4
aget084 = 4
13._at : tvage84 = 6
aget084 = -4
14._at : tvage84 = 6
aget084 = 0
15._at : tvage84 = 6
aget084 = 4
16._at : tvage84 = 8
aget084 = -4
17._at : tvage84 = 8
aget084 = 0
18._at : tvage84 = 8
aget084 = 4
19._at : tvage84 = 10
aget084 = -4
20._at : tvage84 = 10
aget084 = 0
21._at : tvage84 = 10
aget084 = 4
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 10.72309 .4965887 21.59 0.000 9.749799 11.69639
2 | 8.640776 .5055843 17.09 0.000 7.649849 9.631703
3 | 4.078795 1.508292 2.70 0.007 1.122598 7.034992
4 | 10.49164 .498985 21.03 0.000 9.51365 11.46963
5 | 9.414132 .3509857 26.82 0.000 8.726213 10.10205
6 | 5.856959 1.02838 5.70 0.000 3.841371 7.872547
7 | 9.897087 .5224679 18.94 0.000 8.873068 10.92111
8 | 9.824385 .3390492 28.98 0.000 9.15986 10.48891
9 | 7.27202 .6880069 10.57 0.000 5.923551 8.620488
10 | 8.939428 .6173991 14.48 0.000 7.729348 10.14951
11 | 9.871534 .3619756 27.27 0.000 9.162075 10.58099
12 | 8.323977 .5098313 16.33 0.000 7.324726 9.323228
13 | 7.618665 .8506405 8.96 0.000 5.951441 9.28589
14 | 9.555579 .3986018 23.97 0.000 8.774334 10.33682
15 | 9.012831 .4922045 18.31 0.000 8.048128 9.977534
16 | 5.9348 1.239971 4.79 0.000 3.504501 8.365098
17 | 8.876522 .5228362 16.98 0.000 7.851782 9.901262
18 | 9.338581 .5767992 16.19 0.000 8.208076 10.46909
19 | 3.88783 1.774172 2.19 0.028 .4105177 7.365143
20 | 7.834361 .7978954 9.82 0.000 6.270515 9.398207
21 | 9.301228 .7286369 12.77 0.000 7.873126 10.72933
------------------------------------------------------------------------------
. corr recall PredCohAge
(obs=557)
| recall PredCo~e
-------------+------------------
recall | 1.0000
PredCohAge | 0.1570 1.0000
.
. display as result "Eq 10a.2: Fixed Quadratic, Random Linear Model using Years in Study"
Eq 10a.2: Fixed Quadratic, Random Linear Model using Years in Study
. display as result "Controlling for Birth Cohort"
Controlling for Birth Cohort
. mixed recall c.time c.time#c.time ///
> c.aget084 c.aget084#c.aget084 c.time#c.aget084, ///
> || personid: time, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1409.5185
Iteration 1: log likelihood = -1409.2735
Iteration 2: log likelihood = -1409.2726
Iteration 3: log likelihood = -1409.2726
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(5) = 20.50
Log likelihood = -1409.2726 Prob > chi2 = 0.0010
-------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
time | .313228 .112366 2.79 0.005 .0929946 .5334613
|
c.time#c.time | -.0455539 .0149692 -3.04 0.002 -.074893 -.0162149
|
aget084 | -.2972343 .1050947 -2.83 0.005 -.5032161 -.0912525
|
c.aget084#c.aget084 | .0091296 .0183208 0.50 0.618 -.0267784 .0450376
|
c.time#c.aget084 | .0442745 .0207982 2.13 0.033 .0035107 .0850382
|
_cons | 9.340209 .3515655 26.57 0.000 8.651153 10.02926
-------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(time) | .1272023 .0432081 .065367 .2475319
var(_cons) | 12.48375 1.59591 9.716918 16.03843
cov(time,_cons) | -.5963404 .225568 -1.038446 -.1542353
-----------------------------+------------------------------------------------
var(Residual) | 3.940494 .3596468 3.29505 4.712371
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 238.27 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1409.273 10 2838.545 2871.872
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. * Multivariate Test of Birth Cohort Total Effects
. test (c.aget084=0) (c.aget084#c.aget084=0) (c.time#c.aget084=0)
( 1) [recall]aget084 = 0
( 2) [recall]c.aget084#c.aget084 = 0
( 3) [recall]c.time#c.aget084 = 0
chi2( 3) = 9.44
Prob > chi2 = 0.0240
. * Contextual Linear Birth Cohort on Intercept
. lincom c.aget084*1 + c.time*-1
( 1) - [recall]time + [recall]aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.6104622 .1577068 -3.87 0.000 -.919562 -.3013625
------------------------------------------------------------------------------
. * Contextual Quadratic Birth Cohort on Intercept
. lincom c.aget084#c.aget084*1 + c.time#c.aget084*-1 + c.time#c.time*1
( 1) [recall]c.time#c.time + [recall]c.aget084#c.aget084 - [recall]c.time#c.aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.0806988 .0297157 -2.72 0.007 -.1389404 -.0224572
------------------------------------------------------------------------------
. * Contextual Linear Birth Cohort on Linear Slope
. lincom c.time#c.aget084*1 + c.time#c.time*-2
( 1) - 2*[recall]c.time#c.time + [recall]c.time#c.aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | .1353823 .0348185 3.89 0.000 .0671393 .2036254
------------------------------------------------------------------------------
. * Total Linear Birth Cohort on Intercept
. lincom c.aget084*1
( 1) [recall]aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.2972343 .1050947 -2.83 0.005 -.5032161 -.0912525
------------------------------------------------------------------------------
. * Total Quadratic Birth Cohort on Intercept
. lincom c.aget084#c.aget084*1
( 1) [recall]c.aget084#c.aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | .0091296 .0183208 0.50 0.618 -.0267784 .0450376
------------------------------------------------------------------------------
. * Total Linear Birth Cohort on Linear Slope
. lincom c.time#c.aget084*1
( 1) [recall]c.time#c.aget084 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | .0442745 .0207982 2.13 0.033 .0035107 .0850382
------------------------------------------------------------------------------
. estimates store FitCohAgeYIS,
. lrtest FitCohAgeYIS FitFQRLYIS,
Likelihood-ratio test LR chi2(3) = 9.29
(Assumption: FitFQRLYIS nested in FitCohAgeYIS) Prob > chi2 = 0.0257
. predict PredCohAgeYIS, xb,
. margins, at (c.time=(0(2)8) c.aget084=(-4(4)4)) vsquish,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
1._at : time = 0
aget084 = -4
2._at : time = 0
aget084 = 0
3._at : time = 0
aget084 = 4
4._at : time = 2
aget084 = -4
5._at : time = 2
aget084 = 0
6._at : time = 2
aget084 = 4
7._at : time = 4
aget084 = -4
8._at : time = 4
aget084 = 0
9._at : time = 4
aget084 = 4
10._at : time = 6
aget084 = -4
11._at : time = 6
aget084 = 0
12._at : time = 6
aget084 = 4
13._at : time = 8
aget084 = -4
14._at : time = 8
aget084 = 0
15._at : time = 8
aget084 = 4
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 10.67522 .5025292 21.24 0.000 9.690281 11.66016
2 | 9.340209 .3515655 26.57 0.000 8.651153 10.02926
3 | 8.297345 .5367589 15.46 0.000 7.245317 9.349373
4 | 10.76526 .4627276 23.26 0.000 9.858335 11.67219
5 | 9.784449 .3313159 29.53 0.000 9.135082 10.43382
6 | 9.095781 .4854144 18.74 0.000 8.144386 10.04718
7 | 10.49088 .4761396 22.03 0.000 9.557661 11.42409
8 | 9.864258 .3475764 28.38 0.000 9.183021 10.5455
9 | 9.529786 .5433576 17.54 0.000 8.464824 10.59475
10 | 9.852059 .518363 19.01 0.000 8.836086 10.86803
11 | 9.579635 .3824811 25.05 0.000 8.829986 10.32928
12 | 9.599358 .6838897 14.04 0.000 8.258959 10.93976
13 | 8.848809 .6360222 13.91 0.000 7.602228 10.09539
14 | 8.930581 .5117021 17.45 0.000 7.927663 9.933499
15 | 9.3045 .9164275 10.15 0.000 7.508335 11.10066
------------------------------------------------------------------------------
. corr recall PredCohAgeYIS
(obs=557)
| recall PredCo~S
-------------+------------------
recall | 1.0000
PredCohAge~S | 0.1615 1.0000
.
. display as result "Eq 10a.2: Fixed Quadratic, Random Linear Model using Years to Death"
Eq 10a.2: Fixed Quadratic, Random Linear Model using Years to Death
. display as result "Controlling for Death Cohort"
Controlling for Death Cohort
. mixed recall c.tvytdeath7 c.tvytdeath7#c.tvytdeath7 ///
> c.ytdeatht07 c.ytdeatht07#c.ytdeatht07 c.tvytdeath7#c.ytdeatht07, ///
> || personid: tvytdeath7, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1410.224
Iteration 1: log likelihood = -1410.0882
Iteration 2: log likelihood = -1410.0881
Iteration 3: log likelihood = -1410.0881
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(5) = 20.50
Log likelihood = -1410.0881 Prob > chi2 = 0.0010
-------------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------------+----------------------------------------------------------------
tvytdeath7 | .2133189 .1140841 1.87 0.062 -.0102818 .4369196
|
c.tvytdeath7#c.tvytdeath7 | -.0563812 .0159811 -3.53 0.000 -.0877037 -.0250588
|
ytdeatht07 | -.3521827 .12952 -2.72 0.007 -.6060373 -.0983282
|
c.ytdeatht07#c.ytdeatht07 | -.0349326 .0228179 -1.53 0.126 -.079655 .0097897
|
c.tvytdeath7#c.ytdeatht07 | .0846717 .0309499 2.74 0.006 .024011 .1453323
|
_cons | 9.749435 .3681891 26.48 0.000 9.027797 10.47107
-------------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(tvytde~7) | .1168107 .0378687 .0618775 .2205123
var(_cons) | 9.702119 1.366385 7.36188 12.78629
cov(tvytde~7,_cons) | -.0575358 .1559837 -.3632582 .2481865
-----------------------------+------------------------------------------------
var(Residual) | 3.998746 .3578654 3.355413 4.765424
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 239.50 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1410.088 10 2840.176 2873.503
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. * Multivariate Test of Death Cohort Contextual Effects
. test (c.ytdeatht07=0) (c.ytdeatht07#c.ytdeatht07=0) (c.tvytdeath7#c.ytdeatht07=0)
( 1) [recall]ytdeatht07 = 0
( 2) [recall]c.ytdeatht07#c.ytdeatht07 = 0
( 3) [recall]c.tvytdeath7#c.ytdeatht07 = 0
chi2( 3) = 8.76
Prob > chi2 = 0.0327
. * Contextual Linear Death Cohort on Intercept
. lincom c.ytdeatht07*1
( 1) [recall]ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.3521827 .12952 -2.72 0.007 -.6060373 -.0983282
------------------------------------------------------------------------------
. * Contextual Quadratic Death Cohort on Intercept
. lincom c.ytdeatht07#c.ytdeatht07*1
( 1) [recall]c.ytdeatht07#c.ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.0349326 .0228179 -1.53 0.126 -.079655 .0097897
------------------------------------------------------------------------------
. * Contextual Linear Death Cohort on Linear Slope
. lincom c.tvytdeath7#c.ytdeatht07*1
( 1) [recall]c.tvytdeath7#c.ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | .0846717 .0309499 2.74 0.006 .024011 .1453323
------------------------------------------------------------------------------
. * Total Linear Death Cohort on Intercept
. lincom c.ytdeatht07*1 + c.tvytdeath7*1
( 1) [recall]tvytdeath7 + [recall]ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.1388638 .0731693 -1.90 0.058 -.282273 .0045454
------------------------------------------------------------------------------
. * Total Quadratic Death Cohort on Intercept
. lincom c.ytdeatht07#c.ytdeatht07*1 + c.tvytdeath7#c.ytdeatht07*1 + c.tvytdeath7#c.tvytdeath7*1
( 1) [recall]c.tvytdeath7#c.tvytdeath7 + [recall]c.ytdeatht07#c.ytdeatht07 + [recall]c.tvytdeath7#c.ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.0066422 .0176295 -0.38 0.706 -.0411954 .027911
------------------------------------------------------------------------------
. * Total Linear Death Cohort on Linear Slope
. lincom c.tvytdeath7#c.ytdeatht07*1 + c.tvytdeath7#c.tvytdeath7*2
( 1) 2*[recall]c.tvytdeath7#c.tvytdeath7 + [recall]c.tvytdeath7#c.ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.0280908 .0181876 -1.54 0.122 -.0637379 .0075563
------------------------------------------------------------------------------
. estimates store FitCohYTD,
. lrtest FitCohYTD FitFQRLYTD,
Likelihood-ratio test LR chi2(3) = 8.68
(Assumption: FitFQRLYTD nested in FitCohYTD) Prob > chi2 = 0.0338
. predict PredCohYTD, xb,
. margins, at (c.tvytdeath7=(-4(2)6) c.ytdeatht07=(-4(4)4)) vsquish,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
1._at : tvytdeath7 = -4
ytdeatht07 = -4
2._at : tvytdeath7 = -4
ytdeatht07 = 0
3._at : tvytdeath7 = -4
ytdeatht07 = 4
4._at : tvytdeath7 = -2
ytdeatht07 = -4
5._at : tvytdeath7 = -2
ytdeatht07 = 0
6._at : tvytdeath7 = -2
ytdeatht07 = 4
7._at : tvytdeath7 = 0
ytdeatht07 = -4
8._at : tvytdeath7 = 0
ytdeatht07 = 0
9._at : tvytdeath7 = 0
ytdeatht07 = 4
10._at : tvytdeath7 = 2
ytdeatht07 = -4
11._at : tvytdeath7 = 2
ytdeatht07 = 0
12._at : tvytdeath7 = 2
ytdeatht07 = 4
13._at : tvytdeath7 = 4
ytdeatht07 = -4
14._at : tvytdeath7 = 4
ytdeatht07 = 0
15._at : tvytdeath7 = 4
ytdeatht07 = 4
16._at : tvytdeath7 = 6
ytdeatht07 = -4
17._at : tvytdeath7 = 6
ytdeatht07 = 0
18._at : tvytdeath7 = 6
ytdeatht07 = 4
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 10.19861 .3964452 25.73 0.000 9.421596 10.97563
2 | 7.994059 .8406556 9.51 0.000 6.346404 9.641714
3 | 4.671659 1.949581 2.40 0.017 .8505514 8.492767
4 | 10.62445 .3541182 30.00 0.000 9.930395 11.31851
5 | 9.097272 .5170178 17.60 0.000 8.083936 10.11061
6 | 6.452245 1.355488 4.76 0.000 3.795537 9.108954
7 | 10.59924 .3564332 29.74 0.000 9.900647 11.29784
8 | 9.749435 .3681891 26.48 0.000 9.027797 10.47107
9 | 7.781781 .8889718 8.75 0.000 6.039429 9.524134
10 | 10.12298 .3663958 27.63 0.000 9.40486 10.84111
11 | 9.950548 .3540304 28.11 0.000 9.256661 10.64443
12 | 8.660268 .5614494 15.42 0.000 7.559847 9.760688
13 | 9.195673 .4660661 19.73 0.000 8.2822 10.10915
14 | 9.70061 .3854485 25.17 0.000 8.945145 10.45608
15 | 9.087704 .4072816 22.31 0.000 8.289447 9.885961
16 | 7.817312 .7377306 10.60 0.000 6.371387 9.263238
17 | 8.999624 .4639768 19.40 0.000 8.090246 9.909001
18 | 9.06409 .4404178 20.58 0.000 8.200887 9.927293
------------------------------------------------------------------------------
. corr recall PredCohYTD
(obs=557)
| recall PredCo~D
-------------+------------------
recall | 1.0000
PredCohYTD | 0.1514 1.0000
.
. display as result "Eq 10a.2: Fixed Quadratic, Random Linear Model using Years in Study"
Eq 10a.2: Fixed Quadratic, Random Linear Model using Years in Study
. display as result "Controlling for Death Cohort"
Controlling for Death Cohort
. mixed recall c.time c.time#c.time ///
> c.ytdeatht07 c.ytdeatht07#c.ytdeatht07 c.time#c.ytdeatht07, ///
> || personid: time, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1409.7677
Iteration 1: log likelihood = -1409.5484
Iteration 2: log likelihood = -1409.5478
Iteration 3: log likelihood = -1409.5478
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(5) = 19.78
Log likelihood = -1409.5478 Prob > chi2 = 0.0014
-------------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------------+----------------------------------------------------------------
time | .2357403 .1125962 2.09 0.036 .0150559 .4564247
|
c.time#c.time | -.0555024 .0158575 -3.50 0.000 -.0865825 -.0244224
|
ytdeatht07 | -.1468651 .0719932 -2.04 0.041 -.2879693 -.005761
|
c.ytdeatht07#c.ytdeatht07 | -.0076929 .0170878 -0.45 0.653 -.0411844 .0257986
|
c.time#c.ytdeatht07 | -.0232704 .0182938 -1.27 0.203 -.0591256 .0125848
|
_cons | 9.746899 .3929606 24.80 0.000 8.97671 10.51709
-------------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(time) | .1319957 .0439309 .0687484 .2534296
var(_cons) | 12.71236 1.613316 9.9129 16.3024
cov(time,_cons) | -.6415021 .2296561 -1.09162 -.1913844
-----------------------------+------------------------------------------------
var(Residual) | 3.9183 .3566535 3.278077 4.683561
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 240.58 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1409.548 10 2839.096 2872.423
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. * Multivariate Test of Death Cohort Total Effects
. test (c.ytdeatht07=0) (c.ytdeatht07#c.ytdeatht07=0) (c.time#c.ytdeatht07=0)
( 1) [recall]ytdeatht07 = 0
( 2) [recall]c.ytdeatht07#c.ytdeatht07 = 0
( 3) [recall]c.time#c.ytdeatht07 = 0
chi2( 3) = 8.82
Prob > chi2 = 0.0317
. * Contextual Linear Birth Cohort on Intercept
. lincom c.ytdeatht07*1 + c.time*-1
( 1) - [recall]time + [recall]ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.3826054 .1296397 -2.95 0.003 -.6366945 -.1285163
------------------------------------------------------------------------------
. * Contextual Quadratic Birth Cohort on Intercept
. lincom c.ytdeatht07#c.ytdeatht07*1 + c.time#c.ytdeatht07*-1 + c.time#c.time*1
( 1) [recall]c.time#c.time + [recall]c.ytdeatht07#c.ytdeatht07 - [recall]c.time#c.ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.0399249 .0227007 -1.76 0.079 -.0844175 .0045677
------------------------------------------------------------------------------
. * Contextual Linear Birth Cohort on Linear Slope
. lincom c.time#c.ytdeatht07*1 + c.time#c.time*-2
( 1) - 2*[recall]c.time#c.time + [recall]c.time#c.ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | .0877345 .0307001 2.86 0.004 .0275634 .1479056
------------------------------------------------------------------------------
. * Total Linear Birth Cohort on Intercept
. lincom c.ytdeatht07*1
( 1) [recall]ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.1468651 .0719932 -2.04 0.041 -.2879693 -.005761
------------------------------------------------------------------------------
. * Total Quadratic Birth Cohort on Intercept
. lincom c.ytdeatht07#c.ytdeatht07*1
( 1) [recall]c.ytdeatht07#c.ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.0076929 .0170878 -0.45 0.653 -.0411844 .0257986
------------------------------------------------------------------------------
. * Total Linear Birth Cohort on Linear Slope
. lincom c.time#c.ytdeatht07*1
( 1) [recall]c.time#c.ytdeatht07 = 0
------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | -.0232704 .0182938 -1.27 0.203 -.0591256 .0125848
------------------------------------------------------------------------------
. estimates store FitCohYTDYIS,
. lrtest FitCohYTDYIS FitFQRLYIS,
Likelihood-ratio test LR chi2(3) = 8.74
(Assumption: FitFQRLYIS nested in FitCohYTDYIS) Prob > chi2 = 0.0330
. predict PredCohYTDYIS, xb,
. margins, at (c.time=(0(2)8) c.ytdeatht07=(-4(4)0)) vsquish,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
1._at : time = 0
ytdeatht07 = -4
2._at : time = 0
ytdeatht07 = 0
3._at : time = 2
ytdeatht07 = -4
4._at : time = 2
ytdeatht07 = 0
5._at : time = 4
ytdeatht07 = -4
6._at : time = 4
ytdeatht07 = 0
7._at : time = 6
ytdeatht07 = -4
8._at : time = 6
ytdeatht07 = 0
9._at : time = 8
ytdeatht07 = -4
10._at : time = 8
ytdeatht07 = 0
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 10.21127 .3948244 25.86 0.000 9.437432 10.98511
2 | 9.746899 .3929606 24.80 0.000 8.97671 10.51709
3 | 10.64691 .3530452 30.16 0.000 9.954951 11.33886
4 | 9.996369 .3574497 27.97 0.000 9.295781 10.69696
5 | 10.63852 .3553738 29.94 0.000 9.942002 11.33504
6 | 9.801821 .3663432 26.76 0.000 9.083801 10.51984
7 | 10.18612 .3652987 27.88 0.000 9.470144 10.90209
8 | 9.163252 .4295875 21.33 0.000 8.321276 10.00523
9 | 9.289691 .4639689 20.02 0.000 8.380329 10.19905
10 | 8.080664 .6214187 13.00 0.000 6.862706 9.298623
------------------------------------------------------------------------------
. corr recall PredCohYTDYIS
(obs=557)
| recall Pre~DYIS
-------------+------------------
recall | 1.0000
PredCohYTD~S | 0.1559 1.0000
.
. display as result "Eq 10a.4: Fixed Quadratic, Random Linear Model using Years in Study"
Eq 10a.4: Fixed Quadratic, Random Linear Model using Years in Study
. display as result "Controlling for Birth Cohort and Death Cohort"
Controlling for Birth Cohort and Death Cohort
. mixed recall c.time c.time#c.time ///
> c.aget084 c.aget084#c.aget084 c.time#c.aget084 ///
> c.ytdeatht07 c.ytdeatht07#c.ytdeatht07 c.time#c.ytdeatht07, ///
> || personid: time, variance mle covariance(unstructured),
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: log likelihood = -1405.5359
Iteration 1: log likelihood = -1405.3023
Iteration 2: log likelihood = -1405.3017
Iteration 3: log likelihood = -1405.3017
Computing standard errors:
Mixed-effects ML regression Number of obs = 557
Group variable: personid Number of groups = 207
Obs per group: min = 1
avg = 2.7
max = 5
Wald chi2(8) = 28.77
Log likelihood = -1405.3017 Prob > chi2 = 0.0003
-------------------------------------------------------------------------------------------
recall | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------------+----------------------------------------------------------------
time | .2685799 .1132543 2.37 0.018 .0466057 .4905542
|
c.time#c.time | -.0540424 .0158245 -3.42 0.001 -.0850578 -.0230269
|
aget084 | -.2782489 .1054171 -2.64 0.008 -.4848626 -.0716351
|
c.aget084#c.aget084 | .0109324 .0182364 0.60 0.549 -.0248104 .0466752
|
c.time#c.aget084 | .0463432 .0207713 2.23 0.026 .0056322 .0870542
|
ytdeatht07 | -.1231322 .0716018 -1.72 0.085 -.2634691 .0172047
|
c.ytdeatht07#c.ytdeatht07 | -.0104216 .0169713 -0.61 0.539 -.0436848 .0228416
|
c.time#c.ytdeatht07 | -.0279886 .0182238 -1.54 0.125 -.0637067 .0077294
|
_cons | 9.493169 .4270437 22.23 0.000 8.656178 10.33016
-------------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
personid: Unstructured |
var(time) | .1256779 .042687 .064587 .244553
var(_cons) | 12.24373 1.566809 9.527673 15.73405
cov(time,_cons) | -.5892998 .2222487 -1.024899 -.1537004
-----------------------------+------------------------------------------------
var(Residual) | 3.896632 .3547448 3.259846 4.657811
------------------------------------------------------------------------------
LR test vs. linear regression: chi2(3) = 238.05 Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference.
. estat ic, n(207),
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 207 . -1405.302 13 2836.603 2879.929
-----------------------------------------------------------------------------
Note: N=207 used in calculating BIC
. predict PredBothYIS, xb,
. margins, at (c.time=(0(2)8) c.aget084=(-4(4)4) c.ytdeatht07=(-4(4)4)) vsquish,
Adjusted predictions Number of obs = 557
Expression : Linear prediction, fixed portion, predict()
1._at : time = 0
aget084 = -4
ytdeatht07 = -4
2._at : time = 0
aget084 = -4
ytdeatht07 = 0
3._at : time = 0
aget084 = -4
ytdeatht07 = 4
4._at : time = 0
aget084 = 0
ytdeatht07 = -4
5._at : time = 0
aget084 = 0
ytdeatht07 = 0
6._at : time = 0
aget084 = 0
ytdeatht07 = 4
7._at : time = 0
aget084 = 4
ytdeatht07 = -4
8._at : time = 0
aget084 = 4
ytdeatht07 = 0
9._at : time = 0
aget084 = 4
ytdeatht07 = 4
10._at : time = 2
aget084 = -4
ytdeatht07 = -4
11._at : time = 2
aget084 = -4
ytdeatht07 = 0
12._at : time = 2
aget084 = -4
ytdeatht07 = 4
13._at : time = 2
aget084 = 0
ytdeatht07 = -4
14._at : time = 2
aget084 = 0
ytdeatht07 = 0
15._at : time = 2
aget084 = 0
ytdeatht07 = 4
16._at : time = 2
aget084 = 4
ytdeatht07 = -4
17._at : time = 2
aget084 = 4
ytdeatht07 = 0
18._at : time = 2
aget084 = 4
ytdeatht07 = 4
19._at : time = 4
aget084 = -4
ytdeatht07 = -4
20._at : time = 4
aget084 = -4
ytdeatht07 = 0
21._at : time = 4
aget084 = -4
ytdeatht07 = 4
22._at : time = 4
aget084 = 0
ytdeatht07 = -4
23._at : time = 4
aget084 = 0
ytdeatht07 = 0
24._at : time = 4
aget084 = 0
ytdeatht07 = 4
25._at : time = 4
aget084 = 4
ytdeatht07 = -4
26._at : time = 4
aget084 = 4
ytdeatht07 = 0
27._at : time = 4
aget084 = 4
ytdeatht07 = 4
28._at : time = 6
aget084 = -4
ytdeatht07 = -4
29._at : time = 6
aget084 = -4
ytdeatht07 = 0
30._at : time = 6
aget084 = -4
ytdeatht07 = 4
31._at : time = 6
aget084 = 0
ytdeatht07 = -4
32._at : time = 6
aget084 = 0
ytdeatht07 = 0
33._at : time = 6
aget084 = 0
ytdeatht07 = 4
34._at : time = 6
aget084 = 4
ytdeatht07 = -4
35._at : time = 6
aget084 = 4
ytdeatht07 = 0
36._at : time = 6
aget084 = 4
ytdeatht07 = 4
37._at : time = 8
aget084 = -4
ytdeatht07 = -4
38._at : time = 8
aget084 = -4
ytdeatht07 = 0
39._at : time = 8
aget084 = -4
ytdeatht07 = 4
40._at : time = 8
aget084 = 0
ytdeatht07 = -4
41._at : time = 8
aget084 = 0
ytdeatht07 = 0
42._at : time = 8
aget084 = 0
ytdeatht07 = 4
43._at : time = 8
aget084 = 4
ytdeatht07 = -4
44._at : time = 8
aget084 = 4
ytdeatht07 = 0
45._at : time = 8
aget084 = 4
ytdeatht07 = 4
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 11.10686 .5547255 20.02 0.000 10.01962 12.19411
2 | 10.78108 .5898494 18.28 0.000 9.624999 11.93717
3 | 10.12181 .6015213 16.83 0.000 8.942847 11.30077
4 | 9.818951 .4445972 22.09 0.000 8.947557 10.69035
5 | 9.493169 .4270437 22.23 0.000 8.656178 10.33016
6 | 8.833894 .4589082 19.25 0.000 7.93445 9.733337
7 | 8.880874 .6276407 14.15 0.000 7.650721 10.11103
8 | 8.555092 .5854682 14.61 0.000 7.407595 9.702588
9 | 7.895817 .5865837 13.46 0.000 6.746134 9.0455
10 | 11.28102 .5080971 22.20 0.000 10.28517 12.27687
11 | 10.73133 .5421923 19.79 0.000 9.66865 11.794
12 | 9.848143 .5738335 17.16 0.000 8.72345 10.97284
13 | 10.36385 .405629 25.55 0.000 9.568833 11.15887
14 | 9.814159 .3960594 24.78 0.000 9.037897 10.59042
15 | 8.930975 .4646295 19.22 0.000 8.020318 9.841632
16 | 9.796519 .5610477 17.46 0.000 8.696886 10.89615
17 | 9.246828 .5272197 17.54 0.000 8.213496 10.28016
18 | 8.363644 .560475 14.92 0.000 7.265133 9.462154
19 | 11.02283 .5162295 21.35 0.000 10.01104 12.03462
20 | 10.24923 .5522612 18.56 0.000 9.166821 11.33165
21 | 9.14214 .6796063 13.45 0.000 7.810136 10.47414
22 | 10.47641 .4112711 25.47 0.000 9.670335 11.28249
23 | 9.702811 .4085048 23.75 0.000 8.902156 10.50347
24 | 8.595718 .5895146 14.58 0.000 7.44029 9.751145
25 | 10.27983 .6044298 17.01 0.000 9.095165 11.46449
26 | 9.506225 .5756186 16.51 0.000 8.378033 10.63442
27 | 8.399132 .6956384 12.07 0.000 7.035705 9.762558
28 | 10.33231 .5485454 18.84 0.000 9.25718 11.40744
29 | 9.3348 .6216098 15.02 0.000 8.116467 10.55313
30 | 8.003798 .8956613 8.94 0.000 6.248334 9.759262
31 | 10.15663 .4319735 23.51 0.000 9.30998 11.00329
32 | 9.159123 .4740658 19.32 0.000 8.229971 10.08828
33 | 7.828121 .8093797 9.67 0.000 6.241766 9.414476
34 | 10.33079 .7263875 14.22 0.000 8.907099 11.75449
35 | 9.333283 .7225995 12.92 0.000 7.917014 10.74955
36 | 8.002281 .9538848 8.39 0.000 6.132701 9.87186
37 | 9.209447 .652488 14.11 0.000 7.930594 10.4883
38 | 7.988028 .7998544 9.99 0.000 6.420342 9.555714
39 | 6.433117 1.220295 5.27 0.000 4.041383 8.824851
40 | 9.404516 .5368172 17.52 0.000 8.352373 10.45666
41 | 8.183097 .6625265 12.35 0.000 6.884569 9.481625
42 | 6.628185 1.13327 5.85 0.000 4.407017 8.849354
43 | 9.949421 .9380123 10.61 0.000 8.11095 11.78789
44 | 8.728002 .9827554 8.88 0.000 6.801837 10.65417
45 | 7.173091 1.320877 5.43 0.000 4.58422 9.761962
------------------------------------------------------------------------------
. corr recall PredBothYIS
(obs=557)
| recall PredBo~S
-------------+------------------
recall | 1.0000
PredBothYIS | 0.2084 1.0000
.
. ****** END CHAPTER 10a MODELS ******
.
. * Close log
. log close STATA_Chapter10a
name: STATA_Chapter10a
log: C:\Dropbox\PilesOfVariance\Chapter10a\STATA\STATA_Chapter10a_Output.smcl
log type: smcl
closed on: 27 Jan 2015, 17:03:30
------------------------------------------------------------------------------------------------------------------------------------------------------