import os
os.chdir('/Applications/Stata 17/utilities')
from pystata import config
config.init('mp')Chapter 1: Review of linear regression
%%stata
net from http://www.stata-press.com/data/mlmus3/.
. net from http://www.stata-press.com/data/mlmus3/
-------------------------------------------------------------------------------
http://www.stata-press.com/data/mlmus3/
Files associated with text by Sophia Rabe-Hesketh and Anders Skrondal
-------------------------------------------------------------------------------
PACKAGES you could -net describe-:
mlmus3_vol1 Files for Multilevel and Longitudinal Modeling Using
Stata, Third Edition, Volume 1
mlmus3_vol2 Files for Multilevel and Longitudinal Modeling Using
Stata, Third Edition, Volume 2
-------------------------------------------------------------------------------
. %%stata
net describe mlmus3_vol1.
. net describe mlmus3_vol1
-------------------------------------------------------------------------------
package mlmus3_vol1 from http://www.stata-press.com/data/mlmus3
-------------------------------------------------------------------------------
TITLE
do-files, data, datasets
DESCRIPTION/AUTHOR(S)
Sophia Rabe-Hesketh, Anders Skrondal
These are the files used to produce the examples in the book
Multilevel and Longitudinal Modeling Using Stata, Third Edition
ANCILLARY FILES (type net get mlmus3_vol1)
achievement.dta
airlines.dta
alcuse.dta
anorexia.dta
antisocial.dta
army.dta
asian.dta
attend.dta
birthwt.dta
bp.dta
calcium.dta
ch1.do
ch2.do
ch3.do
ch4.do
ch5.do
ch6.do
ch7.do
ch8.do
ch9.do
cigar.dta
cogstyle.dta
coursework.dta
cropareas.dta
dairy.dta
data.cfk
dialyzer.dta
expectancy.dta
ezunem.dta
faculty.dta
family.dta
fat.dta
fife.dta
fringe.dta
garrett.dta
gcse.dta
gpa.dta
grader1.dta
growth.dta
grunfeld.dta
headsize.dta
homework.dta
hours.dta
hsb.dta
instruction.dta
kenya.dta
neighborhood.dta
nitrogen.dta
ols.dta
olympics.dta
papke_did.dta
part3intro.do
pefr.dta
postnatal.dta
productivity.dta
pups.dta
reading.dta
returns.dta
schooleffects.dta
sex.dta
smoking.dta
star1.dta
star_mm.dta
taxprep.dta
twin.dta
videos.dta
wagepan.dta
wheat.dta
-------------------------------------------------------------------------------
. %%stata
net get mlmus3_vol1.
. net get mlmus3_vol1
checking mlmus3_vol1 consistency and verifying not already installed...
copying into current directory...
copying achievement.dta
copying airlines.dta
copying alcuse.dta
copying anorexia.dta
copying antisocial.dta
copying army.dta
copying asian.dta
copying attend.dta
copying birthwt.dta
copying bp.dta
copying calcium.dta
copying ch1.do
copying ch2.do
copying ch3.do
copying ch4.do
copying ch5.do
copying ch6.do
copying ch7.do
copying ch8.do
copying ch9.do
copying cigar.dta
copying cogstyle.dta
copying coursework.dta
copying cropareas.dta
copying dairy.dta
copying data.cfk
copying dialyzer.dta
copying expectancy.dta
copying ezunem.dta
copying faculty.dta
copying family.dta
copying fat.dta
copying fife.dta
copying fringe.dta
copying garrett.dta
copying gcse.dta
copying gpa.dta
copying grader1.dta
copying growth.dta
copying grunfeld.dta
copying headsize.dta
copying homework.dta
copying hours.dta
copying hsb.dta
copying instruction.dta
copying kenya.dta
copying neighborhood.dta
copying nitrogen.dta
copying ols.dta
copying olympics.dta
copying papke_did.dta
copying part3intro.do
copying pefr.dta
copying postnatal.dta
copying productivity.dta
copying pups.dta
copying reading.dta
copying returns.dta
copying schooleffects.dta
copying sex.dta
copying smoking.dta
copying star1.dta
copying star_mm.dta
copying taxprep.dta
copying twin.dta
copying videos.dta
copying wagepan.dta
copying wheat.dta
ancillary files successfully copied.
. %%stata
use pefr.
. use pefr
. %%stata
describeContains data from pefr.dta
Observations: 17
Variables: 5 26 Nov 2004 15:45
-------------------------------------------------------------------------------
Variable Storage Display Value
name type format label Variable label
-------------------------------------------------------------------------------
id float %9.0g
wp1 float %9.0g
wp2 float %9.0g
wm1 float %9.0g
wm2 float %9.0g
-------------------------------------------------------------------------------
Sorted by: %%stata
use http://www.stata-press.com/data/mlmus3/faculty
use http://www.stata-press.com/data/mlmus3/faculty, clear.
. use http://www.stata-press.com/data/mlmus3/faculty
. use http://www.stata-press.com/data/mlmus3/faculty, clear
. %%stata
tabstat salary, by(male) statistics(mean sd n).
. tabstat salary, by(male) statistics(mean sd n)
Summary for variables: salary
Group variable: male
male | Mean SD N
-------+------------------------------
Women | 42916.6 9161.61 128
Men | 53499.24 12583.48 386
-------+------------------------------
Total | 50863.87 12672.77 514
--------------------------------------
. %%stata
graph box salary, over(male) ytitle(Academic salary) asyvars.
. graph box salary, over(male) ytitle(Academic salary) asyvars
. 
svg
%%stata
generate lsalary = log10(salary)
graph box lsalary, over(male) ytitle(Log academic salary) asyvars
histogram lsalary, by(male, rows(2)) xtitle(Log academic salary).
. generate lsalary = log10(salary)
. graph box lsalary, over(male) ytitle(Log academic salary) asyvars
. histogram lsalary, by(male, rows(2)) xtitle(Log academic salary)
. 
svg
%%stata
ttest salary, by(male)
ttest salary, by(male) unequal
anova salary male
margins male.
. ttest salary, by(male)
Two-sample t test with equal variances
------------------------------------------------------------------------------
Group | Obs Mean Std. err. Std. dev. [95% conf. interval]
---------+--------------------------------------------------------------------
Women | 128 42916.6 809.7795 9161.61 41314.2 44519.01
Men | 386 53499.24 640.4822 12583.48 52239.96 54758.52
---------+--------------------------------------------------------------------
Combined | 514 50863.87 558.972 12672.77 49765.72 51962.03
---------+--------------------------------------------------------------------
diff | -10582.63 1206.345 -12952.63 -8212.636
------------------------------------------------------------------------------
diff = mean(Women) - mean(Men) t = -8.7725
H0: diff = 0 Degrees of freedom = 512
Ha: diff < 0 Ha: diff != 0 Ha: diff > 0
Pr(T < t) = 0.0000 Pr(|T| > |t|) = 0.0000 Pr(T > t) = 1.0000
. ttest salary, by(male) unequal
Two-sample t test with unequal variances
------------------------------------------------------------------------------
Group | Obs Mean Std. err. Std. dev. [95% conf. interval]
---------+--------------------------------------------------------------------
Women | 128 42916.6 809.7795 9161.61 41314.2 44519.01
Men | 386 53499.24 640.4822 12583.48 52239.96 54758.52
---------+--------------------------------------------------------------------
Combined | 514 50863.87 558.972 12672.77 49765.72 51962.03
---------+--------------------------------------------------------------------
diff | -10582.63 1032.454 -12614.48 -8550.787
------------------------------------------------------------------------------
diff = mean(Women) - mean(Men) t = -10.2500
H0: diff = 0 Satterthwaite's degrees of freedom = 297.227
Ha: diff < 0 Ha: diff != 0 Ha: diff > 0
Pr(T < t) = 0.0000 Pr(|T| > |t|) = 0.0000 Pr(T > t) = 1.0000
. anova salary male
Number of obs = 514 R-squared = 0.1307
Root MSE = 11827.4 Adj R-squared = 0.1290
Source | Partial SS df MS F Prob>F
-----------+----------------------------------------------------
Model | 1.077e+10 1 1.077e+10 76.96 0.0000
|
male | 1.077e+10 1 1.077e+10 76.96 0.0000
|
Residual | 7.162e+10 512 1.399e+08
-----------+----------------------------------------------------
Total | 8.239e+10 513 1.606e+08
. margins male
Adjusted predictions Number of obs = 514
Expression: Linear prediction, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male |
Women | 42916.6 1045.403 41.05 0.000 40862.8 44970.41
Men | 53499.24 601.9981 88.87 0.000 52316.55 54681.93
------------------------------------------------------------------------------
. %%stata
reg salary market
reg lsalary market.
. reg salary market
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(1, 512) = 101.77
Model | 1.3661e+10 1 1.3661e+10 Prob > F = 0.0000
Residual | 6.8726e+10 512 134231433 R-squared = 0.1658
-------------+---------------------------------- Adj R-squared = 0.1642
Total | 8.2387e+10 513 160599133 Root MSE = 11586
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
market | 34545.22 3424.333 10.09 0.000 27817.75 41272.69
_cons | 18096.99 3288.009 5.50 0.000 11637.35 24556.64
------------------------------------------------------------------------------
. reg lsalary market
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(1, 512) = 107.58
Model | 1.04818106 1 1.04818106 Prob > F = 0.0000
Residual | 4.98852929 512 .009743221 R-squared = 0.1736
-------------+---------------------------------- Adj R-squared = 0.1720
Total | 6.03671035 513 .011767467 Root MSE = .09871
------------------------------------------------------------------------------
lsalary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
market | .3025987 .0291743 10.37 0.000 .2452827 .3599147
_cons | 4.405958 .0280128 157.28 0.000 4.350924 4.460992
------------------------------------------------------------------------------
. %%stata
egen mn_market = mean(market)
generate marketc = market - mn_market%%stata
regress salary market
regress salary marketc.
.
. regress salary market
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(1, 512) = 101.77
Model | 1.3661e+10 1 1.3661e+10 Prob > F = 0.0000
Residual | 6.8726e+10 512 134231433 R-squared = 0.1658
-------------+---------------------------------- Adj R-squared = 0.1642
Total | 8.2387e+10 513 160599133 Root MSE = 11586
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
market | 34545.22 3424.333 10.09 0.000 27817.75 41272.69
_cons | 18096.99 3288.009 5.50 0.000 11637.35 24556.64
------------------------------------------------------------------------------
.
. regress salary marketc
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(1, 512) = 101.77
Model | 1.3661e+10 1 1.3661e+10 Prob > F = 0.0000
Residual | 6.8726e+10 512 134231433 R-squared = 0.1658
-------------+---------------------------------- Adj R-squared = 0.1642
Total | 8.2387e+10 513 160599133 Root MSE = 11586
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
marketc | 34545.22 3424.333 10.09 0.000 27817.75 41272.69
_cons | 50863.87 511.029 99.53 0.000 49859.9 51867.85
------------------------------------------------------------------------------
. Konstantanya saja yang berubah
%%stata
predict yhat, xb%%stata
twoway (scatter salary market) (line yhat market), ytitle(Academic salary) xtitle(Marketability)
svg
%%stata
regress salary male
regress salary male, vce(robust)
regress salary male, robust. regress salary male
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(1, 512) = 76.96
Model | 1.0765e+10 1 1.0765e+10 Prob > F = 0.0000
Residual | 7.1622e+10 512 139887048 R-squared = 0.1307
-------------+---------------------------------- Adj R-squared = 0.1290
Total | 8.2387e+10 513 160599133 Root MSE = 11827
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male | 10582.63 1206.345 8.77 0.000 8212.636 12952.63
_cons | 42916.6 1045.403 41.05 0.000 40862.8 44970.41
------------------------------------------------------------------------------
. regress salary male, vce(robust)
Linear regression Number of obs = 514
F(1, 512) = 105.26
Prob > F = 0.0000
R-squared = 0.1307
Root MSE = 11827
------------------------------------------------------------------------------
| Robust
salary | Coefficient std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male | 10582.63 1031.462 10.26 0.000 8556.213 12609.05
_cons | 42916.6 808.184 53.10 0.000 41328.84 44504.37
------------------------------------------------------------------------------
. regress salary male, robust
Linear regression Number of obs = 514
F(1, 512) = 105.26
Prob > F = 0.0000
R-squared = 0.1307
Root MSE = 11827
------------------------------------------------------------------------------
| Robust
salary | Coefficient std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male | 10582.63 1031.462 10.26 0.000 8556.213 12609.05
_cons | 42916.6 808.184 53.10 0.000 41328.84 44504.37
------------------------------------------------------------------------------
. 1.7 Multiple linear regression
%%stata
tabstat marketc, by(male) statistics(mean sd)Summary for variables: marketc
Group variable: male
male | Mean SD
-------+--------------------
Women | -.0469589 .1314393
Men | .0155718 .1518486
-------+--------------------
Total | -2.96e-08 .14938
----------------------------%%stata
regress salary male marketc
predict yhat2, xb. regress salary male marketc
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(2, 511) = 85.80
Model | 2.0711e+10 2 1.0356e+10 Prob > F = 0.0000
Residual | 6.1676e+10 511 120696838 R-squared = 0.2514
-------------+---------------------------------- Adj R-squared = 0.2485
Total | 8.2387e+10 513 160599133 Root MSE = 10986
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male | 8708.423 1139.411 7.64 0.000 6469.917 10946.93
marketc | 29972.6 3301.766 9.08 0.000 23485.89 36459.3
_cons | 44324.09 983.3533 45.07 0.000 42392.17 46256
------------------------------------------------------------------------------
. predict yhat2, xb
. %%stata
twoway (scatter salary marketc if male==1, msymbol(o)) (line yhat2 marketc if male==1, sort lpatt(dash)) (scatter salary marketc if male==0, msymbol(oh))
(line yhat2 marketc if male==0, sort lpatt(solid)),
ytitle(Academic salary)
xtitle(Mean-centered marketability)
legend(order(1 " " 2 "Men" 3 " " 4 "Women")).
. twoway (scatter salary marketc if male==1, msymbol(o)) (line yhat2 marketc if
> male==1, sort lpatt(dash)) (scatter salary marketc if male==0, msymbol(oh))
Exception in thread Stata:
Traceback (most recent call last):
File "/opt/anaconda3/lib/python3.11/threading.py", line 1045, in _bootstrap_inner
self.run()
File "/Applications/Stata 17/utilities/pystata/core/stout.py", line 176, in run
raise SystemError(output)
SystemError:
. (line yhat2 marketc if male==0, sort lpatt(solid)),
( is not a valid command name
r(199);
r(199);
svg
%%stata
regress salary male marketc yearsdg
ovtest
scatter salary marketc.
. regress salary male marketc yearsdg
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(3, 510) = 367.56
Model | 5.6333e+10 3 1.8778e+10 Prob > F = 0.0000
Residual | 2.6054e+10 510 51087083.4 R-squared = 0.6838
-------------+---------------------------------- Adj R-squared = 0.6819
Total | 8.2387e+10 513 160599133 Root MSE = 7147.5
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male | 2040.211 783.122 2.61 0.009 501.6684 3578.753
marketc | 38402.39 2171.689 17.68 0.000 34135.83 42668.95
yearsdg | 949.2583 35.94867 26.41 0.000 878.6326 1019.884
_cons | 34834.3 733.7898 47.47 0.000 33392.68 36275.93
------------------------------------------------------------------------------
. ovtest
Ramsey RESET test for omitted variables
Omitted: Powers of fitted values of salary
H0: Model has no omitted variables
F(3, 507) = 1.50
Prob > F = 0.2135
.
. scatter salary marketc
. 
svg
%%stata
regress salary i.male marketc yearsdg
margins male, at(marketc=0 yearsdg=10)
margins male, at(marketc=0 yearsdg=0)
display 2040.211+34834.3.
. regress salary i.male marketc yearsdg
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(3, 510) = 367.56
Model | 5.6333e+10 3 1.8778e+10 Prob > F = 0.0000
Residual | 2.6054e+10 510 51087083.4 R-squared = 0.6838
-------------+---------------------------------- Adj R-squared = 0.6819
Total | 8.2387e+10 513 160599133 Root MSE = 7147.5
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male |
Men | 2040.211 783.122 2.61 0.009 501.6684 3578.753
marketc | 38402.39 2171.689 17.68 0.000 34135.83 42668.95
yearsdg | 949.2583 35.94867 26.41 0.000 878.6326 1019.884
_cons | 34834.3 733.7898 47.47 0.000 33392.68 36275.93
------------------------------------------------------------------------------
. margins male, at(marketc=0 yearsdg=10)
Adjusted predictions Number of obs = 514
Model VCE: OLS
Expression: Linear prediction, predict()
At: marketc = 0
yearsdg = 10
------------------------------------------------------------------------------
| Delta-method
| Margin std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male |
Women | 44326.89 639.7602 69.29 0.000 43070 45583.78
Men | 46367.1 444.0554 104.42 0.000 45494.7 47239.5
------------------------------------------------------------------------------
. margins male, at(marketc=0 yearsdg=0)
Adjusted predictions Number of obs = 514
Model VCE: OLS
Expression: Linear prediction, predict()
At: marketc = 0
yearsdg = 0
------------------------------------------------------------------------------
| Delta-method
| Margin std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male |
Women | 34834.3 733.7898 47.47 0.000 33392.68 36275.93
Men | 36874.52 712.6708 51.74 0.000 35474.38 38274.65
------------------------------------------------------------------------------
.
. display 2040.211+34834.3
36874.511
. %%stata
generate male_years = male*yearsdg.
. generate male_years = male*yearsdg
. %%stata
regress salary male marketc yearsdg male_years, robustLinear regression Number of obs = 514
F(4, 509) = 271.67
Prob > F = 0.0000
R-squared = 0.6875
Root MSE = 7112.1
------------------------------------------------------------------------------
| Robust
salary | Coefficient std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male | -593.3088 1014.158 -0.59 0.559 -2585.759 1399.142
marketc | 38436.65 2237.322 17.18 0.000 34041.12 42832.17
yearsdg | 763.1896 64.1736 11.89 0.000 637.1118 889.2673
male_years | 227.1532 77.61123 2.93 0.004 74.67539 379.631
_cons | 36773.64 735.2791 50.01 0.000 35329.09 38218.2
------------------------------------------------------------------------------%%stata
lincom male + male_years*10 ( 1) male + 10*male_years = 0
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
(1) | 1678.223 643.1395 2.61 0.009 414.6882 2941.758
------------------------------------------------------------------------------%%stata
twoway (function Women = 36773 + 763.19*x, range(0 41) lpatt(dash)) ///
(function Men = 36773 + -593.31 + (763.19 + 227.15)*x, range(0 41) lpatt(solid)), ///
xtitle("Time since degree (years)") ytitle("Mean salary"). twoway (function Women = 36773 + 763.19*x, range(0 41) lpatt(dash)) ///
> (function Men = 36773 + -593.31 + (763.19 + 227.15)*x, range(0 41) lpa
> tt(solid)), ///
> xtitle("Time since degree (years)") ytitle("Mean salary")
. 
svg
%%stata
twoway (function Women = _b[_cons] + _b[yearsdg]*x, range(0 41) lpatt(dash)) ///
(function Men = _b[_cons] + _b[male] + (_b[yearsdg] + _b[male_years])*x, ///
range(0 41) lpatt(solid)), xtitle("Time since degree (years)") ytitle("Mean salary").
. twoway (function Women = _b[_cons] + _b[yearsdg]*x, range(0 41) lpatt(dash))
> ///
> (function Men = _b[_cons] + _b[male] + (_b[yearsdg] + _b[male_years])*
> x, ///
> range(0 41) lpatt(solid)), xtitle("Time since degree (years)") ytitle(
> "Mean salary")
. 
svg
%%stata
regress salary male marketc yearsdg i.male#c.yearsdg.
. regress salary male marketc yearsdg i.male#c.yearsdg
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(4, 509) = 279.95
Model | 5.6641e+10 4 1.4160e+10 Prob > F = 0.0000
Residual | 2.5746e+10 509 50581607.4 R-squared = 0.6875
-------------+---------------------------------- Adj R-squared = 0.6850
Total | 8.2387e+10 513 160599133 Root MSE = 7112.1
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male | -593.3088 1320.911 -0.45 0.654 -3188.418 2001.8
marketc | 38436.65 2160.963 17.79 0.000 34191.14 42682.15
yearsdg | 763.1896 83.4169 9.15 0.000 599.3057 927.0734
|
male#|
c.yearsdg |
Men | 227.1532 91.99749 2.47 0.014 46.41164 407.8947
|
_cons | 36773.64 1072.395 34.29 0.000 34666.78 38880.51
------------------------------------------------------------------------------
. %%stata
regress salary male marketc yearsdg i.male##c.yearsdg
lincom 1.male + 1.male#c.yearsdg*10.
. regress salary male marketc yearsdg i.male##c.yearsdg
note: 1.male omitted because of collinearity.
note: yearsdg omitted because of collinearity.
Source | SS df MS Number of obs = 514
-------------+---------------------------------- F(4, 509) = 279.95
Model | 5.6641e+10 4 1.4160e+10 Prob > F = 0.0000
Residual | 2.5746e+10 509 50581607.4 R-squared = 0.6875
-------------+---------------------------------- Adj R-squared = 0.6850
Total | 8.2387e+10 513 160599133 Root MSE = 7112.1
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
male | -593.3088 1320.911 -0.45 0.654 -3188.418 2001.8
marketc | 38436.65 2160.963 17.79 0.000 34191.14 42682.15
yearsdg | 763.1896 83.4169 9.15 0.000 599.3057 927.0734
|
male |
Men | 0 (omitted)
yearsdg | 0 (omitted)
|
male#|
c.yearsdg |
Men | 227.1532 91.99749 2.47 0.014 46.41164 407.8947
|
_cons | 36773.64 1072.395 34.29 0.000 34666.78 38880.51
------------------------------------------------------------------------------
. lincom 1.male + 1.male#c.yearsdg*10
( 1) 1o.male + 10*1.male#c.yearsdg = 0
------------------------------------------------------------------------------
salary | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
(1) | 2271.532 919.9749 2.47 0.014 464.1164 4078.947
------------------------------------------------------------------------------
.