|
|
|
||||
|
|
|||||
NOTE: The code for this page was generated in Stata 11. The graphs from this paper have been replicated in form only. We present them using a demonstration dataset, NOT the datasets used to generate the graphs in the paper.
The code shown on this page includes several user-written commands. Before running the code, you should install:
For details on how to find and download Stata programs, see our FAQ page on the findit command. These examples use the hsbdemo.dta data file which you can obtain from within Stata using the code below.
use http://www.ats.ucla.edu/stat/data/hsbdemo, clear
* table 1
label define yesno 1 "yes" 0 "no"
label value honors yesno
tab honor female, col
+-------------------+
| Key |
|-------------------|
| frequency |
| column percentage |
+-------------------+
honors | female
english | male female | Total
-----------+----------------------+----------
no | 73 74 | 147
| 80.22 67.89 | 73.50
-----------+----------------------+----------
yes | 18 35 | 53
| 19.78 32.11 | 26.50
-----------+----------------------+----------
Total | 91 109 | 200
| 100.00 100.00 | 100.00
* figure 2 on page 758
set scheme lean1
bysort honor female: gen freq = _N
spineplot female honor, text(freq, mlabsize(*1.4)) ///
legend(off) yla(0.1 "Male" 0.9 "Female", noticks axis(2)) ///
ytitle("By Female", axis(2)) xtitle(" ", axis(2)) ///
barall(xscale(off) xlabel(, valuelabels) ///
title("By being in honors class"))
* table 2
sum write read math science
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
write | 200 52.775 9.478586 31 67
read | 200 52.23 10.25294 28 76
math | 200 52.645 9.368448 33 75
science | 200 51.85 9.900891 26 74
* figure 3
ciplot write read math science, hor
* table 3
sum write read math science female honors awards
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
write | 200 52.775 9.478586 31 67
read | 200 52.23 10.25294 28 76
math | 200 52.645 9.368448 33 75
science | 200 51.85 9.900891 26 74
female | 200 .545 .4992205 0 1
-------------+--------------------------------------------------------
honors | 200 .265 .4424407 0 1
awards | 200 1.67 1.818691 0 7
*figure 4, part 2
vioplot write read math science, hor*figure 4, part 1 replace schtyp = 0 if schtyp == 2 collapse (mean) female honors schtyp onewayplot female honors schtyp, title("Means of Binary Variables") xtitle("Means")
* table 4 /* The Stata Journal (2003) 3, Number 3, pp. 245–269 Confidence intervals and p-values for delivery to the end user */ use http://www.ats.ucla.edu/stat/data/hsbdemo, clear tab3way prog female ses, coltot Table entries are cell frequencies Missing categories ignored -------------------------------------------------------------- | ses and female type of | ----- low ---- --- middle --- ---- high ---- program | male female male female male female ----------+--------------------------------------------------- general | 7 9 10 10 4 5 academic | 4 15 22 22 21 21 vocation | 4 8 15 16 4 3 TOTAL | 15 32 47 48 29 29 -------------------------------------------------------------- gen one = 1 collapse (sum) one, by(prog female ses) egen total = sum(one), by(female ses) gen prop = one/total * figure 5 twoway (scatter prog prop if female==1) /// (scatter prog prop if female==0), by(ses, row(3)) /// legend(order(1 "female" 2 "male") row(1)) scheme(lean1) /// ylab(1(1) 3, valuelabels) graph display, ysize(6)
* table 5
use http://www.ats.ucla.edu/stat/data/hsbdemo, clear
regress write math female read
Source | SS df MS Number of obs = 200
-------------+------------------------------ F( 4, 195) = 54.38
Model | 9427.3364 4 2356.8341 Prob > F = 0.0000
Residual | 8451.5386 195 43.3412236 R-squared = 0.5273
-------------+------------------------------ Adj R-squared = 0.5176
Total | 17878.875 199 89.843593 Root MSE = 6.5834
------------------------------------------------------------------------------
write | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
math | .3922032 .0668992 5.86 0.000 .2602643 .5241422
female | 5.521369 .9425592 5.86 0.000 3.66245 7.380288
read | .3185602 .06153 5.18 0.000 .1972105 .4399099
ses | .4862292 .6826549 0.71 0.477 -.8601055 1.832564
_cons | 11.48071 2.925053 3.92 0.000 5.711913 17.24952
------------------------------------------------------------------------------
estimates store m1, title(Model 1)
regress write math female read ses
estimates store m2, title(Full Model)
estout m1 m2, cells(b(star fmt(%8.3f)) se(par fmt(%8.3f))) ///
stats(r2 N, fmt(%6.4f %9.0g)) varlabels(_cons Constant) ///
varwidth(8) modelwidth(10) legend collabels(, none) ///
posthead("") prefoot("") postfoot("")
------------------------------------
m1 m2
------------------------------------
math 0.397*** 0.392***
(0.066) (0.067)
female 5.443*** 5.521***
(0.935) (0.943)
read 0.325*** 0.319***
(0.061) (0.062)
ses 0.486
(0.683)
Constant 11.896*** 11.481***
(2.863) (2.925)
------------------------------------
r2 0.5261 0.5273
N 200 200
------------------------------------
* p<0.05, ** p<0.01, *** p<0.001
clear
capture eststo clear
use http://www.ats.ucla.edu/stat/data/hsbdemo, clear
eststo, title(Model 1): regress write math female read
Source | SS df MS Number of obs = 200
-------------+------------------------------ F( 3, 196) = 72.52
Model | 9405.34864 3 3135.11621 Prob > F = 0.0000
Residual | 8473.52636 196 43.2322773 R-squared = 0.5261
-------------+------------------------------ Adj R-squared = 0.5188
Total | 17878.875 199 89.843593 Root MSE = 6.5751
------------------------------------------------------------------------------
write | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
math | .3974826 .0664037 5.99 0.000 .266525 .5284401
female | 5.44337 .9349987 5.82 0.000 3.59942 7.287319
read | .3252389 .0607348 5.36 0.000 .2054613 .4450166
_cons | 11.89566 2.862845 4.16 0.000 6.249728 17.5416
------------------------------------------------------------------------------
(est1 stored)
eststo, title(Full Model): regress write math female read ses
Source | SS df MS Number of obs = 200
-------------+------------------------------ F( 4, 195) = 54.38
Model | 9427.3364 4 2356.8341 Prob > F = 0.0000
Residual | 8451.5386 195 43.3412236 R-squared = 0.5273
-------------+------------------------------ Adj R-squared = 0.5176
Total | 17878.875 199 89.843593 Root MSE = 6.5834
------------------------------------------------------------------------------
write | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
math | .3922032 .0668992 5.86 0.000 .2602643 .5241422
female | 5.521369 .9425592 5.86 0.000 3.66245 7.380288
read | .3185602 .06153 5.18 0.000 .1972105 .4399099
ses | .4862292 .6826549 0.71 0.477 -.8601055 1.832564
_cons | 11.48071 2.925053 3.92 0.000 5.711913 17.24952
------------------------------------------------------------------------------
(est2 stored)
esttab, label nodepvar nonumber
----------------------------------------------------
Model 1 Full Model
----------------------------------------------------
math score 0.397*** 0.392***
(5.99) (5.86)
female 5.443*** 5.521***
(5.82) (5.86)
reading score 0.325*** 0.319***
(5.36) (5.18)
ses 0.486
(0.71)
Constant 11.90*** 11.48***
(4.16) (3.92)
----------------------------------------------------
Observations 200 200
----------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001
* figure 6
regress write math female read
Source | SS df MS Number of obs = 200
-------------+------------------------------ F( 3, 196) = 72.52
Model | 9405.34864 3 3135.11621 Prob > F = 0.0000
Residual | 8473.52636 196 43.2322773 R-squared = 0.5261
-------------+------------------------------ Adj R-squared = 0.5188
Total | 17878.875 199 89.843593 Root MSE = 6.5751
------------------------------------------------------------------------------
write | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
math | .3974826 .0664037 5.99 0.000 .266525 .5284401
female | 5.44337 .9349987 5.82 0.000 3.59942 7.287319
read | .3252389 .0607348 5.36 0.000 .2054613 .4450166
_cons | 11.89566 2.862845 4.16 0.000 6.249728 17.5416
------------------------------------------------------------------------------
parmest, label list(parm estimate min* max* p) saving(mypars, replace)
+--------------------------------------------------------+
| parm estimate min95 max95 p |
|--------------------------------------------------------|
1. | math .39748258 .26652501 .52844015 1.009e-08 |
2. | female 5.4433699 3.5994204 7.2873194 2.348e-08 |
3. | read .32523894 .2054613 .44501657 2.381e-07 |
4. | _cons 11.895664 6.249728 17.541599 .00004852 |
+--------------------------------------------------------+
file mypars.dta saved
use mypars, clear
encode parm, gen(parmn)
eclplot estimate min95 max95 parmn if parm!="_cons" , hori ylabel(2(1) 4) ///
text(3.5 7 "R-squared = .53" "Adj R-squared = .52" "n = 200", justification(left))
* figure 7 use http://www.ats.ucla.edu/stat/data/hsbdemo, clear tempfile tf1 tf2 parmby "reg write math read if female==0", /// lab saving(‘tf1’,replace) idn(1) ids(male) Command: reg write math read if female==0 Source | SS df MS Number of obs = 91 -------------+------------------------------ F( 2, 88) = 45.09 Model | 4837.46939 2 2418.73469 Prob > F = 0.0000 Residual | 4720.20094 88 53.638647 R-squared = 0.5061 -------------+------------------------------ Adj R-squared = 0.4949 Total | 9557.67033 90 106.196337 Root MSE = 7.3238 ------------------------------------------------------------------------------ write | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- math | .3932125 .1006582 3.91 0.000 .1931754 .5932495 read | .4159189 .0925923 4.49 0.000 .2319113 .5999266 _cons | 7.331651 4.603424 1.59 0.115 -1.816687 16.47999 ------------------------------------------------------------------------------ file ‘tf1’.dta saved parmby "reg write math read if female==1", /// lab saving(‘tf2’,replace) idn(2) ids(female) Command: reg write math read if female==1 Source | SS df MS Number of obs = 109 -------------+------------------------------ F( 2, 106) = 52.10 Model | 3541.83282 2 1770.91641 Prob > F = 0.0000 Residual | 3603.15801 106 33.9920567 R-squared = 0.4957 -------------+------------------------------ Adj R-squared = 0.4862 Total | 7144.99083 108 66.1573225 Root MSE = 5.8303 ------------------------------------------------------------------------------ write | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- math | .419655 .087195 4.81 0.000 .2467825 .5925275 read | .2306052 .0793334 2.91 0.004 .073319 .3878914 _cons | 21.0731 3.370705 6.25 0.000 14.39034 27.75585 ------------------------------------------------------------------------------ file ‘tf2’.dta saved dsconcat ‘tf1’ ‘tf2’ sencode idstr, gene(modtype) sencode label, gene(predictor) lab var modtype "Gender" lab var predictor "Predictor" replace predictor = predictor + .1 if modtype ==2 drop if parm=="_cons" twoway (rcap min95 max95 predictor if modtype==1, hor) /// (scatter predictor estimate if modtype==1) /// (rcap min95 max95 predictor if modtype==2, hor) /// (scatter predictor estimate if modtype==2), /// yscale(range(0 3)) ylabel(1 2, valuelabel) /// legend(order(2 "Model 1" 4 "Model 2") row(1) ring(0)) scheme(lean1) /// xscale(range(-.4 .8)) xlabel(-.2(.2) .6) /// xline(0, lpattern(shortdash))
Below is code for replicating graphs like Figure 6 when using a logistic model or when plotting marginal effects.
*logit model use http://www.ats.ucla.edu/stat/data/hsbdemo, clear logit honors math science read Logistic regression Number of obs = 200 LR chi2(3) = 74.30 Prob > chi2 = 0.0000 Log likelihood = -78.493853 Pseudo R2 = 0.3212 ------------------------------------------------------------------------------ honors | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- math | .1156084 .0318287 3.63 0.000 .0532253 .1779915 science | .0305099 .0294797 1.03 0.301 -.0272692 .088289 read | .0626137 .0267384 2.34 0.019 .0102074 .11502 _cons | -12.53017 1.869501 -6.70 0.000 -16.19432 -8.866016 ------------------------------------------------------------------------------
*margins, post
parmest, label list(parm estimate min* max* p) eform saving(mypars, replace)
+---------------------------------------------------------+
| parm estimate min95 max95 p |
|---------------------------------------------------------|
1. | math 1.1225562 1.0546672 1.1948152 .00028101 |
2. | science 1.0309801 .97309923 1.0923038 .30069377 |
3. | read 1.0646155 1.0102597 1.1218959 .01919541 |
4. | _cons 3.616e-06 9.266e-08 .0001411 2.050e-11 |
+---------------------------------------------------------+
file mypars.dta saved
use mypars, clear
encode parm, gen(parmn)
eclplot estimate min95 max95 parmn if parm!="_cons" , ///
hori xline(1, lpattern(shortdash)) ytitle("") ylabel(2 3 4) xtitle(Odds Ratio)
* marginal effect
use http://www.ats.ucla.edu/stat/data/hsbdemo, clear
logit honors math science read
Logistic regression Number of obs = 200
LR chi2(3) = 74.30
Prob > chi2 = 0.0000
Log likelihood = -78.493853 Pseudo R2 = 0.3212
------------------------------------------------------------------------------
honors | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
math | .1156084 .0318287 3.63 0.000 .0532253 .1779915
science | .0305099 .0294797 1.03 0.301 -.0272692 .088289
read | .0626137 .0267384 2.34 0.019 .0102074 .11502
_cons | -12.53017 1.869501 -6.70 0.000 -16.19432 -8.866016
------------------------------------------------------------------------------
margins, dydx(math science read) atmeans post
Conditional marginal effects Number of obs = 200
Model VCE : OIM
Expression : Pr(honors), predict()
dy/dx w.r.t. : math science read
at : math = 52.645 (mean)
science = 51.85 (mean)
read = 52.23 (mean)
------------------------------------------------------------------------------
| Delta-method
| dy/dx Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
math | .0162469 .0045199 3.59 0.000 .0073881 .0251057
science | .0042877 .0040702 1.05 0.292 -.0036897 .012265
read | .0087993 .003778 2.33 0.020 .0013945 .0162042
------------------------------------------------------------------------------
parmest, label list(parm estimate min* max* p) saving(mypars, replace)
+----------------------------------------------------------+
| parm estimate min95 max95 p |
|----------------------------------------------------------|
1. | math .01624687 .00738807 .02510566 .00032497 |
2. | science .00428767 -.00368968 .01226502 .29213848 |
3. | read .00879933 .00139449 .01620417 .01985544 |
+----------------------------------------------------------+
file mypars.dta saved
use mypars, clear
encode parm, gen(parmn)
eclplot estimate min95 max95 parmn , ///
hori xline(0, lpattern(shortdash)) ylabel(1 2 3) ///
ytitle("") xtitle(Marginal Effect)
Kastellec, J.P. and Leoni, E.L. "Using Graphs Instead of Tables in Political Science". Perspectives on Politics (2007), 5:4:755-771 Cambridge University Press.
UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services