AnIntroductiontoStataforHealthResearchers:

InformationtousersofStata8to12

AnIntroductiontoStataforHealthResearchers,firstedition1waswrittenforStatarelease9,the
secondedition2waswrittenforStata10;andthethirdedition3forStata11.Thethirdeditionisalso,
withfewexceptions,validforStata12,andwearenotcurrentlyplanningafourthedition.Wereferto
thedifferenteditionsofthebookasISHR1,ISHR2,andISHR3.

TheremaybesomediscrepanciesbetweenyoureditionoftheISHRbookandthereleaseofyour
Stataprogram,andinthefollowingwepointtothemostimportantchanges,organizedbythe
structureofthebook.Weareonlypointing,andyouwillprobablyneedtouseStata'shelpfacilitiesto
getmoreinformation.WestronglyrecommendtokeepyourStataprogramupdatedtothelatest
version;Stata11,forexample,toversion11.2.

Toseeallchanges,forexamplefromrelease11to12,type:

.helpwhatsnew11to12

Stataintroducednewfileformatsinreleases8,10,and12.YourStataprogramreadsdatasetsfrom
previousreleaseswithouttrouble,buttheoppositeisnotthecase.45Ifyouwanttocreateadataset
thatcanbereadbyapreviousStataversion,youmayneedtousethesaveoldcommand;seethe
followingtable.

YourversionofStata:

savegeneratesdata
withtheformatof:

saveoldgeneratesdata
withtheformatof:

Stata12

Stata12

Stata8/9

Stata10or11

Stata10/11

Stata8/9

Stata8or9

Stata8/9

Stata7

SomedatefunctionsandformatschangedfromStata9toStata10.WithStata10to12,youcanstill
usetheStata9syntaxusingversioncontrol:

.version9:generatebdate=date(sbdate,"dmy")

SvendJuul,MortenFrydenberg,August2011

JuulS.AnIntroductiontoStataforHealthResearchers.CollegeStation,TX:StataPress,2006

JuulS.AnIntroductiontoStataforHealthResearchers.2nded.CollegeStation,TX:StataPress,2008.

JuulS,FrydenbergM.AnIntroductiontoStataforHealthResearchers.3rded.CollegeStation,TX:StataPress,2010.

AnupdatetoStata11willallowStata11toreadStata12datasets.TheupdateisexpectedinSeptember2011,soStata
11usersshouldstillupdateregularly.

Stata9users:Theunofficialuse10commandreadsStata10/11datasets.Getifbysscinstalluse10.

ChangesfromStata11toStata12
TheinformationisorganizedbythechaptersinISHR3.IntheStata12columnweshowchangesfrom
thedescriptioninISHR3.

Section Stata12
Stata11(asdescribedinISHR3)
1.4

Differentorganizationofwindows,including
anewPropertieswindow.Thedefaultlayout
fitsawidescreen;ifyoupreferanarrower
layout,select:
Edit>Preferences>LoadPreferenceSet>
CombinedLayout

1.4

PasteavariablenamefromtheVariables
windowtotheCommandwindowwitha
doubleclick.

PasteavariablenamefromtheVariables
windowtotheCommandwindowwitha
singleclick.

1.4

Improveddataeditor.IntheVariables
Windowyoucandoquiteafewthings,
includinghidingandreorderingvariables.

5.8

Stata12automaticallycalculatesthememory setmemoryallocatesStatamemory.
needs,andthesetmemorycommandisno
longerneeded.

6.4

Stata12nowimportsandexportsExceland
SASdatadirectly.See:
helpimportexcel
helpimportsas

ImportandexportSASdatasetsusingthe
fdauseandfdasavecommands.

9.2

New,flexiblerenamegroupcommand.See:
helprenamegroup

13.3

Thecontrastcommandmaybeusedasa
handyreplacementfortestparm.See:
helpcontrast

13.3

Theoutputfromregressioncommands
(includinglogistic,stcox,andpoisson)with
factorvariableswillalwaysdisplaythebase
levelifyou,justonce,givethecommand:
setshowbaselevelson,permanently
andwerecommenddoingthattoenhance
transparency.

Bydefault,baselevelsareomittedfromthe
output.

Thesetshowbaselevelsfacilitywas
introducedinStata11.1,June2010.

13.5
14.7

Baselineestimates(baselineodds,risk,rate)
aredisplayedbyeformcommands,like:
logisticandlogit,or
binreg,rr
poisson,irr

Baselineodds,risk,andrateestimatesare
notdisplayedbylogisticoranyothereform
command.

17.12

YoucanexportagraphtoaPDFfile.See:
helpgraphexport

ChangesfromStata10toStata11
TheinformationisorganizedbythechaptersinISHR2.IntheStata11columnweshowchangesfrom
thedescriptioninISHR2.

Section Stata11
Stata10(asdescribedinISHR2)
1.4

DataEditor/Browserimproved.Youneednot DataEditor/Browsermustbeclosedwhile
closeitwhiledoingotherthingsinStata
doingotherthingsinStata

1.4

DofileEditorimproved.Itnowhassyntax
highlighting

1.5

NewVariablesManager

2.2

PDFversionofmanualsincluded

Onlyprintedversionofmanuals

5.3

Newmisstablecommanddescribesmissing
valuepatterns.

Newmifamilyofcommandsperforms
multipleimputationofmissingvalues.

9.3

Ordervariablesalphabeticallyby:
ordervarlist,alphabetic

Ordervariablesalphabeticallyby:
aorder

9.5

Newmergesyntax,e.g.:
usefilea.dta
merge1:1idusingfileb.dta

mergeautomaticallysortsthedatasets.

Theoldmergesyntaxstillworks.

Oldmergesyntax,e.g.:
usefilea.dta
mergeidusingfileb.dta

Thedatasetsmustbesortedbythematching
keybeforemerging

13.1

FactorvariablesisanewconstructinStata
11;itpartiallyreplacesthexi:prefixin
regressioncommandsbutxi:stillworks.

Seeshortdescriptionnextpages.

Usethexi:prefixforcategoricalvariablesand
interactionsinregressionanalyses.

13.3

Stata11.1andlater(June2010):
Theoutputfromregressioncommands
(includinglogistic,stcox,andpoisson)with
factorvariableswillalwaysdisplaythebase
levelifyou,justonce,givethecommand:
setshowbaselevelson,permanently
andwerecommenddoingthattoenhance
transparency.

Bydefault,baselevelsareomittedfromthe
output.

14.3

Youneednotspecifytheschoenfeld()and
basesurv()optionstostcoxtousetheestat
phtestandstcurvepostestimation
commands.

Youmustspecifytheschoenfeld()and
basesurv()optionstostcoxtousetheestat
phtestandstcurvepostestimation
commands.

16.1

Newrandomnumberfunctionnames:
runiform()
rnormal()

CorrespondingStata10functions:
uniform()
invnormal(uniform())

Factorvariables:NewconceptinStata11.

Weusethelbw1.dtadataset:

. cd c:\docs\ishr3
. use lbw1.dta
. list race i.race in 1/3
+-------------------------------+
|
1b.
2.
3.|
|
race
race
race
race |
|-------------------------------|
1. | 2. black
0
1
0 |
2. | 3. other
0
0
1 |
3. | 1. white
0
0
0 |
+-------------------------------+

i.racerepresentsthreevirtual0/1variables.1b.race,2.race,and3.racedonotreallyexistinthe
dataset,butwithsomecommands,likelist,theyseemtoexist.Notethatracemusttakenonnegative
integervaluesonly;factorvariablesdonotworkwithnonintegercodes.

Inregressionanalyses,etc.,theycanbeusedasindicator(dummy)variables;itgoeslikethis:

. regress bwt i.race i.smoke

Source |
SS
df
MS
-------------+-----------------------------Model | 12346897.6
3 4115632.54
Residual | 87568400.9
185 473342.708
-------------+-----------------------------Total | 99915298.6
188 531464.354

Number of obs
F( 3,
185)
Prob > F
R-squared
Adj R-squared
Root MSE

=
=
=
=
=
=

189
8.69
0.0000
0.1236
0.1094
688

-----------------------------------------------------------------------------bwt |
Coef.
Std. Err.
t
P>|t|
[95% Conf. Interval]
-------------+---------------------------------------------------------------race |
1 | (base)
2 |
-450.54
153.066
-2.94
0.004
-752.5194
-148.5607
3 | -454.1813
116.436
-3.90
0.000
-683.8944
-224.4683
|
smoke |
0 | (base)
1 | -428.0254
109.0033
-3.93
0.000
-643.0746
-212.9761
|
_cons |
3334.858
91.74301
36.35
0.000
3153.86
3515.855
------------------------------------------------------------------------------

Theoutputdisplaysthebaselevels,1forraceand0forsmoke.6Withthei.prefix,Stataselectsthe
lowestvaluesasthebaselevel.Weprefer,however,tocontrolthebaselevelsdirectly;thiscanbe
donewith,e.g.,theb2.prefix:

Fromversion11.1ofStata,youcan,onceandforall,givethecommand
.setshowbaselevelon,permanently
Thentheregressionoutputwillshowthebaselevelsdirectly,asintheexampleshere.

. regress bwt b2.race b0.smoke

Source |
SS
df
MS
-------------+-----------------------------Model | 12346897.6
3 4115632.54
Residual | 87568400.9
185 473342.708
-------------+-----------------------------Total | 99915298.6
188 531464.354

Number of obs
F( 3,
185)
Prob > F
R-squared
Adj R-squared
Root MSE

=
=
=
=
=
=

189
8.69
0.0000
0.1236
0.1094
688

-----------------------------------------------------------------------------bwt |
Coef.
Std. Err.
t
P>|t|
[95% Conf. Interval]
-------------+---------------------------------------------------------------race |
1 |
450.54
153.066
2.94
0.004
148.5607
752.5194
2 | (base)
3 | -3.641269
160.537
-0.02
0.982
-320.3599
313.0773
|
smoke |
0 | (base)
1 | -428.0254
109.0033
-3.93
0.000
-643.0746
-212.9761
|
_cons |
2884.317
141.291
20.41
0.000
2605.569
3163.066
------------------------------------------------------------------------------

Wewanttoincludeaninteractionbetweenraceandsmoke;itisdonewiththetermi.race##i.smoke,
or,tocontrolthebaselevelsexplicitly,byb1.race##b0.smoke.Hadweusedasinglecross,
b1.race#b0.smoke,theregressionwouldcomparefiverace/smokegroupswithwhitenonsmokers.

. regress bwt b1.race##b0.smoke

Source |
SS
df
MS
-------------+-----------------------------Model | 14455540.4
5 2891108.08
Residual | 85459758.2
183 466993.214
-------------+-----------------------------Total | 99915298.6
188 531464.354

Number of obs
F( 5,
183)
Prob > F
R-squared
Adj R-squared
Root MSE

=
=
=
=
=
=

189
6.19
0.0000
0.1447
0.1213
683.37

-----------------------------------------------------------------------------bwt |
Coef.
Std. Err.
t
P>|t|
[95% Conf. Interval]
-------------+---------------------------------------------------------------race |
1 | (base)
2 |
-574.25
199.5008
-2.88
0.004
-967.8674
-180.6326
3 | -614.5136
138.2182
-4.45
0.000
-887.2198
-341.8075
|
smoke |
0 | (base)
1 | -601.3654
139.979
-4.30
0.000
-877.5456
-325.1851
|
race#smoke |
2 1 |
250.8654
308.9992
0.81
0.418
-358.7938
860.5245
3 1 |
544.2957
258.8455
2.10
0.037
33.59038
1055.001
|
_cons |
3428.75
103.0218
33.28
0.000
3225.487
3632.013
------------------------------------------------------------------------------

ChangesfromStata9toStata10
TheinformationisorganizedbythechaptersinISHR1.IntheStata10columnweshowchangesfrom
thedescriptioninISHR1.

Section Stata10
Stata9(asdescribedinISHR1)
1.4

Selectwindowingpreferencesby:
Edit>Preferences>ManagePreferences>

Selectwindowingpreferencesby:
Prefs>ManagePreferences>

1.4

Savepastcommandstoadofilebyclicking
somewhereintheReviewwindow.PressCtrl
AtohighlightallcommandsandCtrlCto
copythemtotheWindowsClipboard.Next
openaDofileEditorwindowandpastethe
commandsbypressingCtrlV.

Save past commands to a dofile by right

clicking in the Review window and selecting
Copy Review Contents to Clipboard. Next
open a Dofile Editor window and paste the
ReviewcontentsbypressingCtrlV.

Helpfilesnowhavetheextension.sthlp(but Helpfileshavethe.hlpextension.
Stata10stillunderstandsthe.hlpextension)

5.5

NewinStata10:Dateandtimevariables
withnewformatdescriptors(oldformat
descriptorsstillwork).

Stata9hasdatevariables,butnofacilitiesfor
handlingtimeofday.(Someunofficial
commandsandfunctionsmaybehelpful).

5.5

Inthedate()functionthesequencecanbe
definedas"DMY".

(YoucanusetheStata9"dmy"specification
withversioncontrol;seepage1).

Inthedate()functionthesequencecanbe
definedas"dmy".

6.1

savesavesadatasetinStata10/11format;it
cannotbereadbyStata8or9.Tocreatea
datasetthatcanbereadbyStata8/9,usethe
saveoldcommand

ToenableStata9toreadadatasetcreated
byStata10/11'ssavecommand,installthe
unofficialuse10commandby:
sscinstalluse10

6.4

Stat/Transfer9translatesStata10/11files,
butStat/Transfer8doesnot.

11.10

Newgrapheditor.

13.3

Newexlogisticcommandforexactlogistic
regression

14.2

Newrisktableoptiontostsgraphdisplays
numberatriskbelowthexaxis.

14.5

NewexpoissoncommandforexactPoisson
regression

16.2

Newstpowercommand(samplesizeand
powerforincidenceandsurvivalanalysis)

17.4

Tocreater(varlist):
describe...,varlist

Tocreater(varlist):
describe...,simple

18.7

Newdatasignaturecommand.

ThedatasignaturecommandinStata9.2is
differentfromtheversion10command.

ChangesfromStata8toStata9
TheinformationisorganizedbythechaptersinISHR1.

Section Stata9(asdescribedinISHR1)
Stata8
1.4

YoucanhavemultipleDofile,Viewer,and
graphwindows.

YoucanhaveonlyoneDofileeditorwindow,
oneViewerWindow,andoneGraphwindow

1.4

Datawindow:Stringvariablesaredisplayed
inred,valuelabelsinblue.Youmaytoggle
betweendisplayingvaluelabelsandcodes.

Novisualdistinctionbetweenstringvariables
andvaluelabels.TheDatawindowdisplays
valuelabels,unlessyouopenitby:
browse,nolabel

2.1

NewDatamanagementmanual,[D].

Datamanagementcommandsin[R].

2.3

NewcommandinStata9.1:
adoupdate

6.4

NewcommandsinStata9:
xmlsave
xmluse

8.2

Newfunctionnames(theoldnamesstillwork
inStata9,butnotinStata10):
normal()
invnormal()
normalden()

Oldfunctionnames:

norm()
invnorm()
normden()

8.3

New,lessconfusingegenfunctionnames(the
oldnamesstillwork):
total()
rowmin()
rowmax()
rowtotal()
rowmiss()
rownonmiss()

Oldfunctionnames:

sum()
rmin()
rmax()
rsum()
rmiss()
robs()

9.5

Newsortoptiontomerge.

10.1

Newoptiontocodebook:
codebook,compact

Nocompactoptiontocodebook.Auseful
alternativeistheunofficialsummvl:
finditsummvl

10.3

NewcommandinStata9:
proportion

10.4

NewcommandinStata9:
mean

11.7

Markerandlineoptionssimplified(theold
optionnamesstillwork)

SeeStata8syntaxintablebelow

13

stepwise

sw

Section Stata9

Stata8

13.2

estatgof

13.4

svy:

Surveyanalysiscommandschangedsyntaxin
version9.

14.3

estatphtest

17.1

statsby:

Changedsyntax

17.3

Newfunctionname(theoldnamestillworks
inStata9,butnotinStata10):
strpos()

lfit

stphtest

Oldfunctionname:
index()

Stata8:Optionsfordefiningtheappearanceoflines,barsandmarkers
(theyaremuchsimplerinStata9;seeISHR1table11.2):

Lines

Markers

Element

Color

Pattern

Width

Symbol

Size

Legendetc.

color()

fill

fcolor()

outline

lcolor()

lpattern()

lwidth()

Bars,areas

bcolor()

fill

bfcolor()

outline

blcolor()

blpattern()

blwidth()

Markers

mcolor()

msymbol()

msize()

fill

mfcolor()

outline

mlcolor()

mlwidth()

clpattern()

clwidth()

Connectinglines clcolor()