From 3 to 15: Milestones, dead ends, prospects. A subjective review - PowerPoint PPT Presentation

From 3 to 15: Milestones, dead ends, prospects. A subjective review of Stata’s history Ulrich Kohler ukohler@uni-potsdam.de University of Potsdam Faculty for Economics and Social Sciences German Stata Users Group Meeting June 22 nd 2018 University of Konstanz, Germany 1/31

Inhalt Setting the Scene Milestones (and Dead Ends) Prospects (Why not Stata!?) 2/31

Versions v1 v2 v3 v4 v5 v6 v7 v8 v9 v10 v11 v12 v13 v14 v15 15.0 14.2 14.1 13.1 14.0 13.0 12.1 12.0 11.2 11.1 11.0 9.2 10.010.1 9.1 9.0 64−bit Linux 8.2 8.1 8.0 64−bit Solaris SGI Irix GLLAMM speedup OS X Stata/SE 7.0 6.0 5.0 Mac Win 95 Linux Convex 4.0 DEC Alpha MAC3.1 IBM RS/6000 3.0 DEC RISC 2.1 IC HP/9000 2.05 386/ix 2.0 Sun/Unix BiTurbo 1.31.41.5 1.2 1.1 1.0 5 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 8 8 8 9 9 9 9 9 0 0 0 0 0 1 1 1 1 1 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3/31

Education v1 v2 v3 v4 v5 v6 v7 v8 v9 v10 v11 v12 v13 v14 v15 15.0 IRTUS 14.2 14.1 SWUS 13.1 14.0 Mata 13.0 IVRMUS 12.1 12.0 HEUS 11.2 FEUS 11.1 11.0 MAUS 9.2 10.010.1 SBS Webinar 9.1 BAWS 9.0 64−bit Linux SSG 8.2 8.1 TSUS 8.0 V−Tutorial 64−bit Solaris SGI Irix FPSAUS GLLAMM speedup Blog OS X Stata/SE DMUS 7.0 6.0 ISP 5.0 PDF−doc Mac Win 95 WDAUS Linux MEUS Convex 4.0 IMEUS DEC Alpha MAC3.1 MlLMUS DAUS IBM RS/6000 3.0 VGSG DEC RISC 2.1 ISA IC RMCDV HP/9000 2.05 GLM 386/ix NC 631 2.0 Sun/Unix MLEwS BiTurbo 1.31.41.5 NC 200 NC 152 1.2 NC 101 1.1 NC 151 1.0 5 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 8 8 8 9 9 9 9 9 0 0 0 0 0 1 1 1 1 1 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 4/31

Me v1 v2 v3 v4 v5 v6 v7 v8 v9 v10 v11 v12 v13 v14 v15 psidtools 15.0 IRTUS 14.2 DESUG15 14.1 SWUS 13.1 14.0 DESUG16 Mata DESUG14 13.0 IVRMUS sq V2 12.1 psiduse 12.0 HEUS dirtools 11.2 FEUS 11.1 DESUG12 11.0 MAUS DESUG11 9.2 10.010.1 SBS DESUG10 meresc Webinar 9.1 soepuse BAWS 9.0 nlcorr 64−bit Linux SSG 8.2 DESUG9 8.1 TSUS _gapport 8.0 cm2in V−Tutorial 64−bit Solaris SGI Irix khb FPSAUS GLLAMM speedup DESUG8 Blog OS X sdlim Stata/SE DMUS DESUG7 7.0 6.0 DESUG6 ISP 5.0 DESUG5 PDF−doc Mac sq Win 95 WDAUS DESUG4 Linux MEUS Convex DAUS 4.0 DESUG3 IMEUS DEC Alpha biplot V2 MAC3.1 MlLMUS soepren DAUS IBM RS/6000 mkdat 3.0 VGSG DESUG2 DEC RISC 2.1 NLSUG2 ISA IC outdat RMCDV HP/9000 dsearch 2.05 GLM 386/ix DAMS NC 631 2.0 rgroup Sun/Unix MLEwS hist3 BiTurbo himatrix 1.31.41.5 NC 200 biplot NC 152 cdlplot 1.2 NC 101 me taking NC151 1.1 Me reading SwS NC 151 1.0 5 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 8 8 8 9 9 9 9 9 0 0 0 0 0 1 1 1 1 1 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 5/31

Plan of attack Subjectively picking out ◮ milestones of development ◮ dead ends to learn something on prospects. Of course . . . . . . any statements made here are just personal views. Others have different views. At best, my views are an inspiration for the wishes and grumbles session at the end of the meeting. 6/31

Related ◮ Cox (2005) ◮ . help whatsnew#to# ◮ . ssc hot, author( name ) n(#) 7/31

Inhalt Setting the Scene Milestones (and Dead Ends) Prospects (Why not Stata!?) 8/31

Inhalt Milestones (and Dead Ends) My Milestone Command milestones Other Milestones Dead ends 9/31

Me reading “Statistics with Stata” v1 v2 v3 v4 v5 v6 v7 v8 v9 v10 v11 v12 v13 v14 v15 psidtools 15.0 IRTUS 14.2 DESUG15 14.1 SWUS 13.1 14.0 DESUG16 Mata DESUG14 13.0 sq V2 IVRMUS 12.1 psiduse HEUS 12.0 11.2 dirtools FEUS 11.1 DESUG12 11.0 MAUS DESUG11 9.2 10.010.1 DESUG10 SBS meresc Webinar 9.1 soepuse BAWS 9.0 nlcorr 64−bit Linux SSG 8.2 DESUG9 8.1 _gapport TSUS 8.0 cm2in V−Tutorial 64−bit Solaris SGI Irix khb FPSAUS GLLAMM speedup DESUG8 Blog OS X sdlim Stata/SE DMUS DESUG7 7.0 6.0 DESUG6 ISP 5.0 DESUG5 PDF−doc Mac sq Win 95 WDAUS Linux DESUG4 MEUS Convex DAUS 4.0 DESUG3 IMEUS DEC Alpha MAC3.1 biplot V2 MlLMUS soepren DAUS IBM RS/6000 mkdat 3.0 VGSG DESUG2 DEC RISC 2.1 NLSUG2 ISA IC RMCDV outdat HP/9000 dsearch 2.05 GLM DAMS 386/ix NC 631 2.0 rgroup Sun/Unix MLEwS hist3 BiTurbo 1.31.41.5 NC 200 himatrix biplot NC 152 cdlplot 1.2 NC 101 Me reading SwS me taking NC151 1.1 1.0 Me reading SwS NC 151 1985 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Why did I became a Stata user after reading SwS? 10/31

Why Stata? Command line interface . use ../downloaded/data1, clear . reg incomeR age yedu income Models Humor . mlogit lsat age yedu income endless loop → see “loop, endless” Speed . . . loop, endless → see . set rmsg on “endless loop” . mlogit rep78 foreign r; t=0.04 14:54:11 Support From: "William Gould" <wgould@stata.com> To: statalist@hsphsun2.harvard.edu Subject:Re: statalist: iweights and regress Date: Fri, 30 Jan 1998 10:11:57 -0600 xyz <xyz@abc.de> asked for a clarification on iweights. Stay away from them, I say, because they will invariably surprise you. Let me explain:... 11/31

Inhalt Milestones (and Dead Ends) My Milestone Command milestones Other Milestones Dead ends 12/31

My 14 favorites v1 v2 v3 v4 v5 v6 v7 v8 v9 v10 v11 v12 v13 v14 v15 bayes: psidtools 15.0 IRTUS 14.2 DESUG15 14.1 SWUS 13.1 14.0 DESUG16 npregress: Mata DESUG14 13.0 sq V2 IVRMUS 12.1 psiduse HEUS 12.0 unicode 11.2 dirtools FEUS 11.1 DESUG12 11.0 MAUS DESUG11 9.2 10.010.1 marginsplot DESUG10 SBS meresc Webinar 9.1 soepuse BAWS 9.0 i.group nlcorr 64−bit Linux SSG 8.2 DESUG9 8.1 _gapport TSUS 8.0 mata cm2in V−Tutorial 64−bit Solaris SGI Irix khb FPSAUS GLLAMM speedup DESUG8 graph tw Blog OS X sdlim Stata/SE DMUS DESUG7 7.0 6.0 DESUG6 ISP file 5.0 DESUG5 PDF−doc Mac sq Win 95 WDAUS Linux foreach DESUG4 MEUS Convex DAUS 4.0 DESUG3 IMEUS DEC Alpha MAC3.1 archutil biplot V2 MlLMUS soepren DAUS IBM RS/6000 mkdat 3.0 VGSG net DESUG2 DEC RISC 2.1 NLSUG2 ISA IC RMCDV outdat HP/9000 syntax dsearch 2.05 GLM DAMS 386/ix NC 631 2.0 rgroup Sun/Unix svy MLEwS hist3 BiTurbo 1.31.41.5 NC 200 himatrix biplot NC 152 program cdlplot 1.2 NC 101 me taking NC151 1.1 1.0 Me reading SwS NC 151 1985 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 I’ll give some justifications for these choices. 13/31

Statistical commands svy Describers (like me) need to respect the compexity of samples – especially weights. marginsplot Makes understanding complicated models easy . regress income i.sex##i.emp##c.age##c.age . margins, at(age=(20(5)80) emp=(1,2,3) sex=(1,2)) . marginsplot, by(emp) npregress If you do not believe in homogenous treatment effects, this is for you . . . bayes: In my heart, I am Bayesian. bayesmh were introduced in Stata 14, but Stata 15’s bayes-prefix makes Bayesian analysis (syntactically) easy 14/31

General usability foreach/forvalues By-by endless loops, and by-by clumpsy for . graph twoway A command and a graphics programming language at the same time. Powerful and simple (but sometimes we want it even more powerful and much simpler at the same time.) fvvarlist Factor-variable notation lets you specify complicated models. Use marginsplot to interpret them. unicode America first? Perhaps, but American alone? . display " ✟❡t " ✟❡t 15/31

Programmer commands program Stata wouldn’t be Stata without program . program hello . display "hello, world" . end syntax Parsing made easy . syntax [varlist] [if] [in] file , the core of . esttab . psiduse mata Matrix calculations made easy : b = invsym(X’X)*X’y 16/31

Other net Stata became Web-aware in 1998. It turned out to be a game changer. ssc the command formerly known as archutil made usage of user-written programs easy: . ssc hot . ssc install estout 17/31

Inhalt Milestones (and Dead Ends) My Milestone Command milestones Other Milestones Dead ends 18/31