stata培訓內(nèi)含stata15軟件計量經(jīng)濟學a14-esttab

1、第十四講結(jié)果輸出Extracting results計量經(jīng)濟學服務(wù)中心2017.10提 綱t summary sisticmmarize、tabs)t correlation coefficients (pwcorr、correlaCoefficients/equations Basic syntax and usageestout、esttab、estadd and other useful S dsogout )aStandard errors, p-values, and summary sUse with Word/Rtf/Excel/LaTeXisticsOther Advanced

2、 Ex）les（esttab、outreg2、logout等簡介實證分析的難點在于獲取數(shù)據(jù)、整理數(shù)據(jù)、模型估計、結(jié)果輸出 結(jié)果輸出主要分為樣本分布表定義變量表基本統(tǒng)計量分樣本T檢驗表相關(guān)系數(shù)表回歸分析表各種穩(wěn)健性檢驗表簡介上述各種結(jié)果輸出的表格，就是本講的主要內(nèi)容。主要特色在于你可能遇見結(jié)果輸出中常見結(jié)果輸出主要分為手工錄入VS機器錄入，處理海量數(shù)據(jù)，尤其是相關(guān)系數(shù)矩陣表的時候，能夠省去幾個小時的工足量，并且機器自動生成結(jié)果，不會出錯的；一個文件或者多個文件結(jié)果輸出，工作量更是非常大的，所以我們在回歸結(jié)果輸出的基礎(chǔ)上，直接單獨出來面板數(shù)據(jù)輸出結(jié)果的章節(jié)內(nèi)容；另外還有輸出結(jié)果的分頁以及中文輸出亂

3、碼等問題；表格上面有沒有表頭以及t統(tǒng)計檢驗量，還有置信水平值等問題都需要考慮，而這些問題，這次將直接解決；另外相關(guān)系數(shù)矩陣表輸出、回歸結(jié)果輸出是否符合面的要求，這些都不再是問題，結(jié)果輸出so easy。等版t summary sisticmmarize、tabs)本章節(jié)涵蓋命令有tabs，以及簡潔命令univar，還有輸出結(jié)果的logout，outreg2等對數(shù)據(jù)進行最基本的統(tǒng)計運算就是計算出各描述統(tǒng)計值，常見統(tǒng)計量表為：mean 均值semean sd/ncount 非缺失樣本數(shù)n 同上skewness 偏度kurtosis 峰度sum 求和t summary s常見統(tǒng)計量表為：p1 第一分

4、位數(shù)max 最大值p5 第五分位數(shù)min 最小值p10 第十分位數(shù)range maxminp50 中位數(shù)median 同上sd 標準差variance 方差isticmmarize、tabs)t summary sisticmmarize、tabs)tabs與globaluse E:sadatagdp.dta,clearglobal xx y x1 x2 x3 x4 x5 des $xxtabs$xx, s(mean sd max) columns(s) format(%4.2f)（XX已經(jīng)代表了y x1 x2 x3 x4 x5）sysuse auto, clearsum price weig

5、ht length gear turntabsprice weight length gear turn ,/s(mean sd p5 p25 med p75 p95 max) /format(%6.2f) c(s)t summary sisticmmarize、tabs)tabs和univarssc install logout sysuse auto.dta,clearlogout, save(logout-basic1) word excel replace: sum price mpg rep78 headroomsysuse auto.dta,cleartabsprice mpg r

6、ep78 headroom trunk weight,s(n mean sd minp25 median p50 max)tabs對此命令的功能齊全，但是選項頗多，容易忘記。發(fā)現(xiàn)還有一個比較簡潔的外部命令：univar。ssc instalivarunivar price mpg rep78 headroom trunk weightt summary sisticmmarize、tabs)logoutssc install logoutfindit logoutlogout, save(mfile) word replace: univar price mpg rep78headroom t

7、runk weightsysuse auto, clear logout, save(mytable) word replace:tabsprice weilen mpg rep78,f(%6.2f)ss(mean sd min p50 max) c(s)t summary sisticmmarize、tabs)logoutssc install logoutfindit logoutlogout, save(mfile) word replace: univar price mpg rep78headroom trunk weightsysuse auto, clear logout, sa

8、ve(mytable) word replace:tabsprice weilen mpg rep78,f(%6.2f)ss(mean sd min p50 max) c(s)t correlation coefficients (pwcorr、correlate )相關(guān)關(guān)系中的重要方法相關(guān)分析，可以研究變量之間關(guān)聯(lián)程度和密切程度，而變量之間的關(guān)聯(lián)程度可以用相關(guān)系數(shù)表示。pwcorr use E:sadata新版.dta.dta, clear logout, save(correlate) word excel replace: pwcorr daccrid tm size debt14 ep

9、s, sig logout,save(corr2) word excel replace: pwcorrtm size debt14 epsdacc ridt correlation coefficients (pwcorr、correlate )pwcorrcd E:suse E:saresultsadata新版.dta.dta, clearpwcorr pwcorr pwcorrpwcorrdacc rid tm size debt14 eps dacc rid tm size debt14 eps, sigdacc rid tm size debt14 eps, star(.01)dac

10、c rid tm size debt14 eps, star(.05)help pwcorr_afinditpwcorr_at correlation coefficients (pwcorr、correlate )輸出相關(guān)系數(shù)矩陣到word表格 use E:sadata新版.dta.dta, clear estt correlate dacc rid tm size debt14 eps, matrixesttab using mytable12.rtf, not unstack compre appendsysuse auto, clearsummarize price mpg rep78

11、 foreign, detailoobst correlation coefficients (pwcorr、correlate )高級部分輸出結(jié)果，標*0.01*不是0.001 use E:sadata新版.dta.dta, clear estmatrixt correlate dacc rid tm size debt14 eps, esttab using daqinxueshu12.rtf,star( * 0.10 *0.05 * 0.01 ) not unstack compreappendoobsCoefficients/equations回歸結(jié)果輸出之estoutsysuse a

12、uto,clearreplace price = price / 1000 replace weight = weight / 1000 quietly regress price weight mpgestimates store m1, title(M1)generate forXmpg = foreign * mpgquietly regress price weight mpg forXmpg foreignestimates store m2, title(M2)estout *, cells(b(star fmt(%9.3f) se(par)/ss(r2_a N F chi2 p,

13、 fmt(%9.3f %9.0g) labels(R-squared)/legend label collabels(none) varlabels(_cons Constant)Coefficients/equationsesttabsysuse auto,clearreplace price = price / 1000replace weight = weight / 1000quietly regress price weight mpg estimates store m1, title(M1)generate forXmpg = foreign * mpgquietly regre

14、ss price weight mpg forXmpg foreignestimates store m2, title(Mereturn list2)Coefficients/equationsesttab esttab * using daqinxueshu2.doc,/ title(表1: 使用 esttab 命令生成的表格) / mtitle(模型1 模型2) nonum/ cells(b(star fmt(%9.3f) se(par) / star(* 0.1 * 0.05 * 0.01) sfmt(%9.3f %9.0g) /s(r2_a N F chi2 p, labels(R-

15、squared)legend label collabels(none)varlabels(_cons Constant)Coefficients/equationsesttab多個回歸方程的合并輸出sysuse auto,clearreg price weightest store m1reg price mpg rep78est store m2weightreg price rep78est store m3weight outreg2 m1 m2 m3 usingdaqinxueshu.doc,ss(coef,ts) / adds(Ajusted R2,e(r2_a) replaceC

16、oefficients/equationsesttab多個回歸方程的合并輸出use E:sadata新版.dta.dta, clearreg dacc rid tmest store dareg dacc rid tm sizeest store qinreg dacc rid tm size debt14 eps est store daqinxueshuesttab da qin daqinxueshu using daqinxueshu4.rtf, /mtitles(m1 m2 m3 ) /table)title(This is the Second regresappend star(

17、 * 0.10 * 0.05 * 0.01 ) staraux r2 nogapsCoefficients/equations面板數(shù)據(jù)回歸結(jié)果輸出將檢驗結(jié)果添加在回歸結(jié)果中use E:sclearadata新版.dta.dta, clearset more offuse HYPERLINK http:/w/ http:/w/data/imeus/traffic.dtaxtset sest cleare yearxtreg fatal spircons unrate,fe/固定效應(yīng)估計est store fe/估計結(jié)果xtreg fatal spircons unrate,re/隨機效應(yīng)估計es

18、t store re /估計結(jié)果hausman fe re /進行檢驗Coefficients/equations面板數(shù)據(jù)回歸結(jié)果輸出將檢驗結(jié)果添加在回歸結(jié)果中estadd scalar Hausman_Test = r(p),replaceesttab fe re using daqinxueshu05.doc,scalars(N Hausman_Test) nonum mtitle(FE RE) b(%12.3f)Coefficients/equations面板數(shù)據(jù)回歸結(jié)果輸出 use E:sadataFDI.dta, clearxtreg lngdp lnfdi lnie lnex lnimest store rextreg lngdp lnfd