------------------------------------------------------------------------------- log: C:\Res\Meta\progs\p1.log log type: text opened on: 6 Apr 2002, 13:19:21 . * . * This STATA program produces some standard meta-analysis. . * . set more 1 . * . * Read in, manipulate and summarize the data set . * . use \res\meta\data\data1 . g studyn=_n . label var studyn "Study Number" . g tratio=gamma/segamma . label var tratio "t-statistic for gamma" . g pgamma=ttail(obs,tratio) . label var pgamma "P-value for gamma" . replace pgamma=1.e-20 if pgamma==0. (3 real changes made) . list author studyn gamma segamma author studyn gamma segamma 1. Rose 1 1.21 .14 2. Engel-Rose 2 1.21 .37 3. Frankel-Rose 3 1.36 .18 4. Rose-van Wincoop 4 .91 .18 5. Glick-Rose 5 .65 .05 6. Persson 6 .506 .257 7. Rose 7 .74 .05 8. Honohan 8 .921 .4 9. Nitsch 9 .82 .27 10. Pakko-Wall 10 -.378 .529 11. Walsh-Thom 11 .098 .196 12. Melitz 12 .7 .23 13. Lopez-Cordova and Meissner 13 .716 .186 14. Silvana Tenreyro 14 .471 .316 15. Levy Yeyati 15 .5 .25 16. Nitsch 16 .62 .17 17. Flandreau and Maurel 17 1.16 .07 18. Klein 18 .5 .27 19. Estevadeoral, Frantz and Taylor 19 .293 .145 . d Contains data from \res\meta\data\data1.dta obs: 19 vars: 22 26 Feb 2002 08:00 size: 2,242 (95.0% of memory free) ------------------------------------------------------------------------------- storage display value variable name type format label variable label ------------------------------------------------------------------------------- author str31 %31s Author pubn str20 %20s Publication Venue prefer float %9.0g Preferred/Best (Verbal) Estimate gamma float %9.0g Quantitative Estimate segamma float %9.0g Standard Error of gamma obs float %9.0g No. Observations nonts byte %9.0g 1 for c/s or panel (with N>>T) studies nocountries int %9.0g Span of Countries noyears byte %9.0g Span of Years postwar byte %9.0g 1 for post-WW2 roseauth byte %9.0g 1 if Rose is co-author meanwithin float %9.0g Mean of within study estimates sdwithin float %9.0g S.D. of within study estimates medwithin float %9.0g Median of within study estimates minwithin float %9.0g Min of within study estimates maxwithin float %9.0g Max of within study estimates countwithin byte %9.0g Number of within study estimates trimmax float %9.0g Trimmed Max of within study estimates trimmin float %9.0g Trimmed Min of within study estimates studyn float %9.0g Study Number tratio float %9.0g t-statistic for gamma pgamma float %9.0g P-value for gamma ------------------------------------------------------------------------------- Sorted by: Note: dataset has changed since last saved . sum Variable | Obs Mean Std. Dev. Min Max -------------+----------------------------------------------------- author | 0 pubn | 0 prefer | 19 .9410526 .6867872 0 2 gamma | 19 .684579 .4177448 -.378 1.36 segamma | 19 .2241579 .121654 .05 .529 obs | 19 33448.79 66535.88 190 219558 nonts | 19 .8421053 .3746343 0 1 nocountries | 19 132.0526 78.52705 16 217 noyears | 19 22.42105 15.43559 1 50 postwar | 19 .8421053 .3746343 0 1 roseauth | 19 .3157895 .4775669 0 1 meanwithin | 19 1.010764 .8582899 .1025714 3.902115 sdwithin | 19 .2412389 .3945224 .0589924 1.831724 medwithin | 19 .8299211 .5010568 -.01 1.97 minwithin | 19 .0561579 .847103 -1.53 1.45 maxwithin | 19 6.221421 18.62448 .69 83 countwithin | 19 20.15789 20.9079 4 83 trimmax | 19 1.815211 1.037545 .69 5.44 trimmin | 19 .2118421 .773086 -1.5 1.45 studyn | 19 10 5.627314 1 19 tratio | 19 4.941714 4.927131 -.7145558 16.57143 pgamma | 19 .0660508 .1827615 1.00e-20 .7625544 . * . * Perform the meta-analysis . * . metap pgamma, method(f) Meta-analysis of p-values -------------------------------------------------------------- Method | chi2 p_value studies ----------------------+--------------------------------------- Fisher | 576.56428 1.990e-97 19 -------------------------------------------------------------- . metap pgamma, method(ea) Meta-analysis of p_values -------------------------------------------------------------- Method | . p_value studies ----------------------+--------------------------------------- Edgington, additive | . 6.150e-16 19 -------------------------------------------------------------- . meta gamma segamma Meta-analysis | Pooled 95% CI Asymptotic No. of Method | Est Lower Upper z_value p_value studies -------+---------------------------------------------------- Fixed | 0.772 0.718 0.825 28.437 0.000 19 Random | 0.730 0.580 0.881 9.493 0.000 Test for heterogeneity: Q= 92.334 on 18 degrees of freedom (p= 0.000) Moment-based estimate of between studies variance = 0.069 . metainf gamma segamma, id(studyn) print ------------------------------------------------------------------------------ Study ommited | Coef. [95% Conf. Interval] -------------------+---------------------------------------------------------- 1 | .75439388 .70019108 .80859673 2 | .76913375 .71581489 .8224526 3 | .75782365 .70403385 .8116135 4 | .76828462 .71449482 .82207447 5 | .82220739 .75890207 .88551271 6 | .77449644 .72102231 .8279705 7 | .78465074 .72134542 .84795606 8 | .77081323 .71751523 .82411128 9 | .7710095 .71756369 .82445532 10 | .77453578 .72129035 .82778114 11 | .78466111 .73096889 .83835328 12 | .77251315 .71896398 .82606232 13 | .7727108 .71896058 .82646096 14 | .77373576 .72036338 .82710814 15 | .77473986 .72124863 .8282311 16 | .77546382 .72159809 .82932949 17 | .70282781 .64514363 .76051193 18 | .77427357 .72082776 .82771933 19 | .7888642 .73473293 .84299552 -------------------+---------------------------------------------------------- Combined | .77150418 .71832888 .82467948 ------------------------------------------------------------------------------ . metareg gamma obs, wsse(segamma) Iteration 1: tau^2 = 0 Iteration 2: tau^2 = .01409318 Iteration 3: tau^2 = .07512375 Iteration 4: tau^2 = .09770425 Meta-analysis regression No of studies = 19 tau^2 method reml tau^2 estimate = .0978 Successive values of tau^2 differ by less than 10^-4 :convergence achieved ------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- obs | 4.48e-08 1.16e-06 0.04 0.969 -2.22e-06 2.31e-06 _cons | .7222008 .1004021 7.19 0.000 .5254162 .9189853 ------------------------------------------------------------------------------ . metareg gamma segamma, wsse(segamma) Iteration 1: tau^2 = 0 Iteration 2: tau^2 = .03010489 Iteration 3: tau^2 = .08777508 Iteration 4: tau^2 = .09033508 Meta-analysis regression No of studies = 19 tau^2 method reml tau^2 estimate = .0903 Successive values of tau^2 differ by less than 10^-4 :convergence achieved ------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- segamma | -.9933213 .823494 -1.21 0.228 -2.60734 .6206972 _cons | .9130875 .1771052 5.16 0.000 .5659677 1.260207 ------------------------------------------------------------------------------ . metareg gamma roseauth, wsse(segamma) Iteration 1: tau^2 = 0 Iteration 2: tau^2 = .02397665 Iteration 3: tau^2 = .09391623 Iteration 4: tau^2 = .07073589 Iteration 5: tau^2 = .0759793 Iteration 6: tau^2 = .0746044 Iteration 7: tau^2 = .07495284 Meta-analysis regression No of studies = 19 tau^2 method reml tau^2 estimate = .0749 Successive values of tau^2 differ by less than 10^-4 :convergence achieved ------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- roseauth | .3770075 .1637407 2.30 0.021 .0560817 .6979334 _cons | .5869939 .1004502 5.84 0.000 .3901151 .7838727 ------------------------------------------------------------------------------ . metareg gamma nonts, wsse(segamma) Iteration 1: tau^2 = 0 Iteration 2: tau^2 = .01140247 Iteration 3: tau^2 = .05984409 Iteration 4: tau^2 = .08546313 Iteration 5: tau^2 = .08589089 Meta-analysis regression No of studies = 19 tau^2 method reml tau^2 estimate = .0859 Successive values of tau^2 differ by less than 10^-4 :convergence achieved ------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- nonts | .2387602 .2034955 1.17 0.241 -.1600836 .6376041 _cons | .5384158 .1805667 2.98 0.003 .1845115 .8923201 ------------------------------------------------------------------------------ . metareg gamma nocountries, wsse(segamma) Iteration 1: tau^2 = 0 Iteration 2: tau^2 = .01458143 Iteration 3: tau^2 = .08299151 Iteration 4: tau^2 = .09447728 Iteration 5: tau^2 = .09421696 Meta-analysis regression No of studies = 19 tau^2 method reml tau^2 estimate = .0942 Successive values of tau^2 differ by less than 10^-4 :convergence achieved ------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- nocountries | .0006362 .0010878 0.58 0.559 -.0014958 .0027682 _cons | .6421511 .1654251 3.88 0.000 .3179239 .9663784 ------------------------------------------------------------------------------ . metareg gamma noyears, wsse(segamma) Iteration 1: tau^2 = 0 Iteration 2: tau^2 = .02269689 Iteration 3: tau^2 = .08865432 Iteration 4: tau^2 = .0978047 Meta-analysis regression No of studies = 19 tau^2 method reml tau^2 estimate = .0977 Successive values of tau^2 differ by less than 10^-4 :convergence achieved ------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- noyears | -.0021466 .0056191 -0.38 0.702 -.0131599 .0088667 _cons | .7778385 .1654519 4.70 0.000 .4535587 1.102118 ------------------------------------------------------------------------------ . metareg gamma postwar, wsse(segamma) Iteration 1: tau^2 = 0 Iteration 2: tau^2 = .01098566 Iteration 3: tau^2 = .06575411 Iteration 4: tau^2 = .09711704 Iteration 5: tau^2 = .09749058 Meta-analysis regression No of studies = 19 tau^2 method reml tau^2 estimate = .0975 Successive values of tau^2 differ by less than 10^-4 :convergence achieved ------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- postwar | -.0268468 .2196513 -0.12 0.903 -.4573554 .4036619 _cons | .7457645 .1969708 3.79 0.000 .3597089 1.13182 ------------------------------------------------------------------------------ . * . * Ditto but drop my observations . * . drop if roseauth==1 (6 observations deleted) . sum Variable | Obs Mean Std. Dev. Min Max -------------+----------------------------------------------------- author | 0 pubn | 0 prefer | 13 .7146154 .5992581 0 2 gamma | 13 .5328462 .3849768 -.378 1.16 segamma | 13 .253 .1164281 .07 .529 obs | 13 9024.077 11323.33 190 31010 nonts | 13 .9230769 .2773501 0 1 nocountries | 13 107.6923 82.72886 16 205 noyears | 13 20.76923 13.24861 1 42 postwar | 13 .7692308 .438529 0 1 roseauth | 13 0 0 0 0 meanwithin | 13 .7440238 .5284827 .1025714 1.923333 sdwithin | 13 .1662512 .0984865 .0589924 .4654206 medwithin | 13 .6945 .5040675 -.01 1.97 minwithin | 13 -.1786923 .8489415 -1.53 1.45 maxwithin | 13 1.952077 1.312218 .69 5.73 countwithin | 13 19.15385 22.51609 4 83 trimmax | 13 1.819154 1.236521 .69 5.44 trimmin | 13 -.045 .77003 -1.5 1.45 studyn | 13 12.92308 4.030334 6 19 tratio | 13 3.197564 4.202909 -.7145558 16.57143 pgamma | 13 .0964941 .2166176 1.00e-20 .7625544 . metap pgamma, method(f) Meta-analysis of p-values -------------------------------------------------------------- Method | chi2 p_value studies ----------------------+--------------------------------------- Fisher | 202.92169 2.429e-29 13 -------------------------------------------------------------- . metap pgamma, method(ea) Meta-analysis of p_values -------------------------------------------------------------- Method | . p_value studies ----------------------+--------------------------------------- Edgington, additive | . 3.058e-09 13 -------------------------------------------------------------- . meta gamma segamma Meta-analysis | Pooled 95% CI Asymptotic No. of Method | Est Lower Upper z_value p_value studies -------+---------------------------------------------------- Fixed | 0.802 0.708 0.895 16.785 0.000 13 Random | 0.571 0.318 0.825 4.416 0.000 Test for heterogeneity: Q= 63.139 on 12 degrees of freedom (p= 0.000) Moment-based estimate of between studies variance = 0.155 . * . * Clean up and Close down . * . drop _all . log close log: C:\Res\Meta\progs\p1.log log type: text closed on: 6 Apr 2002, 13:19:25 -------------------------------------------------------------------------------
