Mediation effect with 'medeff' command

[Guilherme Limeira]

Dear Statalist,

I would like your help with the 'medeff' command.

First, I have an unbalanced panel with 200,000 observations.

When I run the 'medeff' with the complete data base, the Stata crashes and I get no results after the bootstrapping - I have also limited the simulations to 10 instead of 1000

However, when I limit the data base to 30,000 observations, the code runs through.

• Is there a way to run the 'medeff' with the entire data base?
• Do you suggest another method to have the mediation effect?
• How do I interpret the results from 'medeff'?

Below follows the commands and the outcomes that I have got

Command:
medeff (regress trucks d_monitoring1 Avgweighttonne Distancedrivenkm Odometervaluekm Avgspeedkmh) (regress fuelconskml d_monitoring1 trucks Avgweighttonne Distancedrivenkm Odometervaluekm Avgspeedkmh), mediate(trucks) treat(d_monitoring1) sims(1000) vce(bootstrap)

Variables:

* Example generated by -dataex-. To install: ssc install dataex
clear
input double fuelconskml byte d_monitoring1 float trucks double(Avgweighttonne Distancedrivenkm Odometervaluekm Avgspeedkmh)
 2.046196761756577 1 177 36.14212014401485           4760.62         823026.73  50.0049076389355
 2.076086226775837 1 199  35.2987334724339 7112.360000000001         830139.09 51.53832187679271
2.0883324555943887 1 198 34.50393177285128            8723.8         838862.89 52.69931083781783
2.0339534318919394 1 197 35.38759871416966           8542.34         847405.23 52.31036054731886
2.0188938064549524 1 120 35.48843649043638           1560.08         796314.13  52.5825351796197
2.2398736838214575 1 141 31.89764655770776           9696.01         806010.14 54.61540844634081
1.8822567958935938 1 165 40.71896048434724           8280.01         814290.15 54.72567612969794
 2.076823386594049 1 177 36.68906020847979 7595.940000000001         821886.09 51.97388512877825
1.8684989045236728 1 199 36.16660892629559           6379.13         828265.22 48.20713759422645
1.9857991284575172 1 198 41.80646047388743           6480.02         834745.24 50.73272106875906
 2.053318145645763 1 197 41.70991937138558           7895.46 842640.7000000001  53.3998196437965
 2.109778615346544 1 120  36.9266674486239           3324.99         791562.48 55.10601840555758
1.8968326335010006 1 141 40.73353486526811           8217.06         799779.54 53.30945395566769
2.1153203120324005 1 165 34.22486140218172           5301.31         805080.85 57.02820226263543
1.9021277561313996 1 177 38.14367879753576           5867.95          810948.8 55.36179089770268
 1.845634217302612 1 199 41.94038775140424            7726.6          818675.4   55.071988595866
  1.93145606586284 1 198 33.92710418695228           6418.75         825094.15 51.69601868506537
 2.141632025138293 1 197 35.00964639297794           8505.77         833599.81 54.97919389821744
2.7913990334617975 1 120 31.26473778790734           3650.48 685231.8200000001 64.99338776761738
 2.279752541186779 1 141 31.01604850535961           7753.37         692984.93   55.167004310643
 2.421752853854634 1 165  31.8703858295387           6744.17         699728.84 58.23925120657065
2.3193162838254064 1 177 34.78475875977016           7304.13 707032.8200000001  56.5265866541123
2.3197217930867873 1 199 35.28367171312994           7297.52         714330.21 50.01584373304395
 2.417144073755543 1 198 31.43627199300906           8628.31         722958.52 55.02982694842468
2.3940575775532986 1 197  32.0076991710799            6948.8         729907.22 53.79612007509532
 2.277955963086126 1 165 29.30511885840647           2315.36         583481.68 51.00442410187061
 2.257455346604768 1 177   28.369176251888           12049.8         595531.48 49.25254782276964
 2.387384317599855 1 199  27.3881464878373          11840.71 607372.1900000001 48.84098528579383
 2.269504840418279 1 198 25.53630764936358          10819.07         618191.13 48.66276101696821
 2.331449353154169 1 197 25.24018289429143          10250.73         628441.86 51.28235387307063
2.5144451813047177 1 120 37.47543978277263           5649.38         700860.76 57.55260812107183
2.2297044828191286 1 141 39.53577512776831           4320.32         705180.99 54.90536304329407
2.0472204747045075 1 165 41.72786845211445            4913.8         710094.62 52.28930192962542
 2.346767279607403 1 177 33.30748251329808           6350.54         716444.73 55.47467090496585
 2.329918241725043 1 199 37.90589230664501           6169.74         722614.39 55.18290480224796
2.2046360917248253 1 198 36.55552342014277            7606.7         730221.01 55.54723441544216
 2.312145066325951 1 197 36.47981094806252           7528.09         737749.09 55.81209198110294
 2.042462308534575 1 120 32.57807207196593           6789.88         769674.38 55.27617772631884
 1.935392338167555 1 141 34.20312926365642          10810.85         780485.23 55.63505836675996
2.0174539169207093 1 165  33.0631975831843           8940.69         789425.92 54.74785679294579
1.9373118551930402 1 177 34.79848184079946           7954.37         797380.13 54.62978606503898
1.9176897332611873 1 199 34.73059061059161           9569.08         806949.21 54.46674002959814
1.8015174395049909 1 198 35.96856844264907 8315.210000000001         815264.42 53.50833421218681
1.9477890501381412 1 197 34.13824452079643           7028.85         822293.27 54.79635324180345
1.0148980073104588 1 135 23.77148599743243           1721.47         539366.15 14.99650815851944
1.0326658738341077 1 152 20.30144602985723 4697.690000000001         544063.78  17.2272528171037
 1.050609474538355 1 154 13.24766375080008              4687         553425.93 20.82234516086438
1.0526409126113803 1 166 19.15582205321625           5817.02         559242.64 19.20787640197974
1.1203148881766523 1 168  19.1259864439842           5064.91         564307.47 20.30975997433667
1.0346843264686794 1 161  17.8872393337191           2263.91         566571.02 15.19085500417512
 1.049685945429878 1 159 9.921397045448018           1044.49         567615.35 19.05115213910788
1.0605254406385103 1 135 17.77290442389219           2455.53         546834.33 16.49285799576107
1.1517241379310346 1 152 20.06855177841141           5362.37          552196.7  17.5773787627702
 1.257003108702234 1 158 17.99228892937229           5895.42         558092.03 22.07627432461108
  1.33012542175021 1 166 16.47353155235893           3737.28 569984.0700000001  16.8164987625928
1.1995740985024403 1 168 19.92544683429264              3301         573284.96 16.48901481619208
1.3025268483890968 1 161 18.05294007009032           5210.42          578495.1 18.80452330827068
1.3099604500471935 1 159 15.31272516698273           2262.21         580757.31 18.42423940781496
1.0499064565750633 1 135 15.47577751906674           2525.34            592111 14.84504448015049
1.0365784126606745 1 152 18.07037141649659           4708.73         596819.67 15.37116467341974
1.2963616750024518 1 158 17.07800269665321           4494.46         601313.81 19.69826453776806
1.0392708731983522 1 154 17.76250358819419           3448.81         604762.25 15.37473809346905
1.0591894704277638 1 166 14.69241880323725           1883.08         606645.04 15.85506798950337
1.0887671338216351 1 168 16.93769157967596           2831.72         609476.24  15.1647833770685
 1.007141033279428 1 161 15.97026088279289           2366.58          611842.3 10.57500335135618
 .9938841806114972 1 135 21.28647220505827           2346.65         568717.25 14.66396575606403
1.2095239842581806 1 158 18.92469794992757           5273.96         574627.75 21.30699229362987
 1.273329619011139 1 154 18.02834062840413           6582.07         581209.25  18.0993086570256
1.3548666448902695 1 166 16.63313821794613           3567.12         584776.23 16.32717327871291
1.2869493357065667 1 168 19.08849921791389            2429.4         587205.34 14.73750549759791
   1.4289293997789 1 161 17.85779298794917           3296.04 590501.2000000001 20.11999002965668
1.3911024534500733 1 159 20.29905113368326           3349.26         593850.46 19.71373775589785
 1.077545876448656 1 135 15.91721356706312           5762.78         628057.47 21.58867890364602
1.0523801744781671 1 152 15.97927636962235           8310.32         636367.79 22.24733779908861
1.1021316308304645 1 158 15.78134054110636          10699.56         647067.35 23.98659385701235
1.1022105266402729 1 154 16.88308431506536          10726.79         657793.55 23.10148600143575
 1.113528332378767 1 166 16.46774766864925           6104.83         663897.42 19.55779534846469
1.0789912946713187 1 168 16.24600493105653           7446.68         671343.97  18.4040943681842
 1.105485905568908 1 159 16.14036245725059           4848.02         683326.72 18.24028694722474
1.1144116533960626 1 135 15.77478575281168           6007.08 622488.7000000001 22.67488151658768
1.1315439963033864 1 152 15.35225087738949 8570.880000000001 631059.0700000001 22.81861243180522
1.1185847027332507 1 158  16.5129938922538          10213.26         641272.31 24.50522392635584
1.1184506685646045 1 154  15.0696108725044          11220.23         652492.46 24.61582920406672
1.1659017300441359 1 166 37.52657357633576           6490.47         658982.93 22.28062553637837
  1.12695667577918 1 168  55.6808115254734           7506.85         666489.62 20.07917349292855
1.1754142698635264 1 161 36.08482491986738            9078.7          675568.3 21.93520242363392
1.1798249430107715 1 159 17.75100006719009           6102.09         681670.38 21.37794174462499
1.0878589953454458 1 135 15.93731854363746           6027.62         594115.12  22.1187819877049
1.0614565437432377 1 152 17.26992176463138           8211.12         602326.09 22.36065980516896
1.0848246937046755 1 158 18.23430768703363          10501.32         612826.91 21.93843618342406
1.1347057834296972 1 154 15.11290075002514          11335.62         624162.53 23.65698722954301
1.1403822010226201 1 166 16.73607407520279           7875.24 632037.7000000001 22.95120326212876
1.1177717413384178 1 161 16.57390835714644            7978.8         646755.93 19.83778198915968
1.2620648768664742 1 159 13.62520743432334 6514.110000000001         653269.77 23.32597868804029
1.0246518851057722 1 135 12.49571212661538           2355.48 586980.0700000001  17.3795545936374
 1.099588989155987 1 152 18.53514552822428           3625.07         590604.98 15.10961728831025
1.2451823319528694 1 158 19.34805801118826           4386.74         594991.59 20.12835435100143
  1.39110638205937 1 154 19.65528398465053 7726.650000000001         602718.21 22.85649253645494
 .8356100148726352 1 168 19.06875109823772           1764.19 605238.4400000001 7.423679482698766
1.2744691629637135 1 161 19.48064661937632           4402.59         609641.03 18.25340725606765
end

Results:

medeff (regress trucks d_monitoring1 Avgweighttonne Distancedrivenkm Odometervaluekm Avgspeedkmh) (regress fuelconskml d_monitoring1 trucks Avgweighttonne Distancedrivenkm Od
> ometervaluekm Avgspeedkmh), mediate(trucks) treat(d_monitoring1) sims(1000) vce(bootstrap)
Using 0 and 1 as treatment values
(running regress on estimation sample)

Bootstrap replications (50)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
.................................................. 50

Linear regression Number of obs = 33,014
Replications = 50
Wald chi2(5) = 5823.39
Prob > chi2 = 0.0000
R-squared = 0.1245
Adj R-squared = 0.1244
Root MSE = 85.4977

----------------------------------------------------------------------------------
| Observed Bootstrap Normal-based
trucks | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
d_monitoring1 | -59.68996 .9123676 -65.42 0.000 -61.47817 -57.90175
Avgweighttonne | -1.249686 .0406836 -30.72 0.000 -1.329424 -1.169948
Distancedrivenkm | .0007154 .0001107 6.46 0.000 .0004985 .0009323
Odometervaluekm | -.000063 2.50e-06 -25.22 0.000 -.0000679 -.0000581
Avgspeedkmh | .0047144 .0124216 0.38 0.704 -.0196315 .0290602
_cons | 303.1544 2.44563 123.96 0.000 298.3611 307.9477
----------------------------------------------------------------------------------
(running regress on estimation sample)

Bootstrap replications (50)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
.................................................. 50

Linear regression Number of obs = 33,014
Replications = 50
Wald chi2(6) = 51894.74
Prob > chi2 = 0.0000
R-squared = 0.6489
Adj R-squared = 0.6488
Root MSE = 0.3008

----------------------------------------------------------------------------------
| Observed Bootstrap Normal-based
fuelconskml | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
d_monitoring1 | .0109973 .0035193 3.12 0.002 .0040996 .017895
trucks | .0002705 .0000164 16.52 0.000 .0002384 .0003026
Avgweighttonne | -.0330538 .0002299 -143.76 0.000 -.0335045 -.0326032
Distancedrivenkm | .0000175 7.68e-07 22.77 0.000 .000016 .000019
Odometervaluekm | -9.54e-07 1.85e-08 -51.55 0.000 -9.90e-07 -9.18e-07
Avgspeedkmh | .0007197 .0004319 1.67 0.096 -.0001268 .0015662
_cons | 3.636518 .0276707 131.42 0.000 3.582284 3.690751
----------------------------------------------------------------------------------
(32,014 missing values generated)
(32,014 missing values generated)
(32,014 missing values generated)
------------------------------------------------------------------------------------
Effect | Mean [95% Conf. Interval]
-------------------------------+----------------------------------------------------
ACME | -.0161522 -.0182105 -.0142628
Direct Effect | .0108793 .0042243 .017732
Total Effect | -.0052729 -.0113257 .0010802
% of Tot Eff mediated | 2.822564 -22.36484 21.19477
------------------------------------------------------------------------------------

[Phil Bromiley]

medeff is user written. It could have sample size limitations. You'd need to either try to modify it yourself or contact the authors.

Mediation is also easy to do in SEM.