Need help with regression with moderation in Stata

[Meike Wenzl]

Hello everyone,

I am not sure whether this is the right forum but I do need some help for my thesis. I have to do a regression with moderator and I am really confused about the command for said regression. I do have a dependent variable, independent variable, a "moderator" variable (an Index I created) and some controll variables. I thought I have to calculate the moderator out of two variables (independent and my "moderator variable" as an interaction term), now my counselor said that I need both the moderator and the interaction term in my regression. So before my command would look like this:

regress DV IV Interaktionterm(IV*Index) control variables

now I am told it has to look like that

regress DV IV moderator (Index) Interaction term control variables

Can someone explain this to me? I thought the moderator IS the interaction term and not some variable alone...

please help me!

[Dick Campbell]

I can't quite tell what you mean in your first equation. If the moderator variable is an index you created and you think that the effect of your IV of interest varies depending on the value of the moderator then you are asserting that there is an interaction between the the IV and the moderator. In that case, the second equation you provide is what you want to do. The point is that both the "main effect" of the moderator variable and the interaction need to be in the model. In your first equation, if I understand it correctly (and I may not) the interaction term is there but not the "main effect." In other words, you appear to have created the product term, but left out one of its components. It would be easier to respond to this if you were specific about the variables in question. For example, is the IV dichotomous, continuous or something else?

[Meike Wenzl]

First of all, thanks for your reply. I really appreciate it.

I have two regression models: the first DV is continoues, the other DV is dichotomous (and thus a logistic regression), the My IV is continoues and measures the change of ROA over a specific period of time. My moderator variable is created by three different aspects and is summarized in an "index" which is supposed to be standardized later (mean=0 and S.D.= 1)
So the interaction term would be my IV * Index

So if I understand correctly, I need to put my moderator variable into the equation as well in order to find out the main effect, right? And then after that the interaction effect.

Now that you explain it to me it seems so logical. I did not think of the main effect in my model before!

[Nick Cox]

Also cross-posted at http://stackoverflow.com/questions/3...ssion-in-stata

NB http://www.statalist.org/forums/help#crossposting

8. May I cross-post to other forums?

People posting on Statalist may also post the same question on other listservers or in web forums. There is absolutely no rule against doing that.
But if you do post elsewhere, we ask that you provide cross-references in URL form to searchable archives. That way, people interested in your question can quickly check what has been said elsewhere and avoid posting similar comments. Being open about cross-posting saves everyone time.
If your question was answered well elsewhere, please post a cross-reference to that answer on Statalist.

[Dick Campbell]

Yes, you have it correctly. You might want to learn about factor notation in Stata if you haven't done so. You will find that it makes the construction and interpretation of interactions a bit easier and allows you to use Stata's margins command effectively.

[Carlo Lazzaro]

Meike:
the following toy-example may be helpful:

. sysuse auto.dta
(1978 Automobile Data)

. reg price c.weight##i.foreign

      Source |       SS           df       MS      Number of obs   =        74
-------------+----------------------------------   F(3, 70)        =     26.20
       Model |   335885357         3   111961786   Prob > F        =    0.0000
    Residual |   299180039        70  4274000.55   R-squared       =    0.5289
-------------+----------------------------------   Adj R-squared   =    0.5087
       Total |   635065396        73  8699525.97   Root MSE        =    2067.4

----------------------------------------------------------------------------------
           price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-----------------+----------------------------------------------------------------
          weight |   2.994814   .4163132     7.19   0.000     2.164503    3.825124
                 |
         foreign |
        Foreign  |  -2171.597   2829.409    -0.77   0.445    -7814.676    3471.482
                 |
foreign#c.weight |
        Foreign  |   2.367227   1.121973     2.11   0.038      .129522    4.604931
                 |
           _cons |  -3861.719   1410.404    -2.74   0.008    -6674.681   -1048.757
----------------------------------------------------------------------------------

The reference command for creating interaction and categorical variables with Stata is -fvvarlist-.

PS: crossed in the cyberspace with Dick's reply.

[Meike Wenzl]

This was very helpful! Thanks everyone!

[Charles Lim]

Meike:
the following toy-example may be helpful:

. sysuse auto.dta
(1978 Automobile Data)

. reg price c.weight##i.foreign

Source | SS df MS Number of obs = 74
-------------+---------------------------------- F(3, 70) = 26.20
Model | 335885357 3 111961786 Prob > F = 0.0000
Residual | 299180039 70 4274000.55 R-squared = 0.5289
-------------+---------------------------------- Adj R-squared = 0.5087
Total | 635065396 73 8699525.97 Root MSE = 2067.4

----------------------------------------------------------------------------------
price | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
weight | 2.994814 .4163132 7.19 0.000 2.164503 3.825124
|
foreign |
Foreign | -2171.597 2829.409 -0.77 0.445 -7814.676 3471.482
|
foreign#c.weight |
Foreign | 2.367227 1.121973 2.11 0.038 .129522 4.604931
|
_cons | -3861.719 1410.404 -2.74 0.008 -6674.681 -1048.757
----------------------------------------------------------------------------------

The reference command for creating interaction and categorical variables with Stata is -fvvarlist-.

PS: crossed in the cyberspace with Dick's reply.

Hi there, can you shed some light on how do we interpret these results shown in the code? Does this mean that weight moderates the relationship between foreign (IV) and price (DV)? Or does this mean that foreign moderates the relationship between weight (IV) and price (DV)?

[Carlo Lazzaro]

Charles:
foreign moderates the relationship between weight (IV) and price (DV), as foreign is the two-level categorical variable.

However, if you invert the interaction terms, the result (as expected) does not change:

. reg price i.foreign##c.weight

      Source |       SS           df       MS      Number of obs   =        74
-------------+----------------------------------   F(3, 70)        =     26.20
       Model |   335885357         3   111961786   Prob > F        =    0.0000
    Residual |   299180039        70  4274000.55   R-squared       =    0.5289
-------------+----------------------------------   Adj R-squared   =    0.5087
       Total |   635065396        73  8699525.97   Root MSE        =    2067.4

----------------------------------------------------------------------------------
           price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-----------------+----------------------------------------------------------------
         foreign |
        Foreign  |  -2171.597   2829.409    -0.77   0.445    -7814.676    3471.482
          weight |   2.994814   .4163132     7.19   0.000     2.164503    3.825124
                 |
foreign#c.weight |
        Foreign  |   2.367227   1.121973     2.11   0.038      .129522    4.604931
                 |
           _cons |  -3861.719   1410.404    -2.74   0.008    -6674.681   -1048.757
----------------------------------------------------------------------------------

[Charles Lim]

Charles:
foreign moderates the relationship between weight (IV) and price (DV), as foreign is the two-level categorical variable.

However, if you invert the interaction terms, the result (as expected) does not change:

. reg price i.foreign##c.weight

Source | SS df MS Number of obs = 74
-------------+---------------------------------- F(3, 70) = 26.20
Model | 335885357 3 111961786 Prob > F = 0.0000
Residual | 299180039 70 4274000.55 R-squared = 0.5289
-------------+---------------------------------- Adj R-squared = 0.5087
Total | 635065396 73 8699525.97 Root MSE = 2067.4

----------------------------------------------------------------------------------
price | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
foreign |
Foreign | -2171.597 2829.409 -0.77 0.445 -7814.676 3471.482
weight | 2.994814 .4163132 7.19 0.000 2.164503 3.825124
|
foreign#c.weight |
Foreign | 2.367227 1.121973 2.11 0.038 .129522 4.604931
|
_cons | -3861.719 1410.404 -2.74 0.008 -6674.681 -1048.757
----------------------------------------------------------------------------------

thanks for your prompt reply, and what is the rationale for you coming to the conclusion that foreign moderates the relationship between weight and price? which numbers do you look at?

[Carlo Lazzaro]

Charles:
see -i.foreign#c.weight- coefficient.

[Anita Mendi]

Hi all,
I am posting here as my query relates to the above discussion. I run the following regression, with desgcont as dummy independent variable and edi as continuous moderator.

xtreg ltobq1 Tesgscr_w i.desgcont##c.edi lROA lage lMC liq_w TD2TC_w dRD dAdv tangi_w ind_totaldyna ind_totalmuni DPS_w i.year i.country_id , fe vce(robust )


Now, I want to plot margin plot
I am not able to run the following commands-

margins desgcont#edi
edi: factor variables may not contain noninteger values
r(452);

. margins edi, dydx(desgcont)
edi: factor variables may not contain noninteger values
r(452);

However, when I run-
margins desgcont, dydx(edi)

I get the following-

Average marginal effects Number of obs = 14,282
Model VCE : Robust

Expression : Linear prediction, predict()
dy/dx w.r.t. : edi

------------------------------------------------------------------------------
| Delta-method
| dy/dx Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
edi |
desgcont |
0 | -.6098575 .1129018 -5.40 0.000 -.831141 -.388574
1 | -.3946498 .10908 -3.62 0.000 -.6084426 -.180857
------------------------------------------------------------------------------

But my moderator is edi and desgcont as my independent. I want the moderator effect of edi on the observed relationship of desgcont on ltobq1.

Below is my dataset. Kindly advice.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(ltobq1 desgcont Tesgscr_w) double(voice edi) float(lROA lage lMC liq_w TD2TC_w dRD dAdv tangi_w ind_totaldyna ind_totalmuni DPS_w)
         . 0         0                  . .827          .         .         .          .            . 0 0         .            .           .         .
         . 0         0                  . .831          .         .         .          .            . 0 0         .            .           .         .
         . 0         0 1.3148020505905151 .835  -5.642459         . 18.180271          .    .02837359 0 0         .    .05637345         1.5         .
         . 0         0                  . .835          .         . 16.794973          .    .06500474 0 0         .    .01704543    .8312098         .
         . 0         0 1.4140313863754272 .834  -4.390955         .  16.73797          .    .02373406 0 0         .   .008995615    .5619081         0
         . 0         0  1.413554310798645 .833  -3.438129         .  18.49665          .      .004596 0 0         .   .010890448    .3612492         0
         . 0         0 1.4960401058197021 .836 -2.2978895         .  18.79071          .            0 0 0         .   .003705731   .18037914         .
         . 0         0 1.4858338832855225 .839 -1.9644064         . 18.986776          .  .0040276586 0 0         .  .0026540486    .1551205 .22918595
         . 0         0  1.365230917930603 .841 -2.3530607         .  19.33193          .     .4330437 0 0         .    .00305615   .12613218  .2564484
         . 0         0 1.3364784717559814 .841  -2.571729         .  19.88039          .    .24398783 0 0         .   .004074698   .13196088         .
         . 0         0 1.3367778062820435 .841          .         . 18.678877          .     .7068889 0 0         .   .005178873   .14802557         .
         . 0         0 1.3349615335464478 .842  -.8791076         .  19.76866          .    .28773746 0 0         .   .003159609    .1605262  .4117825
         . 0         0 1.2977834939956665 .845 -2.2364776         . 19.881395          .     .2926136 0 1         .   .002436926   .14179726         .
         . 0         0 1.3492107391357422 .843  -2.096955         .  20.00941          .     .4795599 0 1         .  .0033650626    .0974995  .6655486
 1.2690134 0         0  1.393738031387329 .842  -2.972159         .  20.47905          .     .6858291 0 1         .  .0018584323   .05612373  1.692598
  1.345607 0         0 1.4061840772628784  .84 -1.0325607         . 21.156185  .52923864    .28635183 0 1  .3152211  .0022524097   .05058207 2.2065618
  1.264119 0         0 1.4363303184509277 .826  -2.112199         .   21.3717   .7239553    .21535115 0 1 .15090446  .0018318768   .04119962 1.9028004
 1.4363217 0         0    1.4219970703125  .83 -2.5424256         . 21.573254  .17945997     .2079216 0 1   .599713   .002114825   .06152667 1.9570347
 1.0028348 0         0 1.3643749952316284  .83  -3.189864         . 21.579594   .1549644     .3401123 0 1  .6715301  .0015089305   .05741762  1.940408
  .5271635 1 23.584824 1.3898249864578247 .828  -2.150362         .  23.40396  .04732602 .00007016443 0 0  .8647873   .003153265    .0272012 1.9679955
 -.0962189 1  36.63683 1.4234733581542969 .816  -2.508994         . 22.919384 .009618208    .00918066 0 0  .7930642  .0026255785  .018094681 2.0806856
         . 1 32.645992  1.335723876953125 .826  -2.751691         .  22.23365  .00954972    .17612337 0 0  .7940513  .0011389385 -.016569408 .05514017
         . 0         0                  . .713          . 1.7917595         .          .            . 0 0         .            .           .         .
         . 0         0                  . .713          .   1.94591         .          .            . 0 0         .            .           .         .
         . 0         0   1.31016671657562 .701          . 2.0794415 18.505644          .    .11593354 1 0         .    .04589926         1.5         .
         . 0         0                  .  .66          . 2.1972246 19.847227          .    .11053365 1 0         .   .014395848    .8182936         .
         . 0         0   1.34456813335419 .663  -2.847201 2.3025851  18.99193          .    .08129847 1 0         .   .010556856    .5174185         0
         . 0         0   1.33822762966156 .691 -1.7617736  2.397895 19.582664          .    .04647166 1 0         .    .01171166    .3183922         .
         . 0         0     1.327751994133 .693 -3.4877245  2.484907 19.318064          .    .03091357 1 0         .  .0045666723   .15775226         .
         . 0         0   1.29452359676361 .689  -4.973392  2.564949  19.01404          .    .31399545 1 0         .   .005308445   .22216317         .
         . 0         0    1.1016218662262 .694  -2.813088 2.6390574 19.810526          .     .2798795 1 0         .  .0017572914    .2420661         0
         . 0         0   1.10981035232544 .724  -2.727222   2.70805  20.12535          .     .2272032 1 0         .   .001289117   .22166982         .
         . 0         0   1.14426684379578 .725          .  2.772589 19.307425          .     .4097062 1 0         .   .003063147   .15797335         0
         . 0         0   1.09720647335052 .732          .  2.833213 18.940722          .     .3132954 1 0         .  .0010577773   .11550122         0
         . 0         0   1.13360571861267 .741  -3.728397  2.890372 18.963652          .    .24296543 1 0         .   .001322156   .10984387         0
         . 0         0   1.12696576118469 .735  -2.973814  2.944439 18.779066          .    .15316957 1 0         .   .008023923   .20600414         .
  .4770776 0         0   1.16158699989319 .735  -2.790675  2.995732   19.2884          .            0 1 0         .  .0043902183   .27245784         .
  .4905892 0         0   1.10301864147186 .739  -2.920944 3.0445225 19.668385  .01944545  .0018395268 1 0  .6568437   .002461872    .2723597         .
  .3779118 0         0   1.07122313976288 .734 -2.9818125 3.0910425 20.086105  .05566336     .3858664 1 0  .6910126  .0016627316   .25756526         .
  .5633031 1 26.357546    1.1098210811615 .731 -2.6347766  3.135494   19.9729  .05493629    .10735536 1 0  .6709687  .0015439098   .19649473         .
  .5146576 1 24.150455   1.11025941371918 .716  -2.483417  3.178054 20.364466  .27102286     .2777991 1 0 .53408873  .0020736663   .19654918         .
  .5450583 1  29.52991   1.05072391033173 .653 -2.9778395  3.218876  20.35435  .25791353    .24525806 1 0 .52176166   .005047116    .2437232         .
  .5267837 1  26.57213   1.04503977298737 .639 -2.8292165 3.2580965  20.48107  .28514445     .2205516 1 0  .4999703  .0038196896    .2596262         .
         . 1 30.972897   .965248048305511 .632  -2.459822  3.295837 20.657724   .3105567     .1880008 1 0  .5110887  .0015620334   .24842477         .
         . 0         0                  . .713          .         .         .          .            . 0 0         .            .           .         .
         . 0         0                  . .713          .         .         .          .            . 0 0         .            .           .         .
         . 0         0   1.31016671657562 .701          .         .         .          .            . 0 0         .    .05309942         1.5         .
         . 0         0                  .  .66          .         .         .          .            . 0 0         .   .019419976     .785805         .
         . 0         0   1.34456813335419 .663          .         .         .          .            . 0 0         .   .014574052    .4774816         .
         . 0         0   1.33822762966156 .691          .         .         .          .            . 0 0         .     .0201837    .2222209         .
         . 0         0     1.327751994133 .693          .         .         .          .            . 0 0         .  .0016127323 -.025849275         .
         . 0         0   1.29452359676361 .689          .         .         .          .            . 0 0         .  .0018322542   .01032796         .
         . 0         0    1.1016218662262 .694          .         .         .          .            . 0 0         .   .003450294   .07531282         .
         . 0         0   1.10981035232544 .724          .         .         .          .            . 0 0         .   .004824695   .14788388         .
         . 0         0   1.14426684379578 .725          .         .         .          .            . 0 0         .  .0026352406   .16918615         .
         . 0         0   1.09720647335052 .732          .         .         .          .            . 0 0         .  .0029167794   .13762626         .
         . 0         0   1.13360571861267 .741          .         .         .          .            . 0 0         .  .0024172966   .08563828         .
         . 0         0   1.12696576118469 .735          .         .         .          .            . 0 0         .   .000935715   .05790803         .
         . 0         0   1.16158699989319 .735          .         .         .          .            . 0 0         .  .0021943143   .07698382         .
         . 0         0   1.10301864147186 .739          .         0         .          .            . 0 0         .  .0015388993   .09800524         .
         . 0         0   1.07122313976288 .734          .  .6931472         .          .            . 0 0         .  .0014409227   .09807293         .
         . 0         0    1.1098210811615 .731          . 1.0986123         .          .            . 0 0         .  .0011403601    .0909178         .
         . 0         0   1.11025941371918 .716          . 1.3862944         .          .            . 0 0         .  .0008310182    .0888484         .
         . 0         0   1.05072391033173 .653          .  1.609438         .          .     .1870041 1 0         .   .002265614     .111275         .
  2.519135 0         0   1.04503977298737 .639          . 1.7917595         .   .5235299    .56063324 1 0 .14539458  .0020122605   .13706179         .
         . 1 10.341302   .965248048305511 .632          .   1.94591  22.71639   .7000328   .066071294 1 0  .2064338  .0009874864   .13558835         0
         . 0         0                  . .731          .         .         .          .            . 0 0         .            .           .         .
         . 0         0                  . .731          .         .         .          .            . 0 0         .            .           .         .
         . 0         0 1.5319653749465942 .732          .         . 13.348972          .            . 0 0         .   .021230195         1.5         .
         . 0         0                  . .729          .         . 13.734136          .            . 0 0         .   .006771991    .8151245         .
         . 0         0 1.4990538358688354 .723          .         .  13.77733          .            . 0 0         .  .0045399484    .5206969         .
         . 0         0  1.547556757926941 .723          .         .  15.70192          .            . 0 0         .    .00582537    .3010304         .
         . 0         0 1.6684545278549194 .722          .         . 16.910791          .            . 0 0         .  .0012538144   .10914017         .
         . 0         0  1.506879448890686 .721          .         . 17.149038          .            . 0 0         .  .0008339151    .1161453         .
         . 0         0 1.4155948162078857 .729          .         . 17.560898          .            . 0 0         .  .0006658428   .11528158         .
         . 0         0 1.3846715688705444 .723          .         . 17.745195          .            . 0 0         .  .0011086015   .12806036         .
         . 0         0 1.4075285196304321 .723          .         .  16.59871          .            . 0 0         .  .0009239631   .11894116         .
         . 0         0 1.4018456935882568 .723          .         . 17.468992          .            . 0 0         .  .0016502803   .08665743         .
         . 0         0  1.352596402168274 .722          .         . 18.333626          .            . 0 0         .   .001199182   .07347444         .
         . 0         0 1.3801445960998535 .706          .         . 18.427645          .            . 0 0         .  .0007724487   .06402074         .
         . 0         0 1.4375053644180298   .7          .         .   18.2111          .            . 0 0         .   .001211901    .0465131         .
         . 0         0  1.453486680984497 .699          .         .  18.60415          .            . 0 0         .  .0012430664  .030537236         .
         . 0         0 1.4123599529266357 .697          .         .  19.13304          .            . 0 0         .  .0008620067 -.012524412         .
         . 0         0 1.4671263694763184 .695          .         . 19.468285          .            . 0 0         .  .0007204496  -.04434409         .
         . 1 10.876996 1.4451552629470825 .725          .         . 20.359314          .            0 0 0         .   .000525026 -.032052368         .
-.08036155 1 10.449095 1.4763754606246948  .74          .         .  19.95873          .            0 0 0         .  .0012046916  -.01103917         .
-.06953597 1  7.284356 1.5260580778121948 .745          .         .  15.62347 .017312652            0 1 0  .9605203    .00092441   .01580063         .
         . 1  7.621446 1.4625053405761719 .747          .         0 15.795263 .018939046            0 1 0  .9657973  .0004207293  .031324685         .
         . 0         0                  . .713          .         .         .          .            . 0 0         .            .           .         .
         . 0         0                  . .713          .         .         .          .            . 0 0         .            .           .         .
         . 0         0   1.31016671657562 .701          .         .         .          .            . 0 0         .   .030969003         1.5         .
         . 0         0                  .  .66          .         .         .          .            . 0 0         .     .0098953    .8180593         .
         . 0         0   1.34456813335419 .663          .         .         .          .            . 0 0         .   .006222823   .52866066         .
         . 0         0   1.33822762966156 .691          .         0         .          .            . 0 0         .   .007972021    .3118774         .
         . 0         0     1.327751994133 .693          .  .6931472         .          .            . 0 0         . .00057498214   .10250276         .
         . 0         0   1.29452359676361 .689          . 1.0986123         .          .            . 0 0         . .00043363945   .10609336         .
         . 0         0    1.1016218662262 .694          . 1.3862944         .          .            . 0 0         .  .0004123272    .1104177         .
         . 0         0   1.10981035232544 .724          .  1.609438         .          .            . 0 0         .  .0008528903   .11749911         .
         . 0         0   1.14426684379578 .725          . 1.7917595         .          .            . 0 0         .  .0004869801   .11225101         .
         . 0         0   1.09720647335052 .732          .   1.94591         .          .            . 0 0         .  .0006318856    .0948219         .
end

[Anita Mendi]

Hi, after reading past threads, I have tried -

margins desgcont#edi
edi: factor variables may not contain noninteger values
r(452);

. egen newedi = group(edi), label
(3080 missing values generated)

. margins desgcont#newedi
factor 'newedi' not found in list of covariates
r(322);

But the margin commands still not running. I appreciate any help in this regard.