High Correlations

[Jana Schue]

Hello everybody,

I run my -pwcorr- and found a lot of high correleations.
Now my question is do I have a problem with these high correlations and do I have to exclude some of my control variables?
Looking at the Variance inflation factor of my IV no high values are found.
DV: totalplatact
IV: gender_diversity; background_diversity; educational_diversity; tenure_diversity
Controls: total_countries itraffic growth frim_age TMT_size
Moderator: dynamism

. asdoc pwcorr ln_totalplatact educational_diversity gender_diversity background_diversity tenure_diversity dynamism total_countries itraffic growth firm_age TMT_size , obs label star(0.05)  fhc(\b) fhr(\b) sig dec(3) save(Corre
> lations table) replace        

             | ln_tot~t educa~ty gende~ty backgr~y tenure~y dynamism total_~s itraffic   growth firm_age TMT_size
-------------+---------------------------------------------------------------------------------------------------
ln_totalpl~t |   1.0000 
             |
             |      861
             |
education~ty |   0.1519*  1.0000 
             |   0.0000
             |      861      861
             |
gender_div~y |   0.0812*  0.0749*  1.0000 
             |   0.0171   0.0279
             |      861      861      861
             |
background~y |   0.1762*  0.1958*  0.1209*  1.0000 
             |   0.0000   0.0000   0.0004
             |      861      861      861      861
             |
tenure_div~y |   0.2596*  0.0801*  0.0528   0.2409*  1.0000 
             |   0.0000   0.0188   0.1218   0.0000
             |      861      861      861      861      861
             |
    dynamism |  -0.0237  -0.0056  -0.0497  -0.0269  -0.0891   1.0000 
             |   0.6265   0.9090   0.3071   0.5797   0.0665
             |      425      425      425      425      425      425
             |
total_coun~s |   0.6757*  0.1378*  0.1118*  0.1079*  0.1415* -0.0789   1.0000 
             |   0.0000   0.0000   0.0010   0.0015   0.0000   0.1041
             |      861      861      861      861      861      425      861
             |
    itraffic |   0.3985*  0.0201   0.0542   0.0162   0.1166*  0.0139   0.3333*  1.0000 
             |   0.0000   0.5556   0.1120   0.6356   0.0006   0.7756   0.0000
             |      861      861      861      861      861      425      861      861
             |
      growth |  -0.1123* -0.0427  -0.0658  -0.0765* -0.0932*  0.3307* -0.3036*  0.0490   1.0000 
             |   0.0016   0.2302   0.0645   0.0316   0.0087   0.0000   0.0000   0.1683
             |      791      791      791      791      791      408      791      791      791
             |
    firm_age |   0.4216*  0.0654   0.0648   0.1108*  0.1239* -0.1780*  0.5681*  0.2145* -0.5719*  1.0000 
             |   0.0000   0.0550   0.0574   0.0011   0.0003   0.0002   0.0000   0.0000   0.0000
             |      861      861      861      861      861      425      861      861      791      861
             |
    TMT_size |   0.3228*  0.2928*  0.1644*  0.3923*  0.3561* -0.0247   0.2036*  0.1267* -0.0688   0.0770*  1.0000 
             |   0.0000   0.0000   0.0000   0.0000   0.0000   0.6111   0.0000   0.0002   0.0529   0.0238
             |      861      861      861      861      861      425      861      861      791      861      861
             |

Thanks in advance and best regards
Jana

[Carlo Lazzaro]

Jana:
regardless their statistical significance for most of them (that may well depend on the sample size), the correlations do not seem to provide any reason of concern.

[Jana Schue]

Hey Carlo,

okay thanks for your help!

[Jana Schue]

Carlo
To note the sample size is 180 groups with 780 observations.
So it is still not a problem?

Best regards Jana

[Carlo Lazzaro]

Jana:
I still do not think that this is a problem with 180 panels.
I'd invoke cluster-robust standard errors and see the Stata output of your regression tio have a clearer idea.