How to generate a table with means and medians by group, their difference, as well as significance levels?

[David Tsu]

I want to replicate the following table:

where the value in each cell is mean (and median in the brackets below it). I tried to generate the first two columns with

table log_mv_assets log_bv_assets log_sales log_1plus_age profit_margin ppe2bv_assets ///
mkt2book rd2sales ad2sales equity2asset vol_asset access, contents(mean median)

however, the response was "too many variables specified". For the third column, what I could do was making the test with commands -ttest- for means and -median- for medians. Is it possible to create the whole table efficiently? Some example data (with only 3 variables) is pasted:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(access log_mv_assets log_sales)
0   2.98085  3.592038
0  2.820389  3.620494
0  3.083743  3.490672
1  5.932678  5.697738
1   6.27632  5.851168
1   6.52489  6.007239
1  6.292146  6.097794
1  6.017305  6.145348
1  5.999619  6.046561
1  5.964461   5.94746
1  6.124185  6.010664
1  6.157353  6.112343
1  6.376326  6.224539
1  6.713152  6.378983
1  7.004667  6.662012
1  6.848636  6.822237
1   6.65363  6.931797
1  6.604878  6.773372
0  3.273743  2.958342
0 3.2325025  2.972975
0  3.229954  3.347058
0 3.6754775  3.594184
0  3.719142  3.701845
0  3.994174 3.5896666
0 4.2457056  3.939774
0  4.770752  4.281654
0  5.003685 4.5231137
0 1.9656384  1.541373
0 1.3545295 1.0009995
0 1.9495597  .9512717
0  2.528778 1.1899753
0  2.466764  1.290609
0 4.7601376 1.8702625
0  4.732373 2.8757656
0 2.3755684  3.175927
0  2.585845  3.311783
0  2.903041  3.556519
0 2.9565756  3.808683
0  5.293977  4.967568
0  5.466451  5.117173
0  5.183933  5.192134
0  5.466319  5.280092
0  5.580708   5.55992
0   6.03099  5.683032
0  6.313433   5.75731
0  6.975975  5.902956
0   7.23284  6.106432
0   7.86315  6.373699
0    8.4355  6.719024
0   8.43554  7.060004
0  8.159143  7.229605
0  8.942077  7.563692
0  9.771268  8.098004
0  2.906737 1.8870697
0  3.470086  4.166743
0   3.58313  4.188123
0 4.4515176  4.685662
0  4.539479  4.784462
0  4.630643  4.919959
0 4.6616216  4.797104
0 4.5151114  4.971603
0  4.423924  4.942442
0   4.37963  4.731803
0 4.3695097   4.82133
0 4.5637465  4.842453
0 3.3484645  2.762286
0  3.329737  2.911263
1  6.688932   6.00607
1  6.875176  6.191094
0  3.108257  2.823638
0 3.1547825  2.977772
0  3.599534  3.769768
0  3.553103  3.816613
0  3.538588  3.877266
0 3.2033534 3.6384015
0  3.165839  3.644353
0 2.9962196 3.6443264
0  2.745709 3.4178905
0  2.547084 3.2804976
0  3.200126 3.5981064
0 3.2984054  3.612241
0 2.9659944  3.519662
1  5.597262  5.944103
1   5.82066  6.208045
0  3.516366  3.234985
0  3.890677  3.403428
0  4.079375  3.867925
0  3.997347  3.872221
0  3.826114  3.512828
0 4.7372813  4.817228
0  5.417011  5.068458
0  5.573306  5.465872
1  5.704102  5.584593
1  5.891969  5.618432
1  6.122792  5.671741
1  6.029373  5.687355
0  6.051079  5.823726
0  6.524072  6.151164
0  6.659762  6.255524
0   6.43774  6.186587
end

[Tiago Pereira]

It would require a bit more of work. I usually accomplish that by writting small do scripts. I don't think there is a command in Stata to do that easily.