Descriptive statistics table generation

[Lang Ding]

Hello everyone,

I need to generate a comprehensive descriptive statistics table based on a panel table as the picture shows(please see the end of the post, thanks):

An extract of the variables are like the follows:

input long audit_fee float ln_AF double total_assets float ln_TA long net_income float(ln_NI busy chTA big_4)
 1         0    7217  8.884194   2058   7.62949 1   2096 0
 2  .6931472    5121  8.541105    479    6.1717 1      . 0
18  2.890372    4040     8.304   -411         . 1      . 0
20  2.995732   10100  9.220291   4352   8.37839 1      . 0
21 3.0445225   16694  9.722805   8022 8.9899435 1   6594 0
24  3.178054  217040 12.287837  -8403         . 1      . 1
25  3.218876    1907  7.553287    -20         . 1      . 0

I have code like this:

ssc install estout
eststo summstats: estpost summarize audit_fee ln_AF total_asset ln_TA net_income ln_NI busy chTA big_4
eststo big_4: estpost summarize audit_fee ln_AF total_asset ln_TA net_income ln_NI busy chTA  if big_4==1
eststo non_big_4: estpost summarize  audit_fee ln_AF total_asset ln_TA net_income ln_NI busy chTA  if big_4==0
esttab summstats big_4 non_big_4 using descriptive statistics.rtf, replace main(mean %6.2f) aux(sd) mtitle("Total sample" "Big Four clients" "Non-Big Four clients")
esttab summstats big_4 non_big_4 using descriptive statistics.rtf, replace cell("mean sd") mtitle("Total sample" "Big Four clients" "Non-Big Four clients")

Now I can't figure out how to add another statistic, namely the median for each variable.
Could anyone help me out with this one?
Thanks a lot in advance.

Sample descriptive statistics:

[William Lisowski]

It seems to me that

estpost tabstat ..., statistics(median)

would be a starting point.

[Lang Ding]

It seems to me that

estpost tabstat ..., statistics(median)

would be a starting point.

William Lisowski
Thank you very much for your suggestion.
Though it still occurs unclear to me how to insert the median statistics (1)into the descriptive statistics.rtf just as the mean and standard deviation values,and
(2)to calculate the median by subgroups "Big Four clients","Non-Big Four clients" and "Total sample"

As you suggested I added:

estpost tabstat audit_fee ln_AF total_asset ln_TA net_income ln_NI busy chTA big_4, statistics(median)

Stata gives me:

Summary statistics: p50
     for variables: audit_fee ln_AF total_assets ln_TA net_income ln_NI busy chTA big_4

             | e(audit~)   e(ln_AF)  e(total~)   e(ln_TA)  e(net_i~)   e(ln_NI)    e(busy)    e(chTA)   e(big_4)
-------------+---------------------------------------------------------------------------------------------------
         p50 |       348   5.852202   253492.5   12.44308       7199   9.585268          1       6836          1

Would you be so kind and give me a further hint?

Thank you.

[William Lisowski]

I'm not an expert on esttab, but a reading of the help file suggests that to achieve what you want you're going to have to move to using estout, the more powerful command on which esttab is based. But I'm no expert on estout either so perhaps someone else can give some guidance.

[Lang Ding]

William Lisowski
Thank you anyway

[Pedro Coelho]

In case someone needs: the median is given by p50:

esttab using file.xls, collabels("Mean" "Median" "N") cells("mean(fmt(3)) p50(fmt(3)) count(fmt(0))")