Problem exporting esttab tables with special characters to Excel

[Mustafa Hussein]

Hello,

I use esttab to create tables of regression estimates, with the following characters for levels of statistical significance: * for P<0.10; † for P<0.05; and ‡ for P<0.01. My code looks like:

esttab model1 model2 using table3, ///
b ci br transform(@*100 @*100) ///
star(* 0.10 † 0.05 ‡ 0.01) staraux ///
replace csv wide plain ///
keep(x1 x2 x3) ///
mti("Unadjusted" "Adjusted") ///
title("TABLE 3- Adjusted vs Unadjusted Estimates")

This worked perfectly with Stata 13 on a Windows 10 platform with Office 2013. I just recently installed Stata 14 (on a new Dell Latitude), with Windows 10 and Office 2016. My significance characters now are exported as " †" and " ‡ ", instead of " † " and " ‡ ", in the CSV file. I am wondering if this has to do with Office 2016 or Stata 14. Really appreciate your insights.

Thanks!
Mustafa Hussein
UW Milwaukee
Zilber School of Public Health

[Christophe Kolodziejczyk]

I think this is an esttab-related issue. Esttab is a user-written program and I am not sure it has been updated in order to deal with Stata 14's change to unicode. But you could ask the author (Ben Jann).
But at least you can use the command unicode translate in order to get the proper unicode signs.

[Mustafa Hussein]

esttab was updated as of Feb 6, 2016. When I use unicode convertfile, I get the characters to change but it mushes together the coefficient and variable name into one cell.

Code:
unicode convertfile table3_1.csv table3_2.csv, dstencoding(UTF-16)

Any advice is greatly appreciated!