asdoc: version update - New features of font, table header formatting, etc. Export results to Word from Stata

[Attaullah Shah]

Version 2.3.3.2, dated Feb 23, 2019, is finally here, thanks for Kit Baum for updating it on SSC. This new version brings significant improvements to existing routines and introduces a few new features. Summary of these is given below and details can be found on my webpage here. If you have not used asdoc previously, I would encourage you to read this half page quick start to asdoc.

To update

ssc install asdoc, replace
help asdoc

New Features

---

1. Font style

2. Formatting the header row and header column

3. Revamped the tabulation commands (really elegant tables now) - Watch YouTube Video

4. Revamped the table command (really elegant tables now)

5. Extending the detailed regression tables [ Read further details here]

5.1 Added confidence intervals to the detailed regression tables
5.2 Added an option for customizing the significance starts
5.3 Added an option for suppressing significance stars
5.4 Added an option for suppressing confidence intervals

6. Adding support for macOS

6. Improving the output from proportion command

7. Support added for the logistic family of regressions

8. Improving table outputs of non-standard outputs i.e. multilevel models

9. eform() option added to nested tables




Future Plans

---

It is now almost a year in developing and constantly adding features to asdoc. I think I would stop here and start working on a premium version of asdoc that would be more flexible in making customized tables. It will also be able to export the output tables to Word, Excel, plain text, CSV, and LaTeX formats. I plan to name the premium version as asdocx.
I would appreciate your suggestions and comments here on this post or sent directly to my email address [email protected]

[Farah Ali]

Dear Dr Shah

I am trying to create an output for my regression table using nest command. As my variables are not much, i would like to report t stat next to the coefficients not below them. I am not able to find the way to this. I have seen the wise command but i dont think thats suitable.

[Attaullah Shah]

If you have a single regression, it is better to report the detailed regression table, which is the default in asdoc. To make that table, you do not need to add the wide or the nest option to the regression command. If you need further customization of the detailed regressions table, see this link.

See the following example:

sysuse auto, clear
asdoc reg price mpg rep78 headroom trunk weight length turn

[Pratap Pundir]

Any plans for supporting longitudinal analyses?

[Attaullah Shah]

Pratap Pundir Can you specify which Stata command are referring to?

[Pratap Pundir]

I mean things such as longitudinal regressions (e.g. fixed effects) with xtreg.

The output is exported, but not in publication ready format as promised. For one, in fixed effects models, what's the point of putting a constant in the table? Also, the table should indicate which fixed effects (e.g. in case of World Bank data, often countries and years are included). Similarly, for the R^2, a publication-ready table needs to show the isolated impact of the x variable, separate from the fixed effects. This is kind of what I expect to see when sending results of a fixed effects model to nested Word table (not talking about the aesthetics, but the content):

Finally, it would be nice if asdoc could:
1. Export to docx, instead of doc
2. Send graphs to the Word document. Internal command putdocx already does this, but if I am using asdoc, I don't expect to create a separate document with putdocx just for the charts...
3. Support for UTF8. At the moment, any letters with accents get messed up, and text export gets confused by common text symbols such as ' and )

[Rodrigo Badilla]

Dear Attaullah Shah

Thanks for asdoc!. I am working with ANOVA analysis.
Lamentably when I try to get an anova table with asdoc I just get a regression table with p-value to each level from independient variable and not like stata result.
I wonder if exist some method to get a similar result to Anova table from stata. Example:

webuse lifeexp.dta, clear
asdoc anova popgrowth  region

stata result

                           Number of obs =      68     R-squared     =  0.4345
                           Root MSE      = .710934     Adj R-squared =  0.4171

                  Source |  Partial SS    df       MS           F     Prob > F
              -----------+----------------------------------------------------
                   Model |  25.2441263     2  12.6220631      24.97     0.0000
                         |
                  region |  25.2441263     2  12.6220631      24.97     0.0000
                         |
                Residual |  32.8527857    65  .505427472   
              -----------+----------------------------------------------------
                   Total |  58.0969119    67  .867118089

With asdoc:

Regression results   
 popgrowth
 Coef.
 St.Err.
 t-value
 p-value
 [95% Conf
 Interval]
 Sig

 1b.region
0.000
.
.
.
.
.


 2.region
1.168
0.218
5.35
0.000
0.732
1.604
***

 3.region
1.405
0.249
5.64
0.000
0.908
1.902
***

 Constant
0.525
0.107
4.90
0.000
0.311
0.739
***



Mean dependent var
0.972
SD dependent var
0.931


R-squared
0.435
Number of obs
68.000


F-test
24.973
Prob > F
0.000


Akaike crit. (AIC)
149.508
Bayesian crit. (BIC)
156.166

Thanks in advance
Regards
Rodrigo

[John Riveros]

Greetings Dr Shah, VECM and VAR results cannot be displayed so easily in the current versions. It would be extremly useful to add this feature to asdoc as it could be able to show a decent output from VAR and VECM estimations.

[Attaullah Shah]

Pratap Pundir These options are available in asdoc, you can read section 4. Regressions

help asdoc

In the following examples, I am using option add() to add additional text, keep() to keep only relevant variables and cnames() to change the column headers. `Option fe and re are options of xtreg command

webuse grunfeld
asdoc xtreg invest mvalue kstock , fe add(Year dummies, no) nest replace cnames(FE)
asdoc xtreg invest mvalue kstock i.year , re add(Year dummies, Yes) nest cnames(RE) keep(mvalue kstock )

[Pratap Pundir]

Thank you!
Are the following also already available?

it would be nice if asdoc could:
1. Export to docx, instead of doc
2. Send graphs to the Word document. Internal command putdocx already does this, but if I am using asdoc, I don't expect to create a separate document with putdocx just for the charts...
3. Support for UTF8. At the moment, any letters with accents get messed up, and text export gets confused by common text symbols such as ' and )

[Attaullah Shah]

Pratap Pundir These options will be included in the premium version, called asdocx. For updates, please visit this page
https://fintechprofessor.com/stata-programs/asdocx/

[Pratap Pundir]

asdoc reg d_ln_co2 l(0/2).d_ln_gdp i.year, vce(cluster country) add(Year FE, Yes) drop(i.year) save($wordfile) nest cnames(FD 2Lag)

returns

func_nested_reg(): 3301 subscript invalid
<istmt>: - function returned error

reg d_ln_co2 l(0/2).d_ln_gdp i.year, vce(cluster country)

runs fine.

asdoc reg d_ln_co2 d_ln_gdp i.year, vce(cluster country) add(Year FE, Yes) drop(i.year) save($wordfile) nest cnames(FD)

also runs fine

[Pratap Pundir]

I suspect this has to do more with the

add(Year FE, Yes)

in the nesting situation, than with anything else.
So, I guess my question should be - if I am adding row info, how do I nest the regression tables?

In your example, you have used the exact same options

add(Year dummies, no)

in consecutive nesting regressions. But it's clearly not working in my case.

[Attaullah Shah]

Pratap Pundir Please confirm that you have the latest version. It should be Version 2.3.3.3

which asdoc

will show the version.
If the version is up to date, then please either post a sample data set that produces the error or send it to me at [email protected]
I checked your example on the following dataset and it works fine.

webuse lutkepohl2
tsset
global wordfile Regressions
asdoc reg dln_inv dln_inc L(0/2).dln_consump i.qtr, add(Year FE, Yes) drop(i.qtr) save($wordfile) nest cnames(CS 2Lag) replace
asdoc reg dln_inv dln_inc L(0/2).dln_consump, add(Year FE, No)  save($wordfile) nest cnames(CS 2Lag)

[Pratap Pundir]

Please find attached my code. The data set is from World Bank, being pulled directly.

which asdoc

returns

Version 2.3.3.3: Changes made March 6, 2019