General:

• Rename main branch
• run-old-versions.md link in deprecation error message 404s

Help file / Options:

• Output options descriptions need to be updated (save, browse)
• The meaning of grpcodes is unclear -- I cannot determine what it does
• "Statistics" options should be listed right after the mandatory options, then, separately, the "data treatment" ones. These are much more important than labelling!
• "F-test" options can be added to the statistics options. What are missing values in F tests? fnoobs should be with the display options.
• What does missminmean() do?
• pairoutput option accepts multiple inputs, such as pairoutput(beta t), which I want to do, but then causes error and returns: option pout_lbl() required without explanation. Recommend to also allow se and ci.

Math:

• Default set of comparison tests (all vs all) is good
• ⚠️ I cannot quite understand what ftest is doing -- is this the omnibus test? If so I would enable it by default.
• ⚠️ SEs and SDs for groups do not display
• pairoutput option could be required, which would remove any issue of default functionality? It could also be called something more intuitive like Stats()
• ⚠️ I am not sure that the nrmd (normalized differences) are correctly calculated. See this link:

The normalized differences approach of Imbens and Rubin (2015) is useful here. This is defined as the difference in means between the treatment and control groups, divided by the square root of half the sum of the treatment and control group variances

• control() option should flip all regression signs (and subtraction order)
• New treatment() option should replace current control() option functionality
• There is no need for a separate fixedeffect() option as long as indicator i. variables can be appropriately included in the covariates() option
• What does covarmissok do? Could this potentially be combined with the requested casewise option somehow, since it appears conceptually similar?
• I would remove the balmiss(), balmissreg(), covmiss(), and covmissreg() options, or at least move them somewhere else ("Advanced"?)... this is quite dangerous....

Display:

• rowvarlabels behavior should be the default, with the option to disable
• Table notes should explain the meaning of non-numerical results like .v.
• The default paired t-tests should also report their SEs by default, with the option of instead producing P-values and, possibly, confidence intervals. These options need to be updated in the help file in any case:

The options nottest, normdiff, pttest, pftest and pboth have been deprecated as of version 7 of iebaltab. See if the
    options pairoutput or ftestoutput have the functionality you need. If you still need the old version of iebaltab,
    for example for reproduction of old code, see this guide for how to run old versions of any command in the ietoolkit
    package.

Misc:

• onerow returns an unhelpful error if the number of observations varies across regressions. The follow, for example, occurs, but the number of obs in the regressions is not reported in the table without the command, so the source of the failure can't be found.

iebaltab price mpg rep78 , grpv(rep78) savexlsx("test.xlsx") replace control(2) total grpc onerow

Option onerow may only be used if the number of observations with non-missing values are the same in all groups
    across all balance variables. This is not true for group(s): [2_1 2_3 2_4 2_5]. Run the command again without this
    options to see in which column the number of observations is not the same for all rows.