lukestein’s Twitter Archive

…in reply to @causalinf
@causalinf @JulianReif My experience: esttab and friends make it very easy to get to a *nearly* production-level table, with a huge amount available off-the-shelf (e.g., indicate for checkmarks on groups of suppressed regressors). But that last bit of customization can be extremely painful.
Permalink On twitter.com ♻️ 4 Retweets ❤️ 9 Favorites 2019 Feb 17 Mood 0

…in reply to @lukestein
@causalinf @JulianReif 2/3 I find texsave and texdoc require more work to get a minimum viable table (manual LaTeX code, looping, subinstr, …). BUT, customization to get from nearly production ready to production ready is easier than esttab.
Permalink On twitter.com ♻️ 4 Retweets ❤️ 5 Favorites 2019 Feb 17 Mood 0

…in reply to @lukestein
@causalinf @JulianReif 3/3 I.e., costs of using esttab are convex, costs of other approaches I’ve tried tend to be concave. Solution I’ve settled on for most work is to use estout to generate a nearly perfect t3.tex and then where necessary manually edit (w. documentation) to make t3_clean.tex. YMMV.
Permalink On twitter.com ♻️ 4 Retweets ❤️ 5 Favorites 2019 Feb 17 Mood +6 🙂

…in reply to @lukestein
@causalinf @JulianReif 4/3. Remember the old saying, “Workflows are like opinions: everybody has one, and everyone else’s stinks.”
On twitter.com ♻️ 7 Retweets ❤️ 8 Favorites 2019 Feb 17 Mood 0