GitHub |
您所在的位置：网站首页 › stata ssc install estout › GitHub |
|
estout
The estout package provides tools for making regression tables in Stata. The package currently contains the following programs: esttab: Command to produce publication-style regression tables that display nicely in Stata's results window or, optionally, are exported to various formats such as CSV, RTF, HTML, or LaTeX. esttab is a user-friendly wrapper for the estout command. estout: Generic program to compile a table of coefficients, "significance stars", summary statistics, standard errors, t- or z-statistics, p-values, confidence intervals, or other statistics for one or more models previously fitted and stored. The table is displayed in the results window or written to a text file for use with, e.g., spreadsheets or LaTeX. eststo: Utility to store estimation results for later tabulation. eststo is an alternative to official Stata's estimates store. Main advantages of eststo over estimates store are that the user does not have to provide a name for the stored estimation set and that eststo may be used as a prefix command. estadd: Program to add extra results to the returns of an estimation command. This is useful to make the the results available for tabulation. estpost: Program to prepare results from commands such as summarize, tabulate, or correlate for tabulation by esttab or estout. To install the estout package from the SSC Archive, type . ssc install estout, replacein Stata. Stata version 8.2 or newer is required. Installation from GitHub: . net install estout, replace from(https://raw.githubusercontent.com/benjann/estout/master/)Main changes: 09apr2023 estout.ado (3.32); ettab.ado (2.1.2) - Windows network paths starting with a double backslash (\\) did not work; this is fixed 10feb2023 estpost.ado (1.2.2) - option elabels() in -estpost correlate- now stores additional info in e(labels) and e(eqlabels) (if option -matrix- has specified); option -nolabel- changes behavior of -elabels()- 10jun2022 esttab.ado (2.1.1) - option -standalone- added 26apr2022 estout.ado (3.31) - when writing RTF format, the horizontal rule between table header and table body was missing if the table body contained only a single physical row; this is fixed 25mar2022 estout.ado (3.30) - estout now internally renames unnamed equations to "__" (rather than Stata's default "_"); this fixes a number of issues (e.g. the issues addressed in v3.25 and v3.26 using a different approach) and it also gives the user more control when addressing coefficients in options such as keep(), drop(), rename() etc.; specifying "_:coef" (or "__:coef") now specifically refers to "coef" in unnamed equations; in the old behavior "_:coef" addressed "coef" in any equation (i.e. same as specifying "coef" without "_:") and there was no possibility to address "coef" in unnamed equations only - rename() did not work with names containing spaces; this is fixed - refcat() did not work with names containing spaces; this is fixed - -unstack- produced erroneous results if equations were not in order; this could happen if order() was used in a way such that the coefficients from an equation were divided into multiple sets interrupted by other equations; the problem is now fixed by enforcing ordered equations in case of -unstack- - when using -unstack-, models for which the first element of cells() was not available were suppressed; this is fixed 24mar2022 estout.ado (3.29) - the unicode translator introduced in 3.27 chopped lines after 200 characters; this is fixed - blanks in coefficient names lead to erroneous results when more than one model was tabulated; this is fixed (Stata 11 or newer only) 24mar2022 estout.ado (3.27) - in Stata 14 or newer, a subroutine is now called that escapes non-ASCII characters if the output format is RTF (the characters are translated to \u#?, where # is the base 10 character code); option -nortfencode- prevents this behavior 22mar2022 estout.ado (3.26) - now using a different approach to fix the order()/unstack issue; this also fixes the second issue in https://github.com/benjann/estout/issues/33 21mar2022 estout.ado (3.25) - order() together with -unstack- lead to erroneous arrangement of the table if there were unnamed equations; this is fixed (see https://github.com/benjann/estout/issues/33) 19may2021 esttab.ado (2.1.0) - added support for (Multi)Markdown through option -md- or -mmd- (or file suffix .md or .mmd); md and mmd are synonyms; the changes are based on the suggestions made by Emanuele Bardelli (see https://github.com/benjann/estout/pull/24) estpost.ado (1.2.1) - pweights and iweights are now allowed by -estpost gtabstat-; proposed by Kye Lippold (see https://github.com/benjann/estout/pull/27) 30apr2021 estout.ado (3.24) - the value of the previous statistic was reported instead of missing, if the model p-value was requested, but the model contained no (nonmissing) F or chi2 value; this is fixed - the last format was used for all CIs if multiple formats were specified in cells(ci(fmt())); this is fixed 17apr2020 - installation files added to GitHub distribution 08aug2019 estout.ado (3.23) and estadd.ado (2.3.5) - changes from 07aug2019 undone because StataCorp will change Stata 16 07aug2019 estout.ado (3.24) and estadd.ado (2.3.6) - Stata 16 introduced support for abbreviations in colnumb()/rownumb(); this new feature could cause problems in estout and in some estadd functions; estout and the relevant estadd functions now include code to enforce old behavior of colnumb()/rownumb() 13jun2019 estpost.ado (1.2.0) - estpost gtabstat added by Mauricio Caceres Bravo 31may2019 estout.ado (3.23) - estout crashed if a coefficient name contained an apostroph ('); this is fixed - style() caused error if used with -estout matrix()-; this is fixed (thanks to Daniel Bela) - indicate() did not work correctly if there were repeated models; this is fixed 01mar2019 estpost.ado (1.1.9): - -svy:tab-: e(lb) and e(ub) were missing in cases where -svy- did not provide e(df_r); this is fixed - computation of p-values and CIs failed if degrees of freedom were larger than 2e17 (or lager than 1e12 in Stata versions < 10) (can happen with -mi estimate-); this is fixed - @span in the erepeat() suboption returned wrong result if labcol2() was specified; this is fixed - element-specific drop()/keep() did not work as expected if coefficient names contained spaces; this is fixed 06feb2016 estout.ado (3.20) - if estout is called without without namelist and the active results are not among the estimates stored by eststo, estout now displays a note (suggested by Nils Enevoldsen) estadd.ado (2.3.5) - the 28jan2016 update of Stata 14 broke estadd.ado; this is fixed (the update caused estadd to remove e(_estimates_name) from the active estimates; a consequence of this was that results were no longer added to the stored copy of the estimates if estadd was applied more than once) esttab.ado (2.0.9) - when tabulating a matrix with more than one column in tex format, only one column was defined in the tabular environment; this is fixed - the alignment() option now allows multiple-column shorthand when tabulating a matrix in tex format; for example, type alignment(*{5}{l}) if the matrix has 5 columns that should all be left aligned - the definition of the tabular environment in tex format did not always contain the correct number of columns if the cells() option was used; this is improved 16jun2015 estadd.ado (2.3.4) - -estadd local- now uses macval() to suppress macro expansion (only when applied to active estimates and not for the display of the results) estpost.ado (1.1.8) - now uses macval() to suppress macro expansion in labels; affects subcommands -tabstat-, -tabulate-, -svy:tabulate-, and -stci- - subcommands -tabstat-, -tabulate-, -svy:tabulate- and -stci- now have an -elabels- option to enforce saving labels in e() - subcommand -tabstat- now uses the actual by() values (instead of 1, 2, 3, ...) if a numeric by() variable is specified and labels are saved in e(labels) 02jun2015: estout.ado (3.19) - estout now supports Unicode in Stata 14 esttab.ado (2.0.8) - added option -[no]float- to suppress/enforce table float environment in LaTeX 20mar2015: estout.ado (3.18) - cell contents longer than 245 characters was chopped off by -file write- when writing the table cells; this is fixed - computation of p-values and CIs from -mi- failed for coefficients for which e(df_mi) was missing; this is fixed (using normal approximation) esttab.ado (2.0.7) - added option fonttbl() to set font table in RTF 02jun2014: estout.ado (3.17) - estout now compiles labels for levels of factor variables and interactions - new option -interaction()- to specify delimiter for interactions - new options -[no]omitted- and -[no]baselevels- to specify whether omitted coefficients and base levels are included; the default is to include these coefficients - the "o." and "b." flags are now removed from coefficient names - estout no longer returns error if models use different base levels esttab.ado (2.0.6) - defaults for interaction() added estout_mystyle.def - entries for -omitted-, -baselevels- and -interaction- added 30may2014 esttab.hlp: - clarified that explicit output format option will override format inferred from file suffix 30may2014 estout.ado (3.16) - estout now picks up e(df_mi) for the computation of confidence intervals and p-values after -mi estimate, post- - estout crashed on models with equation names containing "."; this is fixed - estout now works on Small Stata - estout no longer breaks on wide rtf tables estadd (2.3.3) - the -outcome()- and -split- options in -estadd prchange- did no longer work; this is fixed estpost (1.1.6) - -estpost correlate- did not work in Stata 8; this is fixed 10oct2009 estadd (2.3.1) - -estadd margins- added - -adapt- option of -estadd listcoef- discontinued estpost (1.1.5) - -estpost margins- added - -estpost svy: tab- now correctly returns the subpopulation number of observations in e(obs) if -subpop()- is specified. 06aug2009 estout (3.13): - estout no longer chokes on factor variables in Stata 11 09apr2009 estpost (1.1.3): - -estpost tabstat- now allows -columns(stats)- 25mar2009 estadd (2.2.8): - support for -prvalue- and -asprvalue- from SPost added - -estadd prchange- now adds outcome-specific results and average results by default; new -avg- option to store only average result and -noavg- to omit the average results; -adapt- option that was used in earlier versions to add the outcome-specific results discontinued 11mar2009 estout (3.12): - "z" can now be used as a synonym for "t" in cells() - cells(t(abs)) returned error; this is fixed esttab (2.0.5): - option -z[(fmt)]- added (z statistics) estadd (2.2.4): - improved formulas for coxsnell and nagelkerke (old formulas likely returned 1 in large datasets) estpost (1.1.2): - -estpost tabstat- displayed wrong results (in each column the results from the last column were displayed; returned results were not affected); this is fixed 24feb2009 estpost (1.1.1): - -estpost tabstat, statistics(semean)- crashed in some situations; this is fixed (thanks to Matthew Fiedler for reporting the bug) - -estpost tabstat- now uses label "semean" instead of "se(mean)" - -estpost tabstat- could crash if variables or statistics were repeated; this is fixed 16feb2009 esttab (2.0.4) - -longtable- option added 10feb2009 estout (3.11): - -order()- now adds extra table rows for elements not found in the table - r(coefs) now respects -order()-, -keep()-, and -drop()- 21jan2009 estout (3.10): - missing values (except .y and .z) in table cells are now treated like any other values estadd (2.2.3): - -estadd local- returned error if the specified string contained certain special characters; this is fixed estpost (1.1.0): - support for -svy: tabulate- added - string variables are now allowed with -estpost tabulate- - -estpost tabulate- returned a wrong e(sample) in some situations; this is fixed 08jan2009 estout (3.09): - options in estout.hlp are now fully linked - no longer wrapper for -estimates table-; now incorporates code from est_table.ado; still uses "undocumented" -mat_capp- => can now tabulate models that do not contain any coefs (e.g. stcox without predictors) => name space for coefficients is no longer restricted to names in e(b) => can now tabulate estimates without e(b) or e(V) (e.g. -factor-) => rule for merging e()-vectors: coefficients from e()'s without equations are merged into existing equations (without expanding the name space) if at least one e() does not contain equation "_" and no e() contains equation "_" along with other equations (and e(b) does not contain equation "_"); else: match by equation and expand name space if necessary => returns results in r() (r(coefs), r(stats), r(cmdline), etc.) => new rename() option to rename/match coefficients => el[#] and el[rowname] or el["rowname"] now possible to choose row of e(el) to be tabulated (support for el_# syntax discontinued) => cells(): new -transpose- suboption in -cells()- => new 'var' element (variance) available in -cells()- => existing e(se), e(var), e(t), e(p), e(ci_l), e(ci_u) now take precedence over internal computations (unless -margin- applies to model) - -estout matrix()-, -estout e()-, and -estout r()- now available to tabulate a matrix - new eqlabels(,merge) option => merge eqlabel into varlabel - mutiple values may now be specified in modelwidth() - cells(): parentheses can now be used to bind statistics in a row - -nolabel- option in refcat() added - an individual model can now be included multiple times in a table - now uses abs() in formula to derive SEs of transformed coefficients - now adds e(cmd)="." to current estimates if undefined (so that results without e(cmd) can be tabulated) - no longer error if all coefs dropped - fmt() can now be abbreviated as f() - "\\" was not preserved in model titles taken from label of depvar; this is fixed - @rtfrowdef now generally available (undocumented) - fixed parsing of -cells- from defaults files - fixed bug related to cells(ci(star)) - @modelwidth variable now available in par() suboption in cells() (undocumented) - repeated elements in cells() now have their own set of suboptions - coefficients containing blanks now allowed - fixed alignment bug related to statistics with stars and pattern() - in wide mode, empty columns in equations (due to keep/drop cell-subopts) are now suppressed esttab (2.0.3) - -esttab matrix()-, -esttab e()-, and -esttab r()- now supported - no longer returns r(estout) since -estout- now returns r(cmdline) - -note()- now replaces the standard note - @starlegend is no longer default if user specifies -cells()- - -gaps- is no longer default if user specifies -cells()- eststo (1.1.0) - new -prefix()- option estadd (2.2.2) - support for -prvalue- and -asprvalue- from SPost added (not documented yet) - -estadd prchange- returns results differently; -add(all)- no longer supported; -add()- renamed to -pattern()-; type of main results now returned in e(pattern) - now allows break key in subroutines estpost (1.0.0) - new -estpost- command added to package 09apr2008 estout (2.86) - user parameter statistics in cells() were suppressed under some circumstances if the -equations()- option was specified. This is fixed. eststo (1.0.9) - the -nocopy- option did not work. This is fixed. - parsing in -addscalars()- has been improved (now binds parentheses/brackets and allows quotes) - the -noesample- option returned error if a variable with the same name as the estimation set existed in the dataset. This is fixed. 18feb2008 estout (2.85) - output is now suppressed if -using- is specified - new -[no]outfilenoteoff- option (undocumented) - -style(smcl)- is now default unless -using- is specified - new -[no]smclrules-, -[no]smclmidrules-, -[no]smcleqrule- options (undocumented) - new -note()- option (works like -title()-) - model label row is now suppressed if only the active (unstored) model is tabulated - -varlist()- now takes precedence over the labels generated by -eqlabels(none)- - bug fixed related to the -vacant()- and -unstack- - -se- was sometimes printed as "0" and sometimes as "." if the variance was zero; it is now always printed as "." esttab (1.4.0) - user provided significance symbols (-star()-) are now formatted depending on mode - -addnotes()- is now printed before -legend-; -note()- added eststo (1.0.7) - -eststo drop- now allows wildcards - -eststo dir- now shows list of stored estimates estadd (2.1.4) - new -quietly- option - SPost subcommands added (-fitstat-, -listcoef-, -prchange-, -brant-, and -mlogtest-) - -estadd- now passes through the caller version - lists of added results are now printed (unless colon syntax is used) - output was not suppressed in colon syntax if only one name was specified; this is fixed 29oct2007 esttab (1.3.8) - -title()- and @hline were printed on the same line in -fixed- mode; this is fixed 27sep2007 estout (2.79) - "&" can now be used in -cells()- to join parameter statistics together in one cell - repetitions of elements in -cells()- are now possible - "." can be used in -cells()- to insert empty cells - -fmt()- now allows "%g" and "g" for "%9.0g" and "a" for "a3"; invalid formats (i.e. if not starting with "%") now return error - e()-matrices with multiple rows can now be addressed using name_# in -cells()- - new -topfile()- and -bottomfile()- options - new -labcol2()- option - bug related to -extracols()- and -mgroups()- fixed - -abbrev- does no longer erase "." anymore esttab (1.3.7) - new -oncell- option estadd (2.1.0) - -estadd matrix- now also accepts matrix expressions (e.g. estadd matrix M = A*2) - -estadd matrix- can now be used as -estadd matrix matname- or -estadd matrix r(matname)- (adds e(matname)) - -estadd scalar- can now be used as -estadd scalar scalarname- or -estadd scalar r(scalarname)- (adds e(scalarname)) - -estadd r(name)- now adds, depending on the nature of r(name), scalar or matrix e(name) - -copy- option in -estadd matrix- discontinued - "invalid subcommand" error message added eststo (1.0.6) - now sets e(cmd) to "." if undefined 30aug2007 estout (2.73) - the -relax- suboption in options such as -drop()-, -keep()-, etc. is now documented - -mlabels()- now has a -[no]title- option - -eqlabels("",none)- did not work as advertised; this is fixed - the -none- label suboption can now be specified without comma (e.g. -eqlabels(none)- instead of -eqlabels(,none)-) - new options: -starkeep()-, -stardrop()- - the text specified in -prehead()- etc. and in the -begin()- and -end()- label suboptions is now printed on a single line if it does not contain double quotes - tab characters in defaults files are now allowed - empty table rows did not display correctly in older RTF readers; this is fixed esttab (1.3.5) - -mtitles[()]- now uses estimates titles (if defined) even if -label- is not specified - -addnotes()- now prints just one line if the string does not contains double quotes eststo (1.0.5), _eststo(1.0.3) - the caller version is now passed through estadd (2.0.8) - most functions now support variables with time-series operators 08aug2007: estout (2.70) - inclusion of string e()-macros is now possible in the -stats()- option - Stata 10 uses "e(estimates_title)" instead of "e(_estimates_title)"; this is now recognized - -elist()- and -blist()- in -varlabels()- can now write multiple lines - some bugs fixed esttab (1.3.4) - -cells(none)- now produces better tables - not all -estout- options were always passed through correctly; this is fixed 30may2007: esttab (1.3.2) - bug fixed related to equation lines in -fixed- mode - the -plain- option does not suppress the column labels anymore 22may2007: estout (2.64) - "too many literals" bug fixed - new -wrap- option - the -stats()- option now has a -layout()- and a -pchar()- suboption - first element in -cells()- now sets the default display format (in earlier versions, the default format was set by the -b- element) - the -nofirst- and -nolast- suboptions in -varlabels()- now apply equationwise - new @rtfrowdefbrdrt and @rtfrowdefbrdrb variables (undocumented) - the label_subopts -elist()-, -blist()-, -begin()-, -end()- now also support the @M, @E, @width, @hline, and @rtfemptyrow (undocumented) variables - new -[no]first- label_subopt - new -replace- label_subopt to overwrite -begin()- and -end()- esttab (1.3.0) - -esta- renamed to -esttab- - lines between equations are now supported; new -noeqlines- option eststo (1.0.4) - -esto- renamed to -eststo- 13apr2007: estout (2.58) - robust parsing of cells() if there are spaced between elements and their options - new -cells()- suboption called -pvalue()- esto (1.0.3) - -by- prefix command now supported - new -missing- option to be used with by prefix 16mar2007: estout (2.56) - bug fixed related to -cells(..(drop()))- with -unstack- - new elements for -cells()- called -_star-, -_sign-, and -_sigsign- - bug fixed related to -transform()- - new -[no]asis- option (undocumented) - new -[no]smcltags- option (undocumented) - -style(smcl)- now includes horizontal lines - new -cells()- suboption called -vacant()- - -prefix()- and -suffix()- in -varlabels() now also applies to -refcat()- and -indicate()- - bug fixed related to spaces in -mlogit- equations esta (1.2.6) - improved smcl mode - LaTeX column specifiers are now compiled if -cells()- is specified (-stardetach- not supported) - column labels are now printed if -cells()- is specified estadd (2.0.6) - e(_estimates_name) is now backed up in e(_estadd_estimates_name) while working on the estimates set 23feb2007: estout (2.50) - new @width variable - prefix() and suffix() are now added to labels *after* abbreviation esta (1.2.4) - new -smcl- mode; -smcl- is the new default - bug fixed related to leading spaces in labels in -scalars()- labels (length of the labels is now restricted to 80 or 244, depending on Stata flavor) 04feb2007: estout (2.49) - new -dropped()- option esta - ** was used instead of *** for p |
| CopyRight 2018-2019 办公设备维修网 版权所有 豫ICP备15022753号-3 |