This page summarizes the models, features, and bug fixes that were introduced in each release of stargazer. Please note that I am always looking for comments and suggestions. Do not hesitate to contact me at mhlavac [at] alumni.princeton.edu.
v. 5.2.2:
- very minor update: removed instances in code where if/while statement is used with a condition of length greater than one
v. 5.2.1:
- very minor update: updated author contact information
v. 5.2:
- New models: -- arima (stats) -- censReg (censReg) -- garchFit (fGarch) -- glmrob (robustbase) -- lme (nlme) -- nlme (nlme) -- mnlogit (mnlogit) -- pgmm (plm)
- New features:
-- decimal mark (argument decimal.mark) now gets automatically updated in notes (argument notes)
- Fixed bugs:
-- felm (lfe) working again
-- fixed formatting issues that led to an occasional inability to produce ASCII text output
-- much faster output of multi-model tables with many omitted variables
-- negative numbers are now properly displayed when initial.zero is set to FALSE
-- fixed minor bug that occasionally led to a misalignment of additional lines (argument add.lines)
-- lm() and glm() models now report AIC and BIC
-- underscores are now properly formatted in column and dependent variable labels
-- fixed an issue with the mlogit model that led LR tests to crash the output when no intercept was present
v. 5.1:
- New features:
-- the CRAN description now lists packages whose usefulness is enhanced by stargazer
- Fixed bugs
-- labels that indicate whether variables have been omitted from the model are now correctly assigned (argument omit.labels)
-- argument summary.stat now functions properly
-- out.header and header can both be set to false at the same time
-- minor formatting changes
v. 5.0:
- New models: -- felm (lfe)
- New features:
-- ability to produce HTML/CSS output (argument type)
-- vector and matrix objects are now acceptable (argument ...)
-- additional lines can be included in the tables (argument add.lines)
-- table layout is now fully customizable (arguments table.layout and omit.table.layout)
-- choice over which summary statistics should be reported (arguments summary.stat and omit.summary.stat)
-- users can choose whether and the order in which to report coefficients, standard errors, confidence intervals, test statistics and p-values (argument report)
-- object names can be reported above each column (argument object.names)
-- option to toggle the reporting of dependent variables and model names across several columns (argument multicolumn)
-- reporting of model statistics' degrees of freedom can now be suppressed (argument df)
-- users can now include more than three statistical significance cutoffs (argument star.cutoffs)
-- regression, summary statistic and data frame table can now be flipped (argument flip)
-- colnames and rownames can be included in, or excluded from, data frame content tables (arguments colnames and rownames)
-- argument summary can now have a different value for each table within a single stargazer() call
-- if arguments coef, se, t, p, ci.custom and add.lines only contain a single vector, they will be accepted as though they were a list containing the same vector
- Fixed bugs:
-- fixed bug that caused the number of observations to be misreported in complicated models
-- fixed bug that led to the omission of coefficients from output when multiple models had a single regressor
-- notes now working well for summary statistics and data frame tables
-- font size can now be change even if floating environment is not used
-- fixed compatability issues with latest version of Zelig
-- argument apply.ci now works properly
-- minor formatting changes
v. 4.5.3:
- New features:
-- model objects can now be passed to stargazer in a list (argument ...)
- Fixed bugs: -- fixed problem with printing negative numbers in data frame and summary statistics tables -- corrected formatting of column widths in ASCII text output -- minor bug with lme4 covariate names corrected
v. 4.5.2:
- Fixed bugs: -- much faster output -- now compatible with the updated lme4 package -- corrected a bug that prevented users from including multiple models with a single covariate + intercept -- underscores now print correctly in ASCII text output
v. 4.5.1:
- Fixed bugs: -- fixed a bug involving the calculation of t-statistics from user-given coefficients and standard errors
v. 4.5:
- New models: -- binaryChoice (sampleSelection) -- brglm (brglm) -- coeftest (lmtest) -- heckit (sampleSelection) -- maBina (erer) -- mclogit (mclogit) -- mlogit (mlogit) -- selection (sampleSelection)
- New features:
-- explanatory variables can be ordered in customizable ways (argument order)
-- custom functions can now be applied to coefficients (argument apply.coef), standard errors (argument apply.se), test statistics (argument apply.t), p-values (argument apply.p), and confidence intervals (argument apply.ci)
-- keep explanatory variables and statistics based on regular expressions (argument keep and keep.stat)
-- users can customize confidence intervals (argument custom.ci)
-- ability to have both standard errors and confidence intervals in the same table (argument ci is now a logical vector)
-- different confidence levels for each individual column (argument ci.level is now a vector)
-- ability to omit and keep variables based on their position in the table
- Fixed bugs: -- citation and change log updates -- minor formatting changes
v. 4.0:
- New models: -- aftreg (eha) -- bj (rms) -- coxreg (eha) -- cph (rms) -- dynlm (dynlm) -- errorsarlm (spdep) -- Glm (rms) -- Gls (rms) -- gmm (gmm) -- hetglm (glmx) -- lrm (rms) -- mlreg (eha) -- lagsarlm (spdep) -- ols (rms) -- phreg (eha) -- psm (rms) -- rem.dyad (relevent) -- rq (quantreg) -- Rq (rms) -- weibreg (eha)
- New features:
-- the package can produce ASCII text output, in addition to LaTeX code (argument type)
-- output directly to .tex or .txt files (argument out)
-- column labels (arguments column.labels and column.separate)
-- confidence intervals (arguments ci, ci.level and ci.separator)
-- coefficients and standard errors/confidence intervals can now be reported in the same row (argument single.row)
-- users can choose to omit all empty lines in a table (argument no.space)
-- notes can now be appended to, rather than always replace, the default notes for a given style (argument notes.append)
-- ability to customize the dependent variable caption (argument dep.var.caption)
-- font size can now be changed (argument font.size)
-- comments header (with package and author name, version, date and time) can now be suppressed (argument header)
-- ability to change or disable the floating environment (arguments float and float.env)
-- table placement settings (argument table.placement)
-- customization of column spacing (argument column.sep.width)
-- perl-compatible regular expressions (argument perl)
- Fixed bugs:
-- all columns are now displayed for multinom objects
-- better positioning of tables in the LaTeX document
-- minor formatting changes
v. 3.0.1:
- Fixed bugs: -- corrected an issue that led to problems when printing data frames with multiple decimal places -- some reporting and formatting changes to summary statistics and data frame tables
v. 3.0:
- New models: -- clm (ordinal) -- clogit (survival) -- ergm (ergm) -- glmer (lme4) -- gls (nlme) -- lmer (lme4) -- lmrob (robustbase) -- nlmer (lme4) -- pmg (plm) -- rlm (MASS)
- New features:
-- users can customize coefficients (argument coef), standard errors (se), test statistics (t), and p-values (p)
-- automatic calculation of z-scores and p-values when the user supplies custom standard errors (can be toggled using arguments t.auto and p.auto)
-- ability to set \label{} markers in TeX for each table (using argument label)
-- summary statistics table can now report logical (i.e., dummy, indicator) variables as if they had values of 0 (corresponding to FALSE) and 1 (TRUE). See argument summary.logical.
-- user can choose between rounding to a set number of decimal places (argument digits) and reporting all available decimal places (by setting digits equal to NA)
-- can omit all test statistics
- Fixed bugs: -- package runs much faster -- corrected p-values for polr() and zelig oprobit, ologit models -- coxph now, by default, reports robust standard errors, for consistency with its summary() output -- multinom models now report the first set of coefficients (when multiple sets are present), instead of producing an error -- singularities in model regression no longer disrupt the output -- model formulas can now be symbols -- no more warnings when creating summary statistics tables -- minor formatting changes
v. 2.0.1:
- Fixed bugs: -- stargazer now reports the correct number of observations for plm() models
v. 2.0:
- New models: -- betareg (betareg) -- hurdle (pscl) -- ivreg (AER) -- multinom (nnet) -- plm (plm) -- tobit (AER) -- zeroinfl (pscl)
- New features:
-- direct output of data frames into LaTeX (summary = FALSE)
-- ability to omit the reporting of selected statistics (using argument omit.stat)
-- alignment of columns at the decimal mark (align = TRUE)
-- automatic coordination of star cutoffs and values in regression table notes
- Fixed bugs:
-- argument digits now works properly
-- stargazer updated to reflect, and work properly with, recent changes to Zelig
-- variable names can now contain dollar signs and underscores
-- some minor formatting fixes