dm0101_1: Extensions to the label commands. D. Klein. Stata Journal 19: 867–882.
There have been numerous updates since the first release. These updates include completely rewritten Mata code and the following bug fixes and additions:
elabel list, if a value label contained exactly one extended missing value, incorrectly returned 0 in r(hasemiss); this has been fixed.
elabel rename created characteristics dta[_lang_list] and dta[_lang_c] even if the dataset in memory was monolingual; this has been fixed.
Several help files have been revised.
Several programming commands have been improved and new programming commands (and functions) have been added.
elabel is faster when it is first called.
elabel define and some other subcommands now attach value labels to variables on the fly using new notation: varname
:
lblname.
elabel values has an extended syntax that attaches a value label to each variable.
elabel compare has a new option, assertidentical, that verifies the identity of two value labels.
elabel load has new options lname(), value(), and label(), which are used to define value labels from datasets that were not created by Stata’s uselabel.
elabel recode has two new options: Option separator() is used to combine the labels associated with different integer values; option varlist returns the variable names that have one of the recoded value labels attached.
New (pseudo)function encode() defines value labels from string variables for use with the encode command.
New subcommand elabel duplicates reports or removes duplicate value labels.
New subcommand elabel modify modifies existing value labels and, optionally, defines new value labels as modified copies of existing value labels.
New subcommand elabel swap detaches value labels from variables and, optionally, attaches new value labels to these variables.
st0389_6: Conducting interrupted time-series analysis for singleand multiple-group comparisons. A. Linden. Stata Journal 17: 1024; 17: 515–516; 17: 73–88; 16: 813–814; 16: 521–522; 15: 480–500.
Fixed level() to ensure it is passed on to lincom from the estimation model.
Fixed the date format in the output and the graph to match the date format specified in trperiod().
st0390_2: Generalized maximum entropy estimation of discrete choice models. P. Corral and M. Terbish. Stata Journal 16: 244; 15: 512–522.
The gmentropylogit command has been updated to allow for priors to be used in the model. This change does not affect previous results obtained; it adds only functionality to the method.
st0393_3: Estimating almost-ideal demand systems with endogenous regressors. S. Lecocq and J.-M. Robin. Stata Journal 17: 1024; 16: 244; 15: 554–573.
The aidsills command now allows weights, fweights and aweights.
st0599_1: xtgeebcv: A command for bias-corrected sandwich variance estimation for GEE analyses of cluster randomized trials. J. A. Gallis, F. Li, and E. L. Turner. Stata Journal 20: 363–381.
The following updates have been made to xtgeebcv (Gallis, Li, and Turner 2020) since it was last uploaded. First, xtgeebcv now allows for factor variables and interactions in varlist. Note that with this new functionality, xtgeebcv works only with Stata 16.1 or later. Second, the original version of the program was not correctly accounting for the offset term (if specified). This has been fixed. Finally, the Li and Redden (2015) formulation of the Morel, Bokossa, and Neerchal (2003) (MBN) correction did not include a constant in the computation of φ. Because we used the Li and Redden (2015) formulation in the xtgeebcv program, we also left this constant out. To be consistent with the original MBN formula, we now include the constant. This leads to a negligible change in the standard errors when using the MBN method compared with the earlier version of the program without the constant.
