We add the copulaIMA method which implements the
copula-based Instrumental Variable Model (IMA) proposed by Haschka
(2025)
MAINTENANCE
Fix CRAN notes regarding tests for multilevelIV and
HTML issues in the documentation
REndo 2.4.11
MAINTENANCE
Fix CRAN notes regarding tests for multilevelIV
REndo 2.4.10
NEW FEATURES
The Pstar data generated in copulaCorrection is
improved with the corrections proposed by Qian, Koschmann, and Xie
(2024)
MAINTENANCE
Fix CRAN notes regarding documentation
REndo 2.4.9
MAINTENANCE
Add CITATION file and references to forthcoming JSS publication
Fix missing package alias
REndo 2.4.8
MAINTENANCE
Replace data set dataCopDisCont with a new one
REndo 2.4.7
MAINTENANCE
Fix tests due to changes in package dependencies
REndo 2.4.6
MAINTENANCE
Replace data sets with new ones
Fix tidy HTML warnings in documentation and parsing of news.md
reported in daily CRAN checks
REndo 2.4.5
BUG FIXES
Fix broken correctness tests for latentIV on
M1mac.
MAINTENANCE
Update maintainer email.
REndo 2.4.4
BUG FIXES
Fix documentation to conform to HTML5.
REndo 2.4.3
BUG FIXES
In method multilevelIV matrices could not be
constructed if all groups were of the same size. Thanks to pinson06 for
reporting.
REndo 2.4.2
BUG FIXES
In the copulaCorrection method with a single endogenous
regressor the correlation coefficient rho was wrongly constraint to [0,
1] instead of to [-1, 1]. Further, rho was not included in parts of the
calculations. Special thanks to Fredrik Falkenström for reporting this
observation!
REndo 2.4.1
SIGNIFICANT USER-VISIBLE
CHANGES
Allow using the method copulaCorrection with a single
endogenous regressor
REndo 2.4.0
SIGNIFICANT USER-VISIBLE
CHANGES
Add a formal assumption test for method hetErrorsIV to
detect weak instruments
NEW FEATURES
The log-likelihood function underlying copulaCorrection
case 1 is refactored and implemented using Rcpp and RcppEigen what makes
it considerably faster
Improved checks of user input that disallow any non-finite values in
the data
New data was generated for dataHetIV
REndo 2.3.1
None. Version bump in order to resubmit to CRAN after package was
archived.
REndo 2.3.0
SIGNIFICANT USER-VISIBLE
CHANGES
Add support for predictions through method predict for
every model
BUG FIXES
Fix wrong calculations for the fitted values and residuals in the
enhanced OLS cases of copulaCorrection
Improved documentation
REndo 2.2.1
SIGNIFICANT USER-VISIBLE
CHANGES
To tweak the lmer model fit in multilevelIV to their
linking, users can supply a parameter lmer.control
NEW FEATURES
Warning in confint if there are NAs in the bootstrapped
estimates for copulaCorrection
Updated vignette
BUG FIXES
The coefficient estimates for multilevelIV are more
consistent and independent of data sorting due to different standard
settings for fitting lmer
Bootstrapped parameter estimates for copualCorrection
result in fewer NAs when using L-BFGS-B as optimization method
REndo 2.2.0
SIGNIFICANT USER-VISIBLE
CHANGES
The augmented OLS method in copulaCorrection also
bootstraps parameter estimates
The summary output for results from copulaCorrection
was adapted to reflect that standard errors are bootstrapped
Removed support for the S3 method labels because of
inconsistent behavior across methods
NEW FEATURES
Bootstrapping for copulaCorrection case 1 is now
considerably faster
New data was generated for dataMultilevelIV
BUG FIXES
The sigma matrix in latentIV is constructed as in the
paper by Ebbes et al. what improves results. Special thanks to Jordan
Henderson for investigating and pointing this out!
In latentIV, the parameter for group membership
(theta5) is now transformed back and reported
correctly.
The vcov matrix for latentIV is corrected for the
transformation in theta5.
The bootstrapping in copulaCorrection case 1 now
creates samples of the same length as the original data
The (percentile) confidence intervals and vcov for results from
copulaCorrection now rely on bootstraping
REndo 2.1.0
SIGNIFICANT USER-VISIBLE
CHANGES
The reworked method multilevelIV and accompanying data
was added back to the package
Vignette REndo-introduction was added to showcase
package usage
Users can supply a parameter optimx.args to tweak the
LL optimization to their liking
NEW FEATURES
Method confint was added for methods
latentIV and copulaCorrection
Examples and documentation were improved for all methods
New data was generated for dataHetIV
The default number of iterations for all optimizations was increased
to 100’000
BUG FIXES
To avoid infrequent warnings, the parameter sigma used
in copulaCorrection was constrained to > 0
Various spelling mistakes were fixed
REndo 2.0.0
SIGNIFICANT USER-VISIBLE
CHANGES
Remodeled all methods’ user-interface
Added detailed input checks for every provided parameter
Adapted all visible outputs
Parameter verbose allows to turn on or off
printing
Updated documentation to reflect all changes and added theoretical
background
NEW FEATURES
Formulas support transformations
Improve all code to be more reliable and stable
Provide new example datasets and accompanying documentation for all
methods
Added extensive testing for all aspects of the package
Increased numerical stability for log-likelihood optimization
methods latentIV and copualCorrection
BUG FIXES
Many
DEPRECATED AND DEFUNCT
The multilevel function is temporarily removed from the package due
to ongoing work on it