NEWS.md

In REndo: Fitting Linear Models with Endogenous Regressors using Latent Instrumental Variables

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

REndo documentation built on Sept. 8, 2023, 5:53 p.m.