chandwich 1.1.6

Bug fixes and minor improvements

• Create the help file for the package correctly, with alias chandwich-package.

• In plot.confint() confidence limits are now added to the horizontal axis in cases where only one interval is plotted.

• Activated 3rd edition of the testthat package

chandwich 1.1.5

Bug fixes and minor improvements

• In the vignette references section DOIs are rendered with URLs hyperlinked.

chandwich 1.1.4

Bug fixes and minor improvements

• The function conf_intervals() now stops evaluating the profile loglikelihood when the required confidence limits have been estimated.

• In conf_interals() the initial estimates used in the profile loglikelihood were not reset in between the searches for the lower confidence limits and the upper limits. In some cases this may cause problems for the optimisations. This has been corrected: initial estimates are now reset.

• In confint.chandwich() the argument profile can be used to choose whether to return confidence intervals based on an (adjusted) profile loglikelihood or based on approximate large sample normal theory.

• The anova.chandwich() function has been modified to enable the comparison of nested (adjusted) model objects, where the nesting has not been explicitly created using the argument fixed_pars to adjust_loglikelihood().

• The documentation of conf_intervals() has been updated: the general nature of the argument num is explained and in Details advice is given on what to do if one or more of the confidence limits are not found using the default arguments.

• In the description of the argument parm in the documentation of confint.chandwich(), which_pars has been corrected to parm twice.

chandwich 1.1.3

Bug fixes and minor improvements

• The object returned from adjust_loglik() has an extra attribute loglikVecMLE, which is a vector containing the contributions of individual observations to the independence log-likelihood evaluated at the MLE.

• isTRUE() is used in the GEV example to avoid potential problems owing to the possibility of NAs.

chandwich 1.1.2

Bug fixes and minor improvements

• A typo meant that the text in the Value section of the description of confint.chandwich() was cut short. This has been corrected.

• In adjust_loglik() if p = 1 and the user supplies a scalar H instead of a matrix then an error is thrown by dimnames() just before the results are returned. The calculations are correct, so the code has been modified trivially to avoid the error.

• In the function returned from adjust_loglik the code to perform the horizontal adjustment of the loglikelihood has been changed to x_star <- mle + as.vector(C %*% (x - mle)): using as.vector() avoids a potential warning by ensuring vector + vector, not vector + matrix.

• There were bugs in plot.chandwich(), plot.confreg() and plot.confint() that produced an error if the user tried to set their own legend using the legend argument. These bugs have been corrected.

chandwich 1.1.1

Bug fixes and minor improvements

• The attribute nobs has been added to the object returned from adjust_loglik() and as an attribute to the object returned from logLik.chandwich().

• In compare_models() the parameter names (if any) are passed to the (adjusted) loglikelihod function, in case they are required inside the loglikelihood function.

• There was a bug in the plot method for objects of class "confreg" returned from conf_region(): if the parameters had not been named by the user then ? appeared twice in the console, requiring the user to press return twice before the plot as produced. This has been corrected.

chandwich 1.1.0

New features

• adjust_loglik has an additional arguments mle, H and V that allow the user the option to supply the MLE, the Hessian of the independence loglikelihood (H) and the variance of the vector of cluster-specific contributions to the score vector (V), each evaluated at the MLE, rather than estimating these within adjust_loglik.

• An anova S3 method for class "chandwich" has been added. This compares two or more nested models using adjusted likelihood ratio tests of successive pairs of models, using compare_models().

• A confint S3 method for class "chandwich" has been added. This is based on a fairly trivial call to conf_intervals().

• S3 methods coef, vcov and logLik for class "chandwich" have been added.

Bug fixes and minor improvements

• A bug in compare_models() has been fixed. The bug resulted an error in cases where the argument larger corresponded to a simplication of the full model in which element i of the parameter was fixed but some element i+n, for n > 0, was not fixed.

• Estimated unadjusted (VC) and adjusted (adjVC) variance-covariance matrices of the free model parameters are now available in the object returned by adjust_loglik().

• The documentation of the argument approx to compare_models() has been edited to make it clearer that if smaller is not supplied then approx = FALSE will be used regardless of any value supplied for approx in the call to compare_models().

• If parameter names are supplied to adjust_loglik() (via par_names) but fixed_pars is numeric then the names of the parameters in fixed_pars are now also inferred in the case where a larger model is not supplied via larger. This means that the output from compare_models() will now use the parameter name, rather than the parameter number.

• If a numeric fixed_pars is supplied to compare_models() then the names of the parameters in fixed_pars are inferred, if they are available in the supplied object larger.

• The summary method for class "evpost" is now set up according to Section 8.1 of the R FAQ at (https://cran.r-project.org/doc/FAQ/R-FAQ.html).

• In the Introducing chandwich vignette a typo in the definition of HA has been corrected. The expression given is for the inverse of HA, not for HA.

Chandler-Bate Sandwich Loglikelihood Adjustment

What does chandwich do?

The chandwich package performs adjustments of an independence loglikelihood using a robust sandwich estimator of the parameter covariance matrix, based on the methodology in Chandler and Bate (2007). This can be used for cluster correlated data when interest lies in the parameters of the marginal distributions or for performing inferences that are robust to certain types of model misspecification. Functions for profiling the adjusted loglikelihoods are also provided, as are functions for calculating and plotting confidence intervals, for single
model parameters, and confidence regions, for pairs of model parameters.

A simple example

The main function in the chandwich package is adjust_loglik. It finds the maximum likelihood estimate (MLE) of model parameters based on an independence loglikelihood in which cluster dependence in the data is ignored. The independence loglikelihood is adjusted in a way that ensures that the Hessian of the adjusted loglikelihood conicides with a robust sandwich estimate of the parameter covariance at the MLE. Three adjustments are available: one in which the independence loglikelihood itself is scaled (vertical scaling) and two others where the scaling is in the parameter vector (horizontal scaling).

The rats data contain information about an experiment in which, for each of 71 groups of rats, the total number of rats in the group and the numbers of rats who develop a tumor is recorded. We model these data using a binomial distribution, treating each group of rats as a separate cluster. The argument binom_loglik to adjust_loglik is a function that returns a vector of the loglikelihood contributions from each group of rats. In one-dimensional examples like this the two adjustments using horizontal scaling are identical, but this will not generally hold in more than one dimension.

binom_loglik <- function(prob, data) {
  if (prob < 0 || prob > 1) {
    return(-Inf)
  }
  return(dbinom(data[, "y"], data[, "n"], prob, log = TRUE))
}
rat_res <- adjust_loglik(loglik = binom_loglik, data = rats)
plot(rat_res, type = 1:4, legend_pos = "bottom", lwd = 2, col = 1:4)

Installation

To get the current released version from CRAN:

install.packages("chandwich")

Vignette

See vignette("chandwich-vignette", package = "chandwich") for an overview of the package.