extended documentation

man/bayesmeta-package.Rd

@@ -25,6 +25,7 @@
   errors) and prior information (effect and heterogeneity priors), and
   returns an object containing functions that allow to derive posterior
   quantities like joint or marginal densities, quantiles, etc.
+  The \code{\link{bmr}()} function extends the approach to meta-regression.
 }
 \author{
 Christian Roever <christian.roever@med.uni-goettingen.de>

man/bmr.Rd

@@ -146,12 +146,12 @@
   (\sQuote{zero/one} variables) simply identifying subgroups of studies.
   See also the examples shown below.
 
-\subsection{Connection to the simple random-effects model}{
+  \subsection{Connection to the simple random-effects model}{
   The meta-regression model is a generalisation of the \sQuote{simple}
   random-effects model that is implemented in the
   \code{\link{bayesmeta}()} function. Meta-regression reduces to the
   estimation of a single \dQuote{intercept} term when the regressor
-  matrix (\eqn{X}) has a single column (\eqn{d=1}) and consists only of
+  matrix (\eqn{X}) consists of a single column of
   ones (which is also the default setting in case the \sQuote{\code{X}}
   argument is left unspecified). The single regression coefficient
   \eqn{\beta_1}{beta[1]} then is equivalent to the \eqn{\mu} parameter
@@ -159,22 +159,33 @@
   section below).
   }
 
-\subsection{Prior specification}{
+  \subsection{Specification of the regressor matrix}{ The actual
+   regression model is specified through the regressor matrix \eqn{X},
+   which is supplied via the \sQuote{\code{X}} argument, and which
+   often may be specified in different ways. There usually is no unique
+   solution, and what serves the present purpose best then depends on
+   the context; see also the examples below. Sensible column names
+   should be specified for \code{X}, as these will subsequently
+   determine the labels for the associated parameters later on. Model
+   specification via the regressor matrix has the advantage of being
+   very explicit and transparent; if one prefers a
+   code{\link[stats]{formula}} interface instead, a regressor matrix may
+   be generated via the \sQuote{\code{\link[stats]{model.matrix}()}}
+   function.
+  }
+
+  \subsection{Prior specification}{
   Priors for \eqn{\beta} and \eqn{\tau} are assumed to factor into
   into independent marginals \eqn{p(\beta,\tau)=p(\beta)\times
   p(\tau)}{p(beta, tau) = p(beta) * p(tau)} and either (improper)
   uniform or a normal priors may be specified for the regression coefficients
-  \eqn{\beta}. Accuracy of the eventual computations is determined by the
-  \code{delta} (maximum divergence \eqn{\delta}) and \code{epsilon}
-  (tail probability \eqn{\epsilon}) parameters (Roever and Friede,
-  2017).
-
+  \eqn{\beta}. 
   For sensible prior choices for the heterogeneity parameter \eqn{\tau},
   see also Roever (2020), Roever \emph{et al.} (2021) and the
   \sQuote{\code{\link{bayesmeta}()}} function's help.
   }
   
-\subsection{Computation}{
+  \subsection{Computation}{
   The \code{bmr()} function utilizes the same computational method
   as the \code{\link{bayesmeta}()} function to derive the posterior
   distribution, namely, the \acronym{DIRECT} algorithm. Numerical