region

gaussian by default, or binomial for all y either zero or 1

family

length.out

width of confidence interval; logical TRUE defaults to 0.95

conf

passed to <code><a rd-options="" href="/link/model?package=metaplot&version=0.2.7" data-mini-rdoc="metaplot::model">model</a></code>

Calculates a confidence region. <code>se.fit</code> from <code><a rd-options="stats" href="/link/predict.glm?package=metaplot&version=0.2.7&to=stats" data-mini-rdoc="stats::predict.glm">predict.glm</a></code> is multiplied by <code>z</code> and added or subtracted from fits to give <code>hi</code> and <code>lo</code> columns in return value. <code>z</code> is normal quantile for the one-tailed probablitity corresponding to <code>conf</code>, e.g. ~ 1.96 for <code>conf = 0.95</code>. If non-missing <code>y</code> is only 0 or 1, the model family is binomial and resulting confidence intervals are back-transformed using <code><a rd-options="stats" href="/link/plogis?package=metaplot&version=0.2.7&to=stats" data-mini-rdoc="stats::plogis">plogis</a></code>.

Designs plots in terms of core structure.  See 'example(metaplot)'.
Primary arguments are (unquoted) column names; order and type (numeric or not)
dictate the resulting plot.  Specify any y variables, x variable, any groups variable,
and any conditioning variables to metaplot() to generate density plots, boxplots,
scatterplots, scatterplot matrices, or conditioned plots.  Honors scalar column
attributes 'label' and 'guide' where present to produce fully annotated plots
using concise syntax. Attribute 'guide' is typically units, but may be encoded()
to provide interpretations of categorical values (see 'encode' package).  Utility
unpack() transforms scalar column attributes to row values and pack() does the
reverse, supporting tool-neutral storage of metadata along with primary data.
Supports customizable aesthetics such as such as reference lines, unity lines,
smooths, log transformation, and linear fits. Compact syntax and isolated metadata
promote workflow scalability.

Tim Bergsma

metaplot

Formalized Plots with Systematic Metadata

region function

passed to <code><a rd-options='' href='model'>model</a></code>

Calculates a confidence region. <code>se.fit</code> from <code><a rd-options='stats' href='predict.glm'>predict.glm</a></code> is multiplied by <code>z</code> and added or subtracted from fits to give <code>hi</code> and <code>lo</code> columns in return value. <code>z</code> is normal quantile for the one-tailed probablitity corresponding to <code>conf</code>, e.g. ~ 1.96 for <code>conf = 0.95</code>. If non-missing <code>y</code> is only 0 or 1, the model family is binomial and resulting confidence intervals are back-transformed using <code><a rd-options='stats' href='plogis'>plogis</a></code>.

region: Calculate a Confidence Region

Description

Usage

Arguments

Value

See Also