R/harnik.R

In gustavobio/semeco: Examples of Structural Equation Modelling Usage in Ecology

#' Effects of biological factors on extinction risk in fossil bivalves.
#'
#' Raw data and model syntaxes for the paper by Harnik (2011) on the
#' direct and indirect effects of biological factors on extinction risk
#' in fossil bivalves.
#'
#' @docType data
#' @format A data frame with 108 observations on 8 variables and three character
#' vectors containing model syntaxes.
#' @name harnik
#' @aliases harnik harnik.model.A harnik.model.B harnik.model.C
#'
#' @note
#' The models are in page 2 (figure 1). All models converge. The
#' coefficients in model A come from three independent linear
#' regressions. The dataset is included as it was made available in the
#' supporting information.  Data transformations should be done following
#' the manuscript in order to get the same results. Model C involves
#' groups (superfamilies).
#'
#' Data transformation:
#'
#' 1. Abundance and range extent were transformed with the square root of
#' the arcsine method.
#' 2. Abundance and body size were log-transformed.
#' 3. Abundance, range extent and body size were scaled and centered
#' (we are replicating what was reported in the manuscript, but maximum
#' likelihood estimation expects unstandardized data).
#'
#' @source
#' Article:
#'
#' \url{http://www.pnas.org/cgi/doi/10.1073/pnas.1100572108}
#'
#' Supplementary information:
#'
#' \url{http://www.pnas.org/content/suppl/2011/07/25/1100572108.DCSupplemental}
#'
#' @references
#' Harnik P. 2011. Direct and indirect effects of biological factors on
#' extinction risk in fossil bivalves. PNAS early view. doi:10.1073/pnas.1100572108.
#'
#' @examples
#' \dontrun{
#' data(harnik)
#' harnik.t <- harnik
#'
#' harnik.t$abundance <- asin(sqrt(harnik.t$abundance))
#' harnik.t$range.extent <- asin(sqrt(harnik.t$range.extent))
#'
#' harnik.t$abundance <- log(harnik.t$abundance)
#' harnik.t$body.size <- log(harnik.t$body.size)
#'
#' harnik.t$abundance <- scale(harnik.t$abundance, center = TRUE, scale = TRUE)
#' harnik.t$range.extent <- scale(harnik.t$range.extent, center = TRUE, scale =
#'                                  TRUE)
#' harnik.t$body.size <- scale(harnik.t$body.size, center = TRUE, scale = TRUE)
#'
#' harnik.model.A.1.fit <- sem(harnik.model.A[[1]], data = harnik.t)
#' harnik.model.A.2.fit <- sem(harnik.model.A[[2]], data = harnik.t)
#' harnik.model.A.3.fit <- sem(harnik.model.A[[3]], data = harnik.t)
#' harnik.model.B.fit <- sem(harnik.model.B, data = harnik.t)
#' harnik.model.C.fit <- sem(harnik.model.C, data = harnik.t, group =
#'                             "superfamily",
#'                           group.equal = "regressions",
#'                           group.partial = "duration ~ body.size")
#'
#' summary(harnik.model.A.fit)
#' summary(harnik.model.B.fit)
#' summary(harnik.model.C.fit)
#' }
NULL