mpmsim (version 3.2.1)

calculate_errors: Calculate error (standard error or 95%CI) in elements of a matrix population model.

Description

Given two submatrices of a matrix population model (mat_U and mat_F, the growth/survival matrix and the fecundity matrix respectively) and a sample size, or matrix/matrices of sample sizes, this function calculates the standard error or 95% confidence interval (95%CI) for each element of the matrix. These calculations assume that mat_U is the result of binomial processes (i.e., the survival (0/1) of a sample of n individuals), while mat_F is the result of Poisson processes (i.e., counts of offspring from n individuals), where n is the sample size.

Usage

calculate_errors(mat_U, mat_F, sample_size, type = "sem", calculate_A = TRUE)

Value

A list containing the original matrices and the error estimates (or upper and lower confidence intervals) for the U, F and (optionally) A matrices.

Arguments

mat_U

matrix of mean survival probabilities

mat_F

matrix of mean fecundity values

sample_size

either (1) a single matrix of sample sizes for each element of the MPM, (2) a list of two named matrices ("mat_F_ss", "mat_U_ss") containing sample sizes for the survival and fecundity submatrices of the MPM or (3) a single value applied to the whole matrix

type

A character string indicating the type of error to calculate. Must be one of "sem" (standard error), or "CI95" (95% confidence interval).

calculate_A

A logical argument indicating whether the returned error information should include the A matrix and its error. Defaults to TRUE.

Author

Owen Jones

Details

The output is a list containing the original matrices and matrices showing error estimates or confidence intervals.

See Also

add_mpm_error() which simulates matrices with known values and sample sizes.

Other errors: add_mpm_error(), compute_ci(), compute_ci_U()

Examples

# Set up two submatrices
matU <- matrix(c(
  0.1, 0,
  0.2, 0.4
), byrow = TRUE, nrow = 2)
matF <- matrix(c(
  0, 4,
  0., 0.
), byrow = TRUE, nrow = 2)

# errors as 95% CI, with a sample size of 20 for all elements
calculate_errors(mat_U = matU, mat_F = matF, sample_size = 20, type = "CI95")

# errors as sem, with a sample size of 20 for all elements
calculate_errors(mat_U = matU, mat_F = matF, sample_size = 20, type = "sem")

# Sample size is a single matrix applied to both F and U matrices
ssMat <- matrix(10, nrow = 2, ncol = 2)

calculate_errors(
  mat_U = matU, mat_F = matF, sample_size = ssMat, type =
    "sem"
)

# Sample size is a list of two matrices, one for F and one for U.
ssMats <- list(
  "mat_F_ss" = matrix(10, nrow = 2, ncol = 2),
  "mat_U_ss" = matrix(10, nrow = 2, ncol = 2)
)
calculate_errors(
  mat_U = matU, mat_F = matF, sample_size = ssMats, type =
    "sem"
)