ASMbook (version 1.0.2)

simDat16: Simulate data for Chapter 16: Binomial ANCOVA

Description

Simulate Number black individuals ~ wetness regressions in adders in 3 regions

Usage

simDat16(nRegion = 3, nSite = 10, beta.vec = c(-4, 1, 2, 6, 2, -5))

Value

A list of simulated data and parameters.

nRegion

Number of regions

nSite

Number of sites per region

beta

Vector of regression coefficients

x

Indicator for region number

region

Region name (factor)

wetness

Wetness covariate

N

Number of adders captured at each site

C

Number of black adders captured at each site

Arguments

nRegion

Number of regions

nSite

Number of sites per region

beta.vec

Vector of regression coefficients

Author

Marc Kéry

Examples

str(dat <- simDat16())      # Implicit default arguments

# Revert to main-effects model with parallel lines on the logit link scale
# (also larger sample size to better see patterns)
str(dat <- simDat16(nSite = 100, beta.vec = c(-4, 1, 2, 6, 0, 0)))

# Same with less strong logistic regression coefficient
str(dat <- simDat16(nSite = 100, beta.vec = c(-4, 1, 2, 3, 0, 0)))

# Revert to simple logit-linear binomial regression: no effect of pop (and weaker coefficient)
str(dat <- simDat16(nSite = 100, beta.vec = c(-4, 0, 0, 3, 0, 0)))

# Revert to one-way ANOVA binomial model: no effect of wetness
# (Choose greater differences in the intercepts to better show patterns)
str(dat <- simDat16(nSite = 100, beta.vec = c(-2, 2, 3, 0, 0, 0)))

# Revert to binomial "model-of-the-mean": no effects of either wetness or population
# Intercept chosen such that average proportion of black adders is 0.6
str(dat <- simDat16(nSite = 100, beta.vec = c(qlogis(0.6), 0, 0, 0, 0, 0)))
mean(dat$C / dat$N)        # Average is about 0.6