gwalpha: Geographically Weighted Cronbach's Alpha

Description

This function computes geographically weighted cronbach's alpha (GWalpha) for responses from georeferenced multi-item survey scales.

Usage

gwalpha(
  x,
  data,
  kernel = "bisquare",
  adaptive = TRUE,
  bw,
  ci = FALSE,
  p = 0.95,
  nsims = 1000
)

Value

a data frame with:

gwalpha: local estimates of reliability
gwalpha_u: upper confidence interval if ci = TRUE
coords: coordinates matrix for each responses

Arguments

x: a character vector or numeric vector with the column names or indices of items in data
data: a SpatialPointsDataFrame as defined in package sp, or a sf object defined in package sf
kernel: type of kernel function used to weight responses. Available options: "bisquare" (default), "gaussian", "exponential", or "boxcar"
adaptive: logical; if TRUE (default), the bandwidth bw corresponds to the number of nearest neighbours; if FALSE, bw is a fixed distance
bw: bandwidth for weighting function, can be specified or obtained using bw_gwalpha()
ci: logical; if TRUE, bootstrapped confidence intervals are computed
p: the percentile for the upper confidence interval if ci = TRUE
nsims: number of bootstrap iterations if ci = TRUE

References

Zhang, S., and Z. Li. 2025. “ Geographically Weighted Cronbach's Alpha (GWalpha): An Exploratory Local Measure of Reliability for Scale Construction.” Geographical Analysis 57, no. 4: 758–772.

Examples

Run this code

data(bes2011)
alpha100 <- gwalpha(
  x = c("willHelp", "closeKnit", "trust", "solComProb", "relGroups"),
  data = bes2011,
  bw = 100
  )