code4Analysis.zip (9.16 kB)
Stata and R Code for suppression and homogeneity analyses
softwareposted on 2021-12-23, 18:30 authored by Oliver MillsOliver Mills, Daniel ExeterDaniel Exeter, Nichola ShackletonNichola Shackleton
Stata and R code created by Oliver Mills in the Integrated Data Infrastructure environment, as part of his BSc (Hons) project "The Impact Geographical Scales Have on Homogeneity and Suppression", Department of Statistics, The University of Auckland, 2018
These data were also published in the manuscript "Inter-relationships between geographical scale, socio-economic data suppression and population homogeneity" Published in Applied Spatial Analysis and Policy. Article DOI:10.1007/s12061-021-09430-2
The code loads text files (originally sourced from SQL tables of the 2013 Census tables in the IDI), and merges in geographic identifiers for the scales of interest.
Suppression statistics are calculated for Census variables as a dichotomous variable, in which 'suppressed = 1' if a cell count for a variable is fewer than 6.
Homogeneity statistics are calculated using Intra Class Correlations , via the 'estat icc' command following a Multilevel mixed-effects logistic regression.
Analysis was conducted in Stata, and the accompanying R code reshapes and consolidates results for summary analyses as an excel file.