The table shows two statistics, b-d, representing the breadth and depth of the digital divide, by year and Local Authority. Years 2013/2014/2015 are coded by winter/spring/summer colours. Interventions should be targeted to reduce both breadth and depth.
The breadth statistic is Wagstaff's Generalised Concentration index, for Fixed Broadband Uptake against SIMD deprivation rank, normalised to give a range of ±1. This is a measure of the breadth of the degree to which the digital divide exacerbates existing deprivation.
The depth statistic is Wagstaff's renormalisation of the concentration index for a binary advantage, again for Fixed Broadband Uptake against SIMD deprivation rank. It measures the extent to which those digitally deprived suffer other aspects of deprivation.
The graphs show a form of Lorenz curve, plotting the cumulative proportion of households online against cumulative proportion of households offline, ordered by increasing SIMD rank i.e. (from most deprived to least deprived). In case of perfect equality, this Lorenz Curve would lie on the diagonal line shown in grey.
The area enclosed between the Lorenz curve and the diagonal line of perfect equality, expressed as a percentage of the area of the triangle between the diagonal and the horizontal axis, represents Wagstaff's renormalised concentration index. This represents the depth of the digital divide; higher values represent greater inequality.
The depth statistic can be large, even when most individuals are online. It measures the mean deprivation of those offline, on a scale ranging over ±1. The breadth statistic is 4pq times this depth statistic. It takes its extreme value of +1 when the least deprived half of the population is online and the rest are offline. It decreases as uptake varies from 50%.
The slope of this Lorenz curve at any point is determined by the odds ratio relating the odds of a household at a particular level of deprivation being online to the population odds of being online.
The odds ratio quantifies relative advantage or disadvantage. Where the slope of the Lorenz curve is less than 1, the odds of being online are lower than the population odds; in these digitally disadvantaged sectors of the population the proportion of online households is lower than average. Where the slope is greater than 1 it it is greater.
Typically, the lowest slopes, representing digital disadvantage, with fewer than average online, are found in the initial segements of the curve, representing the lowest SIMD ranks. Steeper slopes, corresponding to a greater than average proportion of households online are found amongst the greater SIMD ranks, representing the least deprived sectors of society.
You can highlight the curves for a particular Local Authority using the selector immediately below, or by clicking in the table on its name. The curves for the remaining authority areas are shown lightly in the background.
© Michael Fourman
Last modified: Mon Jan 18 12:17:02 GMT 2016
Scotland: Fixed Broadband Inequality, a work by Michael P. Fourman, published online at http://http://idea.ed.ac.uk/digiscot/scotland/, is licensed under a Creative Commons Attribution 4.0 International License.
Last modified: Fri Oct 16 12:34:28 BST 2015