Research
Methodology notes
Written for an academic reader. Full detail will accompany the Phase 1 manuscript.
2SFCA (two-step floating catchment area)
The access score is computed in two steps. First, for each facility, a facility-to-population ratio is calculated using all population within the facility's catchment (a 30 or 60-minute travel-time threshold). Second, for each population point, the access score sums the facility-to-population ratios of every facility within reach of that point. This produces a per-capita measure of available capacity, not a raw distance or a facility count.
PCA-based index construction
The vulnerability index standardises each input component (socioeconomic, environmental, disease burden) to a common scale, then combines them using principal component analysis (PCA). The first principal component's loadings — published alongside the index — determine each component's contribution to the final score.
Tercile cut-points
Regions are classified into 'high' and 'low' groups on each axis using tercile cuts by default, with a median-cut toggle available for comparison. Cut-point choice affects which regions fall into the priority quadrant, particularly near the boundary — this is disclosed, not smoothed over.
Multi-scale sensitivity
Where boundary data permits, results are checked across administrative scales (region versus governorate) to see whether the classification is an artefact of the chosen unit size.
MAUP and the ecological fallacy
The modifiable areal unit problem (MAUP) means that results can shift depending on how boundaries are drawn — a caveat that applies to any areal analysis, including this one. Region-level classification also carries an ecological fallacy risk: a conclusion true of a region as a whole is not necessarily true of every community, or every person, within it. Both are stated explicitly rather than implied away by a confident-looking map.
Publications
Manuscript in preparation.