ROUND 2 · POST-EARTHQUAKE RECOVERY
More than a month after the June 24, 2026 earthquake, we segmented 809 affected households into four need profiles to guide humanitarian response where it's needed most.
01 · Introduction
More than a month after the magnitude 7.2–7.5 earthquake that struck north-central Venezuela on June 24, 2026, Equilibrium carried out a second round of the EquiExpress Survey: 809 households across seven states (La Guaira, Distrito Capital, Miranda, Aragua, Carabobo, Yaracuy, and Falcón). Unlike the first round, which focused on the immediate emergency, the instrument for this second measurement was expanded to 15 sections and incorporated modules that only make sense past the first month: livelihood recovery, mental health, gender protection, legal security and documentation, and perceptions of the institutional response.
As in the first round, we applied a cluster analysis to segment surveyed households into distinct need profiles, rather than treating the affected population as a homogeneous block. The result is the same number of groups as in Round 1 — four — but with a more complete picture.
The structure of the crisis did not change — but the nature of the need did. The analysis grouped households into the same number of profiles in both rounds —four— independently. But by incorporating Round 2's new dimensions, Cluster 2 —the "hidden vulnerability" cluster— is now defined with hard evidence: 88% of its households need regular chronic medication, and 52% had not received any humanitarian aid by the time of data collection.
02 · Methodology
The second round was carried out through computer-assisted telephone interviews with 809 households directly affected by the earthquake, distributed across seven states.
The highest silhouette score corresponds to the two-group solution (0.227), but that partition separates households only by severe housing damage or its absence, without distinguishing health, income, or institutional-trust dimensions. Among solutions with three or more groups, the four-cluster solution (k=4, in technical notation) obtained the highest silhouette score (0.085) and produces the most actionable profiles, replicating the same methodological decision made in Round 1.
03 · Comparative finding
Comparing two rounds built with different instruments requires caution, but the methodological consistency allows us to draw three reliable conclusions about how the crisis facing affected households evolved — or didn't.
Relative size of each cluster as % of the sample
The analysis grouped households into four clusters in both rounds, independently, without editorially forcing it in Round 2 as was necessary in Round 1. This suggests that the heterogeneity of needs among affected households is not an artifact of a single survey, but a real and persistent pattern.
Estimated monthly income (US$) by cluster
The relative distance between Cluster 1 (lowest) and Cluster 4 (highest) is nearly identical between rounds (2.0x in Round 1, 1.7x in Round 2). The change in the sample's state mix also affects the overall average.
% of households "not at all satisfied" with the authorities' response, by cluster (Round 2)
Between 46% and 58% of households in each of the four clusters report being "not at all satisfied" with the authorities' response — including Cluster 4, the most materially resilient. Institutional dissatisfaction is not exclusive to households with greater material need, but a common denominator across the entire affected population.
04 · A four-dimension index
To compare clusters that operate on different scales —dollars, days, headcounts— we built a relative need index: each dimension is normalized by dividing each cluster's value by the highest value observed in the sample for that same dimension, so that 1.00 always identifies the most affected cluster on that axis.
Click a cluster to isolate it. Hover over each vertex to see the exact value.
Cluster 1 leads in structural damage, food insecurity, and service failures, but it is Cluster 2 —not Cluster 1— that leads in demographic vulnerability. This confirms, in a single view, this report's central finding: the most urgent need does not always coincide with the most visible damage.
05 · Results by cluster
Ordered from highest to lowest relative need. Explore each profile to see its distinguishing variables and geographic concentration.
06 · Recommendations
More than a month after the earthquake, the response priority is no longer uniform: each cluster requires a different type of intervention.
This report summarizes the main findings. The full report and the interactive dashboard offer the complete detail, question by question.
