Behavioral Data — Full Tables and Methodology

This comment contains the detailed data behind the behavioral summary in the main post. Same cohort, same queries — just the full breakdown.

Cohort

8,156 accounts created between January 2021 and December 2023, classified by self-reported location in profile metadata: Venezuela (3,666), Nigeria (1,492), Philippines (863), Indonesia (732), Cuba (385), USA (816), and UK (202). Venezuelan and Cuban users are further split by whether their first post used Spanish-language tags (hive-196387, cervantes, spanish, etc.), since the Spanish-speaking subcommunities on Hive are organizationally distinct from non-Spanish users in the same countries. USA and UK serve as the affluent-country baseline.

Staking Behavior

Power-up rates are within 4–7 percentage points across all groups. Cuban+Spanish users power up at 42.8% — the highest of any group, well above the USA's 25.9%. The difference is not in commitment but in capital: the average American user stakes 205 HIVE, the average Venezuelan+Spanish user stakes 11. That's a wealth gap, not a behavior gap.

Power-down rates are higher for distressed-country groups (28–39% vs. 16–20% for USA/UK). A Venezuelan user who earns $2 worth of HIVE in a week is making a real economic decision about whether to stake that or pay for food. An American user staking 200 HIVE faces no comparable pressure.

Long-Term Retention

The 2yr and 3yr figures are restricted to cohort members with enough elapsed time (joined before 2024 and 2023, respectively). The 3yr Venezuelan+Spanish figure (13.7%) is based on 1,943 users — a large, stable sample.

Self-Voting

The average self-vote rate is higher everywhere because a small number of heavy self-voters pull up the mean, and that effect is actually strongest in the USA (17.5%) and Indonesia (16.8%).

Voting Diversity

We also computed a "diversity per 100 votes" metric — how many unique authors does a user vote for per 100 votes cast? Every group clusters between 50 and 60. Venezuelan non-Spanish is the most diverse at 60.8, Cuban+Spanish the least at 39.7 (because they vote so much that they naturally revisit authors). There is no signal of concentrated circle-voting in any group.

Social Engagement — Comments Per Post

We also looked at comment balance (comments given vs. received) using ratios rather than raw differences — the raw difference is misleading because small groups like the UK receive comments from the entire platform while only having 202 members to give them. On a ratio basis, most groups fall between 0.80 and 1.05. The USA's high ratio (1.33) is driven by a few heavy commenters; the median American user receives 3x more comments than they write (median per-user ratio: 0.33). The median Venezuelan+Spanish user is more balanced at 0.51.

Content Effort — Post Length

Image usage: Venezuelan+Spanish users include images in 98% of their posts, Cuban+Spanish in 95%. The USA average is 74%.

Combined Engagement

The median tells the most dramatic story: the typical Venezuelan+Spanish user (67 total engagement actions) outperforms the typical American user (11) by a factor of six.

Data: HiveSQL, queried May 2026. Activity data: January 2021 – May 2026. Voting data: January 2021 – December 2025.