AMI v1.5: Argentina and Colombia Added; 39-Country Coverage

TL;DR

AMI v1.5 (May 2026) added Argentina (P=74.57, R=18.0, Q3) and Colombia (P=77.38, R=16.5, Q3). Coverage expanded to 39 countries. Both additions placed in Crisis zone with maxed Spanish-language essay mill and AI submission demand signals.

What v1.5 changed

Version 1.5 was released in May 2026 with three significant updates:

1. Coverage expansion — Argentina and Colombia added (39 countries total)
2. Updated Retraction Watch data — 69,911 records as of April 2026
3. Methodology refinements — improvements to Google Trends signal interpretation for Spanish-language and other Romance-language markets

Why these two countries

Both anchor the Spanish-language pattern

The v1.4 release covered Mexico as a Latin American Spanish-speaking country but did not include the two countries with the most extreme regional demand signals:

• Colombia — anchors the Andean Spanish-language market
• Argentina — anchors the Southern Cone Spanish-language market

Both consistently appeared in academic integrity literature as countries with elevated demand. Including them was a coverage priority for v1.5.

Both demonstrate the Spanish-language structural pattern

Pre-v1.5 hypothesis: the Spanish-language essay mill market is structurally different from other regional markets. Adding Colombia and Argentina would either confirm or challenge this.

The data confirmed the hypothesis decisively — both countries showed D1=100 and D2=100, the same pattern as other high-Prevalence Spanish-speaking countries.

Colombia (P=77.4, R=16.5)

The scores

Colombia immediately entered the top three Prevalence positions, behind only China and (in v1.5 itself) Argentina.

Why so high

The maxed D1 and D2 signals dominate. Colombia's Google Trends data shows essay mill keyword search volume and AI submission tool search volume both at the top of the Latin American distribution.

D6=0 reflects near-zero Retraction Watch presence — Colombian research output is moderate and the misconduct-linked retraction rate per publication is low.

The combination produces the second-highest Prevalence in the dataset.

R=16.5 — among lowest globally

Colombia's R-Score is in the bottom 10 globally. No specific contract cheating legislation, partial detection deployment, minimal disclosure, inconsistent penalties.

The combination of high Prevalence and weak Response places Colombia firmly in Q3 (Crisis zone) — the AMI's primary intervention target.

Argentina (P=74.6, R=18.0)

The scores

Argentina's profile is nearly identical to Colombia's — both score maxed on D1 and D2, both near-zero on D6, similar middle-range scores on D3, D4, D5.

Structurally similar to Colombia

The Argentine and Colombian profiles are the most structurally similar pair in the AMI dataset. Differences:

• Argentina D5 (50) vs Colombia D5 (52) — minimal
• Argentina D3 (12) vs Colombia D3 (14) — minimal
• R-Score: Argentina 18.0 vs Colombia 16.5 — small difference

The two countries share the Spanish-language market dynamic and similar institutional response infrastructure.

Slight R-Score edge

Argentina's R=18.0 is marginally higher than Colombia's R=16.5. The difference reflects:

• Slightly stronger Argentine institutional codes
• CONEAU accreditation framework
• More developed private university integrity infrastructure (ITBA, Universidad de San Andrés, others)

The difference is too small to be policy-significant.

What the additions tell us

The Spanish-language pattern is robust

Colombia and Argentina entering the dataset with the same D1=D2=100 pattern as the existing high-Prevalence Spanish countries (Spain D1=83, D2=100) confirms a structural regional dynamic. Other Latin American countries not yet in the dataset (Chile, Peru, Ecuador, etc.) likely show similar patterns.

Q3 (Crisis zone) is expanding

With v1.5, Q3 now contains 12 countries — the largest single quadrant addition in any single version. The Crisis zone profile is more prevalent than the Best in Class profile in the current 39-country set.

Coverage expansion is feasible

Adding two countries did not require major methodology re-engineering. Google Trends data was available, Retraction Watch data was available, regional integrity literature provided D4/D5 context. The expansion pattern can continue.

Methodology refinements in v1.5

Romance-language signal interpretation

Spanish-speaking countries share a language market that extends beyond country borders. v1.5 refined how Spanish-language Google Trends signals are attributed to specific countries:

• IP-based geographic attribution prioritised
• Search volume thresholds calibrated for Spanish-language markets
• Brand name signal weighting adjusted

The refinements affected Argentina, Colombia, Mexico, and Spain scores slightly relative to v1.4 baseline projections.

Updated Retraction Watch baseline

v1.5 used Retraction Watch data as of April 2026 (69,911 records, 5,390 misconduct-linked). Several countries' D6 scores shifted slightly relative to v1.4 baseline data.

Detection-incidence correction tightened

The correction factor for the detection-prevalence confound (countries with stronger detection report more cases) was tightened in v1.5 using updated Scarfe et al. (2024) detection-rate data.

What comes next

The 39-country v1.5 represents a coverage milestone — the AMI is now positioned to provide cross-country analysis for substantial fraction of global higher education enrolment.

Future versions will:

• Expand to 50+ countries (Chile, Peru, additional Asian, African, European)
• Add expert perception surveys (modelled on CPI Phase 2)
• Address Google Trends signal interpretation more comprehensively (Norway anomaly resolution)
• Improve detection-incidence correction methodology

Sources

• AMI v1.5 release documentation
• Google Trends API (2022–2026)
• Retraction Watch Database, Crossref/GitLab (April 2026)
• Scarfe et al. (2024), University of Reading detection study

Full methodology | Download dataset