Browser Method

Scoring Methodology

How we measure alignment between public statements and IRGC open letter rhetoric

Metric Definitions

Lexical Overlap
Count of significant shared content words (nouns, verbs, adjectives ≥3 chars, excluding stop words) between tweet text and IRGC paragraph.
Low overlap (1–3) is coincidental; 8+ suggests shared rhetorical sourcing.
Rare-Hit Score
Weighted count of uncommon shared words appearing in ≤5% of all tweets in the corpus. Rare words contribute their character length as weight.
Common words like "war" prove nothing. Rare shared terms like "Minab", "paper tiger", or "Zionist capitalists" indicate direct IRGC alignment.
Phrase Match Length
Length (in words) of the longest contiguous phrase from the IRGC letter found verbatim or near-verbatim in the tweet.
A 3-word match could be coincidence; a 7+ word match is functionally a quote.
Engagement Multiplier
(likes + reposts × 2) / 1000, capped at 100. Measures how widely the echoing statement reached the public.
A private journal is different from a tweet seen by 10M people. This measures amplification impact.
Echo Count
Total number of distinct IRGC letter passages a single author mirrors across all their tweets.
1 echo may be coincidence. 12 distinct echoes is systematic alignment.
Propagandist Threat Score
The primary ranking metric. Combines breadth (# of IRGC passages echoed), depth (how closely each mirrors the source), and reach (audience size).
This is the single most significant number for ranking. See formula below.

Composite Resonance Score (CRS)

Per-echo score measuring how closely a single statement mirrors a specific IRGC passage:

Composite Resonance Score CRS = (Lexical × 1.0) + (RareHit × 2.5) + (PhraseLen × 3.0) + ln(1 + Engagement) × 2.0

Significance Classes

CRS values are bucketed into classes calibrated against control corpora (Wendys, Bush Center, Obama):

S — Critical Alignment CRS ≥ 80
Near-verbatim reproduction of IRGC rhetoric with massive reach.
At this level, the probability of independent convergence approaches zero. The author is either directly sourcing from the letter/IRGC media ecosystem, or is so deeply embedded in the same information environment that the distinction is academic. No control account has ever scored this high.
A — High Alignment CRS 50–79
Strong structural and lexical parallels with significant audience.
Clear rhetorical mirroring beyond topic coincidence. The author consistently uses IRGC framing language, not just covering the same events. Multiple rare-word hits and/or phrase-level matches confirm deliberate (or deeply absorbed) alignment.
B — Moderate Alignment CRS 30–49
Clear thematic overlap with some distinctive shared language.
Warrants attention. Could reflect deliberate amplification or heavy consumption of IRGC-aligned media. Either way, the messaging serves IRGC narrative objectives. The 30-point threshold is set at 2× the maximum score any control account achieved, making false positives statistically unlikely.
C — Low Alignment CRS 15–29
Topical overlap with minimal distinctive shared language.
Expected for anyone discussing the same geopolitical events. Not independently significant but contributes to pattern analysis when combined with other C+ matches from the same author. The 15-point floor is calibrated to the 95th percentile of control account scores.
D — Noise CRS < 15
Insufficient evidence of alignment.
Below statistical significance. These matches are expected by chance in any large corpus of political commentary. Shared words are common political vocabulary ("government", "military", "war"). Not reported in primary analysis.

Echo Count Significance

CountPatternInterpretation
1IsolatedSingle topical coincidence. Not significant alone.
2–3EmergingAuthor engages with multiple IRGC themes. Worth monitoring.
4–6ConsistentSystematic alignment across multiple narrative threads. Reliably amplifying IRGC messaging.
7–10ProlificNear-comprehensive coverage of IRGC narrative portfolio. Deeply embedded in IRGC-aligned ecosystem.
11+ComprehensiveMirrors the letter's full rhetorical architecture. Question shifts from “are they aligned?” to “what is the mechanism?”

Propagandist Threat Score (PTS)

The single most significant number for ranking. This is what determines position on the list.

Primary Ranking Metric
PTS = ∑ CRSi × log10(1 + EngagementReachi)
Where the sum is over all n distinct IRGC passages echoed by the author.

This captures breadth (how many IRGC talking points they amplify), depth (how closely each echo mirrors the source), and reach (how many people saw it).

Tucker Carlson (12 echoes × high engagement) scores higher than a niche account with 12 echoes and 500 followers, because Carlson’s amplification has measurably greater impact on public discourse.

Threshold Calibration

Class boundaries were established through statistical calibration:

  1. Control corpus: 10 accounts with no expected IRGC alignment (Wendys, Bush Center, Obama, etc.) scored against all 185 paragraphs.
  2. Noise floor: Controls averaged CRS 5–12, with maximum 18 → established 15 as Class C floor.
  3. 99th percentile: No control exceeded CRS 25 → established 30 (Class B) as statistically unlikely to be coincidental.
  4. Human validation: CRS ≥ 50 matched cases where analysts independently flagged “same source material.”
  5. Class S (≥80): Reserved for rare-word overlap + phrase match + massive engagement — the “smoking gun” tier.
← Back to Browser