Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply Social Identity Theory to analyze how group categorization, identification, and intergroup comparison drive behavior, bias, and conflict. Use this skill when the user needs to explain in-group favoritism or out-group hostility, diagnose organizational silo dynamics, design inclusive team structures, or when they ask 'why do groups polarize', 'how does team identity affect collaboration', or 'what drives us-vs-them thinking'.
.claude/skills/asgard-ai-platform-grad-social-identity/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 61% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 145% | 0% |
| case-04 | ✓→✗ | ▼ Worse | 66% | 0% |
| case-01 | ✓→✓ | = Same ✓ | -31% | 0% |
Social Identity Theory (Tajfel & Turner, 1979) explains how individuals derive self-concept from group memberships. The theory posits a three-stage process — social categorization, social identification, and social comparison — that produces in-group favoritism and out-group discrimination even with minimal group distinctions (minimal group paradigm).
IRON LAW: Social identity is RELATIONAL — it exists only through
comparison with out-groups, and threats to group distinctiveness
trigger identity-protective behaviors. Positive distinctiveness
is a fundamental motive.Key assumptions:
Map the relevant group boundaries in the context:
| Dimension | Indicator | |-----------|-----------| | Cognitive | Self-categorization as group member; "we" language | | Evaluative | Pride, prestige associated with membership | | Emotional | Emotional investment in group outcomes | | Behavioral | Conformity to group norms, in-group helping |
markdown## Social Identity Analysis: [Context] ### Group Map | Group | Salience Trigger | Identification Strength | |-------|-----------------|------------------------| | [in-group] | [trigger] | [High/Medium/Low] | | [out-group] | [trigger] | [High/Medium/Low] | ### Intergroup Dynamics - Comparison dimension: [status/competence/morality] - Perceived status: [in-group vs. out-group] - Identity management strategy: [mobility/creativity/competition] - Threat level: [distinctiveness/status/value threat] ### Behavioral Manifestations - [In-group favoritism examples] - [Out-group discrimination examples] ### Intervention Recommendations 1. [Recategorization or decategorization strategy] 2. [Contact conditions to reduce bias] 3. [Structural change to reduce category salience]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 33,106 | 20,789 | -37% | 1 | 1 | 0% | 5,005 | 3,472 | -31% | 0 | 0 | — |
case-02 | pass→pass | 33,331 | 15,189 | -54% | 1 | 1 | 0% | 4,902 | 3,527 | -28% | 0 | 0 | — |
case-03 | pass→pass | 23,597 | 20,187 | -14% | 1 | 1 | 0% | 3,482 | 4,235 | +22% | 0 | 0 | — |
case-04 | pass→fail | 15,522 | 19,917 | +28% | 1 | 1 | 0% | 2,331 | 3,860 | +66% | 0 | 0 | — |
case-05 | fail→pass | 18,191 | 16,640 | -9% | 1 | 1 | 0% | 2,615 | 3,596 | +38% | 0 | 0 | — |
case-06 | fail→fail | 13,145 | 18,907 | +44% | 1 | 1 | 0% | 1,899 | 4,049 | +113% | 0 | 0 | — |
case-07 | pass→pass | 7,184 | 6,312 | -12% | 1 | 1 | 0% | 1,118 | 2,171 | +94% | 0 | 0 | — |
case-08 | pass→pass | 6,409 | 4,997 | -22% | 1 | 1 | 0% | 952 | 1,960 | +106% | 0 | 0 | — |
case-09 | pass→pass | 5,967 | 6,292 | +5% | 1 | 1 | 0% | 1,009 | 2,122 | +110% | 0 | 0 | — |
case-10 | fail→pass | 12,959 | 13,530 | +4% | 1 | 1 | 0% | 2,048 | 3,305 | +61% | 0 | 0 | — |
case-11 | pass→pass | 12,243 | 20,179 | +65% | 1 | 1 | 0% | 1,880 | 3,289 | +75% | 0 | 0 | — |
case-12 | pass→pass | 17,764 | 36,737 | +107% | 1 | 1 | 0% | 2,927 | 3,724 | +27% | 0 | 0 | — |
case-13 | pass→pass | 5,545 | 5,704 | +3% | 1 | 1 | 0% | 893 | 2,104 | +136% | 0 | 0 | — |
case-14 | pass→pass | 15,825 | 14,483 | -8% | 1 | 1 | 0% | 2,171 | 3,515 | +62% | 0 | 0 | — |
case-15 | pass→pass | 10,235 | 6,506 | -36% | 1 | 1 | 0% | 1,602 | 2,153 | +34% | 0 | 0 | — |
case-16 | pass→pass | 6,196 | 5,366 | -13% | 1 | 1 | 0% | 1,030 | 1,909 | +85% | 0 | 0 | — |
case-17 | pass→pass | 9,938 | 10,031 | +1% | 1 | 1 | 0% | 1,578 | 2,509 | +59% | 0 | 0 | — |
case-18 | fail→pass | 5,907 | 7,917 | +34% | 1 | 1 | 0% | 986 | 2,420 | +145% | 0 | 0 | — |
case-19 | pass→pass | 12,698 | 16,515 | +30% | 1 | 1 | 0% | 2,109 | 3,538 | +68% | 0 | 0 | — |
case-20 | pass→pass | 20,155 | 15,612 | -23% | 1 | 1 | 0% | 2,507 | 3,415 | +36% | 0 | 0 | — |
case-21 | pass→pass | 13,364 | 13,117 | -2% | 1 | 1 | 0% | 2,163 | 2,935 | +36% | 0 | 0 | — |
case-22 | pass→pass | 13,248 | 15,551 | +17% | 1 | 1 | 0% | 2,219 | 3,602 | +62% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +9 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.