AI summary by TwinMind
Why corporate diversity programs fail -- and how small tweaks can have big impact | Joan C. Williams
TED
14m 48s
Starbucks Incident and Traditional Diversity Training Failures
- 2018: Two Black men arrested at Starbucks while waiting for business associate
- Asked to use bathroom, manager ordered them to leave
- Police called, video went viral
- Starbucks response: closed all stores nationwide for 4 hours diversity training
- Workbooks with prompts: "What makes me me and you you?"
- "Understanding our bias: from color-blind to color brave"
- Traditional diversity training assumption: structural racism addressed through conversations about feelings
- Reality: to address structural racism, must change structures
Current State of Corporate Diversity Efforts
- Post-George Floyd: companies feel pressure to deliver on diversity goals but lack effective strategies
- ~$1 billion spent on diversity initiatives
- Basic diversity industrial complex tools don't work:
- One-shot bias training fails (doing anything once won't change culture)
- Employee resource groups/women's initiatives address wrong problem
- Problem isn't with women and people of color - it's with business systems
Root Cause Analysis
- Bias transmitted through basic business systems:
- Hiring processes
- Performance evaluations
- Access to opportunities
- Need to fix business systems, not fix women and people of color
- Analogy: companies facing sales challenges wouldn't just hold conversations about valuing sales
Evidence-Based Approach to Diversity
- Must use same tools businesses use for any problem: evidence and metrics
- WorkLife Law team studied bias in workplace interactions for over a decade
- Five consistent patterns of bias emerge repeatedly
Five Patterns of Workplace Bias
Pattern 1: "Prove It Again"
- Some groups must prove themselves more than others
- Triggered by: race, gender, age, disability, LGBTQ status, social class
- Study example: identical white men résumés with different hobbies
- Polo/sailing résumé: 12x more callbacks than country music/counseling résumé
Pattern 2: "The Tightrope"
- White men need only be authoritative and ambitious to succeed
- Women walk tightrope: seen as abrasive if authoritative, unqualified if not
- People of color assertiveness interpreted as:
- Angry (if Black)
- Hotheaded (if Latinx)
- Untrustworthy (if Asian American)
Pattern 3: "Tug-of-War"
- Bias against group fuels conflict within group
- When room for only one woman/person of color: predictable super-competition within groups
Pattern 4: "Maternal Wall" (Strongest Form of Gender Bias)
- Assumptions about mothers: not committed, shouldn't be, not competent ("pregnancy brain")
- Mothers must prove themselves again after maternity leave
- If they do, seen as bad mothers/bad people and disliked
Pattern 5: Racial Stereotypes
- Asian Americans: great technical skills, lacking leadership potential
- Black professionals: high isolation levels, startling disrespect
- Asian American "too emotional" vs. white man "career-enhancing passion" for identical behavior
Bias Impact by Group
- White women: report 4 patterns of bias
- Men of color: report 4 patterns
- Women of color: report all 5 patterns in substantial proportions
- Black women: report most bias as group
- White men's experience differs from every other group
- White men may encounter bias if first-generation professional or LGBTQ
Real-World Impact Case Study
- Female engineer found male colleague's calculation mistake
- Pointing out mistake violated unwritten rule (women should be modest, self-effacing, nice)
- Research: male experts exert more influence in meetings; female experts exert less influence than female non-experts
- Department's massively negative response led engineer to "smile a lot and bring in cupcakes"
- Company jeopardized mission by allowing gender bias unchecked
Solution: Bias Interrupters
- Evidence-based, metrics-driven tools developed by team
- Metrics help pinpoint where problems occur in hiring pipeline:
- Original candidate pool
- Résumé review survivors
- Interview callbacks
- Interview survivors
- Different problems require different fixes
- Metrics establish baselines and measure progress
Implementation Case Study Results
- Company performance evaluation analysis:
- Only 9.5% of people of color had leadership mentioned in evaluations
- 70 points lower than white women
- Leadership mentions predicted advancement
Two Simple Interventions:
- Redesigned performance evaluation forms
- One-hour workshop showing actual prior year comments, asking which of 5 bias patterns represented
Year 2 Results:
- 100% of people of color had leadership mentioned in evaluations
- White women problem: ~20% had comments about not wanting partnership
- Suspected assumptions, not actual conversations
- Workshop instruction: don't make such comments without actual conversation
- Year 2: only 1 woman received that comment company-wide
Broader Impact and Effectiveness
- Helped over 100 companies make progress toward diversity goals
- Growing evidence bias interrupters work
- Benefits every single group:
- People of color: 30% jump in constructive feedback
- White women: more constructive feedback
- White men: more constructive feedback
- Evidence-based system design helps all groups
CEO Message and New DEI Playbook
- Current systems/culture reflect already-hired workforce
- To replicate workforce: keep doing same things
- To change: use standard business tools
- New DEI playbook:
- Start from evidence
- Gather metrics to establish baselines
- Measure progress
- Keep at it until goals achieved
- CEOs already know these tools (more comfortable than leading conversations about social inequality)
