lecture recording on 03 February 2025 at 19.07.06 PM

Discrimination and Statistical Bias

  • Discrimination often occurs based on visible characteristics rather than individual attributes (i.e., skills, knowledge).

  • People level attributes (aptitude, skills) are more relevant but often overlooked in favor of surface characteristics.

Study Overview

  • A study from the University of Chicago conducted by John Levitt and Yuri Gee examined discrimination based on surface characteristics during interactions (asking for directions).

  • Actors (confederates) dressed similarly but differed in age, gender, and ethnicity to test public response.

Main Variables Assessed

  • Characteristics:

    • Age: 20 vs. 50 years

    • Gender: Male vs. Female

    • Ethnicity: White vs. Black

  • Measures:

    • Percentage of people providing helpful feedback

    • Duration of interactions in seconds

Key Findings

  • Female Confederates:

    • Feedback from the public varied little by age or ethnicity (positive interactions).

    • Consistent helping rates and interaction times.

  • Male Confederates:

    • Black, young males received less helpful feedback (drop by 50%).

    • Shorter interactions with less engagement from the public.

    • Indicates a clear bias against young black males, attributed to stereotypes (statistical discrimination).

Statistical Discrimination Explained

  • Statistical discrimination describes decision-making based on group stereotypes rather than individual merit.

  • Stereotypes lead to preconceived notions affecting interpersonal interactions.

Example Sketch

  • A BBC satire presented biases in security interactions, highlighting how individuals are judged based on appearance or ethnicity, reinforcing stereotypes.

Dressing and Perceptions

  • A subsequent study replicated previous results, showing that individuals dressed in business attire experienced no discrimination, regardless of their ethnic background.

  • It emphasizes the influence of dress on how individuals are perceived and treated in public settings.

Case Study: Trayvon Martin

  • The tragic case of Trayvon Martin illustrates potential dangers stemming from biases based on appearance (wearing a hoodie).

  • The misperception of threat led to fatal confrontations, raising questions about racial profiling and systemic discrimination in society.

Discrimination in the Business Domain

  • Exploratory studies at University of Chicago replicated findings in high-stakes environments like car dealerships.

  • Scenario: Young men role-playing as either a gay couple or friends looking to purchase a vehicle.

    • 75% of the time, the gay couple received worse offers than heterosexual counterparts.

    • Disparity in treatment included fewer offers for amenities like beverages or test drives, indicating systemic bias.

Study on Ethnic Reactions

  • Ethnic diversity among sales staff influenced discriminatory behavior.

  • Hypothesis: White sales staff might possess less experience with discrimination, thus displaying more bias, while minority staff may empathize due to personal experiences.

White vs. Black Male Buying Experience

  • Study showed consistent patterns of discrimination in high-end car purchases, with black males receiving worse offers than white males.

  • Analyzed through the lens of statistical rationalization versus personal dislike.

  • No discriminative behavior noted in lower-priced transactions (e.g., Toyota), suggesting potential bias linked to profit perception.

Group Exercise: Role-Playing for Empathy

  • Students tasked with role-playing discriminatory scenarios, intending to unearth biases based on surface characteristics.

    • This exercise fosters understanding of the underlying assumptions and practices that perpetuate discrimination in hiring processes.

Callback Rates in Employment Applications

  • A study using fictitious resumes (black-sounding names vs. white-sounding names) revealed significant discrepancies in callback rates:

    • White females received a 50% higher callback rate than black females.

    • White males received 30% higher callbacks than black males.

Education Impact

  • Even applicants with advanced degrees experienced bias: white master's degree holders had a higher callback than their black counterparts.

Pricing Discrimination Against Disabled Individuals

  • A study on disabled customers receiving car repair quotes showed they were charged up to 30% higher than able-bodied customers.

Solutions for Addressing Pricing Disparities

  • Potential Interventions:

    • Implement standard pricing policies to prevent discrimination.

    • Encourage customers to mention multiple estimates to rebuild bargaining power.

Conclusion: Addressing Bias and Discrimination

  • Societal biases are deeply embedded; systematic practices are required to counter discrimination, whether in public interaction or business transactions.

  • Future discussions will focus on other forms of discrimination, such as gender differences in acceptance in various domains.

robot