Discrimination In labour Market and Schools

definition of disctimination

identical individuals are treated different due to non productive characteristics

what are two types of Discimination

• taste discrimination

• statistical discrimination

Taste discrimination def

employers may not like certain groups

In which type of economic market can taste discrimination occur

only occur if firms have market power

Statistical discrimination

• employers use membership in a given group as signal about individuals productivity

what is signal extraction problem

difficulty of estimating unobservable ability from noisy signals. In statistical discrimination, it explains why employers might rationally rely on group averages, especially when individual info is imperfect.

What are the two main psychological views on roots of discrimination

• evolutionary (group boundaries enhance coop)

• Unconcious phenomenon (people behave in ways that are unrelated to thier explicit views)

What are the 3 ways to identify discrimination

• decomposition Method

• experimental methods

• implicit association tests

what is the decomposition method

• Explained part – differences in characteristics (e.g. education, experience)

• Unexplained part – differences in how the labour market rewards those characteristics
  → Often interpreted as discrimination

What are the limitations with using wage equations

• conditional mean independance assumption

• selection bias

Why is conditional mean independance assumption a limitation

The conditional mean independence assumption states that after controlling for observable characteristics, group membership (e.g. gender) should not be correlated with unobserved factors affecting wages.
If this assumption fails, the “unexplained” wage gap may capture more than just discrimination — such as preferences, unmeasured ability, or job sorting.

why is selection bias a problem

• only observe those who work not everyone

Why is The invariance of conditional distribution assumption a problem

The invariance of the conditional distribution assumption says that if one group were paid like another, the rest of the wage distribution would stay the same.
If this doesn’t hold (e.g. due to job sorting, selection, or different unobserved factors), the decomposition’s unexplained part may misstate discrimination.

Types of experimental method to asses discrimination

• Audit Studies

• Blind Studies

• Correspondence studies

What are audit studies

• compare labor market performance of identical individuals except for membership in specific group e.g. send cvs with different names same content

Limitations of Audit studies

• don’t give drivers of discrim

• eventhough many characteristics are matched not all

• studies not double blind as auditors know purpose

What is Blind study

• hide appearance of person compare to it they don’t hide

What is correspondence study

• sending ficticious but identical Cvs but mixing the names to show the race of the person applying

cons of correspondence test

• ignores real life importance of networks

• Only at the entry level

• binary outcome and so no inference of employers behavior

Policies to Combat Discrimination

• quotas

• combat pre market differences

• inter group contact

• Debiasing Training

What is main finding of Marianne and Mullainathan (emily and greg)

• large difference in call back rate between whites and blacks

• whites saw larger increase in call back if resume quality increased

What methodology did they use (emily and greg

• took real resumes and made adjustments to ensure employees data was similar

• 2 cities chicago and boston

• call backs recived thorugh mail, email and telephone

Weakness of (emily greg)

• doesn’t directly report race

• newspaper adds used only one channel for job search

other results from emily and greg

• computer skills negatively affect call back

• benefit to living in whiter area

Do stereotypes reduce girls paper main result

• contrary to initial belief male students are discriminated against in israeli schools

• Extent of bias varies by subject and test

• On average bias doubles gender score gap

Do stereotypes reduce girls paper methodology

• students take blind test abd non blind and compare the results of the natural experiment found in the israeli education system

Do stereotypes reduce girls paper explanations

• boys may behave worse in class affecting teacher perceptions

• self fulfilling stereotypes

Immigrants in schools paper Main findings

• teachers with strong stereotypes on IAT gave high achieving immigrants lower grades

• Telling teachers IAT scores was far more affective than general debiasing

Immigrant paper Methodology

• ran IAT test on teachers which recorded time taken to respond to stimuli

• Compare grades of teachers with high and low

• Give teachers opportunity to see results and mark differently

interesting fact on IAT in immigrants study

• even if you tell people how test works they can’t change result

• 80% of teachers underestimate thier level of bias