SOCI 230 Lecture Notes - Lecture 12: Voicemail, Critical Legal Studies, United States Presidential Election, 2016
3 May 2018
School
Department
Course
Professor
Lecture 12: How do we Measure Prejudice and Discrimination? (continued)
February 13th, 2018
! Method 3: statistical analysis
•!Take disparate outcome (e.g., wages)
•!Control for confounding characteristics
•!Goal: see if difference can be “explained” away
•!Residual (what’s left) = discrimination
•!Problems with statistical analysis:
o!Omitted variable bias
! Video: audit study
•!You have a quasi experimental design
•!You send 2 experts out to do some kind of activity – in this case, they’re both looking for
jobs
•!You are controlling for everything
•!You try to control all of the characteristics and the only variable that you manipulate is race
•!Auditors are sent out to find employment
•!Video:
o!Pay attention to how they are treated
o!John is told pleasantly something, whereas for the other audit, Glen, (who is black) he
is told in a way of a lecture and people mention laziness
o!Glen is told the position is already taken, but when John asks, the position is still open
o!When John and Glen try to look for some place to live, the building manager gives
John a relaxed description of the neighborhood area, but tell Glen how the place
needs to be “cleaned” at all times
! Method 4: audit (experimental) studies
•!Manipulating one variable in this case, race
•!The video is far from an academic study, it’s an investigating news show and we might have
to question their reliability for the evidence they found
•!Quasi-experimental design
•!Vary one social characteristics (race) and see if treatment for otherwise 2 equal individual
changes
•!“Clean” estimate of prevalence of discrimination
•!An audit of employment and housing discrimination
•!Study: what’s in a name? (pre-contact discrimination)
o!Sending out CVs in a number of sectors, and we are manipulating the name of the job
candidate. They are assuming that the employers will read the names and assign
particular racial categories to the applicants. The study was conducted in 2010-2011,
the research design was to have matched CVs. They also designated on the CV that
the candidates are fluently bilingual (since this is done in Montreal) and they used
names of real Montreal companies in terms of past experience. They also assigned 2
fake voicemail phone-lines: one line was assigned to the majority group applicant,
and the second line was assigned to the minority group applicant – no accent but you
Document Summary
Lecture 12: how do we measure prejudice and discrimination? (continued) Method 3: statistical analysis: take disparate outcome (e. g. , wages, control for confounding characteristics, goal: see if difference can be explained away, residual (what"s left) = discrimination, problems with statistical analysis: o omitted variable bias. John a relaxed description of the neighborhood area, but tell glen how the place needs to be cleaned at all times. They are assuming that the employers will read the names and assign particular racial categories to the applicants. The study was conducted in 2010-2011, the research design was to have matched cvs. They also designated on the cv that the candidates are fluently bilingual (since this is done in montreal) and they used names of real montreal companies in terms of past experience. They targeted 3 different sectors in terms of employment: private companies, public companies, and non profit organizations. They also controlled for gender. o majority names:
