3- Methods Flashcards
1- Some terminology
Experimental control: The degree to which a researcher can determine the environment in
which a research question is explored. In general, a “lab study” has high experimental control.
Psychological realism: Degree to which a study simulates phenomena as experienced in everyday life. More immersive studies would have higher psychological realism.
Internal validity: The degree to which a study can rule out alternative hypotheses. If a study has many alternative hypotheses that could explain the results other than that proposed by the researcher, the study has low internal validity.
External validity: The degree to which the results of a study generalize to other contexts, samples, and time periods.
A Sample Research Question
How do people “feel” towards Black people?
There is no ‘right’ answer. Each approach has it’s
own strengths and weaknesses. The goal is to
maximize correspondence between a study’s
research question and it’s methods.
2- Self-report (explicit attitudes)
Self-report measures: Assessing attitudes, thoughts or beliefs by presenting a question, several possible responses, and having the participant select the response that they believe
to be most of their own attitudes, thoughts, or beliefs.
One straightforward item:
Which statement best describes you?
-3 = ‘I strongly prefer Black people to White people’
-2 = ‘I moderately prefer Black people to White people’
-1 = ‘I slightly prefer Black people to White people’
0 = ‘I like Black and White people equally’
+1 = ‘I slightly prefer White people to Black people’
+ 2 = ‘I moderately prefer White people to Black people’
+ 3 = ‘I strongly prefer White people to Black people’
Results: 70%=0, 10%=under 0, 20%= over 0
3- Implicit Attitudes
‘Indirect’ measures: Inferring attitudes, thoughts or beliefs from some type of behavior
rather than from self-report..
One example of an ‘indirect’ measure is the Evaluative Priming (EP) procedure.
Evaluative Priming uses the reaction times of identifying positive versus negative words to infer ‘implicit’ racial attitudes.
If certain stimuli facilitate the identification of negative words, then those stimuli are
believed to hold a negative association. If certain stimuli facilitate the identification of
positive words, those stimuli are believed to hold a positive association.
Results (ish): 15% under 0, 30% 0, 55% over 0
Physiological measures:
ex: measure heartbeat….higher when evaluating black person
4- Measuring Behaviour
Many studies try to capture a behavioral measure during a lab context. For example, in an adaptation of a prior study (Haddock, Zanna & Esses, 1994), participants
may be told to imagine that the student government had to make some severe budget cuts for the next year, and that the student government wanted other students’ opinions on how to make budget cuts. Participants would be then shown a list of several organizations and asked to propose some
budget cuts (club sports, music groups, etc.). One student organization would be specifically about Black people (e.g., Black Students Association). The degree to which students cut funding to the Black Students Association could be used as
a measure of how much they dislike Black people.
Intensive behavioural measures:
Some studies go the extra mile by trying to create a high-investment environment. Frank Kachanoff (a McGill alum!) uses a multi-hour long study where participants
join a group, complete computer missions together, and develop an entire culture
by selecting a flag and identifying group snacks.
Then, Frank uses these ‘maximal group paradigms’ to look at questions related to how experiencing lower or higher status impacts group identity and behavior.
A Disconnect Between Predicted and Actual Behavior
Kawakami, Dunn, Karmali & Dovidio, 2009
In one condition, non-Black participants were asked to imagine seeing a White person use a racial slur towards a Black person. They were then asked who they would choose to work
with in a follow-up task. In this hypothetical version, 20% of participants said they would choose the White person.
In another condition, non-Black participants actually saw the interaction take place between a Black and a White person pretending to be fellow participants. Now, when they had to choose a partner for an upcoming task, over 60% chose the White person.
5- Measuring ‘Real’ Behaviour
Other studies use methods that are meant to capture more impactful behavior that are also hard for participants to regulate in the moment.
One example of such a measure is the First-Person Shooter Task (FPST).
Results: rates of shooting an unarmed person were higher when target was black (17%) than white (12%). rates of not shooting an armed person were higher when target was white (13%) than black (7%).
Measuring ‘Realer’ Behaviour
Participants entered the Psychology building and found a note asking them to take a seat in a waiting area. Other participants (actually confederates) were strategically placed around the room. At one
end of a string of empty seats was a Black confederate, at another end was a White confederate.
Researchers analyzed where the participant chose to sit. On average, the (White) participants sat closer to the White than the Black confederate.
6- Archival Analyses: Measuring ‘Real-World’ Behavior
Archival analyses use existing datasets to understand how social forces might influence realworld behavior.
For example, Voigt et al. (2017) coded how police officers spoke to Black versus White drivers following a traffic stop. They found that officers were much more polite and reassuring to
White drivers, and more disrespectful to Black drivers.
In 2011, the US minimum amount of cocaine needed for a felony offense
changed from 50g to 280g. Tuttle (2019) investigated how this change
in policy impacted racial
discrimination in police reports and court sentencing.
Pierson et al. (2020) analyzed over 100 million traffic stops and found that the percentage of stops that
were of Black drivers decreased after
dusk (when it is presumably harder for officers to identify the race of the driver).
7- Audit Studies: Measuring ‘Real-World’ Behaviour
Audit studies: Attempting to combine the experimental control and internal validity of a lab study while also measuring ‘real-world’ behavior. Typically involve finding situations that involve judgment or behavior that may be influenced by social information (such as race, gender, or ethnicity). Fake ads or resumes may be
created, or even trained actors hired, in an attempt to keep everything constant between conditions except social information.
Governments have been known to run audit studies themselves to explore whether antidiscrimination laws are being respected (for example, for equal access to housing).
Butler and Brockman (2011) tested whether elected officials discriminate against their constituents based on race. State legislators in the US were randomly assigned to receive the same email, but the email was made to seem as if it came from ‘Jake Mueller’ or ‘DeShawn Jackson’. Though not a large effect, the same email was more likely to receive a response if from ‘Jake Mueller’ (61%) than from ‘Deshawn Jackson’ (55%).
* This was true only when the email did not signal a
political party affiliation.
Kang et al. (2016) investigated the phenomena of ‘Whitening’ resumes, where minority applicants remove racial cues from resumes to avoid anticipated discrimination. They found real job ads and randomly sent each one a resume that was either low vs. high in
‘Whitening’ cues. Despite having the same qualifications, the
‘UnWhitened’ applicant only received a callback to
10% of jobs, compared to 25% of jobs for the
‘Whitened’ applicant.
Ayres (1991) trained Black and White male and female actors to use the same negotiating strategy when trying to purchase a new car, visiting 90 car dealerships in the process. Negotiators were matched on age, education, and occupation.
On average, White negotiators received a final price offer that was $575 dollars lower than Black negotiators.
* There was also gender discrimination. White women
received worse discounts than White men, and Black
women received worse discounts than Black me.
