Study Design
Explain Experimental Design
Experiments have a high control of setting, they involve manipulating the IVs to determine whether they have an influence on the DV, and they typically have control and test groups.
Experiments allow us to have confidence in the robustness and trustworthiness of causal findings. Very strong in internal validity.
There are field experiments and lab experiments.
What is Quota Sampling?
Non-Probability
Produce a sample that reflects a population in terms of the relative proportions of people in different categories, such as gender, ethnicity, age groups, socio-economic groups, and region of residence, and combos of these categories. AKA "get x number of people in your sample from y group."
Sampling of individuals is not carried out randomly - final selection of ppl is left to the interviewer.
What is a p-value?
The percent change the findings are due to chance.
A low p-value (at or less than 0.05 - at or less than 5% probability) indicates strong evidence against the null hypothesis, prompting its rejection. This means that the findings are likely NOT due to chance (a good thing for you).
What is an index, concept, and indicator?
Index = multiple indicators measuring one concept (all of your likert scale Qs together)
Concept = categories for the organization of ideas and observations but meaning can vary from person to person
Indicator = an instrument (e.g. a question) that captures one dimension of a concept (think of your likert scale Qs)
"Does money buy happiness" study
Cross-sectional
Probability sample (simple random)
Finding emotional wellbeing levels off at 75k but evaluation of life keeps going up
Explain Case Study Design
Involves detailed and intensive analysis of a SINGLE case.
Concerned with the complexity and particular nature of the case in question.
Per the textbook (3.6), can focus on a single community, single school, single family, single organization, single person, and/or single event (among others - key word SINGLE).
What is snowball sampling?
Non-Probability
The researcher makes initial contact with a small gorup of people who are relevant to the research topic. Then, the researcher uses these initial people to establish contacts with others.
AKA getting participants through referrals.
Define univariate, bivariate, and multivariate. What does this mean?
Univariate = one variable at a time
Bivariate = two variables at a time - the relationship between two variables.
Multivariate = three or more variables at a time; to see if the effect is independent.
What is a reactive effect?
People act better when they know they’re being watched
Telles & Lim article
Cross-sectional design
Probability sample (stratified)
Income inequality between whites and browns in Brazil is higher with interviewer classification of race (26%) than with self-classification (17%), whereas black-brown differences hardly change
Explain Cross-Sectional Design
Involves collecting data on a sample of cases and at a single point in time. The SAMPLE represents a cross-section of relevant groups.
The researcher collects a body of quantitative or quantifiable data in connection with 2+ variables. These are then examined to detect patterns of association. See 3.4 of textbook for examples!
What is Simple Random Sampling?
Probability Sampling
Each unit (e.g. person) has been selected by chance and each unit has a known and equal probability of being selected.
What is ecological validity?
Naturalness
Are the findings applicable to people's everyday, natural social settings?
Opposite = artificial setting
Explain/define mean, median, and mode
Mean = average, affected by outliers
Median = is exactly halfway, not affected by outliers, but affected by lots on one side
Mode = the most frequent value but you don’t always have one
Moss-Racusin et al. study
Experimental design
Non-probability sample, quota
Faculty rated male applicants as more competent and hireable, the gender of the faculty did not matter, having preexisting subtle bias against women played a moderating role
Explain Longitudinal Design
A study of some unit across time.
Two main types: panel study and cohort study.
Panel studies: often at the national level. Data collection focuses on at least 2 or more occasions. Cases can include: people, households, organizations, schools, etc.
Cohort studies: Either an entire cohort of people or a random sample is selected as the focus of data collection. The cohort is made up of people who share a certain characteristic (this is what makes it different from panel study), for example, everyone born on a Tuesday.
What is a Stratified Random Sample?
Units are selected at random from a population that has been categorized (put into 'strata').
You pick your criteria, and then you do random/probability sampling within that criteria.
What is reliability and replicability?
Reliability = can you get the same result multiple times?
Replicability = can someone else get the same result in another study?
Explain the deductive vs. inductive relationship between theory & research
Deductive = theory comes first (before data collection)
Inductive = theory comes second (after data collection)
"Pediatricians asking children questions" study
Design = cross-sectional
Non-probability sample (convenience; quota)
Findings = children are asked more if question about more “social” stuff, if older, if white, if father there. No effect of child gender; parent age; doctor age, gender, race
Identify a kind of validity each research design is typically strong in, and why?
Experimental, Cross-Sectional, Longitudinal, Case Study
experimental design - internal
case study – ecological
cross-sectional design - external
longitudinal - internal
Explain sampling bias and sampling error.
Sampling bias = happens when a sample doesn’t represent the population well -- some members of the population are more likely to be selected than others
Sampling Error = happens when the sample and the population differ due to deficiencies in the instruments or process of sampling
Describe at least two things you need to keep in mind when designing research questions.
There are more than 2 options… presuppositions, polarity, social preference, ordering effects (consistency, contrast, additive & subtractive), asking exactly what you want to know, not being confusing, agreement bias, etc.
Give a description and an example of a categorical, ordinal, dichotomous, and interval variable
Categorical – in categories, no order
Ordinal – in an order, not an even distance
Continuous – In an order and equal distance
Binary – two
Cao & Banaji study
Experimental + longitudinal (or repeated measures) design
Non-probability sample (volunteer visitors to Project Implicit — implicit.harvard.edu)
Key finding is that after learning facts about individuals, explicit beliefs correctly aligned with the facts (be they stereotypic or counterstereotypic), but implicit beliefs were immune to counterstereotypic facts and continued to follow the base rate principle