Research Methods Midterm Flashcards
What are the four elements of a theory?
Falsifiability, testable, generalizable, probabilistic (not definite one way or the other)
Inductive vs deductive approaches
Inductive= creates broad conclusions
Deductive= creates imperial analysis
What’s this classes def of sociology?
The scientific study of the social lives of individuals, groups, and societies
Generalizeable
Means how well one conclusion based on one population can be applied to other groups
Agency vs social structure
Agency is an indivials capacity to make and execute decisions
Structure is our social arrangements that facilitate our options
Are sociologists or anthropologists more focused on cultural relativism?
Anthropology’s focus on non-western societies from western researchers focuses on cultural relativism more than sociology, but cultural relativism is relevant to both.
Triangulation/mixed method approach
Using multiple research methods for the same project (like collecting qualitative and quantitative data)
Cross sectional study design
Where people are researched at a specific point in time and place
Repeated cross sectional study design
Studying different people from the same population across different points of time. Also called a trend design. Cannot show changes in individuals.
Panel design
Studying the same people in the same place across different points of time. Cannot show changes in groups.
Longitudinal study designs
Studies built to record data from different points in time
Arrition is
The loss of panel members from death or dropouts (or perhaps ghosting)
Longitudinal vs cross sectional studies
Longitudinal focuses on a timeline approach, where cross sectional is more a coordinate point on an X/Y graph
Longitudinal studies are
Studies that take place over time
Cohort design
Panel design specifying a commonality between the participants (same generation, fought in same war, graduated high school the same year)
Units of Analysis
Refers to the level of social life about which we’re generalizing
Ecological fallacy
When assumptions are made about a micro population from conclusions made from macro research (a country is rich does not mean every individual in it is rich)
Value-free
Is being unbiased in your research (hence, having no values)
Is it possible for researchers to be value free in their work?
No
Subjectivity and reflexivity
S. is when researchers are subject to their social contexts. R. Is when they acknowledge and realign their contexts
Descriptive research asks
What is happening
Explanatory research asks
Why it’s happening
Exploratory research asks
How is this happening?
Here, a theory is
An argument to explain an aspect of social life
Empiricism is
The idea that we can objectify and measure our world
In general, scientists use theory to
Explain, describe, and explore
Give an example of inductive reasoning
An analyst finds married people are healthier and then develops a theory of social support and health
Give an example of deductive reasoning
A sociologist reads Durkhiem’s theories of social relationships and designs a study to test these ideas.
Positivism as a paradigm believes
that you can be empirically certain of something happening or not happening, we can take an objective view of reality
Postmodernism as a paradigm believes
that we’re a mess and have to be skeptical about our principles, bias seeps into the “truth”
What are the four sociological paradigms (according to here)?
Structural functionalism (Durkheim, Parsons) Conflict (Marx), Rational choice, symbolic interactionism (Herbert Mead)
Which paradigm treats society the most like an efficient economy?
Rational choice
What’s an example of spuriousness?
Ice cream and polio are spurious variables, and summer is the confound variable that truly links them
Mediation in variables
A connection: if sleep quality (IV) and academic performance (DV) are studied, alertness is the mediator between the two, its how their connected
Moderation in variables
(Think Discord Mod) is when the strength of the association between two variables is made weaker or stronger by the third variable (can be race, religion, weight, language, nationality). If you studied the link between social media use and lonliess, age would be a moderator since this relationship is stronger among adolescents than the elderly.
Hypothesis of difference example
People above 85 are less likely to vote for Trump than people younger than 85 because they remember fascism in WWII
Hypothesis of association example
Ice cream sales and polio rates will both go up in the summer.
Causal hypothesis
Eating milk will make your bones grow more, meaning your skeleton will outgrow the rest of you
Causal inference
is the degree of confidence a researcher has that the observations made of a test are truly causal.
conceptualization is
turning ideas into variables
Operationalization is
making a procedure out of conceptualized variables
Give an example of a false inductive observation
This building I’m in has the lights on. All buildings have lights on inside them.
Give an example of true inductive reasoning
I observe the sun rising in the east and setting in the west. Therefore, the sun rises in the east and sets in the west everywhere. (If something is true for the member, it must be true for the class.)
Give an example of false deductive reasoning.
I just read an article discussing the daily problems of women wearing makeup; therefore, all women wear makeup. (If something is true for the class, it must be true for the member.)
Give an example of true deductive reasoning
Its my first day on earth, and I just heard from a farmer that all birds lay eggs, and she has some chickens, which are birds. Those chickens must lay eggs.
Human Terrain system is
an ethical dilemma where social scientists are embedded in combat units to study the local culture to report to their leaders (subjects can’t consent be studied; subjects can’t be studied if they know they’re being studied)
Ethics are
A moral system
What’s the Belmont report?
a 1979 publication by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. Emphasizes principles of respect, beneficence, and justice
A vulnerable population is
one that cannot give informed consent (minors, cognitive impairments)
Privacy is
having control over the extent, timing, and circumstances of sharing oneself with others.
Institutional review boards examine
Research being performed on human subjects and ensures it follows all ethical standards
The Nuremberg Trials
persecuted Nazi criminals of WWII and created the Nuremberg Code for ethical human research
The Tuskgee Syphilis Experiment was
a 1932-1972 US public health study of 400 poor Black men in which they were unknowningly infected with syphilis and not given functional antibiotics for them once made available. AKA a total shitshow
The Wichita Jury Study
a 1955 study where jury deliberations were secretly filmed, federal law was passed barring the recording of jury deliberations
The Milgram Obedience Experiment was
a 1960s experiment to see how much pain subjects would inflict on someone else if an authority figure (the experimenter) told them to (the electrocution one). Over half of the participants completed the study.
The Stanford Prison experiment
a 1970s psychological experiment (guess where) testing the extents one would fulfill the roles of “prisoners” and “guards.” It was stopped early.
Tearoom trade study
a 1970 experiment where Humphreys studied tearooms, where men have sex with other men. He violated their privacy.
What did the US Natl Research Act of 1974 establish?
Protection for human subjects against ethical violations in research. Federally funded research must have an Institutional Review Board.
Human subjects research is
any study designed to develop or contribute to generalizable knowledge
T/F There is disagreement about the effecitivess of IRBs
True
T/F Researchers have the legal right to refuse cooperation with law enforcement
Nope
A certificate of confidentiality is
a credential issued by the Natl institutes of Health allowing researchers to protect participants from future requests for data disclosure
Deductive disclosure is
using unique variable combinations to prevent participant info from being deduced
Differential privacy
is when there’s enough noise added to the data so no individual can be recognized in it
Demand characteristics is
the process of research subjects confirming the hypothesis once they’re aware of it (like think of nothing but a pink elephant! Do it! Now!)
The ethical rule for using online deception for research purposes is
researchers cannot study behavior that people reasonably expect to be private
p-hacking is when
researchers choose to present only the statistical models that show as significant
The FPL is calculated by
measuring the annual cost of an adequate diet for a family and multiplying it by three
From lowest to highest, what are the levels of conceptualization?
Concept in hypothesis, dimension, variable, indicator
Social artificats are
aspects of social life that can be counted (letters, tombstones, votes for a politician)
Operationalization is
the process of identifying a plan for measurement
T/F Reductionism and ecological fallacies are the inverse of each other
True
Composite variables are
variables that combine and average a set of items to measure the same concept (GPA is one)
True/False Internal reliability is measured using Cronbach’s alpha, a scale of reliability between 0-1.
True
T/F With 0 being zero and 1 being certain, 0.7 is the convential standard of an alpha for social science.
True
T/F Intercoder reliability is how much different statisticians agree on the reliability of the same data, and CLASS coders should agree about the standard coding norm 90% of the time
False. Only 80% of the time.
What does Cohen’s kappa measure?
Intercoder reliability, measured on a 0-1 scale.
What’s the split-half method?
A measure of robustness by testing the similarity of results by giving two samples two subsets of an item (Like giving one sample apples and the other grapes to see if the population likes fruit)
The test-retest method is used for what?
The measure’s robustness: if respondents’ results are similar at multiple times, the measure is reliable
T/F Pilot testing in sociological studies is exactly what it sounds like
True
internal validity of the study vs internal validity of the measures
study= how well the independent variable is proved to impact the dependent variable
measures= how well the measures accurately capture concepts
T/F Face validity is how well something is valid on its face
True
What are the two types of criterion-related validity?
Concurrent validity and predictive validity. Concurrent measures against existing measures and predictive measures against something it should be correlated with
External validity is
How well the findings of the study can be applied outside of its immeidate discipline
What’s an example of content validity?
A statistics exam having questions on all materials covered in class and having zero spongebob essay requirements
What’s an example of construct validity?
A questionnaire measuring student anxiety levels asking about fatigue, irritability, and nausea and not if they hallucinate has high construct validity. More theoretical.
Convergent validity and discriminant validity
Both are a dimension of validity gauging whether concepts should be (convergent) grouped together or whether they should NOT be (discriminant)
Here, a vignette is what?
A hypothetical situation where people are asked to put themselves in another’s shoes (an empathic story)
A probability sample is
a randomly selected sample
A population parameter characterizes what?
some quantitative aspect of a population, a “true measurement.”
Discuss nonprobability samples
They’re samples that are chosen deliberately, not randomly. Convenience sample (reaching for whats closest), quota sample, voulnteer sampling, and judgemental sampling are all examples
What’s the convential confiednece level?
95% (There’s a 95% chance that 41% of Black people xyz)
What’s a sampling frame?
A list of all population members
What’s the oversample?
A group that is deliberately represented at higher rates in the sample than its frequency in its population to compensate for underrepresentation in the population
Postsurvey weighting is when
a sample is weighted in the study to match the population it reflects
Variable vs case-oriented research
variable= limited data (variables), plenty of cases
case= small number of cases, plenty of data on those cases
T/F Choosing a single case to study is considered sampling
true
Sequential sampling is when
case-based researchers choose what data to collect next based on the data they’ve already collected
Snowball sampling is indeed like rabbit hole sampling, a term I made up.
Yep!
Big data is kept in administrative records, and is not like big pharma or big glasses (my personal conspiracy theory)
Yep!
What are modes of administration?
How surveys are conducted (face to face, mail, online)
