exam prep Flashcards
ontology
nature of the social world
epistemology
what can we know about social phonomena
does positivism use induction or deduction?
induction
observation => theory
does logical positivism use induction or deduction?
deduction
theory => observation
abduction: what is it? which approach?
abduction is selecting the most simple explanation that best explains something, and is used by scientific realism.
analytic review
- summarize
- evaluate
- conceptualize
grounded theory process
- coding
- sorting
- memo writing
key elements of causality
- non spurious
- temporal ordering
- spatial and temporal contiguity
- covariance
measurement validity - content validity
does the conceptualization and operationalization match, does the concept cover what it’s. necessary for the theoretical definition?
measurement validity - criterion/ construct validity
does the measurement actually measure what it seeks to measure?
convergent validity
how closely a test is related to other tests that measure the same (or similar) constructs.
concurrent validity
does it correlate with something that’s happening at the same time?
challenges of selection
5
- selection bias
- outliers
- non-equal size: heterogeneity
- historical contingency (joint history)
- path dependency (stable trends)
rolling cross section (longtidunial)
dynamic changes and trends 1 huge sample divided into groups, and interviewed at different times
methodological issues in surveys (3)
- surveys usually focus on predicting, rather than explaining
- surveys cannot identify causation, but can only imply causation. However, in panel studies making causal explanations is more possible.
- It’s important to not make inferences accross levels. For example, “individualistic fallacy” is generalizing from the micro (individual) level to the macro (group, general) level, and “ecological fallacy” is making inferences from the macro (group, general) level to the micro (individual) level. These types of fallacies should be considered while conducting survey research.
solutions to methodological issues in surveys (4)
- randomizing questions
- writing balanced questions (like giving both pro and con statements)
- pre-testing (to see if questions are easy to understand)
- monitoring and verifying to see if there’s problems
types of probability sampling (3)
- simple random sampling
- strafied sampling
- multistage cluster
types of non probability sampling (4)
- convenience
- snowball
- theoretical
- purposive
types of response biases (2)
self selection
non response
conditions of probability sampling (3)
- every unit has equal chance of being chosen
- an observer can’t predict which units will be chosen
- sample must include all possible combinations of units
difference between stratified random sampling and systematic random sampling?
- In systematic sampling, the list of elements is “counted off”. That is, every kth element is taken.
- Stratified sampling also divides the population into groups called strata. However, this time it is by some characteristic.
multistage cluster sampling
random sampling in several stages from big to small.
is quota sampling probabilistic, or not probabilistic?
not probabilistic
quota sampling
having quotas of subgroups so you can compare them
(but doesn’t represent the actual distribution)
- if the quota is ARBITRARY and DOESN’T REFLECT REALITY => NOT STRATIFIED
- if quota is REPRESENTATIVE and REFLECTS REALITY => STRATIFIED
three types of comparative historical analysis
- parallel demonstration of theory
- contrast of contexts
- macro causal analysis
the article by ulriksen and dadaulari is an example of
indepth process tracing and single case study for theory testing.
influential cases are used for..
theory testing
(influential cases should confirm a theory)
pathway case selection
cases that allow the researcher to isolate certain causal mechanisms- to prove that this and nothing else leads to the effect.
difference between cohort and panel
in a panel study, random selection is done once, and we can track individual-level changes, and control for the differences that may occur within different samples.
in a cohort study, the researcher does random selection multiple times in different time periods. This alllows the researcher to observe changes at the aggregate level. in a cohort study, you can see how age affects certain variables, or if a critical event influenced all segments of the sample, or if attitudes are tied to specific birth cohorts like milennials etc.
when is qualitative comparative analysis useful?
medium to large n comparative research
- too many cases to study in depth, but not enough cases to study qualitatively
what’s data reduction?
related to: semi/un-structured interviews: systematically summarizing and interpreting responses
assumptions of experiments
- same unit different treatment (counterfactual-unit homogeneity)
- conditional independence (iv is independent of dv)
- no endogeneity, yes exogeneity
- no selection bias
- no omitted variable bias
what’s matching? (under random assignment)
random assignment that’s sensitive to the demographics of participants
solomon four group design
four groups:
2 pretest- posttest
2 posttest only
=> to both verify that random asignment was able to create equivalent groups while having the advantage of posstest only.
what’s the interaction effect examined in mintz?
participants x certainty
syntactical unit of content
words, sentences, paragraphs
referential unit of content
events, people
thematic unit of content
topics
3 V’s of big data
volume
velocity
variety
data mining
focusing on data correlations instead of theory by analyzing large volumes of information.
key assumptions of inferential statistics
- sampling from complete target population
- simple random sampling with perfect response rate (or non response completely random)
- no nonsampling erorr
sampling distribution
if you draw an infinite number of samples of the same size, you will get a sampling distirbution.
necessary conditions for causality (4)
- covariance
- temporal ordering
- spatial and temporal contiguity
- non spurious
