week 1-8 Flashcards
NORMATIVE RESEARCH
study of what OUGHT to be
EMPIRICAL RESEARCH
study of what IS
ONTOLOGY
what is the nature of the social world and what are its constituent parts
EPISTEMOLOGY
what is knowable
Positivism
social world same as natural world - can use same methods
Ontological assumptions-P
no difference between social and natural world -objective reality
Epistemological assumptions-P
scientific knowledge is limited to what we can actually observe -law like generalisations
INTERPRETIVISM
scoial world not same as natural world
ontological assumptions-I
scientific knowledge about social world can only be gained through interpreting meanings
EPISTEMOLOGICAL ASSUMPTIONS-l
scientific knowledge about the social world can only be gainede through itnerpreting meanings
OPERATIONALISATION
theories involve concepts - CONCEPTS DEFINED TO MAKE THEM MEASURABLE
ECOLOGICAL VALIDITY
artificial experimental settings MAY NOT GENERALISE TO WORLD
POPULATION VALIDITY
experiments often involve UNREPRESENTATIVE subject pools
REACTIVITY
people may change behaviour when they know they are being observed
observational research design
researcher does not have control over values of the INDEPENDENT VARIABLE - good for description , explanation , variability in terms of internal validity
experimental research design
experiments , a research design in which researcher both CONTROLS AND RANDOMLY ASSIGNS VALUES OF THE INDEPDNENT VARIABLE
internal validity vs external
degree to which we can be confident that a study identifies the causal effect of the independent on the dependent variable
EXTERNAL-degree to which findings can be GENERALISED to other contexts
types of observational design
TIME SERIES - over time , not over diff units
CROSS SECTIONAL-different countries , not over diff units
PANEL - over time and cross units - but same unit over time
REPEATED CROSS SECTIONS- over diff cross units and over time buyt not same units
TIME SERIES CROSS SECTIONAL-over time over cross units , and same units - SAME AS PANEL
LARGE C DATA
numbers quantitative scientific realism high standardization
scientific realism
small c number meaning , quantitative +qualitative numbers stats words high or average standardization
interpretivist small c
data expressed as meaning , qualitative , words , low standardisation
LEVELS OF MEASURMENT
NOMINAL- categories without any natural order e.g type of political system
ORDINAL-data arranged in a meaningful order , but intervals between rankings may not be equal, e.g level of interest
INTERVAL-numeric scales with equal intervals between values but no true 0 point e.g put yourself on scale 1-10
RATIO - TRUE 0 point , numeric scales with equal intervals e.g number of protests
LARGE C RESEARCH
less intensive study of a large number of cases using quantitative methods- avoids sample selection bias - selection on dependent variable
-typically deductive
-inc potential for generalizability
-inc ability to identify causal effects
-less useful for inductive research , interpretivist
SAMPLING strategies
- probability sampling- random - leverages law of large numbers
-non probability sampling
SMALL C RESEARCH
-intensive study of a single case or small number of cases
-qualitative methods
-good for causal mechanisms , inductive research , thick description
case selection in descriptive case studies ?
TYPICAL AND DIVERSE CASES
case selection in large c research - total population sampling
-select all cases in a population e.g census
high external validity
simple random sampling / probability sampling
-leverages law of large numbers
representative
all cases drawn from pop with same probability randomly
stratified random sampling -probability sampling
populations divided into relevant strat then drawn at random from different strat
non probability sampling- convenience sampling methods
convenience sampling - volunteer , snowball
quota sampling - surveys
-cannot be generalised to population
but high internal validity if rely on experimental designs to rule out confounders
Ways of tackling confounders in observational studies
-statistical control
-most similar and most different designs
-time changes observations
observational research design ? what is it good for ?
Researcher does NOT have control over values of independent variable
-DESCRIPTION QS AND EXPLANATIONS QS
experimental research designs
researcher both controls and randomly assigns values of the INDEPENDENT VARIABLE to the participants
-deductive
types of experiments
laboratory , field ,survey
special case : natural experiments
observational design
TYPICAL CASES
one or multiple cases which represent a larger population well on important features
-strategy when goal is description
diverse cases
-populations often diverse
case selection in explanatory case studies
-inductive studies
-testing of causal hypothesis
-mechanism studies
inductive studies - case section techniques
explanatory search for new explanations for a phenomenon
EXTREME CASES , DEVIANT , MOST SIMILAR , MOST DIFFERENT
CASE SELECTION:
most similar design + most different
extreme and deviant cases
crucial cases
1- minimum of two cases which similar in terms of background conditions but differ in terms of X OR Y
2-extreme - unusually values in Y
deviant -one or multiple cases which deviate from a common causal pattern
3-Most LIKELY - EASY test that if fails cast as strong doubt on theory
,least likely = HARD test that if passes provides strong evidence for theory
causal hypothesis - common case selection techniques
DOES X CAUSE CHANGE IN Y
crucial cases , most similar cases , most different cases
mechanism studies
exploration or testing of causal mechanism
common case selection - pathway cases
