midterm 1 Flashcards
6 steps of the research process
- forming a question
- reviewing relevant literature
- advancing a hypothesis
- constructing a theoretical framework
- defining concepts, variables and relationships
- creating a research design
empirical vs. normative and how they interact
empirical- research what is
normative- research what ought to be
they cross paths because empirical is shaped by normative thinking and normative is validated by empirical evidence
positivism
ontological- there is no difference between the natural and social world
epistemology- we can gain an objective understanding of the social world through observation
methodology- we can find law-like generalizations that the social world runs on
historical context of positivism
behavioral revolution- was the application of positivism and empiricism to the field of political research, and encouraging researchers to go beyond stats and numbers (i.e voting) there an objective way through application of methods to understand the behaviour of individuals.
classical vs. logical positivism
- used in positivism
- classical-> inductive looks “in” to answer questions about the world, goes from a specific case to making a generalization
- logical-> deductive, is looking to a general theory to explain a specific hypothesis
Popper argument against inductive reasoning
observation is limited so therefore we can’t make generalizations without having observed everything, but we can falsify and confirm laws; black swan; stick to logical reasoning
how do we use deductive reasoning to find general laws
deductive-nomological model: something can be explained when attached to a law like generalization; deducing from a theory to explain a particular event
hypothetico-deductive model: treating a law-like generalization like a hypothesis and testing it to verify the law; deduce a hypothesis from a law and test it.
scientific realism
ontological- the social and natural world is the same
methodology- you learn about society through observable and unobservable
methodology- causality; for unobservable, we look at their observable outcomes
causality in scientific realism
they focus on causality to explain unobservable vairalbes
what is causal mechanism
- the system or process by which an effect is produced
3 types of causal mechanisms
- environment- social is influenced by external forces
- cognitive- changes in human’s perceptions
- relational- interpersonal connections among people
what is the structure-agency problem
how do we relate individuals to instiutions; is society deducible to individuals or are individuals part of a bigger collective; which avenue should analysis take; individual or institutions
how do you verify theories of unobservable phenomena
‘interference to best explanation’- which every has the best explanation/justification is the right one.
interpretivism
ontological- the social world and the natural world are two separate entities, and the social world doesn’t exist independent of our perception s
epistemology- can’t use the same methods used in the natural science because they don’t apply
methodology- we understand human behaviour through interpretation and finding meaning behind action
argument of interpretivists
brute data isn’t enough, you need to interpret and look the believes, values and meanings taking place behind the action
posivitivism vs. interpretivism
both use the hypothetic- deductive model
can we have separate facts and values in social science; max weber
max weber- you can have separate questions of fact (because it is different from knowing the facts of something and knowing the value), in this way researchers can focus on questions of fact.
at the same time it is impossible because values influence virtually every aspect of research process
can avoid this by fact and value mixing by stating which was made by observation, logical deduction and empirical analysis- here we make references to values without making value judgements
can we have separate facts and values in social science; Ernest Nagel
one way to diminish the effects of influence of value is recognizing that you have a bias, in order to combat it you have to acknowledge its existence
different influences of value
- enthography-> getting so immersed in the community you are studying, you identify with them and loose your sense of objective judgement
- Rosenthal effect-> expectations of the researcher affect the conclusions of the observation
- Heinseberg effect-> people change their behaviour when being watched.
- self-fulfilling prophecy-> telling people about the theory which makes them act the way the theory suggests
- values within social science research->
Thomas Kuhn value argument
nothing is without value or bias, even researchers have an agreed upon way to conduct research and fall into a norm (society encourages conformity): the cycle: pre-science (fact finding/no paradigm), turns into normal science, a crisis comes which starts a new paradigms, and leads to a new norm)
Imre Lakatos reaction ot Kuhn’s agrument
Kuhn doesn’t recognize the rational and methodology involved in research and boils it down to mob-mentality. Lakatos says that there is a method: there is a core of info that scientists agree not to challenge (negative heuristic), and that those are the basis of the foundation then the questions around the core they question, change as necessary (positive heuristic)
3 criteria for a good research question
it is significant, it is researchable, has not yet been answered
how to find a research question
- look for wholes in existing argument
- look for topics that are in debates (politically significant)
- survey literature for other questions that have been asked in that field
- research grants suggest research questions
research vase
reps. different parts of the research process and question formulation
top- a broad question or topic relating to the subject
middle- specific researchable question
bottom- conclusion of the research question and what it tells us about the roader question or topic
types of questions
you have to make sure your asking the questions you want answers to
- descriptive question: characteristics of something and how to works
- explanatory question: causes of something, what caused it and why something exists or happened
- predictive question: future outcomes of current conditions
- normative question: what ought to be done to bring about certain outcomes
characteristics of unanswerable questions
- they beg other questions
- present a false dichotomy
- ask about fictional events
- are metaphysical
- are tautology
what are the functions of a literature review
- explains why a definite answer to your question doesn’t exist
- sets the stage for your own argument
what are the stages of a literature review
- analysis of the literature
- summarizing the literature
- what still needs to be done
3 criteria for creating a hypothesis
- it is appropriate to the type of question being asked
- adds to existing knowledge
- it is clearly and fully specified- states all the factors that must be considered in order to answer the question
2 functions of a hypothesis
confirmatory research- test an existing theory
exploratory research- act as a guide to research process
theory
attempt to make sense of the world by indicating that some factors are more important than others and specifying relations among them.
types of theories
grand theory- theories that are collectively exhaustive and attempt to explain all observed social phenomena
theories of middle range- theories that attempt to understand a specific section of social life
deductive theory- hypothesis is deduced from a theory
inductive theory- hypothesis leads to a general conclusion
grounded theory- theory grounded in data
empirical theory- concerned with questions answerable with empirical data
normative theory- concerned with right and wrong
3 step process to theory
proposition-> idea/hunch that needs investigation
hypothesis-> tentative argument you wish to develop
theory-> defines the factors that are involved in answering the question and how they relate with the outcome
theoretical framework
set of assumptions on which the variables connect and have significance
criteria and importance of variables in a hypothesis/argument
- without clearly defined variables and their relations a hypothesis cannot be tested,
- independent and dependent variable and proposition
intervening variable + types
- variable that affects the relations between x and y
- moderating variable: variable that affects the relations between x and y
- mediating variable: a variable that transmits the effect form the independent to dependent variable.
2 types of relationships between variables
correlation- x and y are going in the same direction
causation: change in x causes changes in y
4 criteria for causality
- x precedes y
- x and y are correlated
- causal mechanism can be identified through links between x and y
- correlation between x and y is not spurious
research design
constructing a design or strategy for undertaking research for investigating the research in a coherent and logical way
criteria for good research design
- specific on research technique used during data collection
- shows the logic they used to come to their conclusions
- evidence confirms hypothesis, and gives a convincing test to the hypothesis
- decrease threats to internal and external validity
(someone should be able to retrace your steps and end up at the same conclusions)
internal vs. external validity
internal- confidence that x caused y not, z caused y
external- to what extent can your research be applied generally
reliability
on repeat will we get the same results
3 types of studies in research design
small-n, large-n, single-n
small n vs. large n
small- more detailed and good for understanding casual mechanisms because large n studies may be too large that you may see correlation but you arent’ sure whether what is causing it is what the theory says is causing it , depth gives it internal validity, hard to have external validity because its a stretch to use small number of cases to apply to the world generally
large n- external validity because of large cases more likely to apply more broadly
historical design
helps to understand certain events and their influence on others, by contextualizing the events to understand the relations between events, and explain how timing and sequence of actions affect the nature of certain events
experimental design
researcher has direct control over the subject and its environment through lab experiments, field work, and natural occurrences
experimental design trade off
the more controlled the experiment i.e lab the more internal validity the less external validity
experimental data
when there researcher can directly intervene in data results
cross-sectional design vs. longitudinal design
cross- sectional- looks at the differences between people and places at a specific point in time; doesn’t look at changes within individuals
longitudinal- shows how individuals development changes over time and answering questions on dynamics of change
two type of longitudinal design studies
cohort- looking at individuals from a particular section
panel- nation wide random sample and collects infor on it at two or more points in time
historical process research vs. comparative historical research
historical- a series of events in a single case to understand how the process changed over time
comparative- multiple events in many cases and how process in time and between places
purpose of historical research
to answer the questions of the present by looking at how the past helps answer them, and even predict the future
important gathering strategy
set boundaries for the data you need to collect
temporal (what time), spacial (where from), actors that are relevant, variables that are important to arriving to your answer
3 elements of ethical research
- voluntary participation- can withdrawal without penalty at anytime
- informed consent- purpose, process, risks and benefits of the study
- privacy- securing the identity of the participant
data research process
- conceptualization- clearly define the term/defintion has to be valid
- operalization- determining the indicators of the topic, when will we know the topic is there
- measurement- how do we measure these indicators
3 types of validity
- face validity- is there broad consensus on this measure
- content validity- to what extent are the indicators
- construct validity- comparing it to other valid measures
selection bias
choosing cases that directly give you the outcome you want
good data=?
transparency; allowing for verification and replication of data
what two elements does a good argument require
rhetoric: skill of persuasive speaking or writing
research: system of investigation in order to establish facts and reach new conclusions
history of “science” in political science
science was first applied to political science with Aristotle and Plato who used methods and systems to understand how the world works while trying to build the concept of the perfect society
natural vs social science
natural science is found to be more objective because it studies inanimate objects
criteria for a thesis
- significant
- contestable
- specific/focused
- clear, reader should tell exactly what you plant to argue
- assertions are to be supported with evidence
reliability vs. validity- on outcomes
reliability- outcomes are consistent and can be repeated
validity- outcomes meet objectives
reliability vs. validity- proof
reliability- proof is found in past events
validity- proof is found in ability to predict future events
reliability vs. validity- variables
reliability- limited number of objective variables
validity- broad number of diverse variables
what’s to evaluate facts/evidence
- authorship
- date/currency
- relevance
- intended audience
what is statistics
scientific methods of collecting and analyzing numerical data
types of stats methods
descriptive- seek to describe what the data is saying without making inferences
inferential- seek to reach conclusions that go beyond simple interpretation
predictive- draw predictive conclusions from statistical data
questions for evaluation statistics
- what is the source
- possibility of bias
- margin of error
- age of statistic
- purpose
pitfall of making generalizations
- its hard to know where to start and stop with a social phenomena
- social changes can occur that change the results
qualities of ethical research
- honesty/transparency
- objectivity
- competence- accurate and efficient
- respect for participants and colleagues