week 4: concepts and indicators

concepts vs indicators

concepts are abstract, indicators are concrete

t/f indicators are not observable

false → indicators ARE observable

what makes a concept good

• parsimony

• depth

• coherence

• differentiation

parsimony

how simple and straightforward is the concep?

depth

does my definition capture enough of whats important?

coherence

all attributes have to logically be related and fit together

differentiation

definitions must ensure the concept is sufficiently distinct from related concepts

validity

does the indicator capture what we mean by the concept

reliability

does the indicator produce consistent results

3 dimensions of validity

• face validity

• content validity

• construct validity

face validity

is it reasonable and intuitive? at face value, does the indicator seem to capture what the concept is about?

content validity

does it capture the most important parts of the concept?

construct validity

does it operate in line with our theoretical expectations?

• what is happening with that indicator in the world; does it operate in the way the theory would predict

the core of all research projects

concepts and indicators; if you define concepts poorly or pick the wrong indicators, your results will be bad

causal research designs

ask whether some factor or combination of factors produces the outcomes of interest

interpretive research designs

asks how do people experience and give meaning to their social and political circumstances or specific political outcomes

similar features of both interpretive and causal design systems (but not by quantitative projects)

• research process is more flexible

• qualitative research follows an iterative rather than sequential process

• expectations about causality might be revised

how is the research process more flexible

as long as you know what analyses you need to run, bc the computer does it and you have to interpret what the computer produces. theres more work happening at the front end (figuring out what u want to know and what data you should collect)

iterative process

data collection and data analysis happen at the same time in a repeating, cyclical loop rather than in separate, rigid steps

• Research goes back and forth between stages, using what you learn to revise and improve the research.

what does it mean for causality to be revised

you can change your predictions of causality

• but if you pre register your studies, then you cant.

• if you revise, you must be transparent about it

difference between causal and interpretive design systems

• causal research asks “why” questions

  • why do some new democracies prosecute former regime officials while others dont

    • independent variable → causal mechanism → dependent variable

• causal and interpretive have different explanatory goals

is causal mechanism a variable

not a variable, it is a link. the process through which x produces the outcome y

causal explanatory goals

causal researchers want their explanations to apply beyond the cases theyve studied

how do causal researchers ensure their explanations apply beyond the cases theyve studied

identifying scope conditions: “i expect that my findings will hold in any other cases so long as they fulfill these conditions”

interpretive explanatory goals

interpretive researchers want explanations that offer contextualized understandings of the meanings people attach to their circumstance, to other peoples behaviour, and to their own. they explain outcomes by virtue of understandings how things are experienced, justified, or incentivized

• ultimate goal is to understand how individuals create meanings or understand thier context and how these meanings lead to behaviour

what do causal research deigns include

a causal hypothesis that specifies the independent variable(s) with causal factors rigorously conceptualized with clear indicators

• clear indicators are necessary so your hypotheses can be tested against the evidence

• evidence = data from texts, interviews, field research notes

  • all should support ur causal model

causal research design assume

a social and political reality is independent of the researcher

• the researcher and the world they’re studying are inseparable

• when researchers disagree over descriptions of events, teh source of disagreement is likely over conceptual definitions and indicators

• reality exists independently of the words we use to describe it

• describing something as violent doesnt mean it actually is violent. what does being violent mean

interpretive reseacrh begins with the assumption that

we construct our political and social worlds through our language and concepts

• a researchers own identity can plat a role in the research process

  • insider outsider relationship (canadian researcher interviewing canadian mps may receive less information bc of their assumed background knowledge and all)