Please enable JavaScript.
Coggle requires JavaScript to display documents.
Module 3: Concepts & Measures - Coggle Diagram
Module 3: Concepts & Measures
Concept: an abstract idea that represents or symbolizes a quality
Explicit: clear statement of what the term is and how it is defined. Consistent use of the term throughout work
Familiarity: established definition in academic literature and/or a commune usage of the word.
Contextually Appropriate: must be appropriate to the context
Parsimony: short and to the point
Differentiation: clear boundaries ; tells us what it is and what it is not
Variable: concrete representation of a concept
Ordinal level variable
ranking relative to other categories, but precise distance between categories not known
categories are mutually exclusive
Example: Education (in categories)
Interval level variable
Ranking have precise (or known) distance between categories
Categories are mutually exclusive (can't be in more than one category)
More precise than ordinal
Example: age (in years)
Nominal level variable
Categorical
categories are mutually exclusive
Cannot be ranked
Least precise level of measurement
Example: Gender
Indicator (Measure): observable evidence that is used to describe a dimensions of a variable
simple concept, fewer measures
Complex concept, more measures
Operationalization: abstract to meaasureable
Assessing Measures
Measurement error
Random error
found in all samples because the full range of possible respondents cannot be included
amount of random error can be estimated in a probability sample
inaccuracies caused by factors that are not systematic or intentional
Non-random error
inaccuracies caused by factors that are systematic or intentional
Bias
Reliability: the extent to which the measurement of a variable yields consistent results
The lower the random error, the higher the reliability
Designing Reliable Measures
Increase precision
Use multiple indexes
Use careful design
Measure one thing a time
Exhaustive categories
Mutually exclusive categories
Measurement validity: the extent to which a measurement of a particular concept matches its operational definition
The lower the non-random error, the higher the measurement validity
Designing Valid Measures
Ensure clear conceptual definition
Increase number of measures
Increase abstraction
Is the instrument reliable?
Things to see in literature:
Test-retest reliability
Inter-coder(-rate) reliability: used for observer reported (as opposed to self-reported data)
Internal consistency reliability: similar responses to items measuring same concept
Is the instrument valid?
Things to look out for in literature
face validity: do items seem to measure the concept?
Construct validity: asses if the instrument really measures the construct (and not another related construct)
-- factor analysis
Is the instrument culturally sensitive?
Measurement equivalence
linguistic equivalence
translation and back translation
conceptual equivalence
metric equivalence
"the same answers or score in two cultures might not depict the same level of the construct being measured" --Rubin 2008, 243
What to look for as readers:
discussion of equivalence (comparative research)
consider whether important groups within sample may interpret questions differently