3 SRM (OPERATIONALIZATION) Flashcards
Research and methods
Research – the process of discovering answers to
questions using systematic procedures.
❖ Methods – systematic procedures for performing a
task and/or achieving a goal.
QUALITATIVE and QUANTITATIVE methods.
➢Quantitative – collecting data with predetermined
instruments and analyzing statistically.
➢Qualitative – collecting emergent data and
focusing on words and meanings.
❖Quantitative research
➢ Deductive
➢ Studies variables and their
relationships
➢ Based on meanings derived
from numbers
➢ Collects numerical and
standardised data
➢ Structured inquiry
➢ Samples are larger and
probabilistic
➢ Macro-level
➢ Easier to replicate
➢ Results more “objective” and
generalizable
➢ Uses statistics for analysis
and inference
➢ Detached researcher
➢ Difficulty with complex
causality
❖Qualitative research
➢ Inductive
➢ Studies cases
➢ Based on meaning
expressed in words
➢ Smaller samples, guided by
theory, purposive (non-
probability)
➢ Micro-level
➢ Flexible inquiry
➢ Sensitive to context, process
and complexity
➢ Seeks in-depth
understanding
➢ Focused on authenticity
➢ Less replicable
➢ Involved researcher
➢ Limited by researcher’s
characteristics
Factors to consider when choosing approach - qualitative vs quantitative
❖Research question
❖What exactly are we trying to find out?
➢Make standardized, systematized comparisons,
sketch contours and dimensions, account for
variance? – Quantitative.
➢Study this phenomenon in detail, holistically, in
its context, find out its meaning for the people
involved, figure out the process? – Qualitative.
❖Knowledge payoff
❖Which approach will give more useful
information?
➢Relatively superficial and rational responses –
quantitative.
➢Below the surface, emotional responses – qualitative.
❖Which approach fits our population?
➢Sparse description of a large number of cases –
quantitative.
➢Rich, detailed description of a smaller number of cases
– qualitative.
✓Practical considerations
❖Time.
❖Money.
➢Qualitative – relatively high cost per respondent and
relatively low cost per project.
➢Quantitative – relatively low cost per respondent and
relatively high cost per project.
❖Availability of samples.
❖Availability of data.
❖Access to situations.
➢Physical-geographical.
➢Cognitive.
❖Participants’ cooperation.
✓Literature
❖What have others used to research this topic?
❖Do I want to go along with the literature or offer
alternatives?
✓Why combine quantitative and qualitative approaches?
❖Triangulation – check findings against
each other → enhance validity.
❖If the study is mostly qualitative,
quantitative research can:
➢help in the choice of subjects;
➢help strengthen otherwise ungeneralizable
findings;
❖If the study is mostly quantitative,
qualitative research can:
➢provide background information;
➢be a source of hypotheses;
➢help create scales and typologies;
➢help explain relationships between variables
and interpret findings.
✓Qualitative research
❖Fieldwork
❖Interview
➢ From semi-structured to
unstructured
➢ From semi-standardized to
unstandardized
❖Focus groups
➢ Guided group discussions –
quite popular in political
campaign & marketing research
❖Ethnography
➢ Observing a group in its natural
environment for a prolonged
period of time
➢ Primarily non/participant
observational and interview data
❖Grounded theory
➢ Start with observations and
continuously refine their
categorization and explanation
until you have a theory
❖Discourse analysis
❖Process tracing
✓Quantitative research
❖Survey
➢Sampling is key
Representative
Probabilistic
➢Structured interviewing
➢Standardised questionnaires
❖Content analysis
➢Counting keywords/key
phrases in a text
❖Secondary statistical analysis
✓Experiments
Most scientific
Quasi-experiments/ Field experiments
✓Action/participant research
✓Case study
❖Multiple sources/ types of data
collection methods
❖Variable
❖Variable – an empirical characteristic/
quality that can take 2 or more different
values in different situations.
❖To establish a causal connection between IV & DV:
➢IV and DV must co-vary
➢IV should precede DV in a time sequence
➢No other factor can be a possible cause of the change in
DV – control for effects of other variables
➢A logical explanation – theory – why the IV affects DV
Other variables could affect the relationship between the DV
and IV:
❖Control – variables that are held constant/ neutralized/ eliminated
to not have a biasing effect.
❖Extraneous/ nuisance/ confounding – variables in the research
environment that are not controlled. Dangerous for validity, must
be taken into account when interpreting findings.
❖Intervening – link the IV and DV but are not directly observable,
must be inferred. More complex causality. IV for DV and DV for IV.
❖Mediating – surfaces after IV started and before the effect is
observed, helps model the process.
❖Moderating – modifies the relationship between IV and DV by
moderating the strength of the effect of intervening variables, can
be measured and taken into account. The relationship between IV
and DV would be different in degree without it.
❖Measurement
❖Measurement – assignment of numbers or other
symbols to characteristics/attributes of objects
according to a pre-specified set of rules.
❖Your choices in measurement are inseparable
from whether this is a quantitative or a qualitative
study.
❖Operationalization
❖Operationalization – reducing the abstract
notions to observable behaviour and/or
characteristics that can be measured in tangible
ways. NOT describing correlates!
❖Process of operationalization:
➢Define the construct you want to measure:
❖Dimensions
❖Elements
➢Think about the content of the measure –
items/questions that will actually measure the
concept that you want to measure.
➢Create the instrument, such as a questionnaire
with a response format that will be measured on a
scale.
➢Assess reliability and validity of the measure.
❖Nominal – classification
✓Mutually exclusive and exhaustive.
✓No intrinsic value or comparability, just grouping of
differences. Even if coded in numbers, those are just labels
and cannot be used arithmetically.
✓Few statistics apply to nominal variables (mode, frequency,
some measures of association).
✓If independent variables, may be converted into dummy
variables in statistical analysis.
✓Used for non-ranking data (gender, religion, race,
hometown, political affiliation, candidate preference, study
programme).
