Please enable JavaScript.
Coggle requires JavaScript to display documents.
Research (Operationalization (Reliability and Measurement validity…
Research
Operationalization
Steps:
- Define concepts
- Identify specific aspects of concepts
- Formulate indicators for every aspect
- Develop measurement instrument that can measure variables
Key elements
- Operationalization process
- Concept
- Indicator
- Measurement instrument
- Variable
-
Measurement error
Radom error
Systematic error
- constant error: the same for each measured object
- correlated error: dependant on values of other variable
Measurement levels of variables
- Nominal
- Ordinal
- Interval
- Ratio
Research design
Experimental design
-
Way
- manipulation
- random assignment of research units
- control of all other influences
Cross sectional design
Types of hypothesis
- non-relational hypothesis
- correlational
- developmental
- causal
- measuring one or more variables at one moment in time
- data collection at one point time
- no random allocation of research units to groups
Longitudinal
- measuring one or more variables at multiple points in time
Types
- trend design: every time a different sample from the same population
- panel design: one sample from the population is measured repeatedly over time
- cohort design (a special type of panel design): panel that consists of research units that have something in common (e.g. birth year, year of graduation, etc.)
2 or more measurements
- prospective
- retrospective
- quasi-longitudinal / trend / repeated cs
Types of hypothesis
- non-relational hypothesis
- correlational
- developmental
- causal
Case study
Characteristics
- Case study = phenomenon + case(s)
- A “full” understanding of a phenomenon is to be achieved
- One case or a few cases is/are investigated
- Multiple data sources are used
Case
The case is the unit of analysis, where the phenomenon of
interest can be studied
- Place or location (village, island, lake)
- Organization or group (company, Greenpeace, family, Ajax
football hooligans)
Phenomenon
= all the activites of interest that take place within the case
Sampling takes place at three levels
- selection of the case(s)
- selection of data sources within each case
- selection of objects within data sources (people, documents)
Types
- descriptive, explanatory
- Single, multiple
-
Research objective
Type
- Practice-oriented/ Applied research
first component (in relation to problem) only, second components not provided
Help to increase the profits of a supermarket
- Theory oriented / Fundamental research
first component in general terms only, followed by several second components
To help solve a knowledge problem (fill knowledge gap)
Finding out the effects of political system on economic welfare
Investigate if there is a relation between education and income
Assess if the Dutch population has increased over the last decade
The research objective describes your motivation to perform the research
In other words, why do you want to perform your research project and what is needed?
- why? (solve problem, either practical or fundamental)
- what? (need knowledge to solve problem)
Two components
Objective in relation to the problem
- the why part of your research motivation
- states the willingness to help solve the problem
Objective in relation to the research
- the what part of your research motivation
- states the knowledge that is needed
-
Sampling
Random sampling
Simple random sampling
Systematic sampling
Stratified random sampling
- strata are sub-groups of units that possess specific characteristics(male/female, Dutch/Germans/French)
- random sample of units taken from EACH stratum
Cluster sampling(one stage or multistage)
- operational population geographically dispersed into custer
- clusters are places where research units are found(schools, cities, hospitals)
- random selection of SOME cluster
Non-random sampling
- Accidental / convenience sampling
- Quota sampling: the researcher takes a certain amount of units per category or value of the selection variable (like stratified sampling, but non-random)
- Volunteer sampling
- Handpicked / purposive / expert sampling: the researcher selects who can provide the best information to achieve the study objectives (e.g., experts)
- Chain / snowball sampling
Steps:
- Defining the operational population
- Selecting a sampling frame
- Choosing a sampling method
- Deciding about the sample size
Data collection
Asking questions
-
Closed-ended interview / questionnaire
Content analysis
- documents are the research units
- selected ‘texts’ are coding units, can also be film fragments, photograph elements
- key concept
- Data of open-ended interviews also have to be analyzed with a content analysis!
Observation
-
Structured
- Coding or systematic observation
- data are quantitative
Types
- Participant vs. Non-participant
- Covert vs. Overt
Reliability & Vilidity
Problems with internal vilidity
- History
Differences between pre- and post-tests due to external events
- Maturation
Differences due to the passing of time (e.g., becoming more experienced, hunger, tiredness, boredom)
- Testing
Differences due to being tested repeatedly (e.g., pre-test – remembering answers; having had time to form an opinion, etc.)
- Instrument decay or instrumentation
Differences due to change in the way of measuring over time (e.g., change in questionnaire or in interviewer)
- Selection
Confounding effects due to initial differences between the groups
- Mortality / dropout
Effects due to loss of certain types of participants over time
Problems with external validity
- Unrepresentativeness
The causal relationship found in the experiment cannot be generalized to other populations and places, e.g., lab experiment with students only
- Reactive effect of pre-testing
Pre-testing sensitizes people to the effects of the
- intervention Artificiality
Effects found in a controlled and artificial lab setting may not generalize to the real world
Hypothese
- state expectations about reality
Types
- Non-relational
- Correlational
- Developmental (Y changes with time)
- Causal
The last 3 can be further defined to
- Directional hypotheses
- Non-directional hypotheses
Prerequisites for causality
1.Co-variation
2.Time order
3.Rationale (plausible theory)
4.Non-spurious relation (elimination of third variables)
5.Logic (DV is capable of being varied, e.g. sex cannot be changed)
Regulative cycle of problem solving
- Problem identification
- Diagnosis
- Design plan
- Implementation
- Evaluation
-