5 - Theory, Problem Definition And Tesearch Design Flashcards
A theory
or model proposes relationships between abstract concepts and allows for généralisations based o individual situations and facts. It can be expressed graphically as a conceptual framework
Abstract concepts
Consists on different: - specific - objective - measurable Behaviours
Research design
A framework for the collection and analysis of data
Research method
A technique for data collection
Conceptual framework
Links abstract concepts and turn them into measurable variables in order to make sense of their relationship (operationalising)
Turn vague concepts into specific, objective and measurable data by defining variables in terms of procedure used to measure it
Decide on an operational definition
- provide valuable data
- convincing the audience
- use several operational definitions
Theoretical framework
Offers the nature and relationships between the measurable variables in a diagram.
Independent variable
Manipulated
Dependent variable
Measured or observed
Moderating variable
Gender, Age, affect in between
Mediating variable
Almost a dv
Issues with intact groups
- How subjects were allocated in groups
- What happened in the groups
- The length of time groups pre existed prior to interventions
Blind studies
Don’t know if they receive treatment or placebo
Can be double blind when the administrator doesn’t know as well.
Problem definition to theory
1- management problem: open question focusing on outcome presenting the problem like an opportunity
How can we attract better qualified IT staff?
2- research problem: what information you need
Why don’t our job ads attract well qualified IT staff?
3- research questions: break down research pb
What reach do we have ? Most effective channel?
4- hypothesis: questions turned into hypothesis
Types of research studies
- cross-sectional research studies: gather data one time
- longitudinal research studies: collect data several times
Exploratory studies
When the study area is vague. The initial step is qualitative methods. Quick and cheap but should not guide more than the direction of the research as they lack experimental control + adequate sampling
Descriptive studies
Record what is observed. It provides quite accurate information but casual links cannot be established
Correlational studies
Observe the degree to which certain things tend to co-occur or relate to each other.
No IV are being manipulated.
It doesn’t imply causality.
Experimental studies
The researcher manipulates the level of IV and observed any corresponding change in DV.
The goal is to advance knowledge but costly and time consuming.
How to have similar groups ?
- random assignment: external factors are randomly distributed among both groups
- paired groups: subjects that match on important factors
- pre-testing: for initial group difference
- same group twice
Different experimental group design
- random independent group design
- pre-post random group design
- factorial design
- quasi-experimental group design (doesn’t meet control or random)
- one group pre-post design
- non-equivalent group comparison designs (not random)
In between group design
We have influence of individual differences on the treatment
