What is an experimental research statergy
Where the independent variables are systematically manipulated to examine the effects on the dependant variable.
Thus it examines the differences between two or more groups.
The goal is the explain and establish a cause and effect relationship between two variables.
What is a quasi-experimental research design
It attempts to establish the same goals as an experimental research design however it is done without random assignment. Due to this we cannot be sure that the groups are otherwise identical thus the causality of the effect on the DV from the IV cannot be certain.
What is the non-experimental research strategy.
This involves examining and describing a relationship between two variables without explaining the relationship. This strategy is good for prediction.
Due to not direct examining of the relationship you cannot establish a cause and effect relationship as there is no manipulation.
These include cross sections and logitudinal designs.
What are the quasi/non/experimental designs used for
Comparing two or more groups of scores
What is the correlational research strategy and what is it used for
Where the relationship between two or variables are measured for each participant. This is useful for prediction purposes however this does bring up the directionality problem and the third variable problem.
What is the descriptive research strategy
This strategy simply seeks to measure individual variables and provide a description of the variables in relation to a certain group. There is no comparison or correlation merely just the description.
What is internal validity
Internal: validity of the entire research study and refers to how well the study addresses the research question.
- IV directly changes DV and there is no other cause: high
- there are alternative causes for the relationship: low
What are the threats to internal validity?
Hint: OSMHRMTIFED
- Ambiguous ordering, aka. Unclear if B proceeds A or A proceeds B.
- Selection bias. Differences may occur in groups due to selection bias
- Maturation/development: regular development or natural changes could appear as treatment or condition effects.
- History: other events that happen at the same time as manipulation may cause change over time.
- Regression to the mean: people who have extreme scores aka. Very high depression then their scores will regress towards the mean which may appear to be a treatment/condition.
- Mortality: withdrawal with change results
- Testing: repeatedly taking a test can result in changes that may appear as a condition effect. Aka. Just answering questions about depression may change future answers.
- Instrumentation: changes across life of study in how its measured (equipt.) or how its interperated/observed.
- Floor or ceiling effects: unable to detect an effect due to measurements being too high or low and no longer susceptible to change.
- Experimentor/ subject effects: someone may be bias
- Diffusion of treatment: groups may not be completely separated thus aspects fo the intervention/treatment/manipulation may affect control.
What is external validity
Refers to the ability of a study to be generalised to other populations.
What are the three main concerns about external validity
- How well a sample generalises to the population
- How well they relate to other studies
- How well the research study findings generalise to the real world. Aka. Lab vs. real life
What are the threats to external validity
- Interaction of causal effect with units. Ie. teaching styles in uni may not apply to primary schools
- Interaction of the casual effect over variations, an effect may not hold when it is varied i.e. full treatment vs. partial or when combined w other treatments.
- Interaction of the causal effect with the outcome.
- Interaction of causal effect with settings.
What is a between subjects design, matched subject design and what is within subjects design
Between: each participant is measured under one condition
- no history, maturation or time effects
- easier to recruit and maintain
- least powerful
Within: each is measured under multiple conditions
- controls for inter-individual differences
- most powerful
Matched: incorperates the advantages of both within and between subjects. They only go through one treatment but are matched based on similar characteristics.
What is power
The probability of detecting differences between treatment conditions if differences are present
