Flashcards
SP300: Research design
Cards in this deck (31)
- 1Why are truly random samples good?- All people have an equal chance of being selected
- This means they are more likely to represent the population findings would be generalised to - 2What is random cluster sampling?Everybody in a random cluster of the population is researched (so a school is randomly selected, but everyone in that school participates).
- 3Why is non-random sampling good?Convenient
Helps fill a quota
Snowball: participants recruit other participants (if they are part of a sub-group) - 4What is WEIRD and why is a problem?Western
Educated
Rich
Democratic
Majority of knowledge of psychology comes from these. - 5How are WEIRD samples different from those in other countries regarding the Muller-Lyer illusion.More likely to add extra cm to make the same length; other countries perceive the lines (correctly) to be the same length.
- 6From controlled to uncontrolled, list the types of research psychologists use.Experimental, quasi-experimental, correlational, descriptive.
- 7What are the three conditions to infer causation?Covariance (or correlation)--> Between X and Y
Temporal precedence--> X before y
Exlusion principle--> X is the only cause. - 8How do we ensure temporal precedence?The design of the study must allow or manipulation of the IV.
- 9How do we ensure the exclusion principle?The IV must be isolated from all other variables in the design--> there are no confounds.
- 10How can we ensure covariance?The IV must change when the DV changes and this is observed.
- 11How do we enable casual inferences to occur?Internal validity is created by ensuring the IV is manipulated as to ensure temporal precedence
and
The IV is isolated to ensure no other potential causes of the effects occur. - 12How do we ensure IV validity?- Should reflect the hypothetical construct (construct validity)
- Have good generalisation (external validity)
- Be applicable to real life (ecological validity) - 13How do we design a powerful study?Minimise error by ensuring there is a critical amount of the IV (not too much or too little manipulation).
- 14What are the key differences between a within and between subjects design?Participants: more are needed for between
Order effects: potentially for within
Powerful: within
Individual differences: controlled more with within
Diffusion of treatment: seen more in between. - 15Why do we use manipulation checks.To see if the manipulation changed the IV and look at the reliability of the manipulation.
- 16What is a DV, and what are some examples of them?DV: a measurement which is intended to demonstrate, through the manipulation of an IV, the validity of the a hypothetical construct.
Examples
Behaviours (from self-report, eyemovement, confidence).
Psychophysciological/ brain activity
-Scans (PET, MEG, FMRI)
- Heart rate. - 17How do we ensure a powerful DV?It should be
Sensitive: responds to small changes it is measuring
Reliabile: will have little error from the DV. - 18What factors influence the power of a DV?Validity can be reduced by:
Ceiling and floor effect
Practice
Fatigue - 19What kinds of graph skews do we get from ceiling and floor effect?Ceiling (too easy): Negative skew
Floor (too hard): positive skew. - 20How to deal with inconsistent perfromance in trials?Include practice trials
Have multiple trials (increase reliability)
Don't measure for too long (due to fatigue)
Account for order effects by counterbalancing.
Account or carrover effects (change in psychology from previous experimentation) - 21What is a latin square/ permuation?Latin square
Each group will only occur once in each coloumn/ row relative to how many controls there are to counter order effects. - 22What is the difference between a true experiment and a quasi experiment?Random allocation: cannot be done in a quasi
Temporal precedence: cannot established in a quasi
Infer effect: cannot be done a quasi
Quasi has a naturally occuring rather than manipulated variable. - 23How to ensure good quasi experiment design?Conditions are maximally different (Low IQ vs High IQ)
This will also maximise internal validity if matched.
Try to avoid demand characteristics if IV pre-test is used. - 24What is the problem with picking those at each end?In a normal distribution, scores regress towards the mean (meaning scores may either get closer or further away from initial score due to their extremity).
- 25Quasi experiments look at time scales too, how can these be done. Explain.Cross-sectional (sample at a set point taken from different populations regarding time- such as ages)
Longitudinal: taken from the same person, but sampled over different times (IE during development) - 26What are cohort studies?A longitudinal study in which several people with the same characteristics are studied over time.
- 27Which is better, correlational or quasi experiments?Correlational can be more powerful as it allows a full range of scores but quasi allows a greater control over things such as extraneous variables.
- 28What is descriptive research?Observing and describing behaviour/ mental processes, where these happen and who demonstrates them.
May not be an experiment and may not even have a hypothesis.
- Field Surveys
-Observations
-Case studies - 29What are field surveys?Research done in a natural environment, where a person naturally lives or where they are not in a lab based place with internal manipulations.
- 30What types of observation studies are there?Naturalistic
Structured
Participant observation (become part of the experiment)
Covert/overt. - 31Why is descriptive research useful?-Allows insight where not much research lies
- May be difficult to manipulate in an experiment due to ethics
- Forms the basis for new research.
- Allows to see the basics of some phenomenon.
Information
- Description
- Psychology (Sp300) Flashcards on SP300: Research design , created by chantel Smith on 07/03/2018.
- Created by
- chantel Smith
- Views
- 1
- Rating