RESEARCH METHODS
- Created by: laurasmith96
- Created on: 03-01-14 15:53
View mindmap
- RESEARCH METHODS
- Experiments & Techniques
- Laboratory Experiment- provide the highest level of control over variables.
- + repeatability of procedures + equipment and measuring sophisticated + control over variables to give cause and effect
- -Artificial resulting in loss of ecological validity - Demand characteristic - Evaluation apprehension or social desirability.
- Field Experiments- carried out in natural environment of those being tested.
- +Improved ecological validity +Reduction of demand characteristic
- -Controlling variables difficult -Generalising to other situations hard -Cost is often more than lab -Use of technical equipment harder than lab.
- Natural Experiment- a study where conditions cannot be controlled by the experimenter because the independent variable is naturally occurring.
- +Demand characteristic reduced because less aware of being studied. +Sometimes only ethical way of investigating.
- -Area of study may be rare so opportunities and sample may be limited. -Experimental control reduced so extraneous variables decrease validity.
- Naturalistic Observation- behavior observed and recorded with no manipulation of variables. Either overt (p's aware) or covert (unaware). Categories are used to produce quantitative data.
- +For young children is only way to gather information. +Behavior is natural and demand characteristics reduced if covert. +Good realism and ecological validity,
- -Behavior can be misinterpreted - difficult to replicate as lack of control and unique situations -Overt leads to demand characteristics.
- Questionnaire- P's answer a set of questions in writing.
- +Self report allows actual opinion. +Large amounts of data can be collected quickly. +Less influence of experimenter.
- -Researcher bias with leading questions. -Illiterate people are excluded so sample may be bias. -Social desirability.
- Interviews- researcher asks participant questions face to face.
- +Self report allows actual opinion. +Nearly anyone can take part. +Rich, detailed, qualitative data. +Flexible and easier to tailor questions to explore complex issues.
- Case Studies- indepth studies of an individual/ small group, undertaken in real life context, using interviews/ observations to collect qualitative data.
- +Rich & interesting data with high degree of realism. +Challenge and change theories.
- -Objectivity may lack as relationship forms between researcher and p. -Bias with picking from large amounts of data. -Problems with repeating and reliability as unique, hard to generalise.
- Laboratory Experiment- provide the highest level of control over variables.
- Ethics
- Informed Consent- should be given by P's. May conflict with researcher needs so Prior General Consent- p's agree in advance to take part in experiment but not know the aims. Presumptive Consent- asks similar group for consent. Debriefing also allows to withdraw data.
- Confidentiality & Privacy- no personal details should be given away. Important in socially sensitive situations. Photos or observations should not be taken without consent.
- Deception- should be avoided to not infringe rights of P. may be appropriate but consider: cost benefit analysis, use ethical committees, give right to withdraw and debriefing.
- Protection From Harm- P's should not feel a greater level of psychological discomfort than in daily lives. Debriefing should return P to this state or right to withdraw.
- Right To Withdraw- before study all P's should be aware they have right to withdraw from themselves and data from the study.
- Aims & Hypothesis
- Aim- explains why the study is being carried out, Hypothesis- a statement predicting what will occur so it can be tested.
- Non Directional/ Two Tailed- doesn't predict direction of results, just states a difference.
- Null Hypothesis- states there will be no difference,
- Directional/ One Tailed- predicts the direction of results eg. faster, bigger.
- Terms
- Validity- whether it sets out to measure what it was supposed to, legitimate, representative,
- External Validity- can the data be generalised beyond setting: across people, places (ecological), time (temporal)
- Internal Validity- can results be linked to direct manipulation of the IV. Not from extraneous variables, investigator effects, demand characteristics. Standardised instructions and random sample should be used.
- Operational Definitions- are ways of measuring mental processes by using behavior eg: memory thorugh numbers of words recalled.
- Independent Variable- the variable the researcher changes to see if it has an effect on the DV. Dependent Variable- the variable measured to see if the IV has had an effect.
- Validity- whether it sets out to measure what it was supposed to, legitimate, representative,
- Sampling
- Target Population- group of people who share characteristics researcher needs for study. Sample used which is representative of this population.
- Random- p's equal chance of being picked from list (sampling frame). Could toss a coin, names from hat, number to each p, computer.
- +if done right, representative sample achieved, -Hard to identity all members of a large sample.
- Volunteer- p's volenteered to be in study by responding to advert. +Easy method. -May not be representive of population, bias.
- Experimental Design
- Independent Group Design- each P is selected for one group by randomization. +No order effects +No lost participants +Can be used when repeated measures inappropiate.
- Matched Pairs Design- two separate groups of people who are matched on one to one basis. +Controls Individual differences +No order effects +Used when repeated measures is inappropriate.
- Repeated Measure Design- the same participants for both groups. +Requires fewer participants. +Controls individual differences. -Order effects -P guess purpose.
- Order Effects- where P may be better in one condition due to tiredness, practice, bordem etc. Overcome by counter balancing- half group do one condition and then swap.
- Experiments & Techniques
Comments
No comments have yet been made