Types of Research
Types of Reasearch
There are 2 types:
QUALITATIVE- This is where the data being found is extracted from participants experience e.g. interviews and questionairs
QUANTATIVE- This is where the data measuring behaviouris numerical
Lab Experiment- Quantative
1) Ideal form of an experiment as there is possiblitiy of all variables being controlled, even extraneous. 2) Replication is good, the IV is the variable that is manipulated by experiment, DV is not measured
1) Total control never possible, results may be effected e.g. extraneous variables, experimenter bias, demand characteristics, volunteer bias, sample bias 2) Artificial situation, results may not generalise to real life, no Ecological/external validity
1) Informed consent not always possible 2) may not be able to exercise the right to withdraw 3) Participants should not be subject to stressful or negative manipulations
Field Experiment- Quantative
1) Greater Ecological Validity in general 2) Technique aviods experimenter bias and evaluation aprehension because the participants are unaware that they are part of an experiment.
1) Inevitably, extraneous variables are harder to control 2) Some design problems remain, such as sample bias 3) More time-sonsuming and expensive than Lad experiments
1) It is not possible to gain informed consent or to debriefing 2) Participants may be distressed by the experience (especially because the experience isn't real)
natrual exp 1
Natrual Experiment- Quantative
1) Only way to study cause and effect in certain situations e.g. where there are practical and/or ethinical objections to manipulating the variables (looking into effects of deprivation) 2) Greater ecological validity in general 3) Participants unaware of being studied, technique avoids experimenter bias
1) Cannot establish cause and effect as so many other factors may influence the DV 2) Not easy to replicate, may not be possible at all 3) A Lack of control reduces internal validity. 4)Can only be used when conditions are very natrual, such conditions are not always possible to find 5) May be unaware thier being studied and show improvements just due to this
natrual exp 2
1) May involve withholding treatment from one groupd, as when new educational programmes are being tested, the research requires one group not to have the new teaching method 2) May be unaware they are taking part, issues of informed consent 3) Researchers need to be sensitive to the problem of participants in unfavorable circumstances that may surround the behaviour being studies e.g. deprivation
Correlation Analysis- Quantative
1) Can be used where experimental manipulation would be unethical or impossible 2) Indicates possible relationships between co-variables, might suggest future research ideas which would look at casual relationships 3) Can rule out casual relationships - if two variables are not correlated then one cannot cause the other
1) Does not establish cause and effect 2) Relationship may be due to extraneous variables
1) cause inferences shouldn't be made but often this misinterpretation of the data does happen 2) This is especially important with relation to socially sensitive issues such as IQ, which tend to rely on correlation.
sturctured interviews 1
Structured interviews- Quantative
1) Can collect informention about people's feelings and attitudes, which cannot be obtained through observation or experiments 2) Requires less skill than unstructured interview 3) Can be conducted on the telephone
1) Interviewers expectations may influence the interviewers performance (halo effects, confirmatory bias, raical/sexual, ageist prejudice) 2) People often dont know what they think and therefore thier answers are influenced by suggestion and responce bias 3) Method relies on self-report, which is open to problems such as soical desirability bias4) In comparison with unstructured interviews, data collected will be restrictedby pre-determined sets of questions
sturctured interviews 2
1) Deseption may be necessary 2) questions may concern sensitive and personal issues 3) Confidency and privacy must be respected
Unstructured interviews- Quantative
1) Rich data can be obtained
1) Requires well trainned interviewers which makes it more expensive to produce reliable answers 2) Interviews may not be comparable because different interviewrs ask different questions (Low inter-interviewer reliability). Reliability may also be affected by the same interviewer behaving differently on different occassions 3) It is more affected by interviewer bias
Questionaire survey- Quantative
1) Large amounts of data can be collected at relatively little cost - in terms of time and money
1) Answers may not be truthful (social desirability bias) 2) Only Sutable for certain kinds of participants - those who are literate and willing to spend time filling in a questionaire, creating a biased sample 3) designing questionaires requires considerable skill.
Natrualistic obseravtion 1
Natrulaistic observation- Qualatative
1) Observation is need to establish possible relationships2) Offers a way to study behaviour where there are ethical objections to manipulation variables 3) Gives a more realistic picture of spontaneous behaviour. Has high ecological validity- if two variables are not correlated then one cannot cause the other.4) If the observer remains undetected, the method aviods most experimental effects, such as experiemnt bias and evaluation aprehension
1) It is not possible to infer cause and effect 2) It is difficult to replicate and therefore you cannot be certain that the results are a "one off" 3) It is not possible to control extraneous variables 4) Observer bias mean the observer see's what he wants to see 5) Where participants know they are being watched (disclosed observations) they may behave unatrually. Even non-participant observers, by thier mere presence, can alter a situation
Natrualistic obseravtion 2
1) Undisclosed observations preclude the right to informed consent 2) Disclosed observations may affect the individuals observed, creating distress
Case Studies- Qualatative
1) Only option when behaviour is rare 3) Provides insights from an unusual perspective and in rich data3) Relate to real life (High ecological validity)
1) Usually involve recall of earlier history and therefore are unreliable 2) Close relationships between experiemnter and participant introduces bias 3) Cause and effect are difficult to establish 4) Not a rigorous methodology: oten unstraucuterd and unreplicable 5) Limited samples, lack generalisation (Low in ecological validity)6) Time-consuming and expensive
1)Confidentiality and privacy must be protected. Individuals should not be named
A Directional / One Tailed Hypothesis - Predicts a particular direction eg Higher.
Non Directional / Two Tailed Hypothesis - States there will be a difference.
A Null Hypothesis - States that there will be no effect / change.
Random Sampling – This is a sample in which every member of the target population has a chance of being selected. Tossing a coin, picking names out of a hat etc, could do this. BUT it is almost impossible to have a truly random sample, as a total list of the target population cannot truly be identified.
Systematic Sampling- This is a modified version of random sampling that may involve selecting every 10th person from a phone book or selecting every 5th house in a street. BUT there may be participants that refuse to take part.
Opportunity Sampling – This involves using who ever is available and willing to take part. BUT this is not a representative sample and is usually bias, i.e. all one gender, friend’s / colleagues of the researcher.
Volunteer or self-selected sample - participants volunteer themselves, suffer from volunteer bias as participants are usually more motivated than participants who are randomly selceted.
Stratified sample - Population is divided into sections or strata in relation to factors considered relivant, e.g. social class/ age. Participants are then randomly selected out of the strata
Quota sample - Also uses stratified methods but the sample is not randomly determined, the research seeks participants with satisfy each criteria.
Independent variable (IV) - the one which is manipulated to see the effects
Dependent variable (DV) - the variable which is being assessed
Levels of measurement - variables are measured at different levels of detail, each level expresses more information about what is being measured
Nominal - data is in categories, e.g. grouping of people to waht foods they prefer (Italian, American, Indian ect.)
Ordinal - data is ordered e.g. people put out food in order of thier preference. The difference between each item is not the same.
Interval - data is measured using equal intervals
Ratio - There is a true zero point, e.g. most measures of physical quantities