Research Methods

Sampling
1. Opportunity Sampling- Taking a sample from people who are available at the time the study is carried out. + FAST AND EASY - MAY BE ONE AGE RANGE, NOT GENERALISABLE
2. Random Sampling- Everyone gets an equal chance of being chosen to take part in experiment e.g.put all of the names of students in a hat and draw out a name + EQUAL CHANCE/ FAIR - NOT GENERALISABLE, INDIVIDUAL DIFFERENCES
3. Self- Selected Sampling (volunteer) - this consists of participants becoming part of a study because they volunteered to do so , from an advert e.g. milgrams study + MORE SUITABLE PARTICIPANTS - TAKES LONGER
Experimental Design
1. Independent Measures design (Independent groups)- Using different participants for each condition in the investigation + CAN BE QUICKER TO CARRY OUT, LESS LIKELY TO RESPOND TO DEMAND CHARACTERISTICS, ONLY DESIGN THAT ALLOWS COMPARISON - TWICE AS MANY P'S NEEDED, P VARIABLES MAY CONFOUND RESULTS, DIFFICULT TO KEEP VARIABLES CONSTANT
2. Repeated Measures design - The same participant taking part in each condition of the IV. + PARTICIPANT VARIABLES WILL NOT CONFOUND RESULTS, FEWER P'S ARE NEEDED - RESULTS MAY BE INFLUENCED BY ORDER EFFECTS E.G. BOREDOM, P'S ARE LIKELY TO GUESS THE AIM OF THE STUDY AND RESPOND TO DEMAND CHARACTERISTICS.
3. Matched pairs design - P's are matched on a factor important to the experiment e.g if it is a memory experiment, P's will be tested on memory before hand and grouped with other people with similar results.+ REDUCES THE EFFECT OF SOME KEY VARIABLES, P'S ARE LESS LIKELY TO GUESS THE AIM OF THE STUDY - DIFFICULT TO ESTABLISH MATCHES, REQUIRES TWICE AS MAY P'S
Controlling extraneous variables
1. Standardisation - Keeping everything the same for participants. This can be achieved by adopting a standardised procedure for all participants - instructions that are exactly the same for everyone.
2. Counterbalancing - May have to use this method if using repeated measures design but wish to reduce order effects e.g. boredom or practice. Half of the p's do condition A then B and the other half do condition B then A. The idea is that if the people get better on the second test this order effect will be balanced out, 50% of p's do better in condition A and 50% do better in condition B
3. Randomisation- Participants are allocated to one situation or the other randomly (determined by chance)
4. Participant reactivity- This is something that can affect the results of the experiment, how p's respond to the experimental situation.DEMAND CHARACTERISTICS. EXPERIMENTER EFFECT (INVESTIGATOR EFFECTS). prevent happening by single and double blind control.
Ethics
1. Consent
2. Protection from distress
3. deception
4. debriefing
5. withdrawal from investigation
6. Confidentiality
7. observational research -consent
8. Giving Advice
9. Colleagues
Pilot Study - a trial run of research with a small number of p's to allow researchers to make necessary adjustments and to save wasting valuable resources
Types of Data
1. Quantitative Data- Focus on numbers and frequencies rather than meaning and experience. It provides information that is easy to analyse statistically and fairly reliable Associated with the scientific approach, and are criticised for not providing an in depth description.
2. Qualitative Data- this is concerned with describing meaning rather than statistics. they lose reiiability however gain validity. Provide an in depth and rich description.
Reliability - how consistently it measures
1. Internal reliability - how consistent a method measures within itself. if measures of measurements were not standardised they would distort final results.
  1. Internal reliability could be checked for test items by the split half method - correlating the results with the other half and gaining a positive correlation.
2. External reliability - how consistent something measure over time when repeated. Should have the same results when repeated with the same people under the same conditions.
  1. External reliability could be checked for test items by the test/re-test method. correlating the results of the test conducted on one occasion with the results of the test conducted on a later occasion - gaining a high positive correlation.
3. Inter-rater reliability - two raters provide consistent or similar responses. The ratings for each observer are correlated to check for agreement. It is a method of assessing the reliability of a set of measurements or ratings such as in an observation.
Validity - does it measure what its supposed to measure.
1. Face validity - examining whether a measure looks like it measures what its supposed to measure.
2. Concurrent validity - comparing a new method or test with an already well-established one that claims to measure the same variables.
3. Predictive validity- whether a model allows us to make predictions.
4. Internal validity- what happens in a study - whether the independent variable really had an effect on the dependent variable.
5. External validity - whether the findings of the study can be generalised beyond the settings of the study. e.g. populations, location, measures used, times.
Observational studies
1. Controlled observation - this involves the recording of spontaneously occuring behaviour, but under conditions controlled by researcher. +CONTROL OF CONFOUNDING VARIABLES, LEADS TO INCREASED REPLICABILITY AND RELIABILITY. - OBSERVED BEHAVIOUR MAY BE ARTIFICIAL DUE TO THE SETTING AND P'SKNOWING THEY ARE OBSERVED, DEMAND CHARACTERISTICS, OBSERVER BIAS.
2. Naturalistic observation - this involves the recording of spontaneously occuring behaviour in the participants own behaviour. + BEHAVIOUR PRODUCED IS MORE NATURAL AND VALID, APPROPRIATE IN SITUATIONS WHERE THE EXPERIMENT MAY BE INAPPROPRIATE. - HIGHER RISK OF OBSERVER BIAS , DIFFICULT TO HIDE OBSERVERS, REPLICATION IS DIFFICULT, EXTRANEOUS VARIABLES ARE POORLY CONTROLLED, OBSERVER BIAS.
3. Participant observation - this is a technique used when a researcher takes part in the study posing as an 'ordinary person', usually concealing their true identity and function from the real p's + BEHAVIOUR IS LIKELY TO BE NATURALISTIC, LIKELY TO PRODUCE VALID RESULTS, USEFUL WHEN EXPERIMENTS WILL BE UNETHICAL, RESEARCHER GAINS FIRST HAND EXPERIENCE. - DATA RECORDING RELIES ON MEMORY WHICH COULD BE DISTORTED OVER TIME, EMOTIONAL INVOLVEMENT FROM RESEARCHER MAY CLOUD OBJECTIVITY, PROBLEM WITH GENERALISABILITY.
4. Behaviour sampling methods
  1. Time sampling - the observer decides on a time interval, such as 5 minutes every hour and makes a note of particular behaviours during that time.
  2. Event sampling - observer counts how many times a specific behaviour occurs. better when less p's and they dont know they are being observed.
Content analysis
1. 2 parts to it - first is an interpretative aspect which involves deciding which categories are meaningful in a study in terms of what you are researching.the second is a mechanical aspect which involves organising and subdividing the material, counting how frequently words, images, idea, actions etc occur.
Presenting qualitative data
Self report methods- allows the participant to provide information on him/herself usually in the form of an interview of questionnaire. for example, p's may be asked to keep a sleep diary over the course of a week. + ALLOWS RESEARCHER TO GATHER INFORMATION ABOUT VARIABLES WHICH MAY BE DIFFICULT TO INVESTIGATE IN OTHER WAYS BECAUSE THEY ARE VERY PERSONAL OR OCCUR RARELY., ENABLES RESEARCHER TO GATHER BOTH QUANTITATIVE AND QUALITATIVE DATA DEPENDING UPON THE TYPE OF QUESTIONS ASKED. - PARTICIPANTS MAY NOR RESPOND IN A TRUTHFUL. WAY DUE TO LACK OF UNDERSTANDING/ UNPREPARED TO TELL THE TRUTH/ MAY NOT KNOW ANSWER TO QUESTION, RISK OF RESEARCHER BIAS.
1. interviews - face to face or conducted electronically. range from structured to unstructured. A structured interview involves researcher asking p's a pre- determined set of questions, semi-structured is when p's are asked pre- determined questions but researcher can add spontaneous questions or allows p''s to elaborate. unstructured interviews involve researcher supporting p as they talk about what they need to, used to help client rather than gather info. + ALLOWS RESEARCHER TO CLARIFY QUESTIONS IF THEY ARE AMBIGUOUS , SEMI STRUCTURED INTERVIEWS ALLOW THE RESEARCHER TO DIGRESS OR PROBE FURTHER IF THE NEED ARISES BECAUSE OF THE LESS STRUCTURED FORMAT, ALLOWS RELATIONSHIP TO BUILD, STUCTURED INTERVIEWS ARE EASY TO STANDARDISE . - RISK OF RESEARCHER BIAS, P MAY BE RELUCTANT TO OPEN UP AND TELL TRUTH, TAKES LONGER TO GATHER INFO.
2. questionnaires - paper or electronically, open ended questions allow p's to respond freely and in depth. Closed questions restrict responses p's can give.fixed questions - likert scales involves indicating how much they agree using a rating. Semantic differential type scales include p's placing a mark on a scale - numbered or bipolar. + TAKES LESS TIME TO GATHER DATA, P'S MAY BE MORE HONESTAS A ANSWERING ANONYMOUSLY, RESEARCHER DOES NOT NEED TO BE PRESENT. - AS RESEARCHER IS NOT PRESENT, P'S CANNOT QUESTION ANY QUERIES, RISK OF POSITION RESPONSE BIAS , FIXED QUESTIONS INCLUDE A FALSE IMPRESSION OF PRECISION.
3. Validity and self reports.
Correlation
1. Advantages - correlational analysis can demonstrate an association/relationship between variables and how strong that relationship is. It can also demonstrate the type of relationship ie positive or negative. It is a technique of statistical analysis which can be applied to data obtained by other means.
2. Disadvantages - No matter how strong the correlation is, it does not suggest causation., Such analysis can only measure straight line relationships. The technique is subject to any problems associated with the method by which data is obtained.
Case Studies
1. Alternative methods

Next up

Research Methods

Description

Resource summary

Media attachments

Similar

	Created by Livvi Maglio about 9 years ago