Research methods Flashcards
What is meant by ‘the results were significant at p<0.05?
There is a less than 5% chance that there is no real difference between the conditions.
What are the 2 ways to test validity?
- Face validity - whether the measure appears to test what it claims to.
- Concurrent validity - when a new measure is devised (e.g., new IQ test or new questionnaire), we can correalte the results of the new measure with the equivalent current measure.
What is a case study?
A study of one particular individual or a small group of individuals. It may report descriptive information about the person relating to events in the individual’s past as well as to significant events which are currently occuring.
Strengths of case studies?
- capture and describe changes which happen overtime rather than a ‘snapshot’ of an experiment
- data gathered are qualitative (richer source of info)
- reminds us that all individuals are unique
- allows us to study things which we may not be able to in any other way e.g., Clive Wearing, can’t inflict brain damage just for a study
- a single case study which contradicts a theory is enough to cause that theory to be altered
Limitations of case studies?
- method is often non-standardised which can prevent replication
- data are retrospective. Recall of events and experiences might be prone to confabulation (replacing fact with fantasy)
- researcher’s beliefs might affect the analysis of the behaviour studied
- difficulty of generalising any findings to explanations of the behaviour of people in general
What is content analysis?
A technique for analysing data according to themes or categories
What are the 4 key stages of content analysis?
- coding - decide on categories which are going to be used
- work through transcripts/recordings
- count the number of times the categories our in the material you’re analysing
- inter-rater reliability may need to be used to check data
What are the 4 stages of inter-rater reliability?
- use anohter researcher to carry out the same analysis
- agree the categories to be used beforehand
- the second researcher reads the transcripts/views the recordings independently and tallies the number of times each category occurs.
- correlate both sets of results and a positive correlation of +0.8 or above menas the data has good reliability.
Evaluation of content analysis (strengths and limitations)
Strengths:
- allows researchers to study people they would have little/no access to. (e.g. magazine)
- there are fewer ethical issues as there is no or little direct contact with participants
- good external validity as it uses material which has been naturally generated
Limitation:
- might lack construct validity (the degree to which a test measures the contruct or concept that it’s supposed to measure - thinking you’re measuring one thing when you’re actually measuring something else)
How does thematic analysis differ from content analysis?
It involves the gathering of qualitative data (no counting involved, just seeing whether themes/categories are there or not), but uses the same types of material as content analysis (transcipts, recordings). Advantages are same for content analysis as thematic analysis
Limitations to thematic analysis?
- potential misinterpretation of thoughts and behaviours as researchers have little/no contact with people
- interpretation is subjective because it’s based on opinion and therefore may differ from one researcher to another
What are the 2 ways to test reliability?
- test-retest reliability: used when participants are completing a test or questionnaire. The same group of participants complete the test/questionnaire and then complete it again after a suitable time period. Results are then correlated
- Inter-observer reliability: used in observational studies. Behavioural categories are clearly defined and agreed beforehand and sometimes a pilot study will be carried out to check both observers are recording behaviours in the same way. The observers recrod the data in tally chart independently, then correlate results.
How to improve relability in:
- questionnaires
- interviews
- observations
- experiments
Questionnaires = test-retest. If this fails to show good reliability, the questionnaire will need to be reviewed and questions withdrawn or rewritten.
Interviews = the same interviewer should be used each time. Training should also be given to avoid leading or ambiguous questions. Structured interviews with set questions are more likely to be reliable.
Observations = behavioural categories need to be properly operationalised to avoid any misinterpretatoin or overlap. There should also be training for observers to ensure they’re recording behaviour in a similar way.
Experiments = method, instructions, procedure should all be standardised to ensure participants receive the same treatment
What is the difference between internal and external validity?
internal = whether the results are due to manipulation of the IV rather than another variable.
external = whether the results can be generalised to real life settings, the wider population, over time and to other cultures.
What are the 4 types of external validities?
- ecological
- cultural
- temporal
- population
How to improve validity in:
- experiments
- questionnaires
- observations
- qualitative research
experiments = standardisation which reduces investigator effects. Single and double-blind studies which reduce demand characteristics
questionnaires = lie scale (truthful indicator) which highlights any inconsistency and social desirability. Anonymity can also increase validity
Observations = covert observations means behaviour observed is more authentic and more valid. These types of observations have better ecological validity but more ethical issues. Clear, objective, behavioural categories which don’t overlap can also increase validity.
qualitative research = rich in detail, provides in-depth explanations for behaviour meaning it has greater validity. Triangulation can help enhance validity, which is hwne other poeple involved with the participant are interviewed (e.g., friends, family etc)
Difference between field experiments and natural experiments?
Field - carried out in natural settings, possible to manipulate the IV but oftne participants are unaware they’re taking part (an experiment anywhere outside a laboratory)
Natural - not possible to manipulate the IV. It is a situation being recorded that already exists e.g., going to the lunch hall and measuring the amount of boys and girls that walk in.
What is a quasi-experiment?
There is an IV but it cannot be manipulated e.g., gender, age, height, intelligence, music ability. You can’t put a female participant in a male group.
Evaluate the use of independent groups as an experimental design
Strength: same materials can be used in each condition. No order effects like practice or fatigue because different participants are used in each condition. Participants likely won’t know aim of the study either as they are only exposed to one condition.
Limitation: requires more participants. There may be participant variables which differ in each condition and they might confound the results (e.g., IQ, memory, personality)
Evaluate the use of repeated measures as an experimental design
Advantages: eliminates participant variables and fewer participants are needed.
Disadvantages: there may be order effects like practice or fatigue. This means a participant may improve on their performance in the second condition. Or fatigue could affect the participant and they perform worse. Participants may work out the aim of the study which may lead to demand characteristics. As participants take part in both conditions, the same set of materials cannot be used which presents the problem of finding 2 sets of material which are equivalent so they can be compared.
Evaluate the use of matched pairs as an experimental design
Advantage: reduces participant variables in each condition. There are no order effects, participant remains naiive about aim of the study meaning demand characteristics are unlikely.
Disadvantages: very time consuming, sometimes challenging to match participants on particular variables. It’s impossible to eliminate all participant variables even with identical twins.
What is samping bias?
When the sample does not reflect the characteristics of the target population