Research methods revision notes (Yr2)
Analysis and interpretation of correlations
What is a correlation
- A mathematical technique which looks at the relationship between two continuous
variables/co-variables
What graph do you use to display correlation results
- Scatagram
What is a correlation coefficient
- A number that represents the strength and direction between co-variables
What is the range of a possible correlation coefficient
- -1 to 1
- where +1 represents a perfect positive correlation and -1 represents a perfect negative
correlation
At what point is a result considered strong
- Above 0.5
Case Studies
“ An in depth study that gathers a lot of detail about one person or a small group or event.”
- Often focuses on unusual events
- Usually involves the production of qualitative data, using interviews, observation and/or
questionnaires
- Although if a person is subject to psychological testing this may produce quantitative data
- Case studies tend to be longitudinal and may involve gathering additional data from family
and friends
Strength of case studies
- Rich & Detailed Qualitative Data -
- Avoids Practical/ Ethical issues - e.g. can study brain damage using people like patient HM
- ‘A single pebble starts an avalanche’ - can spark interest into new domain and generate new
hypotheses (undermine established theories)
Limitation of case studies
- Lacks generalizability - as based on such a small sample size
- Subjectivity - difficult for researchers to stay subjective (observer bias)
- Inaccurate recall - personal accounts from participants and their family and friends may be
prone to inaccuracy and memory decay, especially if childhood stories are being told.
- Lack of replication
- Time consuming
For example
- One issues of using case studies is that it is difficult to generalise the results to a wider
population
- For example, case studies are usually conducted on individuals with specific problems;
- It makes it exceptionally difficult (if not impossible) to generalise the results from Patient X to
a healthy population.
Para on issues with subjectivity with case studies - Genie Wiley
One limitation of case studies is that there may be issues with the subjectivity of researchers. For example, the case of Genie Wiley the ‘wild child’
who was a feral child who was a victim of severe abuse, neglect, and social isolation. Who after a few months in recovery at the hospital was
fostered by Jean Butler, her special teacher at the hospital. The hospital didn't approve of this however as they believed that it would interfere with
the study. As expected, although her intentions were good, Butler's care did interfere with the overall study and therapy Genie needed and she
refused to cooperate with some of the specialists Genie needed to see. Some doctors went as far to speculate that Butler was keeping Genie's
progress to herself in hopes that it would one day make her famous. The authorities later decided that she was not the best fit to gain full custody
of Genie and she was moved to a different home. This is a prime example of how over longitudinal case studies the personal connection that
researchers have with the patients make it difficult for them to stay subjective. This therefore undermines the validity of the study and thus sheds
doubt on their results.
,Content Analysis
What is content analysis?
- A type of observational research for systematically summarising and describing any form of indirect
content.
- It converts qualitative data into quantitative data
- It involves drawing up coding categories and counting how often these categories occur.
- The aim is to summarise and describe communication in a systematic way so overall conclusions can
be drawn.
Content analysis can be done on any type of human communication qualitative data:
- Spoken interaction - such as conversation or speech/presentation
- Written forms - such as texts or emails
- Broader examples form the media - such as books, magazines, TV programmes or films.
How to Conduct Content Analysis
1. Sampling - the researcher must decide what material to use
2. Prewriting: the researcher familiarises themselves with the material -
3. Coding categories - now decide how to analyse material
4. Pilot it - pilot the coding categories which you have created
5. Tally - organise and categories the data, record the number of occurrences
6. Compare - check for reliability by comparing with another researcher (inter-rater reliability)
- Some data sets to be analysed may be extremely large and so there is a need to categorise this
information into meaningful units.
- This may involve simply counting up the number of times a particular word or phrase appears in the text
to produce a form of quantitative data.
- For example, examining TV adverts to see how often men and women are depicted in ‘professional
roles’ (at work) or ‘familial roles’ (at home).
Thematic Analysis
- A way of summarising qualitative data
- Unlike content analysis the data remains qualitative.
- The researcher looks for themes that recur in narratives and summarises the main themes using
quotations from the evidence as support.
- Theme is an idea, explicit or implicit, that is recurrent (keeps ‘cropping up’)
- Themes may then be developed into broader categories
How do I do a thematic analysis?
- Often starts with collecting the data and transcribing it (writing it out).
- The researcher familiarised themselves with it - reading it several times.
- The researcher looks for recurring themes and patterns which run through, and puts them into
categories.
- Once satisfied that the themes cover most aspects of the data they are analysing, they may collect a
new set of data to test the validity of the themes and categories.
- Assuming these explain the new data adequately the researcher will write up the final report, typically
using direct quotes from the data to illustrate each theme.
For instance,
- People with mental health issues may be misrepresented in newspapers as ‘a threat to the well being of
our children’ or as ‘a drain on the resources of the NHS’.
- Such themes may then be developed into broader categories such as ‘control’, ‘stereotyping’ or
‘treatment’ of people with mental health issues.
Evaluations of content and thematic analysis
Strength
- It can circumnavigate (get around) ethical issues. E.g. much of the information is already in the public
domain so permission is not needed, can also be used to access data of a sensitive nature
- High external validity
- It is flexible in the sense that it can generate both qualitative and quantitative data.
Weakness
- Indirect. It is analysed outside of the original context - liable to misinterpretation or researcher bias,
especially when more descriptive forms of thematic analysis are employed
- However modern psychology reflexivity reduces this issue
, - Reducing rich, qualitative data to numerical values reduces the validity, as data may be omitted. Wider
picture is lost.
Reliability
Reliability
- This is the extent to which a test produces consistent findings every time it is done
- External reliability: produce same results each time the test is carried out
- Internal reliability: the consistency of the measure within itself
Ways of assessing external reliability
1. Test retest
- Test retest - do the same test with the same participants again at a different time.
- Example - for a reliable test such as an IQ test, if someone took an IQ test on a particular day
then you would expect the same result on a different day
- However, there must be enough time between the occasions to ensure that the participants
cannot recall their answers to the question but not so long that their attitudes, opinions or
abilities may have changed
- With quantitative results, a correlation can be made to test how reliable the test or
questionnaire is. If the correlation coefficient exceeds +.80, then the measuring instrument is
considered reliable.
2. Inter-observer reliability
- Relevant to observational methods. Different researchers may interpret the same situation in
a different way - observer bias.
- The recommendation therefore is that observations should be made in teams of at least two.
- Firstly, a pilot study may be run in order to check that the observers are applying behavioural
categories in the same way; observing the same event, or sequence of events, but record
their data independently
- The data collected by the two observers should be correlated to assess reliability (whether the
data has a +.80 correlation coefficient)
- Similar methods are used for other types of research:
- Content analysis = inter rater reliability
- Interviews = inter interviewer reliability
Ways of assessing internal reliability
- Usually associated with measures such as attitude scales or personality tests
- If a test you do has people scoring really high on one part but not the other then it has low
internal consistency.
Split half method
- This involves splitting a test into two and having the same participant doing both halves of the
test. If the two halves of the test provide similar results this would suggest that the test has
internal reliability.
- Examines whether a measure is consistent with itself.
