Clinical Methods Flashcards
similarities and differences between HCPC and BPS
Similarities between HCPC and BPS:
Focus on Ethics: Both emphasize ethical conduct, confidentiality, and professional integrity.
Standards of Practice: Both ensure psychologists work within their competence and keep skills up-to-date.
Client Protection: Both aim to safeguard service users, promoting trust and well-being.
Accountability: Both require psychologists to follow guidelines for responsible behavior and report concerns.
Differences between HCPC and BPS:
Purpose:
HCPC: A regulatory body that legally ensures practitioners meet required standards and can remove unfit professionals.
BPS: A professional body that promotes psychology and offers membership, support, and guidance.
Membership:
HCPC: Registration is mandatory to practice as a psychologist in the UK.
BPS: Membership is voluntary but beneficial for career development.
Focus:
HCPC: Regulates all health professionals, not just psychologists.
BPS: Focuses exclusively on psychology and advancing the field.
Disciplinary Actions:
HCPC: Can impose legal restrictions on practice.
BPS: Can only take internal actions, like removing membership, without legal authority.
describe primary data
Involves data collected by the researcher(s) first hand from the source. It may present original thinking or new information. Data gathered can be quantitative or qualitative depending on what type of data the researcher needs to gather.
Many psychological studies usually gather primary data. Questionnaires, observations, content analysis and experiments are all ways to gather primary data. The purposes may be to obtain a first-hand “picture” of a group or society, or to test a hypothesis (an untested theory).
strengths and weaknesses of primary data
Primary data
Operationalisation is done with the research aim in mind, so there is likely to be validity with regard to the aim.
More credible than secondary data, because they are gathered for the purpose with chosen research method, design etc.
Weaknesses Expensive compared with secondary data because data gathered from the start.
Limited to the time, place and number of participants etc., whereas secondary data can come from different sources to give more range and detail.
what is secondary data
Secondary refers to data that already is available to psychologists because it was already collected by others, but psychologists use the results for their own research. Secondary data consists of a very wide range of material collected by organisations and individuals for their own purposes, and include sources as complex as official government statistics at one extreme and as personal as diaries at the other. These data include written material, sound and visual images. Such material can be from the present day or historical data. For example, government statistics from a census can inform researchers about the number of females living alone. The internet is a good source of secondary data, many published statistics and figures are available on the internet are either free or can be purchased cheaply for a small fee.
strengths and weaknesses of secondary data
Relatively cheap compared with primary data, as they are already collected.
Can be large quantities of data, so there might be detail.
Can be from different sources, so there is a possibility of comparing data to check for reliability and validity.
Likely to be gathered to suit some other aim, so may not be valid for the purpose of the study.
When analysed to be presented as results, there may have been subjectivity.
May have been gathered some time before, so not in the relevant time period.
describe longitudinal studies and give examples
A longitudinal approach is not a research method as such; it is a way of carrying out a study.
‘Longitudinal’ means taking place over a period of time rather than at one moment in time.
hankin et al 1998, goldstein 1988
strengths of longitudinal studies
Longitudinal designs are good because they follow the same people or person over a period of time so there are
no individual differences that might affect the results. The participants each time are the same and are often from the same cohort, so they are likely to have had very similar experiences, at least in some ways. If research focuses on
someone with a mental disorder which is specific to them, then it makes sense to follow the course of their illness to
take into account individual differences.
Another strength is that it is a good way of finding out how we develop – in fact if someone is going to study how someone or something develops over time, they will by definition be using a longitudinal design. Development is hard to study any other way, so if a researcher wants to see how a mental disorder affects someone’s functioning over time, then they will use a longitudinal.
weaknesses of longitudinal studies
A difficulty is keeping the participants in the study for long enough to draw conclusions about the course of a disorder or about the issue being studies. The nature of mental health may mean that participants are more prone to drop out, loss of motivation or becoming uncontactable. Participants are likely to drop out of the study as they might move away or they may decide they no longer want to take part.
There are also potential ethical difficulties. For example, following someone or a group of participants over time can be more intrusive than studying them just once. A longitudinal study tends to be about someone’s development so the data gathered might also be intrusive. If someone has already consented to be part of a study they might find it hard to refuse later. This can be especially the case for those with a mental disorder, who may be classed as more vulnerable because of their disorder, which is likely to affect their functioning.
explain simply how rosenhan 1973 and gottensman and shields 1966 are examples of primary research
Primary research refers to studies where researchers collect new data firsthand rather than relying on existing data.
Rosenhan (1973): This was a primary research study because Rosenhan and his colleagues directly gathered data by pretending to be patients with schizophrenia and getting admitted to psychiatric hospitals. They observed and recorded how they were treated, making their findings original and firsthand.
Gottesman and Shields (1966): This was also primary research because they conducted their own twin study to investigate the genetic basis of schizophrenia. They gathered and analyzed data from twin pairs, making their conclusions based on newly collected evidence.
Both studies involved direct observation, experimentation, or data collection rather than summarizing past research, making them examples of primary research.
Recap on your contemporary study for Schizophrenia, Carlsson et al. (2003). Explain below why secondary data may have been better than primary data this area of study in schizophrenia.
In Carlsson et al. (2003), secondary data was better than primary data because the study was a review of existing research rather than collecting new data. Here’s why secondary data was beneficial in this case:
Broader Perspective – Carlsson used results from multiple studies, including brain scans and neurotransmitter research, to get a more complete understanding of schizophrenia. This provided stronger evidence than a single experiment.
Saves Time and Resources – Instead of conducting new experiments, Carlsson analyzed already existing high-quality studies, making the research more efficient and comprehensive.
Ethical Considerations – Studying neurotransmitter function often involves invasive procedures or experiments with drugs. Using secondary data avoided the ethical issues of exposing participants to potential harm.
More Reliable Findings – By reviewing multiple studies, Carlsson could compare results and identify consistent patterns, increasing the reliability of conclusions about the role of dopamine and glutamate in schizophrenia.
Overall, secondary data allowed Carlsson et al. (2003) to provide a well-supported and ethical review of schizophrenia research without the limitations of collecting new data.
examples of longitudinal studies in clincial psychology
Hankin (1998): Conducted a longitudinal study on depression in adolescents, tracking how their symptoms changed over time to understand risk factors.
Goldstein (1988): Studied schizophrenia over time, following patients to see how factors like gender affected the course of the disorder.
Goldstein (1988). Make a note of how you can use this as evidence in the section about the reliability of the DSM:
Goldstein (1988) can be used as evidence for the reliability of the DSM because she tested the consistency of schizophrenia diagnoses using the DSM-III. She re-diagnosed patients previously diagnosed with schizophrenia using updated criteria and found a high level of agreement, supporting the inter-rater reliability of the DSM. Additionally, her study showed gender differences in schizophrenia, which highlights how the DSM can classify symptoms consistently across different groups.
This supports the idea that the DSM provides a reliable way to diagnose mental disorders, as similar conclusions were reached when using standardized criteria.
Why might secondary data (meta-analysis) be better for studies such as Goldstein (1988)?
Increased Sample Size – A meta-analysis combines multiple studies, providing a larger and more diverse sample, making findings more generalizable.
Greater Reliability – By analyzing multiple studies, researchers can see consistent patterns, reducing the risk of individual biases or errors affecting the results.
Saves Time and Resources – Instead of collecting new data, researchers can use existing studies to draw conclusions more efficiently.
Stronger Evidence – Comparing results from different studies improves the overall validity of conclusions about schizophrenia and DSM reliability.
what is a cross sectional methods + examples
A cross-sectional design is one where data are collected at one moment in time, over a short period – it provides a
‘snapshot’ of something. Like longitudinal designs, cross-sectional designs tend to be used to look at development
in some way, but instead of following the same person or people over time, they focus on getting different people at
the same moment in time. The difference in the people is what will be of interest.
Becker et al. (2002): Studied the impact of Western media on eating disorders in Fijian girls by comparing data before and after TV was introduced. Since data was collected at specific points in time rather than following individuals long-term, it is a cross-sectional study.
Wijesundera et al. (2014): Investigated mental health in medical students by collecting data at one time to compare stress levels among different year groups. Since it didn’t track changes over time, it is also a cross-sectional study.
strengths of cross sectional studies
A cross-sectional study provides a useful way of studying something that might take a long time to study naturally and can be done in a short space of time, which can be more efficient in practical terms. If a longitudinal design is not possible, perhaps for ethical reasons or because results are needed quickly in order to affect policy and practice,
then a cross-sectional design is a practical alternative.
–
- A cross-sectional design can be cheaper than a longitudinal design. A cross-sectional design also
requires less commitment in terms of time from a researcher or a team of researchers than a longitudinal
design. The researcher sets up the study, gathers the data and then writes the study up before moving on.
Requiring less time commitment is a strength in itself, as is the possibility of reduced costs of researchers.
weaknesses of cross sectional studies
There might be a cohort effect because the study looks at different people at the same moment in time and
those people will belong to a different cohort, there may be individual differences in the participants. Masellis et al (2003) were unable to monitor and follow their participants through the course of their OCD, therefore the data gathered on QoL only reflects individual experiences at that moment in time and does not account for other external factors that may increase or decrease the participants QoL scores.
Cross-sectional designs are not good at finding out the causes of something like a mental disorder because they
are descriptive research. Because they are a snapshot at one moment in time they are unlikely to include any
historical information about a patient or participant, and they do not gather any information about the
future either. They are not useful for seeing the course of a mental disorder, or how it began, what might
have caused it, or how treatment might work for an individual. In Masellis et al. (2003) the ratings done by each individual are likely to rely on how they felt at that moment. As the symptoms of the obsessions and depression come and go in OCD, one moment in time only captures the quality of life at that time.
what is a cross cultural research method
Cross-cultural psychology refers to studying people’s behaviour and thoughts across different cultures to see what is common across cultures and what is culturally specific.
A cross-cultural design is used when researchers want to look at a particular behaviour or pattern of thinking between different cultures. In order to do this, they compare data from the cultures they are interested in. The researchers may not always gather data themselves from the different cultures; they may use data already gathered about one culture and compare it with data from another culture that looks at the same thing.
If behaviour or way of thinking is found to be the same across cultures it might be argued that it comes from human nature and not from upbringing (nurture). However, if behaviour/way of thinking is different in different cultures it might be thought that it came about because of environmental influences in the different cultures. This is an argument about what is universal in humans and what is not.
etic vs emic
An “etic” approach studies a culture from an outsider perspective, trying to identify universal patterns across cultures, while an “emic” approach focuses on understanding a culture from the inside, considering the meanings and interpretations held by the people within that culture
examples of cross cultural research methods
Tsuang et al. (2013), explored the relationship between ‘schizotypy’ and handedness using students in Taiwan and looked at how this relationship stands up between different cultures (positive schizotypy is associated with being non-right handed). Most previous studies that linked non-right-handedness with schizotypy were done in Western populations. They found that fully left-handed participants had the highest score for positive schizotypy. This link between left-handedness and schizotypy was evident despite social pressure on being left-handed, which meant that handedness is not a result of cultural pressure on handedness, therefore, apply cross-culturally based on schizotypy. This was supported with many Western world studies as well as in Japan, and Taiwan. The idea of left- or mixed-handedness accompanying positive schizotypy was true across cultures.
Mandy et al. (2014) chose to test the DSM-5 diagnosis of autism spectrum disorder. They suggested that the USA and UK testing had supported the new diagnosis of autism spectrum disorder and they wanted to see if the diagnosis would generalise to other cultures/countries. They then compared a Finnish sample with UK participants. They found that the DSM-5 model fitted well in Finnish autism spectrum disorder. The autism phenotype diagnosis did not fit so well in Finland but fitted in the UK. They concluded that
there might be cross-cultural variability in the milder autism diagnosis, though not perhaps for the autism spectrum disorder diagnosis.
strengths of cross cultural research methods
Cross-cultural designs allow generalisations between cultures to build a body of knowledge. For example,
if schizophrenia is diagnosed using the ICD-10, which is used in many different cultures and countries, then
knowing that ‘schizophrenia’ is found universally is important. Studies like Tsuang et al. (2013) that suggest
that schizotypy goes with non-right-handedness universally can help to show that schizophrenia is a
universal disorder and so classification systems like the ICD-10 or the DSM can be used cross-culturally.
–
Improves Validity of Psychological Theories – Testing theories in multiple cultures ensures they are not culture-bound and remain scientifically reliable.
weaknesses of cross cultural research methods
In order to compare data, studies have to use the same method and procedures. An issue in using the same method cross-culturally is that what is understood in one culture might be different in another. For example, questionnaire items in one culture might not suit norms and ideas in another culture. If the method affects the findings, then that affects any conclusions about cross-cultural similarities and differences.
–
Different cultures may interpret questions, behaviors, or psychological concepts differently, making it hard to compare results fairly.
describe meta analysis
Meta-analysis means an analysis of analyses! It is a way of using results from different studies, about the same issue, and studying them as a whole to look for an overall picture about that area of study. It is more a technique of analysis than a research method.
If a number of studies separately find the same answer, and then those studies are analysed together then that answer becomes stronger as the studies support one another. A meta-analysis can also help to adjudicate where studies find different answers. In clinical psychology one use of a meta-analysis is in looking at the effectiveness of a treatment. If a treatment is evaluated by different studies, drawing their findings together to look at the effectiveness of that treatment can be beneficial.
describe publication bias
Publication bias can affect a meta-analysis. This refers to the tendency of journals and publications not to publish results that have negative or non-significant results. This might not be deliberate, it might be that such dissertations are not put forward for publication, for example, or it may be that another study is then carried out to look for more significant results by the researchers. Publication bias can lead to Type-I errors.
examples of meta analysis
Stafford et al. (2015) carried out a meta-analysis to look at treatments of psychosis and schizophrenia in children, adolescents and young adults. They searched for any study that compared any drug, psychological or combined treatment for psychosis or schizophrenia that looked at children, adolescents or young adults. They excluded studies that involved fewer than ten participants. They assessed the studies to look for bias and graded the studies for the quality of their results. They used 27 trials, which had 3,067 participants in total. This highlights how a meta-analysis can include a much larger sample of participants than other methods.
strengths of meta analysis
An advantage of meta-analysis is that it is a way of finding out trends about an issue and any relationships that might exist. A meta-analysis involves statistical analysis and because it uses more data than single studies (the meta-analysis combines the results of studies) then the power of the statistical result is larger than the result of the single studies.
Smaller samples in studies become a large sample in a meta-analysis, so there can be better generalisability of findings. Also, studies may be in different cultures, which might enable the meta-analysis to draw conclusions and to generalise about the universality of an issue (if the results from the different cultures relate to one another).
A meta-analysis can help to uncover patterns that can then be the subject of further research. This can help in building a body of knowledge.
weaknesses
A difficulty is that the studies that the meta-analysis draws on are unlikely to be identical in their research method, procedure, sampling and decision-making. Putting the results of different studies together needs careful decision-making to make sure that the data that are compared are fairly comparable. Stafford et al. (2015) had to discard results in order to choose just the data that could be compared, for example.
Badly designed studies that are used in a meta-analysis lead to a poor meta-analysis, which is why Stafford et al. (2015) chose to grade the quality of the studies they used. It can be argued that if there is doubt about the quality of a study, such as there being bias, then that study, if used, is going to give bias to the results of the meta-analysis.
A meta-analysis will draw on studies that have been published in an area of study, and this means unpublished studies of course will not be used. This might show publication bias and distort the findings of the meta-analysis. Studies that show negative or non-significant results are less likely to be published. Ferguson and Brannick (2011) found in a review of meta-analysis studies, that approximately 25% of showed a publication bias issue.
evaluate the use of meta analysis in carlsson et al 2003
Recap of Carlsson et al. (2003)
Carlsson et al. (2003) conducted a review (meta-analysis) of existing research on schizophrenia, focusing on the role of neurotransmitters like dopamine and glutamate. The study analyzed findings from various sources, including brain scans and drug trials, to challenge the traditional dopamine hypothesis and suggest a glutamate-dopamine interaction as key in schizophrenia.
Evaluation of Meta-Analysis in Clinical Psychology (Using Carlsson et al., 2003)
Strengths:
Increased Reliability – By reviewing multiple studies, Carlsson could identify consistent patterns, strengthening the validity of neurotransmitter theories in schizophrenia.
Time and Cost Efficient – Instead of conducting new experiments, Carlsson used existing data, avoiding the time and ethical issues of working with vulnerable patients.
Reduces Bias of Single Studies – Using diverse data sources minimizes the risk of individual study errors affecting the overall conclusions.
Ethical Advantage – Avoids exposing new participants to potentially harmful drug tests or invasive brain scans.
Weaknesses:
Data Quality Issues – The reliability of a meta-analysis depends on the quality of the original studies; if the studies used have flaws, the conclusions may be unreliable.
Lack of Control Over Variables – Carlsson relied on different studies with varying methodologies, making it harder to ensure consistency in findings.
Potential Researcher Bias – Selecting which studies to include could introduce bias, as researchers may favor studies that support their hypothesis.
Findings May Become Outdated – As new research emerges, previous meta-analyses may become less relevant, requiring continuous updates.
Conclusion
Meta-analysis, as used by Carlsson et al. (2003), is a powerful tool in clinical psychology, allowing for a broader, more ethical, and reliable understanding of schizophrenia. However, its effectiveness depends on the quality and consistency of the original studies, making careful selection and interpretation essential.
what are case studies and what is a nomothetic/idiographic approach
Case studies are interested in individuals and in detail about them. They can be said to have an idiographic approach, which means focusing in detail on a topic and on individuals.
Case studies are used to find out detail about something rather than for building cause and effect understanding. They describe someone or a small group as much as they explain them. Case studies are called a ‘research method’ here, though they are not quite the same as other research methods, like experiments. This is because case studies use other research methods like interview observation or questionnaire to gather data about the case they are interested in.
They gather mainly qualitative data, because such data give the richness and detail that a case study is looking for. A case study will often use more than one way of collecting data, such as questionnaires for people who know the individual being focused on, or experiments on brain damaged patients, for instance, as well as observations and interviewing to find out more. This means that a lot of the data are qualitative but there can be quantitative data too.
Triangulation can be used to look for reliability and validity. Triangulation means taking data from at least two different sources and checking whether the data agree. If the same data is gathered using different methods, then the data has retest reliability. The data are also likely to be valid, because if data match and some data at least come directly from an individual, then the data are likely to be measuring what they claim to measure, which gives the data validity. The process of triangulation add credibility to the results of a study.
Nomothetic -
This approach focuses on large groups of people and seeks to establish general laws that can be applied across a population. Nomothetic methods include experiments, surveys, and statistical analysis.
Idiographic -
This approach focuses on the individual and seeks to understand their unique characteristics. Idiographic methods include case studies, interviews, and observations.
strengths of case studies
Case studies are good at finding out detailed and in-depth information about an individual or small group, and such detail means rich data to draw conclusions from. The advantage here is that the case study allows depth, which is hard to find using other research methods. Experiments and surveys, for example, gather focused data, looking at a specific variable rather than wider and richer data focused on the whole individual or group. Case studies allow that richness, variety, and detail, where other methods often do not.
Case studies can look at rare or unique conditions, where samples of such individuals are by definition going to be difficult to find. If a wide sample of a certain type of person is not easily available, meaning data cannot be collected in a more general way, a case study is a way of gathering valuable insights about a rare occurrence. For example, if someone has specific brain damage, then knowing a lot about that case can be crucial in understanding the effects of brain injuries and improving treatment options.
As case studies tend to use more than one way of collecting data, triangulation can be used to test for both reliability and validity and also to show consistency in the findings.
weaknesses of case studies
A main difficulty is that the focus on a single individual or small group means that generalizing the findings to others is not really possible. The sample is so limited it is not going to reflect the diversity of the wider population, and so results may be limited in their usefulness.
Another limitation is that a case study is often carried out by one researcher (such as Freud’s (1909) Little Hans) or a small team, who will get to know the individual who is the focus of the case study very well. This might not cause overt bias, but it is possible that there is subjectivity in the data gathered, and the findings might be affected by the researcher’s input. This is especially the case in clinical psychology, whereby many diagnoses and conclusions are influenced by the individual interpretations of the researcher, leading to potential bias in understanding complex psychological conditions.
describe the use of interviews in clin. + example
An interview is a conversation between two or more people where questions are asked to obtain information about the interviewee. Interviews can be divided into two rough types, interviews of assessment and interviews for information. Interviews involve face-to-face situations or telephone contacts in which the researcher orally solicits responses.
The interviewer will find out about the personal data required for the study, such as gender, age, marital status, employment status and other details of relevance. The interviewer will use some standardised instructions at the start of the interview so that the respondent is aware of ethical issues such as confidentiality and the right to withdraw. The respondent will also be told something about the interview and what the purpose is.
Interview schedules must be planned out well in advance so that all areas are fully operationalised and an aim or hypothesis formulated. The way of recording the data can be either recorded or written down, whatever format the full conversation it must be fully TRANSCRIBED after the interview. This can be very time consuming in many ways.
Goldstein (1988) looked at differences in how males and females experience schizophrenia, using both primary and secondary data from interviews. Goldstein used trained interviewers to gather data about the symptoms of the patients and she used questionnaires administered by an interviewer to gather information about their past histories, family structure and previous experiences.
types of interview
Structured Interviews
Definition: These are highly controlled interviews where the researcher asks a set list of predefined questions in a specific order.
Advantages: Easy to replicate, reliable, and quick to analyze as responses are standardized.
Disadvantages: Lack of flexibility, may limit the depth of answers, and the interviewee has less freedom to express themselves.
Unstructured Interviews
Definition: These are more flexible and conversational, with no set list of questions. The interviewer may have a general topic in mind, but the conversation can flow freely.
Advantages: More in-depth and rich responses, allowing the researcher to explore new topics as they arise.
Disadvantages: Time-consuming, harder to analyze, and results can be influenced by interviewer bias.
Semi-Structured Interviews
Definition: These combine elements of both structured and unstructured interviews. The interviewer has a list of key questions but can follow up on responses and ask additional questions.
Advantages: Balances structure with flexibility, allowing for depth while still covering specific topics.
Disadvantages: Can be inconsistent across interviews, making analysis more difficult than structured interviews.
strengths of interview
Unstructured interviews are useful for obtaining qualitative data because there can be exploration of the issues and the respondent is able to use their own words and ideas. The qualitative data offered by interviews can give more detail, depth and validity that quantitative data. This may be beneficial in clinical psychology when looking at issues such as the experiences of mental health stigma, effectiveness of treatments, or how an individual functions in their wider social circumstances.
Data tends to be more valid in interviews than questionnaires because the respondent can use their own words and issues can be explored face-to-face (or telephone) with the interviewer. This represents the experiences of mental health issues much more realistically and allows the individuals to expand on their points.
Interviews enable a large amount of data to be collected which is descriptive and may give a better picture of what is going on in real life so are valid to what is being studied.
Interviews give access to information which is not available through direct observation, such as what individuals think and feel about certain topics which again makes it a more valid method.
weaknesses of interviews
The interviewer may affect the findings because of the way questions are asked, the way he or she is dressed or other characteristics. This is known as interviewer bias. If an individual with a mental health problem is being interviewed they may feel uncomfortable or nervous and not give reliable responses.
There might be subjectivity involved in analysing interviews. When categories and themes have to be identified from in-depth and detailed data, usually involving a transcript, the researcher may allow personal judgements and experiences to affect their analysis. This links to the concept of clinical interviews for diagnosis, where the judgements of subjective concepts, like failure to function adequately, are determined by clinicians with their own social norms, values and beliefs. Objectivity is required to build a body of knowledge, so subjectivity must be avoided where possible in research.
describe thematic analysis
Thematic analysis involves looking at qualitative data and picking out themes and patterns in the data.. Themes are the patterns and they help to describe the overall data, often in relation to a specific research question.
Themes/patterns are identified from coding, they then become categories and the data can be examined carefully, allocating the data to the different categories.
Thematic analysis takes place by:
● familiarising oneself with the data
● generating codes
● looking for themes in the codes
● reviewing the themes
● defining and naming the themes
● producing the report.
The researcher must be very familiar with the data in order to generate the codes and also then to identify themes in the codes that suit the particular research question. Thematic analysis involves generating codes, which can come from theory, from how the researcher understands their data or from previous research. Thematic analysis is a main tool for analysing qualitative data and can be used in other ways to analyse qualitative data, including grounded theory.
issues of thematic analysis and example
Issues of thematic analysis
Issues of thematic analysis include how ‘long’ a pattern or theme should be, so that it is put into a category and coded. A pattern would tend to suggest there are quite a few instances of that point or theme, and that it would be quite a large theme, so that a lot of the data would fit into it. That would be efficient in terms of reducing the size of the qualitative data. However, it is not about counting the instances a pattern
arises or how much of the data would fit into that pattern. It is not about bringing quantity into the data. A pattern that did not appear often but was judged to be a theme would be accepted by a researcher.
The researcher would judge whether an idea is a theme according to the research question and importance of that pattern. A researcher would look for data that captured something in relation to the research question, rather than for ideas that appeared ‘most often’, for example. This would be thematic analysis driven by theory. A researcher might want to capture all the data that has been collected, in which case themes are more likely to be ideas or features of the data that appear often. This would be thematic analysis driven by the data (and more like grounded theory). There are no rules about how large a theme must be to be a ‘theme’.
Content analysis involves categorising qualitative data and counting instances of categories. Content analysis is very similar to thematic analysis but is more likely to involve quantity (such as frequency of an idea or feature in the data).
Lim et al. (2014) used thematic analysis of the evidence found in other studies looking at placement (work) and support as an intervention for those with schizophrenia or schizoaffective disorder. They say they used a literature search using the internet and found 358 articles. Lim et al. (2014) followed steps suggested by Braun and Clarke (2006). The steps were familiarisation with the data, generating initial codes and then building and reviewing themes. To do the coding they coded each study ‘line by line, from the introduction to the conclusion’. They used 627 excerpts from the data and had 417 codes initially. Then using thematic analysis of the codes and grouping excerpts, checking the data and going back to the themes, they put the 417 codes into 40 categories. The first author did the initial coding, then the second author did the initial coding (separately) and the third author reviewed so that there was consensus. Themes and patterns came from the codes.
The researchers follow the steps put forward for doing thematic analysis, such as starting with familiarisation with the data and then coding carefully before building themes. The separate authors act as coders so that reliability can be shown. Themes and patterns come from the coding, which is how thematic analysis works.
strengths of thematic analysis
● Thematic analysis is a flexible way of analysing
qualitative data. It can be done using categories
and themes that come from theory, or it can be
used to generate categories and themes from data
directly. Thematic analysis does not need a theory
to drive the analysis and, unlike grounded theory,
it does not need to derive a theory from the data;
it can do either.
● Another strength is that it is a way of
maintaining a richness in data and yet summarising
a large amount of qualitative data in a manageable
way.
● Thematic analysis allows insights that are not
anticipated by the researcher and presents correct and detailed data in a way that does not lose its richness. Participants can be collaborators, for example, and can verify the data once they are analysed, which maintains richness. Participants do not have to be trained to understand the data after the analysis; the data are accessible to everyone as they are written in a clear format, again maintaining the richness and detail.
weaknesses of thematic analysis
● A weakness is that it is hard to say how the analysis is done in general terms. There is a tendency to think of themes emerging from the data, which in a way they do, but if a researcher sees that their data show repetition about a certain phrase or feature when many interviews are considered, then that phrase or feature seems to ‘emerge’, but it’s because the researcher has noticed the feature or phrase. Ely et al. (1997, cited in Braun and Clarke, 2006) say it is thinking about the data and creating links that gives the themes; they do not ‘reside in the data’.
● If data come from a questionnaire or an interview, they come from questions put to the participants, and that is often the case. It can be hard to find themes in the data that are not driven by the questions in the questionnaire or interview, as that has already guided the data offered.
● Thematic analysis can be said to have subjectivity. Those analysing the data are likely to make judgements, for example, about what makes data form into a ‘theme’. Science is about objectivity so that there is no bias in data. However, using more than one researcher to analyse the data can reduce subjectivity.
describe grounded theory
Grounded theory is a way of analysing qualitative data and means not using existing theory but finding theory from the data – the theory is grounded in the data. The goal of analysis is to generate a useful theory from the data – a theory that explains that data.
Grounded theory is a research method as much as a method of analysing data, because it focuses away from the positivist tradition of ‘doing science’. Science is about taking a theory, deriving a hypothesis from the theory, testing the data empirically (from the senses) and then accepting, adapting or rejecting the theory because of what was found.
Grounded theory focuses in the opposite direction and avoids the adoption of a theory before looking at the data; it involves looking at the data first before arriving at a theory. The researcher using grounded theory looks at the data, sees themes and ideas that are repeated, uses them as categories and groups the data accordingly. After that a theory is put forward.
Note that when grouping the data and using coding, grounded theory is no different from thematic analysis. It is just that grounded theory is a method that does not analyse data with a theory already in mind.
how does grounded theory take place and example
How grounded theory takes place:
- First, coding is done. Coding involves taking text in small pieces and putting the text into a heading that suits it.
- Second, codes are collected into concepts, which groups the data together. The idea is to see how the coded data might fit together and in a way to generate concepts that are ‘wider’.
- The theory comes from collecting the categories together; it is about forming a model that can explain the data.
Grounded theory uses both inductive and deductive thinking:
● Deductive thinking refers to going from a theory
to ‘deduce’ what will arise from that theory. An
example is ‘if all men are mortal, and Henry is a man, then, by deduction, Henry is mortal’. It would be the job of the researcher to check the claims, and if the claims are true then the conclusion must be true.
● Inductive thinking refers to the ideas arising from empirical evidence. The idea is empirical observations are carried out and then a general
theory is derived.
Grounded theory focuses on inductive reasoning but there is deductive reasoning too. In grounded theory the data drive the concepts and from the concepts come the theory, which uses induction. However,
checking back from a possible theory to the concepts and back again to build the theory can involve deduction. Grounded theory is about developing a theory from the data – it involves explaining the data, not simply describing it.
Coldwell et al. (2011) looked at how people with the diagnosis of psychosis or schizophrenia might contribute to their family. They interviewed six people diagnosed with schizophrenia and six who had a relative with schizophrenia, and then analysed their data using grounded theory. The researchers kept a reflexive journal and memos and analytic auditing was done, so that the reliability of the analysis could be checked. As themes began to emerge they were explored using follow-up interviewing with the participants so that there was more understanding of the categories that were emerging from the data. This is the process of deriving theory that is grounded in data.
strengths of grounded theory
● There is validity in that the coding is done
carefully using meaning from the data in chunks,
and then codes are developed into wider concepts,
going backwards and forwards from the codes to
the concepts in order to build the concepts. This
means that the participants’ own thoughts and
feelings are used to drive the analysis, which
means the concepts should measure what they
claim to measure, which gives validity.
● Grounded theory uses specific terms to explain
how it is done and is explicit in its guidance, which,
when it comes to deriving theory from qualitative
data, is not done in any other way. Thematic
analysis can analyse qualitative data and content
analysis also can analyse qualitative data and give
categories, but it is grounded theory that then
goes on to develop a theory to explain the data.
Grounded theory allows the creativity that is
needed to move from data about people to draw
up ideas about how to explain that data.
● Another strength is that using grounded theory
allows the richness and detail of the qualitative
data to survive the analysis. The concepts that
come from the coding bring the richness of the
data to the analysis, rather than losing the detail.
weaknesses of grounded theory
● It is perhaps impossible to code and categorise data into concepts without some theory in mind. We cannot put out of mind theories when analysing qualitative data.
● Another similar criticism is that the search is for the shared meanings of individuals that have given the data and yet the way grounded theory is carried out has a formula and is prescribed. This means it would appear to go against the idea of encapsulating rich meaningful data in the meanings of the participants (rather than those of the researcher(s)).
● It could be argued that it is not appropriate to ignore previous research to aim to generate a completely new model or theory from empirical data. The original question is likely to have come from previous research and previous findings, probably including previous theory.
● Engaged theory is another idea. It focuses on generating concepts from data, as does grounded theory, but engaged theory uses existing theory to help in the analysis.
● It is not possible perhaps to generalise from the findings of a study using grounded theory to say that any theory that comes from that data can explain all examples of what is being understood. The data will come from a specific culture in a specific time-frame, using specific participants. Generalising is, therefore, limited.
overview of measures of central tendency
Measures of central tendency: Where is the middle of your data?
Mode.
This tells you what the most common score is. To work it out, count how many times each value appears in your data. The most common is your mode. If two values appear the same number of times, then you have two modes. The mode is most useful when you have data that is categorical or ordinal, for example if your data is ‘yes/no’ questions.
Median.
To work this out, put your data in order, and count half way through. The value right in the middle is your median. If two values are in the middle, work out what is half way between them: that is your median. This is a particularly useful statistic if you have extreme scores.
Mean.
To work this out, add up the data and divide it by the number of values you have. For example, if ten participants gave you a speed estimate, add up the ten estimates and divide this by ten. This is the most common way to describe the middle of your data set.
Measures of dispersion: How spread out are your data?
Range.
You can easily work out the range of your data by subtracting the bottom score from the top score. This is a very basic way to describe the spread of your data.
Standard deviation.
This is usually more informative. The standard deviation is calculated based on the average distance from the mean of your data: if you have a lot of extreme scores, you will get a higher standard deviation; if your data are very stable and clustered closely around the mean, you will get a lower standard deviation. What is most useful for you to look at is the difference between the standard deviations of the groups in your study. If there is a big difference in the standard deviation, this means one group has more variation in their scores.
which statistical test and when?
Chi-squared
To test the significance of the association/relationship.
· To test if two categorical variables are linked to one another
· It can be used with nominal data
· It can be used with data that does not have normal distribution
· The data must only fit ONE category of the contingency table
Spearman’s rank correlation co-efficient
To test the significance of a correlation.
· To see if there is a correlation between two ranked scores
· Describes the strength and direction of the relationship
· Uses ‘at least’ ordinal data
Wilcoxon
To test the significance of the difference.
· A repeated measures design will have been used
· Level of data is at least ordinal
· The data has a normal distribution
Mann Whitney-U test
To test the significance of the difference.
· An independent measures design will have been used
· Level of data will be ‘at least’ ordinal
