RM1

quantitative data

involves gathering numerical data through experiments, observational studies and surveys with closed-ended questions

begins with a testable prediction from prior research (hypothesis)

generalisable to whole population - looking for norms and patterns

useful in capturing the big picture - larger, representative samples

measure and quantify phenomena, in a structured way

finds relationships between variables (cause and effect)

described as positivist - basis of ‘hard’ sciences

theory testing and deductive

hypotheses are tested using statistical analysis

• allows us to establish generalisability of findings due to underlying assumptions of these tests

produces explanatory theories and descriptive/inferential statistics

quicker to analyse statistically

objective (subjectivity introduces bias, which threatens analytic validity)

• requires measures for bias to be controlled/eliminated

useful at outlining differences

stepping stone towards complete understanding of the single reality

‘unnatural’ as quantification makes the collection of naturalistic data difficult

quantitative research questions

cause

leads to

effect of x on y

relationship/difference between x and y

how does x impact y

extent to which x predicts y

qualitative data

non-numerical data - words/images/observations/language/meaning

good for the smaller picture - smaller samples (even case studies)

attempt to provide deep, rich descriptions of people’s meaning (depth over breadth)

not able to generalise (instead aims to say something specific about a certain phenomena/cohort of people and produces complex accounts)

seeks patterns, but explores difference and diversity

researcher’s focus can be broad

no predictions about findings

more subjective (subjectivity is an asset)

values creativity, reflexivity and novel forms of data

• reflexivity is a tool to value and harness subjectivity

takes time using interviews and focus groups with open-ended questions

theory generating and inductive (some researchers do not believe in reading literature before carrying out interviews)

can be combined with quantitative

• preliminary to large scale experimental work

• added on to large surveys to acquire deeper understanding

useful at understanding why there are differences

part of a rich tapestry of understanding

related to philosophical approaches

qual research is the first step to quantification

rejects positivism - adopt relativist position of no fixed ‘reality’

take the postmodernist perspective

reality is constructed socially/individually (ideographic)

qual researchers argue critical realism

• accepts there is a ‘reality’ out there, but at best we view it though infinite windows, that distort reality in some way

• observers come to observations with expectations and baggage (culture, interests, perspectives

qualitative research questions

the understandings of x

making sense of x

personal/lived experience of x

qualitative data analysis

systematic examination of non-numerical data

involves transcription

• audio/video recordings converted into written text in preparation for simpler analysis

• allows familiarisation (transcribed data can be read through many times to immerse oneself)

involves coding

• labels (codes) are assigned to segments of data that relate to themes and identify overarching patterns within data

• interpretation of themes and how they relate to each other

qualitative data collection can sometimes be analysed quantitatively

random sampling

leads to a representative sample, confident in generalising results to entire population

uncommon in psychology

simple random sampling - every person in population has equal chance of being picked

stratified sampling - population divided into meaningful groups and then simple sampling conducted on each group

non-random sampling

leads to less representative sample

saves time and money though

more common in psychology

often only practical option when population is large

voluntary sampling - members self-select to participate in research

• snowball sampling - participants get friends/family to participate

convenience sampling - members who are easy to reach are asked to participate

size of sample determines…

extent to which you can generalise findings

probability of a chance finding

larger sample = more power to detect an effect, if it exists

how big a sample should be, depends on…

size and homogeneity of population

nature of variables measured

required precision of results

how confident you want to be about results

true/randomised experiment

involves experimental manipulation of IV

manipulation in IV = change in DV

• however changes in DV may reflect biases and random error

randomisation of participants to each condition

• matched/block randomisation can be used for equal number of participants in each group

• randomisation of experimenters too, if multiple (for experimenter differences)

controlled lab setting

use of a control group and experimental group

standardisation of procedure

• only experimental manipulation should vary

• time of day, time since eaten/slept, researcher behaviour

considers ecological validity and how it can be generalised

experiment - experimental design

the only way to explore causal relationships

manipulate (systematic variation) one variable and see if it affects a second variable, keeping all others constant

random allocation

control for chance differences between groups

each member has equal chance of being allocated to either group

important when establishing cause and effect

independent variable (IV)

variable you manipulate/change

has 2+ conditions/groups

eg. exercise vs no exercise

dependent variable (DV)

variable you measure

eg. reaction time

quasi studies

researcher cannot randomly allocate participants to conditions of the IV, due to already existing factors

eg. when IV is biological sex

extraneous variables

other variables potentially affecting the DV

need to control as many as possible

participant variables - eg. age, gender, education, stress

situational variables - eg. lighting, weather, background noise

they’re classed as confounding variables if they differ systematically with the IV, as well as affecting DV

how to prevent extraneous variables

match conditions on key variables eg. balance out age, sex, education etc

• cannot always predict all possible EVs before though

standardised procedure

randomisation of the sample to conditions (ensures equal dispersion of all EVs)

demand characteristics

cues that lead participants to change their behaviour

they sometimes guess the aim/hypotheses and act accordingly to support/not support the hypotheses (consciously or unconsciously)

not just a reaction to researcher’s behaviour, but possibly the testing environment setup

• eg. the presence of a release form vs not, in a chamber study influenced how the participants reacted to the sensory deprivation

how to prevent them:

• deception - conceal study’s hypothesis, what measurements are of interest (debrief at earliest opportunity though, and highlight withdraw)

  • tricky due to ethical issues

• use measures that are hard to control eg. reaction time, physiological responses

• blind techniques (double blind studies)

experimenter effects

the experimenters/researchers desire to support the hypothesis affects their behaviour (consciously or unconsciously)

they shouldn’t know which condition each participant has been assigned to to prevent effects

acquiescence bias

tendency for participants to positively agree to all items presented on a scale

to overcome this, most scales include a mixture of positive and negative worded items

types of experimental deisgn

between groups

within groups

related design helps to control for the variation between individuals, which affects their performance in the different conditions

between-groups design

independent measures / uncorrelated groups

compare different participants in different conditions

preferable design when order effects are likely

use unrelated statistical tests

advantages:

• no carry over effects (avoids one condition contaminating the other)

• process is quicker for participants (less likely to get bored/drop out)

disadvantages:

• individual differences have greater effect

  • to overcome this, randomly allocate participants or use a matched pair design when a particular variable may influence result

• need more participants

within-groups design

repeated measure

compare same participants in both conditions

use related statistical tests

advantages:

• effect of individual differences reduced

  • eg. a person’s pre-existing tendency to make lots of errors applies equally to all conditions

• fewer participants needed

disadvantages:

• boredom, fatigue (so mistakes may be greater in the second condition)

• participants practice the task (practice effect)

• carry over/order effects likely and conditions can contaminate each other

  • to overcome this, use counterbalancing (so both orders occur equally frequently)

• carryover/asymmetrical/differential transfer

  • effect of an earlier condition affects a subsequent one (not equally for all orders)

counterbalancing

deals with order effect in within groups design (repeated measures)

sample is split into half

one half completes conditions in one order, the other half completes it in the reverse order

null hypothesis

manipulation of IV will have no effect on the DV

no difference between the conditions

we assume this before conducting tests

experimental/alternate hypothesis

manipulating IV will cause a change in DV

a difference between conditions

we want to support this, but only know after we conduct tests

direction of effect

all experimental hypotheses must predict a difference

either directional (one-tailed)

or non-directional (two-tailed)

typically determined by prior research

operationalising variables

being clear on what each variable is and how to quantify the DV

code of ethics

contains the professional standards members should uphold

provides a framework for guiding decision-making of all members

ethical reasoning is subject to competing biases

desire to retain autonomy (self-regulation of members by professional body)

maintaining good reputation

based on 4 key principles

• respect

• competence

• responsibility

• integrity

research ethics committees (REC) review research proposals

ethical principles - respect

psychologists value the dignity and recognise the worth of all persons

with sensitivity to the dynamics of perceived authority/influence over clients and differences (eg. social status, ethnics, gender)

with particular regard to people’s rights

• including those of privacy and self-determination

all humans are worthy of equal moral consideration

members consider privacy, confidentiality, respect, communities and their shared values, impacts on the broader environment, issues of power, consent, self-determination and the importance of compassion (empathy, generosity and courage)

ethical principles - competence

ability to provide specialist knowledge, training, skill and experience to a professional standard

psychologists value the continuing development and maintenance of high standards of competence in their professional work

the importance of preserving their ability to function optimally within recognised limits of their knowledge and skills

members consider possession of appropriate skills, limits of their competence and caution in making knowledge claims

ethical principles - responsibility

psychologists value their responsibilities to clients, the general public and to the profession

professional autonomy

trust of others is not abused

members consider professional accountability, responsible use of their knowledge and respect for the welfare and living things and the world

ethical principles - integrity

being honest, truthful, accurate and consistent in actions, decisions and methods

setting self-interest to the side and being objective

psychologists value clarity and fairness in their interactions with all persons, and seek to promote integrity in all facets of their scientific and professional endeavours

includes not fabricating any data, and being honest and accurate regarding any claims you make in reports

members consider openness, unbiased representation, fairness, avoidance of exploitation and maintaining personal/professional boundaries

key responsibility/ethical issues

informed consent

deception

protection from harm and discomfort

debriefing

confidentiality

informed consent

all studies require a participant information sheet, written in appropriate lay-language

• includes what the study involves (purpose, procedures, duration) and the right to refuse/withdraw at any time without giving reasons

  • get participants to create unique ID code, so if wanting to withdraw they can email this to stay anonymous

• also researcher’s contact details

after reading the PIS, they can give informed consent and sign the form if they choose to participate

deception

intentionally misleading participants

shouldn’t be used unless necessary

must explain the deception as early as possible (in debrief)

opportunity for participants to withdraw data after being made aware

monitor response during debrief, and make sure they are not upset

protection from harm and discomfort

potential for harm must be made explicit in the PIS

obligation to protect participants from harm

includes physical and psychological harm (eg. stress, anxiety, fear, embarrassment)

can put resources in place for support if harm does occur

debriefing

debrief sheet presented to participant at end of study

• includes summary of study’s true aims

• has resource contact details if needed

ensure participant’s well-being

opportunity to ask questions

verbal discussion about debrief contents before ending study

confidentiality

participant data should remain confidential and anonymous (stated on PIS)

• confidentiality - researcher knows participant identity, but takes steps to avoid it being known by others

• anonymity - researcher doesn’t know identity

need to inform participants if confidentiality cannot be guaranteed

use of unique ID codes so participants can remove data

separation of data files and names/identifiable information

• a legal issue, as well as ethical

key requirements in ethics

obtain informed written consent

avoid deception

protect participants from harm (avoid research likely to cause distress)

right to withdraw

maintain confidentiality and anonymity

debrief

adults only (18+), and no vulnerable populations

ceiling effects

unable to detect any possible effects because the upper range is restricted

eg. task is too easy

floor effects

unable to detect any possible effects because the lower range is restricted

eg. task is too hard

generalisability

extent to which results can be applied to a wider population

eg. study only used undergrads, results can be applied to wider population, but they’re younger, more educated and intelligent so be cautious with wording

structure of report

title

abstract

introduction

method (design, participants, materials, procedure)

results

discussion

references

appendix

ethical considerations

referencing

APA 7

citations include surnames and year of publication only

with 3+ authors, shorten to ‘et al.’

surname, initials (year), title of article, title of journal, volume (issue), page numbers

writing an introduction section

explains study rationale and justifies research question

includes why topic is important, what work has already been done, how to build on this knowledge and aims

use funnel order

• start off general

• put most relevant literature later on

make clear how study builds on previous research

writing a methods section

report method in past tense, don’t justify reasons here

allows reader to replicate research

design

• type of design (within/between groups)

• variables being measured (IV, DV)

• different conditions

participants

• sampling strategy

• sample size

• important demographics/characteristics eg. sex, age

• any inclusion/exclusion criteria eg. language, vegan

materials

• only include specialised equipment

• only say what was used, not how

procedure (steps taken to run study)

• precisely describe what was done from start to end

• be clear, only include relevant information

• explain how participants were assigned and reasons behind choices

‘Method’ heading is bold, centered, no underline (like all headings)

subheadings are to the left, bold, no underline

writing a results section

report data with analysis (no raw data) and describe results briefly

don’t give rationale of why, but state direction of results

include descriptive and inferential stats (in this order)

presenting table:

• introduce and describe table in text before it is presented

• include table number and label above it (to left side)

• statistical letters (M, SD) are italicized

• all values are rounded to 2 decimal places, except p-values

presenting figure:

• same rules as table, but label goes below the figure

• label axis meaningfully, with units of measurements

• only use black, white, grey

writing a discussion section

summarise aims and key findings (no statistics here)

say if hypothesis was supported

offer explanations for findings with previous research used in introduction

say implications of research (theoretical and practical)

critical evaluation of research (limitations and strengths)

suggest areas for future research

conclusion

research purpose

quantitative

• recording and understanding objective truth

• seeking explanatory models/theories

• often reductive

• hypothesis testing

qualitative

• focused on meaning

• understanding situated meaning and meaning-making practices

the big theory position

a positive paradigm

involves ontology and epistemology

• this creates research paradigms

assumes a single, objective reality that can be measured and understood through quantitative methods

there is a ‘clear truth’ through observation

ontology

concerned with what reality is

this examines the nature of reality and existence

eg. single vs multiple truths/realities is constantly debated

• single = one objective truth

• multiple = people have differing perspectives on same event = many valid interpretations

epistemology

the theory/study of knowledge/reality and how we know/understand/examine it

knowledge can be measured using reliable tools and designs, best suited to solve the problem

reality needs to be interpreted to discover underlying meaning

positivist vs realist

research paradigm / structure of scientific revolutions - Thomas Kuhn

science is guided by paradigms

combination of ontology, epistemology and methodology

a dominant way of doing/thinking about something in a certain field

• a typical example/pattern/model of something

they come with their own set of tools and ways of measuring things, which is refined over time to provide explanatory power

when they change, there usually are significant shifts in criteria determining problems and solutions

overtime we repeatedly refine our tools and measures to provide explanatory power, but bit by bit a picture is built telling us our framework for understanding a phenomenon is wrong

• a new idea is put forward and gains traction, and a scientific revolution occurs

• new paradigm might have whole new set of tools, ideas and ways of measuring things, making it comparable to the previous paradigm

crucial that we think of science being measured against a common set of standards to decide what idea/explanation is better

3 most common paradigms

positivism

• ontology - 1 single reality/truth

• epistemology - knowledge can be measured

constructionism

• ontology - multiple realities/truths

• epistemology - reality needs to be interpreted

pragmatism

• ontology - reality/truth is constantly debated

• epistemology - knowledge should be examined using the best tools

orientation to truth

quantitative

• singular truth

• a world knowable through systematic observation and experimentation

• positivist or post-positivist

qualitative

• multiple truths, situated/life-embedded truths, partial truths

• partially knowable world

• meaning and interpretation as situated practices

• non-positivist or constructionist (multiple)

researcher role

quantitative

• impartial observer of object of study

• unbiased reporter

• objectivity valued

• subjectivity threatens single, objective truth

qualitative

• situated interpreter of meaning

• subjectivity valued

outsiders

researchers who don’t share similar experiences/backgrounds with the group under study (not members of the group)

advantages:

• objectivity, detachment and distance occurs

• independent observations not available to the insider eg. by asking naive questions

• meanings and perspectives not obvious to insider

• participants might reveal sensitive information, they wouldn’t reveal to the insider (more open)

disadvantages:

• lack of understanding of experiences

• unaware of cultural and social norms of group

• pathologizing the other

• descriptive and shallow analysis

• difficult access to group

• need a trusted informant/gatekeeper

insiders

researchers who share similar background, believes and experiences as the researched group, or they’re a prior group member with existing relationships within the community

advantages:

• understanding of the communities history, culture, interactional styles and shared meanings

• insight into the matter = easier development of research question and interview schedule

• easy access to group, recruitment and rapport

• depth and breadth of understanding, unavailable to outsider

disadvantages:

• overidentification with the insider role and overinvolvement may compromise ethics

• illusion of ‘sameness’ of experience of others

• loss of intellectual and emotional distance = biased analysis

• negative impact of existing relationships

straight/direct replication

repeating same exact method

can only confirm the original findings completely

or disconfirm them to some extent (increases confidence in original findings, but does nothing to further understanding of topic)

when carrying out replication studies, most find the original fails to include all necessary detail to enable precise reproduction

used to fight fraud, statistical error

partial replication

including variations not part of original study

offers possibility that something new will be learnt (extra value as a consequence)

there is more to be gained from investigating new questions generated by original study, than from straight up replicating it

role of replication in research

has a paradoxical position in psychology

the means of scientific progress, but obstacles get in the way

seen as a mundane, uncreative process lacking in originality (bad career move)

failures blamed on methodological shortcomings of replication, rather than original study

researchers biased in favour of significant results

• a study with non-significant results is unlikely to be published, even though it contradicts original study

if a study is replicable using different contexts/procedures, this is better evidence that original findings are not just reliable, but also robust

ways of generating hypotheses

detailed description of phenomenon

attention to relevant theories

deductions from theories

everyday issues

new social, technological or biological developments

antecedent behaviour consequences model (ABC)

predicting behaviour

alternative explanations of findings

temporal order of variables

more realistic settings

conflicting/inconsistent findings

methods of qualitative data collection

observation/ethnographic field notes

case studies

interviews and oral histories (semi-structured)

focus groups

diaries (written, video)

media/meta-data (newspapers, magazines, TV)

documents/archives

internet data

naturally occuring data

ethnography

immersion in a particular field

examine a group/phenomena for an extended period of time

observing behaviour, listening to conversations

active participation eg. asking questions

issues:

• gaining access = ethics (overt/covert)

• having key ‘informants’ or gatekeepers

• can be structured or unstructured field notes

• video record

diaries and documentaries

often used in health psychology

participant set a task to complete

could have been produced prior to the research

documents (letters, autobiographies)

internet-mediated research

publicly available data from the internet growing area of research

data taken off public websites eg. online support groups

may have issues of ethics, so best practice to seek consent

naturally occurring data

some qualitative researchers argue that data should be ‘naturally occurring’ eg. mealtime conversations

data has been produced without the intervention of a researcher

discursive psychologists have suggested a move away from interview data

interviews

opportunity to hear participants talk about a particular aspect of their life/experience

questions function as triggers that encourage talk (need prompts too)

non-directive, although interviewer drives the research question

balance between control and freedom for participants to go ‘off-track’

• generates novel insights

most widely used method of data collection in qualitative research

compatible with several methods of data analysis

easier to arrange than other methods, but not always easy to conduct

requires careful preparation and planning

issues to consider:

• what questions to ask, to get at research question focus

• who/how to recruit

• where to interview, or online

• how to record and transcribe the interview

includes structured, semi-structured and unstructured

structured interview

similar to a questionnaire - a pre-set list of questions

easy analysis, quick, easy process

more standardised

may be lack of detail in responses

unstructured interview (in-depth, qualitative)

driven by the participant, not researcher

time-consuming

more flexible and interactive

not a fixed agenda/questions

semi-structured interview

includes an interview schedule as a guide, including questions

also allows expansion on answers too for extra detail

interview schedule

guides the interview, not dictates

forces you to think explicitly about what to cover

enables thinking about potential difficulties and sensitive areas

researchers tend to argue that rapport is important

important to frame questions in a way participants will understand

aim is to hear their story, so don’t have to rigidly stick to questions

rapport

establishing atmosphere of trust

constructing a semi-structured interview schedule

think about broad range of themes to cover

put questions from these themes in a logical order

think of appropriate questions related to each area, and sequence them

think of prompts

if covering a sensitive issue, leave these til later

to establish rapport:

• sensitive questions in middle

• lighter questions at start and end

types of interview questions

descriptive – want participant to provide an account of something

• ‘please describe to me what happened’

structural – how does participant organise their knowledge

• ‘why do you think that happened’

contrast – ask participant to make comparisons between events and experiences

• ‘could you please compare these 2 events’

evaluative – what are their feelings towards someone/thing

• ‘how did it make you feel’

probing – ‘can you explain that more’

interview questions to avoid

closed questions (produce yes/no answers eg. did you like what happened)

double-barreled questions (confusing and forgetful eg. what did you think about that and why)

leading questions (with evaluative/emotional value eg. sexism is horrible so what do you think of his words)

jargon

conducting an interview

take schedule as a guide, become familiar with it before

relax participant before, don’t rush

check recording equipment is working

be prepared for questions to be answered before you ask them

respond to what is said, and monitor effect on interviewee

keeping interview records

research diary useful tool

systematic labelling of data

reminder of event

start of analytic procedure

informs later interviews

need demographic details of participants (age, sex, ethnicity, occupation)

focus groups

groups of 4-8 people, recruited under some remit

useful for informal group discussions that are focused on a particular topic/set of issues

based around a focus group schedule - its’ job is to facilitate group discussion between the participants

collective views of the group can be expressed

evidence suggests the group context facilitates personal disclosures

good for eliciting people’s own understandings/viewpoints, and for observing how these are advanced, elaborated and negotiated in social context

issues:

• how many in a group? how many groups?

• how to recruit

• need an organised and engaging schedule

• what questions and tasks to include

• confidentiality is tricky

• permission (ethics)

they can’t be indicative of population characteristics

transcriptions made after of the group discussion (audio/video)

• analysed using the broad principles of grounded theory, or discourse/conversation analysis

ethical considerations in qualitative research

informed consent, right to withdraw, confidentiality, anonymity, security of data

working typically closer with participants’ words, audio and video, so more care is needed

we present data in reports so need to use pseudonyms and change any identifying features during transcription stage

each mode of data collection has its own challenges:

• able to access the right people?

• are participants particularly vulnerable? eg. kids, NHS patients, criminals

• have the skills to handle difficulties and develop appropriate protocols? eg. if become upset

• check if participants want to stop/withdraw

thematic analysis

one the the simplest methods of data analysis

researcher identifies themes which reflects the data

data familiarisation is key

flexible, compatible with any research paradigms

provides a rich, detailed and complex account of data

a method for identifying, analysing and reporting patterns/themes within data

minimally organises and describes data set in rich detail

a rich thematic description of your entire data set = reader gets a sense of them predominant/important themes (complexity is lost here but rich description achieved)

includes inductive and theoretical

inductive thematic analysis (data driven)

bottom up approach

focus on the data

driven by data, less by researcher’s interest/prior reading of topic

don’t necessarily link directly to interview schedule themes

no pre-existing coding frame (research question can evolve through coding)

theoretical thematic analysis

top down approach

driven by researcher’s interest and prior reading of an identified gap in literature

may link more closely to interview schedule focus

rich description of dara and more detailed of analysis of some aspect of the data

reflexive thematic analysis - 6 stages of analytic procedure (Braun and Clarke)

familiarisation of dataset

coding

generating initial themes

developing and reviewing themes

refining, defining and naming themes

writing up

familiarisation of dataset - step 1 of reflexive TA analytical procedure

deep, intimate knowledge of your dataset - immersion

critically engaging with data and reading it more than once

transcription of data

questions to consider:

• how does the person make sense of whatever they’re discussing

• why might they be making sense of things in this way

• which different ways do they make sense of the topic

• how ‘common sense’ or socially normative

• what assumptions do they make in describing the world

• what kind of world is revealed

reflexive TA

• why might I be reacting to the data in this way

what does my interpretation rely on

note making

• note ideas around data and hand scribble on hard printed copy

• voice recognition software to comment

• additional document for each interview, research diary etc

• brief, systematic overall familiarisation of whole data set

• potential patterns

code

meaningful piece of transcript/data

initial coding - step 2 of reflexive TA analytical procedure

preparing for coding

• forms the building blocks of analysis

• codes capture specific meanings within a dataset of relevance to research question

• succinct labels that evoke data content

• code label = summary of analytic idea

• codes can be summative/descriptive/conceptual

systematic process

• involves reading each data item closely, line by line, tagging all segments that is potentially relevant, and giving it a code label

• some may be tagged to different codes

• insight and rigour, avoid cherry picking

codes should connect to more than one segment of data

• idea is to capture repetition of meaning

• code can be useful even if it only occurs once, as themes can be developed from multiple codes

subjective process of interpretation and meaning-making

multiple coders can be useful to gain richness, but not essential

technologies of coding

• handwrite code levels on sticky notes on printed transcripts

• use comment box in review mode

• computer assisted qualitative data analysis software

inductive data coding

data driven

dataset as starting point

always shape what we notice about the data

deductive data coding

researcher/theory driven

dataset provides foundation for coding and theme development, but reflect theoretical/conceptual ideas the researcher seeks to understand through dataset

existing theories and concepts might provide lens the researcher can make sense of data through

semantic coding

participant driven

exploring meaning at surface level of data

semantic codes capture explicitly expressed meaning (participant expressions)

initial coding often semantic

latent coding

researcher driven

deeper, more implicit/conceptual meaning

sometimes quite abstract from obvious content

themes

capture shared meaning, united by a central organising concept

expression of shared/similar ideas/meanings across different contexts (semantic or latent level)

unite a topic, rather than a shared meaning/idea

generating themes - step 3 of reflexive TA analytical procedure

analysis

• generative, circular process

• series of choices made

• cluster together potentially connected codes into candidate themes

  • these can change for the final version of themes

• if there is a core idea, likely to be a theme

• consider what story they tell about dataset in addressing the research question

good themes are distinctive and capture something meaningful

• are coherent

• have a central idea/organising concept that meshes the data codes together

• has clear boundaries

if there is a separate construct within the theme, develop a subtheme

thematic maps help for visual people

• helps identify themes and subthemes

• theme tables common way

avoid clustering codes into themes by answers to a question

• this constrains ability to notice patterned meaning across data set

• prevents exploring those not immediately obvious

developing and reviewing themes - step 4 of reflexive TA analytical procedure

thematic mapping helps to review tentative themes

identify boundaries

is there enough meaningful data to evidence themes - are they nuances, complex and diverse?

are the data contained within each theme too diverse, wide-ranging?

does the theme convey something important - if not, rework and discard some

refining, defining and naming themes - step 5 of reflexive TA analytical procedure

theme definition

• write a few sentences that clarify theme (take home point)

• what is it about, what is the boundary, what’s unique about each theme, what does each one contribute to overall analysis

naming themes

• should convey gist

• short phrases that captures essence of theme

poorly named themes misrepresent the data

problems – themes might be too descriptive so would not capture the deeper meaning in them or over-interpretative so they’re not rooted in the data (codes, quotes and transcripts)

writing up report - step 6 of reflexive TA analytical procedure

when to link analysis to wider literature:

• early reading = narrow analytic field of vision = focusing on some aspects of data at the expense of other potential crucial aspects

• engagement with literature can enhance analysis by sensitising you to more subtle features of the data

• no one right way to read and incorporate it

pitfalls

• don’t forget to analyse data

• avoid all analysis being based around your interview schedule

• do the themes work? Is it convincing? Not too much overlap between themes?

• mismatch between data and analytic claims – ground claims in data

• address research question (which can be adjusted after data analysis, especially in data driven TA)

different types of research

experimental (aim is cause and effect)

cross-sectional (aim is looking for relationships)

qualitative (aim is in-depth understanding)

observational (aim is initial investigations)

structured observations

researcher’s don’t observe everything

certain predetermined behaviours and observed that are relevant to the research question

likely to use a table and tally when certain behaviours occur

quantitative data

observer consistency

objective, but may miss out on important details

unstructured observations

researchers recording the behaviour they can see

observer monitors all aspects of the phenomenon that seems relevant

appropriate to identify key components of the problem and to develop hypotheses

qualitative data

rich, detailed, potentially subjective