Research Methods Flashcards
Research Ethics
The moral principles guiding research from its inception through to completion and publication of results
4 principles of BPS
Respect
Competence
Responsibility
Integrity
Respect
inherent worth of all humans should be respected.All worthy of equal moral consideration
Respect for participants privacy
Competence
Working within the recognized limits of your knowledge, skill, training, education and experience
Responsibiltiy
Value responsibility to people, general public as well as profession of science and psychology
Responsibility to avoid harm
Responsibility to not misuse or abuse your contribution to society
Integrity
value honesty, clarity, fairness in all interactions
Informed consent
Participants need to know what theyre signing up for Clear consent from participant that they consent to particpating, after having heard what experiement is after having heard their role in experiment/task and siginifxance of stuy
Right to withdraw
Allowed to withdraw at any point in experiment
Briefing
Time commitment
Aim of project
Risks outlined
Confidentiality and anonymity conditions
Contact details of researcher
Deception
Not telling whole truth to participant when giving briefing in order for participant demand, placebo or any other kind of … to occur.
Inappropriate if it leads to anger or discomfrot in patients
Must be done as soon as study is completed
Debriefing
Full info is communicated
Explain rationale for deception
WW2 experimentation
Eg Josef Mengele: Angel of Death
Experimentation on inmates often resulting in mutilation or death
Nuremburg Code of Ethics (1947): When conducting medical research risks against benefits. Cant torture, cant result in ppls deformit or death
Informed consent
Milgram study (experiement no5)
“Study on Memory”
Standford Prison Experiment (1971)
Cost vs time tradeoff (Online data collection)
Either you will have to spend a lot of money hiring someone who knows the tech, programming etc or spend a lot of time learning and doing this yourself
Challenges: Stimulus presentation and multiple submissions
(Online data collection)
Is the stimulus being presented and viewed in the way you intended/want?
Multiple submissions form participant because of forgetfulness or gained incentives can distort results
Challenges: Data quality
(Online data collection)
Missing values : Researcher must decide whether to keep participants data or leave it depending on how relevant it is to research question –> validty
Outliers:
-Straightlining (giving same response to all questions, take into account completetion time
-Are they real outliers?
Characteristics of Big Data (3Vs)
VOLUME: large data size
VARIETY: Unstructured, different data types
VELOCITY: data generated fast, real time
Dealing with Big Data
1) Extracting data:
Data scraping: extraction of data from web sources and structuring it into a more convenient format.
AND/OR
2) Analysis :
Data mining: analyzing large datasets to uncover trends and valuable insights, no data gathering or extraction.
Data collection: Reactive vs Non-reactive
Reactive: people are aware that they are being studied (study, experiment, interview…)
Non-reactive: People are not aware that they are being studied (observational: on social media, in shopping center…)
Internet-mediated Research (IMR)
Research where researcher is physically absent
Reactive: eg online survey
Non-reactive: eg. app generated data, search engine history, social media posts…
Challenges of Big Data and IMR
Right for privacy:
Where is user expecting to be observed? Privacy setting change? Terms and conditions of platform?
Informed consent and withdrawal:
Deleted post/account count towards research/in data?
Does researcher notice?
Confidentiality and Anonymity: Sensitive topics? Risk of embarassment?
Literature Review vs Systematic Review
Systematic Review: focused on specific research question
Literature Review: broad and descriptive research question
- Plan the review protocol
(Conducting a systematic Review)
-a complete “recipe” of how you will conduct your research so someone else can follow exactly
- Explain and justify each step
- Search the literature
(Conducting a systematic Review)
Define SEARCH TERMS on RELEVANT DATABASES and DATES
So youre choosing time frame
Using key/relevant words
- Screen your search results/findings
(Conducting a systematic Review)
Papers youve found meets inclusion criteria?
Go hrough titles and absteact to decide whethe rto include
You want Randomised controlled trials in your research then kick out the research that doesnt have it. Do this for the resz of your inclusion criteria
- Appraise the quality of studies
(Conducting a systematic Review)
Bias/quality assessment
- Synthesize the evidence
(Conducting a systematic Review)
Write out about bias/quality, results, exclusion etc
Basically everything you did but putting it onto word
Meta-anaylsis
Should only be done in the context of final step of a systematic review (but optional step)
Types of statistics: descriptive and infernetial
Descriptive:
sumarrise and desrcibe data set
Concrete and known values
Inferential:
Use probability to infer/ draw conclusiosn about larger population from smaller samle data
Abstract/estimated
- Nominal
(Levels of measurement of data)
Just names categories eg gender, age, monolingual etc
- Ordinal
(Levels of measurement of data)
Categories and their RANK
Eg 1st, 2nd and 3rd place or
disagree, slightly disagree, agree etc
But “distance” between values are not known
eg whether 1st and 2nd place were 1s or 10s apart
- Interval
(Levels of measurement of data)
Categories are ranked on a SCALE
So the quantities are meangful because you know the interval between them and greater range of stats can be calculated
- Ratio
(Levels of measurement of data)
Categories are ranked, on a scale and has a TRUE ZERO, but there is no negative values
eg height
Can be used in ratios
Discrete and continuous data
Discrete: Nominal and Ordinal
Dataset has fixed number of values
Eg categories monolingual and bilingual or on likert scale 1 2 3 4 5 and no 1.5
Continous: interval and ratio
theoretically there are an infintie number of values.
What do descriptive statistics describe
-central tendency: most typical/represntaitive socre
-dispersion: how much do values vary around central value
oImportant to look at and rpeort both
Mean:
(Measures of central tendency)
Most senstive measure–> senstibe to outliers
Eg averga incoeme in UK will seem inflated if you take into account the few billionaires and caluclate the mean. Real value will lie somewhere closer to the median value
Thats why its important to visualize your data!
Median
(Measures of central tendency)
Point/value that divides data values in equal halves
Mode
(Measures of central tendency)
Value that occurs most often
Range
(Measures of spread/dispersion)
difference betwen highest and lowest value
Interquartile Range, IQR
(Measures of spread/dispersion)
Box plot–>
Line up data into fourths
Lower quarter and upper quarter together make the interquartile range, this accounts for 50% of values
The outliers are added to the box in the form of “whiskers”
Variance
(Measures of spread/dispersion)
- Calculate mean
- Each value - mean
- Square value from step 2.
- Add up all the values
- Divide by the number of values - 1
hard to interpret because doesnt show spread in terms of original units
Standard deviation
(Measures of spread/dispersion)
Is the square root of variance
Get better sense of spread of values than with variance
The larger the SD, the larger the variance/ data is more dispersed
Normal distribution
Mean, median and mode are the same value
Most interval meausures should produce normal distributed data
Skewness
Measure of asymmetry in a distribution
Positive skew = less values at the positive end, values clustered at negative end
Negative skew: less values at negative end, values clustered at positive end
Central tendency will be further in the direction of the skew : mean> median > mode
In box plot, median line will no longer be in the middle
Kurtosis
Refers to the PEAK and DISPERSION of the data–> how “squished” it is
Platykurtic (“flat) : values are more dispersed so curve is flatter
Leptokurtic (“lept”) : distriution is more peaked
Probability
Number of ways the event could arise / number of possible outcomes
In statistics, probabilty expressed from 0 to 1
1 = 100%
0.5 = 50%
0.05 = 5%
Factors leading up to qualitative research
Epistemology–> what is knowledge
Making observations, gathering very many of them and then coming up with a conclusion
Positivism: Observations reflect reality
Can we as scientists observe the truth?
Post-positivism: the truth is out there as long as we develop a good enough questionnaire/experiment
Limitations:
-no matter how good the experiment, you cant make an unbiased judgement
-importance of context
-truth and experiences are not culturally relevant around the world
Constructionism and Contextualism
Constructionism: Our perception of reality is very much determined by our community/social factors
Contextualism: Truth is not universal and its relative to context
Eg take an image of a sportsperson, crop out the body and only keep face–> without its entire context its doesnt give the same impression so a different judgement is easily made
What is Qualitative research
Shift away from positivism
Use naturalistic setting where people express their views and thoughts that researchers can anaylse
-Inductive research
-Through interviews, focus groups, diaries etc
-Experiential–> naturally occuring beliefs, behaviour and narratives
-Take as long as necessary
-Rich, detailed knowledge of topic area (not limits to how much you can gather unlike quantitative which has strict limits according to the hypothesis)
Phenomenology
(Qualitative approaches)
Interested in how people experience a particular phenomena–> interested in their personal perception, how did different people perceive the same thing?
how does the chocolate bar taste to you?
Grounded Theory
(Qualitative approaches)
You have/start with a general question, then you collect more and more data
Inductive approach: focus group arrives, you look at inital data. Keep gathering data that gets more and more specific until you get a to a specific quesition. You build you data from the intial focus group
The goal of grounded theory is to develop a theory that is grounded in the data, rather than being imposed on the data from a preconceived theoretical perspective.
Discourse Analysis
(Qualitative approaches)
Focus on language
HOw does language construct our social reality: HOW people say things, how LOUD theyre speaking, their TONE etc–> minute details of language and behaviour
Small q or Big Q
Small q: quantitative questionnarie but with open ended questions
Big Q: open ended, inductive recount/exploration and experience
Qualitative research process
1.Planning
Literature review
Rationale
Ethical approval
- Data COllection
Recruitment
Organise focus groups/ interviews
transcription - Data Analysis and write up
Coding
Data anaylsis
Reflective anaylsis
Full report
1.Experience
(Different types of Qualitative Research Questions)
People from a particular group talk/recount their experience of something related to this “group” they belong in
Eg how do service users with BIPOLAR DISORDER experience RECOVERY and SELF MANAGEMENT
- Accounts of practice
(Different types of Qualitative Research Questions)
exploring and understanding how individuals or groups carry out specific activities, tasks, or practices within a particular context.
- Understanding/Perceptions
(Different types of Qualitative Research Questions)
How people UNDERSTAND their own experiences and PERCEIVE the world around them
Eg how do unusual sensory experiences affect people with AS (austism spectrum) in their lives
- Influencing Factors
(Different types of Qualitative Research Questions)
Asking people more directly “what do you think impacts your decision to …”
Eg factors that influence parents’ food purchasing behaviours
- Interviews
(Qualitative Data Collection)
A conversation with a purpose
not question-answer question-answer but a flowing conversation.
-Specialized and specific sampling:
Not randomized because questions can be very specific to sample and if eg backgrounds of participants is very differnet its hard to be cohesive
Ethical issues: Sensitivity, confidentiality and allowing participant to guide researcher into whatever they feel comfortable talking about. Questions should be relevant to research but also not too intrusive
Transcript: Word for word recording. Can takw this back to particiapnt and ask if they are okay/comfrotable with it being in the research
Reflexivity: Work out if you as researcher have been biased, log your challenges, how you felt, did you mb agree to heartily to smth they said because it was smh u also felt strongly about etc
Conducting the interview (Qualitative Data Collection)
Wording question in a way that feels conversational and relaxed and starting the interview with general questions (not necessarily to do with research question yet)
Indirect questions: “what do you think other people might think of…” and they might follow up with their own experience
Allow silence, give time to think about answer and good active listening! Playback what you think the person is saing to make sure youve understood them corrwctly and they have a chnace to answer you
Prompts can also be used rather than questions: Photoelicitation oe vigenttes
Problems/struggles : Interviews
(Different types of Qualitative Research Questions)
- participant can go off topic and make the whole interview about a specific topic that they are very passionate about, making it difficult for you to steer towards the quesions towards the research quesiton you want
- time and resource intensive
Focus Groups
(Different types of Qualitative Research Questions)
Recorded group interview
Study the social interaction, where interviewer is faciliatator
Good for conscinetiousb ess issues where people dont agree and you can see how responses are elicited and expressed–> do views get modified through group discussion?
Naturalistic social interaction (bit researcher has less control so have to be careful abou going off on tangents)
Pros and Cons: Focus Groups
(Different types of Qualitative Research Questions)
+Good for conscinetiousness issues where people don’t agree and you can see how responses are elicited and expressed–> do views get modified through group discussion?
+Naturalistic social interaction
-Researcher has less control so have to be careful about going off on tangents)
+Efficient because you can get data from 4-6 people
-can be difficult to transcribe: who is speaking?
+Empowering for people, collective sense
-People, especially when they dont know eachother, have a higher tendency to agree with eachother rather than have a discussion about it
Online Interview and Focus Groups
(Different types of Qualitative Research Questions)
+ better accessibility
-technical hitches, making it less naturalistic than face to face
Synchronous and Asynchronus
Audio/Visual
Text
Diaries
(Different types of Qualitative Research Questions)
Record experience over a period of time. Clear instructions that diary entry completed at specific intervals and as repsonses to certain events
+Ppl recounting their feelings in the moment rather than reflecting on it later on
Disadvtange:
Drop out rates higher
No control/idea of environment of person while writing
Online data collection
(Different types of Qualitative Research Questions)
+ access larger amount of people
+Talking about sensitive issues is easier because anonymity
+practical and accessible
-trustworthiness: do people really meet the screening charactersitics/what reseacher is searching for
-digital divide: no access to people thjat dont have access to online survey eg more likely to get info on people who are more well off and younger generation who know how to use tech
Ethics and qualitative research
-Participatory Approach: Generally in positivist/quantitive research ur the expert.
Participant is the expert in the qualitative approach–> they have made the experience of a certain situation so they know more about it that u do. “Partnership between participant and researcher” , they’re more involved less passive
Informed consent: at the end to check in and ask if they’re still okay with the data they have produced to be used. Consent at the beginning and at the end of experiment/interview
Anonymity: not just about removing their names, but also about other people, places etc that are mentioned because they can be recognized
Qualitative data anylsis
Whatever method u use to collect data your gather, you’re making a story/ narrative
Starts with research question
Then data collection (through whatever methods/means u decide to use)
With this data, regardless of the research question or method used, you’re creating a NARRATIVE or story
Text to data (transcription happening) and then PATTERN BASED ANALYSIS: similarities, differences that come up again and again in the text to identify themes and sub themes
Transcription
Orthographic: Everything that has been said and who said it
Jeffersonian: what was being said and how (voice goes high, how they laugh etc)
No correction of incorrect grammar, include all the “ums” and “ah” and “like”s
Engaged with making transcript as soon as possible so you can pick it up correctly
Grounded Theory Anaylsis
You have resrach question but not many prior assumptipns as to what youre going to find.
Line by line with data and see what patterns are emrginng
Transcipt– line by line coding– focused codes (codes relevant to research question)– neagtive sampling l(ooking for evidcne against that code)– theoriteicyl
limitations of qualitative research
difficult to replicate
lack of transparency
practical challenges
