Research Methodology Flashcards
What is research?
The systemic investigation into a study of materials & sources in order to establish facts & reach new conclusions
What comes under experimental research?
- hypothesis
- variables
- deductive approach
- objective
- quantitative data
- cause & effect
- statistical analysis
What comes under non-experimental research?
- open research question
- no defined variables
- inductive approach
- subjective
- qualitative data
- identify meanings/experiences
- interpretive analysis
Define experimental research
Determining the cause of something by isolating hypothesised causes & comparing the controlled results.
What are the 2 paradigms?
- ontology - what is reality
- epistemology - understanding of knowledge
What is equality of opportunity?
Everyone gets equal regardless of circumstances
What is equality of outcome?
Try to get final outcome at same level
What are the 4 philosophical worldviews?
- positivist
- constructivist
- pragmatic
- transformative
What do positivists believe?
- one reality
- causes determine effects of outcomes
- reductionist
- theory verification
What do constructivists believe?
- theory generation
- no single reality- how individuals interpret experiences
- inductive
What do pragmatic believe?
- caused by actions, consequences and situations;
- mixed methods
What is the transformative viewpoint?
- includes critical theorists & participatory action researchers
- change orientated -research focused with policy & political change to confront social oppression
Order within triangle of evidence (starting from top)
Systemic review (meta analysis) Randomised controlled trials Controlled clinical study Retrospective / prospective cohort Case report Expert opinion
What is a hypothesis?
A prediction about a specific event or relationship between variables
What is a null hypothesis
No relationship between the variables (Ho)
What is an alternative hypothesis
Makes a a prediction that there is a difference between the variables (H1)
What is a non-directional hypothesis?
Stating there is a difference but not the direction of the difference
What is deductive reasoning?
- starts with a hypothesis
- facts need to be correct to make logical conclusion
What is inductive reasoning?
Broad generalisations- conclusions made from false information
What is abductive reasoning ?
Bringing together certain information together to make most logical reason
Define temporal Order
One variable precedes another in time = one variable affects/predicts another variable
What is the independent variable
Variable you change
What is the dependent variable?
Variable you measure. Depends on independent variable
What is a predictor variable? (Antecedent)
Variable used to predict an outcome of interest
What is a outcome/criterion or response variable?
Outcome of predictor variable
What is a confounding variable?
Additional variables having an affect e.g. age
What is extraneous variables?
Any variable other than the independent or dependent variable. They can have a negative effect on study
What is a mediating variable?
Used to explain relationship between independent & dependent variable
What is a moderator variable
Increases or decreases the direction/strength of the relationship between independent & dependent variable
What is a control variable?
Variable that is kept the same
What are the 3 types of design/approaches to research
- quant
- qual
- mixed methods
What are the 3 parts of mixed methods?
- explanatory sequential (quant to qual)
- exploratory sequential (qual to quant)
- convergent (quant & qual together)
Wha is reliability?
Consistency
Getting same results
What is validity?
Obtaining correct answer
Accuracy
Can be replicated
3 breaths of coverage
- micro-level = provides explanations to small bits of time, space or numbers
- meso-level = theories link micro & macro levels
- macro level = explain larger groups e.g. social institutions
What is evidence based practice?
Interpreting individual clinical expertise with best available external evidence from systematic research
What are the key principles of qualitative methods?
- understanding context
- understanding people
- understanding interaction
Quantification without qualification is….
Irrelevant
Qualification without quantification is…
Inaccurate
What are the 3 types of research dimensions?
- experimental
- quasi-experimental (manipulation present but NO randomisation
- non-experimental
What are 4 research designs?
- descriptive (documents conditions, attitudes or characteristics of individuals or groups
- exploratory (focuses on relationships among these factors)
- predictive (development of systems to predict criteria of interest)
- explanatory (testing hypothesis & explaining phenomena)
What are the 4 research timeframes?
- retrospective (past)
- prospective (present)
- longitudinal (over time)
- cross-sectional (one point in time)
What are non traditional designs?
- case studies
- historical research
- methodological research (development, testing, evaluation)
- meta analysis for quant & meta-synthesis for qual
- evaluation research e.g. how well policy is implemented & uselfull
- needs assessment e.g. needs of group/organisation
- action research (strategy brings about social change)
What are the 4 types of qualitative designs?
- phenomenology (human experience)
- ethnography (observe in natural habitat, social culture)
- grounded theory (observation & interpretation of phenomena, construction of theory through analysis of data, analysed through coding)
- philosophical enquiry (rational reflection upon experience, all about perspective & how you view it)
What are data collection methods for qual research?
- observation
- questionnaires
- interviewing
- focus groups
- field notes
- journal entries
- photography
- documents & records
What are the types of sampling?
- convenience
- quota
- purposive/theoretical/judgemental
- snowball
How do you know when you have a big enough sample size in qual research?
Data saturation
Types of data analysis for qual data?
- narrative (subjective view)
- descriptive (objective view)
- thematic (groups/themes)
- coding
- graphic (visual representation)
What are some strengths of qual research?
- data collected in naturalistic setting
- useful for studying small sample to describe complex phenomena
- can conduct cross case comparisons & analysis
- useful in determining idiographic causation
Limitations of qual research?
- might not be generalisable to other settings
- difficult to make predictions
- more difficult to test hypothesis
- may have less credibility
- data collection & analysis time consuming
- results more easily influenced by researchers biases & idiosyncrasies
What are examples of probability sampling?
- simple random e.g. names out a hat
- systematic random e.g. every 3rd person
- stratified random - same proportion of sample
- random cluster
Examples of non-probability sampling
- quota
- purposive
- volunteer (self selection & snowball)
- Haphazard (convenience)
What are the 4 types of data?
- nominal (data follows simple naming system to indicate commonality)
- ordinal (data follows a rank order by their position on a scale)
- interval (data had equal degree of difference between each position)
- ratio (data can be compared as multiples of one another e.g. age)
What is parametric statistics?
- normal or expected distribution
- homogenous (looks same on both sides of curve)
- independent
- data type = interval or ratio
- measured using mean
What in non-parametric statistics
- unexpected or not a normal distribution
- data = nominal or ordinal
- measure = median
What is level of significance?
- expressed as numerical probability value (p)
- p value equates to the % likelihood that a result is due to chance
What does p<0.05 mean?
Probability less than 5%
What would the p value be if the probability was less than 1%?
P<0.01
What does p<0.001 mean?
Probability is less than 0.1%
What is a confidence interval (Cl)?
Range of scores within which the true parameter might exist e.g. 95%
What is the relationship between the alpher & p value?
Alpha value - sets the significance bar
P value determines whether you will get over the bar or not
If alpha = 0.05
If P is < than alpha (less than 5% of what you observe is due to chance)
If P is > than alpha (more than 5% due to chance)
What value is between 1 standard deviation of the mean?
68% of data
How many standard deviations is 95% of data?
2 standard deviations from the mean
How much of the data lies within 3SD of the mean?
99% confidence interval
What % is the median of a non-parametric curve
50%
What type of error is a false positive?
Type 1 error
What type of error is a false negative?
Type 2 error
What is the likelihood ratio?
Opposite of betting e.g. 1:1 = 50%, 1:4 = 80%
How do you work out post test odds?
Pre test odds x likelihood ratio = post test odds
How to calculate inter-rater reliability
Kappa = (Po -Pe)
———-
(I - Pe)
Po = observed proportion of agreement Pe = proportion expected by chance
Used to determine observer bias
What is external validity
Being able to generalise research to other settings
What is internal validity?
Ability to draw link between treatment & dependent variable
If null hypo is true but you reject null hypo what is it?
False positive
If null hypo is false but you reject null hypo what is it?
True positive
If null hypo is true but you fail to reject null hypo, what is it?
True negative
If null hypo is false and you fail to reject null hypo, what is it?
False negative
Give examples of biostatistical tests
- chi square
- t test
- Mann- Whitney u test
- wilcoxson signed-rank test
- ANOVA (analysis of variance)
- regression
- correlation
- kappa
What is cluster randomisation?
Random sample from groups is selected
What is sham controlled?
Sham surgery (placebo surgery) = fake surgical intervention that omits the step thought to be therapeutically necessary
Which of these words might a quant study start with?
Do Explore Why If Understand
Do
Is probability sampling random?
Yes
Which of these is NOT a type of probability sampling? Cluster random Simple random Columnar random System random
Columnar random
Which is not a type of non-probability sampling?
- indirect sampling
- quota
- snowball
- purposive
- convenience
Indirect
True or false - surveys are useful for collecting retrospective data only
False
True of false - construct validity explores the structure of the tool & uses cross-referencing of questions to determine participant honesty?
False
True or false- content validity can be determined if one can review the tool & suggest what is being measured?
True
True or false - external validity explores how the sample represents the parent population
True
True or false? Criterion-related validity compares 2 different data collection instruments
True
Reliability can be tested using Cohen’s kappa - what level and above is considered acceptable?
0.6
Give an example of nominal measurement
Male/female
What does a p value of 0.04 suggest after 2 groups were statistically compared?
There is a difference between the groups
What concept is not a trustworthiness characteristic of qualitative research?
Correlation
Which philosopher developed phenomenology?
Heidegger
Can triangulation enhance truthfulness in a qual research?
Yes
Should informed consent always be gained in qual research?
No, but where possible
What are the 4 ways of knowing
Knowing-that (theoretical knowledge found in books & research journals that can be widely applied to different situations
Knowing-how ( practical knowledge gained from training sessions and also from past experiences)
Knowing-why (a deeper understanding of a situation that allows practitioners to work outside of standard procedures)
Knowing-who (personal knowledge of other people & ourselves that enables practitioners to apply general knowing-that & knowing-how to specific & unique cases
What is intuitive knowing?
Knowing the right thing to do without fully realising how or why you know it
What is hypothetico-deductivism?
Attempting to disprove a theory
Key principles of clinical governance
- pt & public involvement
- risk management
- clinical audit
- staffing & staff management
- education, training & continuing personal & professional judgement
- clinical effectiveness & use of clinical information to inform practice
What are the components of evidence based practice?
- research
- clinical expertise
- pt preferences
- resources
Define snowball sampling
When participants recommend someone else to participate (nominated sampling)
What is triangulation
Using different sources of data &/or different collection methods to look at same phenomenon but from different perspectives
What is purposive sampling?
Selection of participants who are believed to be able to give rich, robust info about phenomenon being researched
What is transferability?
Whether the findings are relevant to other settings or to other groups of people
What is dependability?
Whether findings would be consistent if the study was carried out again with similar participants or similar setting
What is confirmability?
If research findings can be tracked back to original data
What is content validity?
When each item on questionnaire is examined for relevance
What is criterion-related validity?
How well one measure predicts an outcome for another measure.
What is construct validity?
Degree to which test measures what it claims to be measuring
What is ecological validity?
Extent to which findings are applicable to participants natural social settings
What is external reliability?
Consistency over time
What is reliability coefficient
Indicates degree of reliability by describing strength of relationship between variables
Give an example of interval measurement
Temperature
Give an example of ordinal data
Average, Good, excellent
Common tests for exploring categorical data?
Chi-square and fishers exact test
Tests used to explore continuous data?
Mann-Whitney U
Kruskal-Wallis (non-parametric)
One & two tailed tests (parametric)
What is the Hawthorne effect?
Changing behaviour because aware you’re being observed
What is attention/interest thoughts?
Sentences with the purpose to keep the reader on track
What is intercoder agreement?
When 2 or more coders agree on codes used for the same passages in the text
What is nominal data?
Named data, which do NOT overlap in categories
What is ordinal data?
Data is placed in order or scale
What is ratio data?
Equal & definitive ratio between each data
Has an absolute zero as point of origin (can NOT be negative numerical value)
What is interval data?
Data which is measure along a scale, with equal distance between each value.
What is Mann-Whitney U test?
Non-parametric test
Compares differences between 2 independent groups when dependent variable is either ordinal or continuous
What is a T-test?
Normal distribution
Test if there is a significant difference between the means of 2 groups
What is chi square test?
Determine whether there is a significant difference between the expected frequencies & observed frequencies in one or more categories
What is cohens kappa test?
Used to measure inter-rater reliability of qualitative items
When would a 1 tailed test be used?
Directional hypothesis
When would a 2 tailed test be used?
Non-directional hypothesis
What is a single blinded procedure
When participants aren’t aware of study hypothesis
What does a manipulation check measure?
Whether the study successfully manipulated the independent variable
What is specificity?
Ability to identify people without disease
What is sensitivity?
Ability to correctly identify those with a disease
What is law of averages?
Outlier which inaccurately represents the data
What are the 3 central tendencies?
Mean
Median
Mode
Define standard deviation?
Measure of amount of variation of set of values from the mean