qualitative and quantitative methods, bias and scientific considerations Flashcards
what is qualitative data?
provides understanding of experiences, perceptions, motivations, behaviours based on description and observation
what is collected from qualitative data ?
- spoken words, written words, pictures, visual, documents ; observations described in words
what is the starting point of qualitative data?
- situation the researcher can observe
what is the goal of qualitative designs?
- pts comfortable with researcher so honest and forthcoming
- allows researcher to make robust observations
what are examples of methods that collect qualitative data?
- semi- structured interviews
- focus groups
- participant observation
what are interviews?
- variously directed convo based on language and interaction
- can be structured or semi structured
what are the advantages of interviews?
+ avoids pre- defined categories and opens up to surprising outcomes
+ helps refine research questions for future
+ interaction between research and researched population
what is quantitative data?
based on numerical data and usually seeks to establish casual relationships and/ or differences between variables
what does quantitative data involve?
- observations measured in numbers
what is the starting point of quantitative data?
- testable hypothesis
what is the goal of quantitative data?
- findings can be repeatedly tested by others
- variables are defined and statical analyses test relationship/ differences between vs
what are examples of methods that collect quantitative data?
-surveys
- randomised control trials
- lab experiments
list some differences between qualitative and quantitative
qualitative more subjective and in depth due to no stats test whereas quantitative is objective and has a wider breadth of info due to stats analysis
what is the difference in reliability and generalisability of the two data sets?
qualitative can be valid if researcher is skilful and presents rigour but is less generalisable quantitative is valid depending on measurement device and is more generalisable
what is the aim?
- statement describing the purpose of the study
what is the research question?
- specific question delineating the primary aim of research study
what is the hypothesis?
- proposed explanation for phenomenon; specific and testable statement about the possible outcome of scientific research
what does PICO stand for?
- population
- intervention
- comparison
- outcome
what is the null hypothesis?
no significant difference/ relationship in results so any difference is due to sampling error or chance
- no effect hypothesis
what is the alternative hypothesis?
any observed significant differences/ relationships are not due to sampling error/ chance
- the effect hypothesis
what is an in- depth mechanistic study?
- small sample and in- depth measurements also allow for detailed look at intervention effects which might inform design of larger studies in future
what is conflict of interest?
- situation in which a person is in a position to derive personal benefit and unfair advantages from actions/ decisions made in their official capacity
what is researcher bias?
- any factors e.g. investment in product that might influence researcher to favour certain results
why is conflict of interest a problem?
- compromises ability of researchers to conduct research and report results in accurate and impartial manner
what are researchers required to do?
- be transparent about funding sources and any potential conflicts of interest
what is selection bias?
individuals recruited are selected in a way that isn’t random so it doesn’t represent the population, which limits generalisability
what is self- report bias?
- error introduced with self report data that prevents accurate assessment of relationships between variables
what is recall bias?
differences in ability of pts to recall variable being measured means erroneous conclusions made ; data from one group may be more accurate than other
what is reporting bias?
differences between reported and unreported data; may select positive favourable results which don’t represent whole set of data
what is publication bias?
tendency for journals to publish studies with positive results; lack of transparency may lead to ineffective or harmful interventions/ drugs
what is confounding?
error in effect of an exposure on an outcome due to influence of another factor
what is residual confounding?
error remains after controlling for confounding so incorrectly conclude that a given exposure causes an outcome
how can you minimise bias in research ?
- journals enforce disclosure requirements and penalise anyone who doesn’t
- prevent industry influence
- develop accurate measures to reduce self- report and recall bias
- more funding for large scale RCTs
- require journals to publish negative findings
- don’t imply causation when results are indicative of associations
what is health inequalities?
preventable, unfair and unjust differences in health status that arises from unequal distribution of social, environmental and economic conditions which determine risk of getting ill, ability to prevent sickness and access treatment
what is deprivation?
damaging lack of material benefits considered to be basic necessities in a society
what is the overall relative measure of deprivation?
- index of multiple deprivation
what are some exmaples of health inequalities?
- ethnicity e.g. black people 2-4 x likely to develop type 2 diabetes
- family member has disease
- age e.g. older
- gender e.g. men more likely to be obese
- physical activity e.g. if not active your more likely to develop chronic diseases
why do some samples not reflect population diversity?
- cost, practicalities, convivence
- lack of diversity in area where research is
- research question targeted to limited demographic
- protect vulnerable groups
- optimise positive findings
what are the implications of studies with low diversity?
- findings not applicable to those under- represented
- those who aren’t represented may be deprived of benefits
- messages generated from research can lead to lack of awareness of disease risks
- findings don’t reflect complex ‘ lived experiences’ of diverse populations
- policies created could be harmful to those not represented
is it always necessary to recruit diverse study samples?
no as it depends on research question and whether representation matters
what should be done if the treatment effects are expected to be different for various groups?
- sub group analysis
what are other considerations?
- time of day; circadian rhythms cause organ systems to be active at different times
- seasonal changes e.g. vitamin D exposure
- shift working increases chance of cardiovascular disease
-exercise affects glucose response - taking blood samples e.g. type of blood tube
