PPS RCS Quantitative and Qualitative Research Overview Flashcards
What is quantitative research?
Measure numerical data in sample using specific study design, analyse, generalise results & findings to population.
What is qualitative research?
Understanding underlying reasons, opinions, & motivations; non-numerical data.
What happens in an experimental study?
Researcher deliberately influences events.
Records effects of intervention
Randomised Controlled Trial (RCT);
What happens in an observational study?
No intervention.
Observe & record behaviours, symptoms, attitudes.
Can be done over a period of time
Cohort study;
Cross-sectional study (surveys).
Describe the questions asked, risk factors, outcome and data collection involved in a prospective cohort study
Exposed to risk factor when recruited?
Follow participants through time, develop outcome?
Statement: Association between risk factor and outcome.
Risk Factor: FOR EXAMPLE Frequency & temperature tea drinking:
Less than weekly, Weekly, Daily (Warm, Hot, or Burning hot)
Outcome: Oesophageal cancer: Yes versus No.
Data Collection:
Recruitment different times (2004 to 2008)
Length of exposure: Variable (average 9.2 years)
Data: Time from recruitment untilcancer (if developed
Each risk factor category: Total number of years of exposure
Frequency cancers per 1000 years exposure (at risk).
What is a challenge when it comes to collecting data for a prospective cohort study over long time periods?
Recruitment can be at different times, eg someone could be exposed to substance being investigated years before another person joining the study (length of exposure variable)
Exposure for each person will also change with time, will not be consistent for many years
How do we mitigate the varying lengths of exposure in a study?
Total up the number of years of exposure so when data is collected can establish how many incidences occurred for the total years of exposure
eg. Cancer cases (n) divided by number of person years (pys) to get = cases per 1000 pys
What can be done if someone is not regular in their intake/exposure to risk factor being investigated?
These risk factor categories are quite crude, and it’s not always clear when there are people who deviate from the average, ie drink more than twice a week or are erratic in their exposure
If there are more men in a study, they are in the study longer and have more cases than women, are they more likely to develop the condition than women?
LOOKING AT RAW DATA:
So what we don’t have in here is other potential contributing factors, eg: if men were older, more likely to smoke and drink alcohol etc –> those other risk factors for cancer of the oesophagus may be in some way contributing to this data and that may lead to an increased association THEREFORE it looks like there’s a higher risk of getting cancer if you’re a man compared to being a woman. This is due to CONFOUNDING
What is confounding data?
Mixing of effects’ wherein the effects of the exposure under study on a given outcome are mixed in with the effects of an additional factor (or set of factors) resulting in a distortion of the true relationship.
In statistics, a confounder (also confounding variable, confounding factor, extraneous determinant or lurking variable) is a variable that influences both the dependent variable and independent variable, causing a spurious association.
How do we investigate populations in research?
Not physically possible to investigate the entire population so SAMPLE = Estimate population.
Population value(s): Cases cancer per 1000yrs follow-up (tea drinking categories). Fixed yet unknown (dont actually know whats happening in population): so we need Population parameter(s)
Sample estimate(s) = Estimate population parameters
What is the problem with population parameters?
we take the sample from the population to estimate population parameters –> we make inferences back into the population –> presents many challenges in research because we don’t know how accurate this sample is, how accurately it reflects what’s happening in the population.
Why is deciding what the population is a challenge?
Who are, or what is, the population?
Defined geographically, or by time?
ACCORDING TO EPIDEMIOLOGISTS:
Theoretically infinite: Statistically (important) –> Defined by study inclusion criteria.
Question: Can results be generalised?
Usefulness of study;
Rest of China? Internationally?;
Depends on sample characteristics.
What is the complex sampling method?
Need natural clusters of people (cluster sampling)
so stratify by region (to be representative of general population) -> urban vs rural, select a number of each, 100+ administrative units (communities and villages) and invite eligible adults to participate
What is an issue with complex sampling?
The selected regions and units are mostly based on convenience (convenience sampling)
eg. Just sampling from one hospital, one area
