What are two main types of study design? (broadly)
Descriptive studies (studies which describe things) Analytic studies (studies which test hypotheses)
What are two types of analytic studies?
Experimental studies -eg; randomised controlled trials
Observational studies eg; cohort studies, case control studies
What is the definition of a population?
Complete set of entities of elements or units or people that we wish to describe or make inference about
What is a sample?
A subset of a population
What is a census?
A complete enumeration of a population. Rare.
What is a sampling frame?
List of items in a population from which a sample is drawn
Rarely coincides with the entire population of interest:
- Telephone numbers
- Electoral role
Often don’t exist
- All people with depression
- All potential users of a new drug
Even without a list, we can ensure an ‘unbiased’ sample if every individual has the same chance of being drawn.
What is random sampling?
- Choose the sample in such a way that every individual in the population has a known chance of being selected
- In a simple random sample, everyone has an equal chance of being chosen
- This method is the best way of obtaining a sample which is representative of the population.
What are two types of error?
- Systematic error (bias)
- Random error (chance)
Describe random error (chance)
- Due to natural biological variability
- Increasing the sample size will reduce the random fluctuations in the sample mean
- Statistical methods allow us to quantify the influence of random error on our estimate
Describe systematic error in a descriptive study (bias)
- Due to aspects of the design or conduct of the study which systematically distorts the results
- Cannot be reduced by increasing the sample size
What are some types of sampling? (4)
Probability sampling (those below are examples)
- Simple random sampling
- Stratified random sampling
- Cluster random sampling
Describe simple random sampling
- The simplest form of probability sampling
- For a finite population of size N draw a sample size n such that each possible sample has the same probability of being selected
Describe stratified random sampling
- Much of statistical design theory is about controlling variation
- Statified sampling is useful when the population comprises several groups of similar individuals
(A stratum is a population sub-division of similar units)
-Take s ample random sample from each stratum
Describe some benefits of stratified sampling
- More precise estimate for the same sample size
- Can take different sized samples from different strata (A device for reducing overall variability)
- Useful if you are interested in the results for each stratum and some of the strata are small
What are two types/variants of stratified sampling
Probability proportional to size
- Everyone has the same chance of being selected
Equal numbers from each strata
- Those in smaller strata are more likely to be selected
Describe cluster sampling
The population may be composed of similar and naturally occurring groups
- Cluster samples take a random sample of groups
- In a single stage cluster sample, everyone is included in the study
- In a two stage cluster sample, there is a simple random sample of one person from each cluster. Probability of someone being in the study depends on the number in their cluster
(similar to stratified)
What are two important characteristics of a RCT
- Randomisation/random allocating. Means every study unit has the same chance of being in each group. This allows some of the underlying model fitting assumptions to be met for a more straightforward analysis
- Control group
What is confounding?
Confounding is a distortion of the association between exposure and outcome caused by the presence of a third factor.
What are traits for a confounder?
A confounder is a variable which causes the distortion (confounding). To be a confounder, a variable must be both:
- Associated with the exposure (independent of outcome); and
- Associated with the outcome (independent of the exposure)