SAMPLING STRATEGIES 27/10/16 Flashcards
What is a sample?
collection of data drawn from a population
e.g. patients, clinicians
Why is sampling important?
- who or what is included in your studies can affect your answers to your studies
Why do you sample?
- cost and utility problems
2. accessibility issues
What is total finite population we wish to know about from which your sample is drawn from?
target population
What are the units/participants drawn from the target population that constitute our data set?
study sample
What needs to be done/taken into account in a study sample?
- need to be extrapolated to your target population or greater
What does the including/excluding criterias from a study sample impact on? (2)
- target populations
2. ability to extrapolate
What are the most important biases to be aware of?
those from inclusion/exclusion criteria
What is a complete sample?
entire study population
Unstratified random sampling?
where every member of the target population has the same chance of being sampled
What is stratified random sampling?
- random sample from the target population with in strata i.e. groups such as sex,age, disease onset
- with in each stratum has the same chance of being sampled
What are the pros and cons of complete sample?
PRO
- no bias introduced by design
CONS
-expensive
What are the pros and cons of stratified random sample?
PRO
easy to design and conduct
CONS
- smaller groups may be under-represented
What are the pros and cons of stratified random sample?
PRO
- representative population
- unequal sampling strata improves power for rare strata
CONS
- population may not be easily divisible into strata
- strata may not be known until after sampling
What steps do you need to take into consideration when sampling? (6)
- what is your target population
- what are yous ambling limitations?
- what is your sampling scheme?
- what sample size will you aim to collect given the limitations?
- how will check if your sample is adequate
- what could go wrong and how can you improve?
