Quanti - Sampling Process Flashcards
What is a sampling plan?
The overall plan for recruiting subjects to a study.
What is target population
The entire aggregation of cases (e.g. human), to which the researcher intends to generalize findings.
Defined by a set of eligibility criteria. (inclusion and exclusion criteria)
What are inclusion and exclusion criteria?
Inclusion: Primary traits of the population that will qualify
someone as a subject.
Exclusion:
- Factors that would preclude someone from being a subject.
- Factors potentially interfere the interpretation of results
Purpose of eligibility criteria
- Create homogenous sample that is of researcher’s interest.
- Assure the subject’s ability to participate in a study
- Eliminate the characteristics which would interfere results interpretation (Control the
extraneous factors)
What is accessible population?
- Portion of the target population, to whom the researcher can gain access to.
- Represents the target population as closely as possible
e.g. Target population: All Chinese patients with heart failure who meet the eligibility
criteria in Singapore
Accessible population: All Chinese patients with heart failure in NUH in Singapore
who meet the eligibility criteria.
What is sample?
A portion of the accessible population.
- reflects the relevant characteristics and
variations of the accessible population
What are some key considerations for a study sample?
- Representativeness
The extent the sample i similar to the accessible population. - Sampling bias
Systematic over/under-representation of some segment of the population in terms of a characteristic relevant to the research question
What constitutes sampling bias?
Non-response bias:
- when a significant segment of the accessible population refuse to participate.
- can be detected by comparing the characteristics
of the recruited subjects and the non-respondents.
E.g., Investigation of HRQoL of patients with Heart Failure
Questionnaire survey:
– 400, 210 completed the questionnaires, while 190 were not responded.
– One possible reason is that those with no response may experience more severe signs and symptoms
What does the researcher study in a descriptive research?
And what does he draw a conclusion about?
The researcher studies a sample.
He draws a conclusion about the target population.
What is sampling?
The process of selecting a portion of the accessible
population to represent the target population.
What is probability sampling?
- Random selection of subjects from the population.
- Each element in the population has an equal chance of being selected.
What is non-probability sampling?
- Select subjects with nonrandom methods.
- No way of estimating the probability of each subject’s
being included
What are the 3 approaches of probability sampling?
Simple random sampling
Stratified random sampling
Systematic samplling
What is simple random sampling?
- Process of randomly selecting a sample
from the accessible population - Step 1: Establish a “sampling frame”:
- The actual list of all the elements in the accessible population
- Step 2: Randomly selected the required number to the study by
- Random table
- Lucky draw
Advantages and disadvantages of simple random sampling
Advantage: Any difference between sample & population is only by chance.
Disadvantages: Time-consuming, full list of elements of the population might not be available
Random assignment vs random sampling
Random assignment (subjects randomly assigned to intervention groups): a signature of an experimental
design (RCT)
Random sampling: method of selecting people for a study, is not a signature of an experimental design.
In fact, most RCTs DO NOT involve random sampling
What is stratified random sampling?
Process of selecting a sample to identify subgroups in the population that are represented in the sample.
Step 1: Define strata of concerns
- Strata = mutually exclusive segments of a population based on specified characteristic (e.g. gender, education levels)
- Characteristic of the population that have prominent influence on outcome was considered
Step 2: Randomly selected the desired number of elements from each strata
- Proportionate : the elements from each stratum is proportional to the size of the strata in the population.
- Disproportionate: the proportion of elements from each stratum is not equal across the included strata.
Advantages and disadvantages of stratified random sampling
Advantage: Prevents the possibility that some subgroups of population are under-represented.
Disadvantage:
- Require more labor than simple random sampling.
- Impossible if information about stratifying variables is unavailable
What is systematic sampling?
A process of selecting every kth case from some pre-existing list of the accessible
population.
Step 1:Define sampling interval:
- the standard distance, in terms of number of
individual in a list, between the selected individual
- Number of population / number of desired subjects
Step 2: Recruit the element at each sampling interval
e.g. Population: 5000
Desired sample: 50
Sampling interval: 5000/50= 100
Recruit every 100th individual as listed in a list
Advantages and disadvantages of systematic sampling
Advantage: A convenient and efficient method to achieve
the similar effect of simple random sampling.
Disadvantages: Can only be applied if there is a list for the
accessible population.
What are the 3 approaches of non-probability sampling?
- Convenience sampling
- Consecutive sampling
- Snowball sampling
What is convenience sampling?
Select the most conveniently available people as participants.
e.g. Stopping people in the street for interview/Post up advertisement for subject recruitment
What is consecutive sampling?
Recruit entire accessible population within a defined period of time.
e.g. A study examined the effectiveness of dressing with duoderm gel and wet-to-dry dressing in promoting healing of stage III wound. All patients who admitted to the surgical wards with a stage III wound over a 6-month interval were recruited. By flipping a coin, they were assigned to either to receive dressing with duoderm gel or wet-to-dry dressing.
What is snowball sampling?
Relies on early sample subjects to refer others who meet the study eligibility criteria, to participate in the study.
Accessible population are those with special traits who are difficult to identify.
Advantages and disadvantages of convenience sampling
Advantages: Convenient and practical
Disadvantages: Potential sampling bias due to the possibility of self-selection by the researcher or the subjects.
Probability vs non-probability sampling
- Prob:High likelihood of obtaining a
representative sample
Non-prob: Rarely representative of the target
population - Prob: Less likely to subject to sampling bias
Non-prob: More likely to subject to sampling bias - Prob: Expensive & demanding
Non-prob: Practical and economical
How do you evaluate a sampling design?
- Is target/ accessible population being identified &
described? - Are eligibility criteria specified?
- Are the sample recruitment procedure clearly described?
- Does the sampling method yield representative sample?
Any justification for its use? - Are possible sample biases identified, e.g. response
bias? - What is the source of possible sampling error?
Sampling error vs sampling bias
Sampling error: difference between population parameter and sample statistic (DUE TO CHANCE)
e.g. difference between mean PSLE score in people staying in Clementi and national mean PSLE score.
Sampling bias: DUE TO SYSTEMATIC ERRORS in sampling process
e.g. difference between RGPS mean PSLE score and national
mean PSLE score.
