Chapter 5: Selecting Research Participants Flashcards
target population
population that generally share at least one characteristic
accessible population? cons?
portion of the target population consisting of individuals who are accessible to be recruited
cons: must be cautious about generalizing results to target population. the accessible population may differ too greatly from target populations and thus the conclusions based on study results only apply to the study population./
representativeness
degree to which a sample mirrors/ resembles the population of interest
biased sample
characteristics of individuals in the sample are DIFFERENT from those of the target population.
sampling bias
when researchers create a biased sample by selectively choosing desirable participants (ex/ by choosing subjects with slight heat conditions to test a heart drug instead of incorporating people with severe cardio problems–> would make the drug look like its super effective)
Law of Large Numbers
the larger the sample size, the more likely values taken from the sample will match actual values for population.
relationship between representativeness and sample size
representativeness increases in relation to the square root of the sample size.
Probability sampling:
sample when each element has a known probability of inclusion; investigator has no discretion regarding the inclusion or exclusion of an element.
conditions of probability sampling
1) exact popultaion must be known
2) list of individuals in population must be known
3) each individual must have a specified probability of selection
4) random selection: every possible outcome is equally likely
Non probability sampling
odds of selecting a particular individual is not known because the researcher does not know the population size, or the list of the popultion members are not known
participants are thus selected based on availability or researchers judgement
non probability sampling leads to a ______ risk of producing a biased sample
an increased risk of producing a bias sample
probability sample methods
1) simple random sampling
2) Systematic sampling
3) stratefied random samplin
4) proportionate stratified random sampling
5) cluster sampling
6) combined strategy sampling
non-probability sampling methods
1) convenience sampling
2) quota sampling
simple random sampling
when all members of the population of interest have an equal and independent chance of being seslcted
pros and cons of SRS
pros; totally random, effective and practical way to create a REPRESENTATIVE SAMPLE, can estimate sampling error
cons; impossible without a complete up to date list of members of the population, expensive if population is dispersed, impractical if the population is extremely large, may not give samples you need (minorities) or equal # of people (males vs females)–> it is possible to obtain a distorted sample
systematic sampling?
type of probability sampling; list all the individuals in the population, randomly selecting a starting plan, and selecting every nth element from a list of the total population.
is independence princriple still intact in systematic sampling?
no, but there is still a high degree of representativeness, and spread sample more evenly amongst the population
stratefied random sampling? process?
used when populatin consists of identifyable GROUPS but is homogenous in the characteristics you are studying, allows adequate representation of all strata.
process:
1) identify subgroups
2) select an equal random sample from each subgroup
3) combines subgroups into one sample
pros and cons of stratefied random sampling?
pros; guarantees sample will contain relatively large group representing each subgroup (ex/ equal number of males and females), results of each statum may be of interest and can be analyzed separately, or compared with each other, reduced sampling error
cons;
- may not have enough information to stratify
- more complex to organize and analyze results
- difficult to measure sampling error
- may produce DISTORTED PICTURE of overall population
- groups are represented equally in a sample but likely not equally distributed in the popultaion
proportionate stratified random sampling?
allows the proportion of individuals with a given characteristic n the population to be reflected in the sample–> used in political surveys, more intensive and more work.
Cluster sampling?
population members clustered in pre-existing and natural groups, and random clusters are chosen.
example of cluster sampling?
a researcher wants to study 3rd graders, and needs 300 3rd grade students.
-instead of choosing 300 students in the city (target population) one by one, researcher can choose 10 classrooms at random, each with 30 students.
pros and cons of cluster sampling?
pros; quick and easy, treatments can be done in groups
cons; less likely to represent whole population, imprecise if homogenous clusters, larger sampling error than SRS
convenience sampling? pros and cons?
type of non-probability sampling methods; data collected are from participants that are accessible, available and willing to participate in the study.
pros; easy, less expensive, less time consuming
cons; weak form of smapling, no attempt to know the target population, no attempt to use random selection because its based on willingness to participate, little control of reprsentativeness.
Quota sampling
type of non-probability sampling methods; availability sampling with constraint that there is a SET quote for each category–> kind of like stratified random sampling.
