Class 6 Sampling 2022 Sp
Lit reviews due now
Grading this afternoon and tomorrow
Final paper due 4/8
Grades posted Sunday after grading papers and peer reviews
To do: CITI Certificate
Next class on April 1, 2 weeks off
Generalizability, population, sample, strata defined
Probability vs. nonprobability sampling explained
Sampling involves studying a sample to infer about the population
Target population, accessible population, sample defined
Bias and error in sampling discussed
Ways to minimize sampling error
Inclusion and exclusion criteria in sample selection
Careful consideration of discriminatory language
Detailed criteria for breast cancer study
Inclusion and exclusion criteria listed
Subject recruitment methods discussed
Various ways to recruit subjects mentioned
Sampling methods: probability and nonprobability
Different sampling techniques listed
Simple random sampling explained
Every member has an equal chance of selection
Systematic random sampling described
Every nth person chosen from a list
Stratified random sampling defined
Ensures representation of important subgroups
Oversampling explained
Example of stratified sampling in Head Start program
Stratification based on state (rural and non-rural)
Sample represented 5% of Head Start sites and 3% of enrolled children
Select sub-groups (clusters) instead of individuals randomly.
Useful when the population consists of clusters or units (e.g., class).
Disadvantages include homogeneity within clusters requiring a large number of clusters.
Can employ stratified cluster sampling.
Falls under probability sampling if the number of clusters is known, otherwise under non-probability sampling.
Example: Candies divided into 4 clusters of 5 candies, randomly choose 2 clusters.
Used in large national surveys.
Involves drawing a stratified random sample, then selecting units within the stratification, and finally randomly choosing participants.
Total number of subjects is unknown.
Selects individuals believed to provide the best information.
Not random, similar to focus groups.
Useful for locating hard-to-find participants.
Participants refer the researcher to others.
Based on trust and confidentiality.
Subjects selected based on characteristics without randomization.
Sample may be stratified.
Example: Selecting the first 50 women who walk through a cafeteria.
Non-random with limited representation.
Often involves a captive audience like a class or team.
Subjects play a role in experiencing the intervention and providing information for observations and outcomes.
Larger samples reduce sampling error.
Guidelines for sample sizes in different types of studies.
Consider increasing sample size for greater variability between groups or small differences.
Email surveys may require a 40-50% increase in sample size.
A comparison of different sampling methods and their outcomes.
Understanding various
Lit reviews due now
Grading this afternoon and tomorrow
Final paper due 4/8
Grades posted Sunday after grading papers and peer reviews
To do: CITI Certificate
Next class on April 1, 2 weeks off
Generalizability, population, sample, strata defined
Probability vs. nonprobability sampling explained
Sampling involves studying a sample to infer about the population
Target population, accessible population, sample defined
Bias and error in sampling discussed
Ways to minimize sampling error
Inclusion and exclusion criteria in sample selection
Careful consideration of discriminatory language
Detailed criteria for breast cancer study
Inclusion and exclusion criteria listed
Subject recruitment methods discussed
Various ways to recruit subjects mentioned
Sampling methods: probability and nonprobability
Different sampling techniques listed
Simple random sampling explained
Every member has an equal chance of selection
Systematic random sampling described
Every nth person chosen from a list
Stratified random sampling defined
Ensures representation of important subgroups
Oversampling explained
Example of stratified sampling in Head Start program
Stratification based on state (rural and non-rural)
Sample represented 5% of Head Start sites and 3% of enrolled children
Select sub-groups (clusters) instead of individuals randomly.
Useful when the population consists of clusters or units (e.g., class).
Disadvantages include homogeneity within clusters requiring a large number of clusters.
Can employ stratified cluster sampling.
Falls under probability sampling if the number of clusters is known, otherwise under non-probability sampling.
Example: Candies divided into 4 clusters of 5 candies, randomly choose 2 clusters.
Used in large national surveys.
Involves drawing a stratified random sample, then selecting units within the stratification, and finally randomly choosing participants.
Total number of subjects is unknown.
Selects individuals believed to provide the best information.
Not random, similar to focus groups.
Useful for locating hard-to-find participants.
Participants refer the researcher to others.
Based on trust and confidentiality.
Subjects selected based on characteristics without randomization.
Sample may be stratified.
Example: Selecting the first 50 women who walk through a cafeteria.
Non-random with limited representation.
Often involves a captive audience like a class or team.
Subjects play a role in experiencing the intervention and providing information for observations and outcomes.
Larger samples reduce sampling error.
Guidelines for sample sizes in different types of studies.
Consider increasing sample size for greater variability between groups or small differences.
Email surveys may require a 40-50% increase in sample size.
A comparison of different sampling methods and their outcomes.
Understanding various
