Please enable JavaScript.
Coggle requires JavaScript to display documents.
Sampling Selecting the Right People to Represent the Bigger Picture -…
Sampling
Selecting the Right People to Represent the Bigger Picture
Why Sampling Matters
Population
The entire group of people, organizations, objects, or events I want to understand.
Sample
A smaller group selected from the population to participate in the study.
Unit of Analysis
The specific person, group, organization, or object being studied.
Statistical Inference
Using information from the sample to draw conclusions about the population.
Feasibility
Researchers use samples because studying an entire population is often too expensive, difficult, or time-consuming.
Representativeness
A strong sample reflects the important characteristics of the population.
Generalizability
When the sample is representative, the findings are more likely to apply to the larger population.
The Sampling Process
Define the Target Population
Clearly identify who or what the study is meant to describe.
My research example:
Underserved youth participating in mentoring and life skills programs.
Identify the Correct Unit of Analysis
The unit must match the research question.
My research example:
If I am studying youth development, the youth should be the primary unit of analysis.
Choose a Sampling Frame
The sampling frame is the accessible list or group from which the sample is actually selected.
My research example:
Enrollment lists from participating mentoring programs or schools.
Check the Sampling Frame
The sampling frame may leave out members of the target population.
My research example:
Using only one school could exclude youth from other neighborhoods or school systems.
Select a Sampling Technique
The researcher chooses between probability and nonprobability sampling.
Draw the Sample
Participants are selected from the sampling frame using the chosen method.
Evaluate Representation
The researcher must ask whether the final sample reflects the target population.
Probability Sampling
Every unit has a known, nonzero chance of being selected, and random selection is used at some point.
Simple Random Sampling
Every possible sample has an equal chance of being selected.
Strength:
Reduces researcher selection bias.
Example:
Randomly selecting 100 youth from a complete program enrollment list.
Systematic Sampling
The researcher selects every (k)th person after choosing a random starting point.
Sampling Ratio
The selection interval is calculated using: k=N/n
N represents the sampling frame size, and (n) represents the desired sample size.
Example:
If the list contains 500 youth and I need 100 participants, I would select every fifth person after a random start.
Stratified Sampling
The population is divided into nonoverlapping subgroups called strata. A random sample is then selected from each group.
My research example:
Divide participants by age group, gender, school level, or program location.
Proportional Stratified Sampling
Each subgroup appears in the sample in the same proportion that it appears in the population.
Preserves the population’s overall composition.
Nonproportional Stratified Sampling
Some smaller groups are intentionally selected in greater numbers.
Makes sure small or underrepresented groups have a meaningful voice.
Cluster Sampling
The population is divided into naturally occurring groups, often based on location. Some clusters are randomly selected.
My research example:
Randomly select several schools and study the participating youth within those schools.
Strength:
Makes geographically spread out populations easier and less expensive to study.
Limitation:
Selected clusters may differ from clusters that were not selected.
Matched Pairs Sampling
Participants from two groups are matched according to similar characteristics.
My research example:
Match youth in a mentoring program with youth of the same age, grade, and background who are not in the program.
Multistage Sampling
Two or more probability methods are used in stages.
My research example:
Select school districts, then schools, then grade levels, and finally students.
Probability Sampling and Generalizability
Probability sampling usually provides the strongest foundation for applying results to the larger population.
Nonprobability Sampling
Participants are selected without random selection. Some members of the population may have no chance of being included.
Convenience Sampling
Participants are selected because they are easy to reach or readily available.
My research example:
Surveying youth who attend one mentoring workshop.
Strength:
Quick, affordable, and useful for pilot studies.
Limitation:
The participants may not represent all underserved youth.
Proportional Quota Sampling
The number selected from each group reflects the population’s proportions.
Quota Sampling
The population is divided into groups, and participants are selected until a set number is reached for each group.
Expert Sampling
Participants are selected because they have specialized knowledge or experience.
My research example:
Interviewing mentoring program directors, school counselors, youth advocates, and life skills instructors.
Nonproportional Quota Sampling
A minimum number is selected from each group, even when the sample proportions do not match the population.
Allows smaller groups to be included in meaningful numbers.
Snowball Sampling
The researcher begins with a few qualified participants and asks them to recommend others.
Strength:
Helpful when the population is difficult to identify or reach.
Strength:
Helpful when the population is difficult to identify or reach.
Sampling Bias
Bias occurs when some members of the population are more likely to be included than others.
Limited Generalizability
Results from nonprobability samples usually cannot be confidently applied to the entire population.
Appropriate Uses
Nonprobability sampling can still be useful for:
Pilot studies
Exploratory research
Hard to reach populations
Specialized knowledge
Qualitative research
Studies without an available sampling frame
Statistics of Sampling
Response
A response is the measurement or answer provided by one sampled unit.
Frequency Distribution
Shows how often different responses appear in the sample.
Normal Distribution
A bell-shaped distribution in which most observations fall near the center, and fewer appear at the extremes.
Sample Statistic
A value calculated from sample data.
Examples:
Sample mean and sample standard deviation.
Population Parameter
The actual characteristic of the entire population.
A sample statistic is known from the collected data, while the population parameter usually remains unknown.
Sampling Error
The difference between a sample statistic and the true population parameter.
Sample Size
As sample size increases, sampling error will generally decrease if the sample is selected appropriately.
Sampling Distribution
The distribution created by calculating the same statistic across many different samples.
Standard Deviation
Measures how much individual responses vary within one sample.
Standard Error
Measures how much a sample statistic varies across multiple samples.
Confidence Interval
An estimated range in which the population parameter is expected to fall.
The 68, 95, 99 Percent Rule
(in normal distribution)
About 68 percent falls within one standard error
About 95 percent falls within two standard errors
About 99 percent falls within three standard errors
Precision
A smaller standard error produces a narrower confidence interval and a more precise estimate.