Chapter 5: Sampling Flashcards
Research Details and Design
What is sampling?
An examination of a portion of a larger group of people, places, or things in a research study
It influences the extent to which results can be generalized to other, different populations (external validity).
What is the population?
The population is the entire set o fpoeple, places or things to which you will generalize results. The population is made up of elements (individual units)
What is the study population?
The study population is the group from which you select your sample.
What is a sampling frame?
The sampling frame is the list of all elements that guides your sample selection from the sample population.
How do you determine an adequate sample size?
A common method used is to use statistical power calculations. Power calculations determine the sample size needed to identify treatment effects or attribute differences to study interventions or treatments with a particular level of confidence.
Power can be defined as the probability of rejecting the null hypothesis.
What is effect size?
Effect size is the quantification of differences between two or more groups.
Example: you can estimate a 10% difference in weight loss between a control group and an intervention group.
How many categories are there of sampling methods?
Two.
1) Probability sampling is used to estimate the extent to which the sample represents the larger population.
2) Non Probability sampling - a complete sample frame may not exist, so individuals may not have an equal and independent chance of being selected. You cannot estimate the effect of sampling error, which may reduce representativeness and generalizability to larger populations.
What are the four main types of probability sampling?
Simple Random Sampling - elements are selected similar to a coin toss and nothing but chance influences the selection
Systematic Random Sampling - a starting point is chosen, and then every Nth element is chosen thereafter.
Stratified Random Sampling - draw a sample to accurately reflect a population by some defined characteristics
Cluster Random Sampling -similar to stratified random sample, in that groups are the primary selection
What are the four main types of non-probability sampling?
Convenience sampling - availability sampling method takes advantage of easy-to-access people, places, or things to study.
Purposive sampling - based on some prior knowledge about the sample population. You set criteria for inclusion or exclusion
Quota sampling - setting proportional goals based on sample characteristics. This method is a way to demonstrate representativeness of those characteristics to the larger population.
Snowball sampling - snowball sample gets larger as the research project advances. You identify a small number of participants, and during data collection, and ask them to identify others who are relevant to your study.
What is sampling or selection bias?
The misrepresentations of some population characteristic due to the method of selecting the sample.
What is non response bias?
Nonresponse bias may occur when you do not real all members in your sample. Nonresponse bias may also be due to differences in characteristics of those who response and those who do not.
What is sampling error?
Sampling error is a result of sampling bias and is defined as the difference between the results from a sample population when compared to results from the whole population.
What should I consider when choosing a sample size?
How large to answer research question (Power & Effect size)
What size fits my research budget?
What sample size can I access?
What analysis will I be conducting with the data?
