week 6
MC17
Tappen2eChapter8.pptxChapter 8
Sampling
Sampling
Sampling involves decisions about who or what will be tested, observed, or interviewed in your study (Morse, 2007)
Key questions to address:
Who should and should not be included?
How many should be included?
Probability
Probability is the likelihood that an event or a condition will occur
You can express probability in terms of the chance the event will occur or in percentages
Levels of Significance
Levels of significance are the difference that will be accepted as too large to be attributed to chance
These levels are set by the researcher at the outset of a study
Probability Samples
Probability samples are formed to ensure that each subject has an equal chance of being included so an unbiased sample can be used
Probability Samples
A sampling design explains how the subjects are chosen and should include:
Number of subjects
How they will be assessed, screened, and selected
Inclusion and exclusion criteria
Probability Samples
Random selection is accomplished by having:
Identification of all possible participants
Every potential participant is given an equal chance of being selected
Probability Samples
Variations of random sampling include:
Stratified: randomly select from each stratum
Cluster: sample groups rather than individuals
Multistage: sample from multiple sets of clusters
Nonprobability Sampling
Reasons why researchers use nonprobability samples are:
Limited resources for developing an accurate sampling frame or purchase lists of potential subjects
Information needed to identify all potential subjects is not available
Nonprobability Sampling
Reasons why researchers use nonprobability samples are:
Limited number of subjects
Subjects are difficult to find or difficult to persuade to participate in study
Subjects do not complete study
Experimental mortality
Nonprobability Sampling
Types of nonprobability samples include:
Quota sampling: select a specified number of participants from each group
Convenience sampling: enroll those who are available
Snowball network or referral sampling: begin with known individuals and ask them to refer others who meet selection criteria
Tracking and Reporting Sample Development
In order to improve the reporting of randomized controlled trials (RCTs), the Consolidated Standards of Reporting Trials (CONSORT) were developed
A flow diagram that can be used for tracking sample development
CONSORT Flow Diagram
Source: Altman, D.G., Schulz, K.F., Moher, D., Egger, M.. Davidoff, F., Elbourne, D., Gøtzsche, P.C., & Lang, T. (2001). The revised CONSORT statement for reporting randomized trials: Explanation and elaboration. Annuals of Internal Medicine; 134(8), 663-694.
Example of Flowchart
Source: Buchbinder, R., Osborne, R.H., Ebeling, P. R., Wark, J.D., Mitchell, P.M., Wriedt, C., Graves, S.D., Staples, M.P., & Murphy, B. (2009). A randomized trial of vertebroplasty for painful osteoporotic vertebral factures. The New England Journal of Medicine, 361(6), 557-568.
Types of Errors in Quantitative Research
A type I error occurs when a null hypothesis that is true is rejected
A type II is when we fail to reject a false null hypothesis
Power Analysis Using Effect Size
The power of a statistical test is the probability that it will yield a statistically significant result
An underpowered study is when the sample is too small and leads to a type II error
An overpowered study is when the sample is too large and leads to a type I error
16
Power Analysis Using Effect Size
Effect size is the estimated magnitude of the phenomenon under study
An effect size calculation indicates the strength of the relationship between the independent and dependent variables
The equation is: d =( XC – X1) / SD pooled
17
Equivalence of Effect Size to Correlation Coefficient
| d | r | |
| Small | .20 | .10 |
| Medium | .50 | .30 |
| Large | .80 | .50 |
Power Analysis Using Effect Size
If the null hypothesis is not true, then the effect size will be greater than zero
The larger the effect size, the greater the degree to which the phenomenon is shown
The larger the effect size is, the greater the power will be so a smaller sample is needed
Purposeful Sampling
The sampling done for qualitative studies is called purposeful because it is directed by the purpose of the study, not by statistical calculations
It is also called purposive, judgmental, or theoretical sampling
20
Types of Sampling Used in Qualitative Designs
Case study: the case or cases selected are unusual in some way
Ethnography: select a culture, subculture, or ethnic group of interest; begin with “big net” and then narrow it down
21
Types of Sampling Used in Qualitative Designs
Phenomenology: select subjects who experienced the phenomenon under study
Grounded theory: selection of subjects and sources of data based on their ability to contribute to the evolving theory
22
Sampling Strategies
Extreme cases
Intense experiences
Maximum variation sampling
Negative instances or confirming and disconfirming cases
23
Sampling Strategies
Homogeneous sampling
Criterion sampling
Stratified purposeful sampling
24
Evolving or Iterative Sampling
Reasons why the sampling strategy may be altered during the study:
Saturation
Scope
Variation
Verification
25
Purposive Sample Size
Single or multiple cases
8 to 12 participants for focus groups
Theoretical saturation
Redundancy
26
