The Surprising Power of Extreme Case Sampling
Location
2073
Format Type
Paper
Format Type
Paper
Start Date
14-1-2017 2:10 PM
End Date
14-1-2017 2:30 PM
Abstract
Extreme case sampling is often a very effective way to determine if relationships exist variables in large data sets. Extreme case sampling is an exploratory qualitative research method that allows content experts to discover relationships in the data without the restrictions of having a priori assumptions of what kinds of relationships exist. In this presentation examples of the successful use of extreme case sampling will be given. Seven characteristics of a study that make extreme case sampling an ideal research method are (1) the variables have a linear or non-linear continuously increasing or decreasing relationship to each other, (2) potential predictor variables are well-defined but potential outcome variables are unclear, difficult to measure, or difficult to assign values to, (3) a "double sampling" procedure can be performed where small number of cases from a representative sample of the population can be obtained, (4) scores for a predictor variable or composite of predictor variables can be assigned to all of sample subjects, (5) there are one or more theories or models that can be evaluated by a person or persons with content expertise, (6) these theories or models have not already been confirmed with more definitive quantitative studies, and (7) the possible relationships between predictor variables and outcome variables may be complex and/or there are a large number of potential predictor variables. Even if only a few of these seven criteria are met, extreme case sampling can be the preferred research method.
The Surprising Power of Extreme Case Sampling
2073
Extreme case sampling is often a very effective way to determine if relationships exist variables in large data sets. Extreme case sampling is an exploratory qualitative research method that allows content experts to discover relationships in the data without the restrictions of having a priori assumptions of what kinds of relationships exist. In this presentation examples of the successful use of extreme case sampling will be given. Seven characteristics of a study that make extreme case sampling an ideal research method are (1) the variables have a linear or non-linear continuously increasing or decreasing relationship to each other, (2) potential predictor variables are well-defined but potential outcome variables are unclear, difficult to measure, or difficult to assign values to, (3) a "double sampling" procedure can be performed where small number of cases from a representative sample of the population can be obtained, (4) scores for a predictor variable or composite of predictor variables can be assigned to all of sample subjects, (5) there are one or more theories or models that can be evaluated by a person or persons with content expertise, (6) these theories or models have not already been confirmed with more definitive quantitative studies, and (7) the possible relationships between predictor variables and outcome variables may be complex and/or there are a large number of potential predictor variables. Even if only a few of these seven criteria are met, extreme case sampling can be the preferred research method.
Comments
Breakout Session H