Jittered Sampling
- A sampling approach that slightly perturbs (jitters) sampled locations to reduce clustering.
- Helps mitigate spatial autocorrelation, which can cause biased results and incorrect conclusions.
- Improves representativeness and reliability of spatial samples.
Definition
Section titled “Definition”Jittered sampling is a method of sampling that is used to reduce the effects of spatial autocorrelation in the data and is commonly used in spatial data analysis.
Explanation
Section titled “Explanation”Jittered sampling works by introducing small random displacements to the locations of selected sampling units. By avoiding perfectly spaced or evenly distributed sampling units, jittered sampling reduces the chance that samples will cluster spatially. Reducing clustering mitigates spatial autocorrelation, which otherwise can lead to biased results and incorrect conclusions. This increases the representativeness of the sample and improves accuracy and reliability in analyses of spatial patterns and relationships.
Examples
Section titled “Examples”Plant distribution in a forest
Section titled “Plant distribution in a forest”A researcher randomly selects a number of plots within a forest and then jitters the locations of the plots slightly to reduce the chance of clustering. The plots are not perfectly spaced and are not evenly distributed, which can help to reduce the effects of spatial autocorrelation when measuring abundance and distribution of plant species.
Crime rates in a city
Section titled “Crime rates in a city”A researcher randomly selects a number of neighborhoods and then jitters the locations of the neighborhoods slightly to reduce the chance of clustering. The neighborhoods are not perfectly spaced and are not evenly distributed, which can help to reduce the effects of spatial autocorrelation when measuring incidence of different types of crime.
Use cases
Section titled “Use cases”- Spatial data analysis where spatial autocorrelation may bias results.
- Studies of spatial patterns and relationships in observational data.
Notes or pitfalls
Section titled “Notes or pitfalls”- Spatial autocorrelation can lead to biased results and incorrect conclusions if not addressed.
- Jittered sampling reduces clustering of sampling units; failing to reduce clustering can decrease sample representativeness and reliability.
Related terms
Section titled “Related terms”- Spatial autocorrelation
- Sampling
- Spatial data analysis
- Clustering