Skip to contentScholarGate
LibraryBookshelfDeskReview StudioAssistant
Sign in
On this page
IntuitionHow it worksWhen to use itStrengths & limitationsCommon pitfallsApplicationsFrequently asked🔒 Read the full methodSourcesRelated methods
Cite this pageSpotted an issue on this page? Report or suggest a fix →
Home›Survey Methodology›Adaptive Cluster Sampling
Process / pipelineSampling design

Adaptive Cluster Sampling

Also known as: Adaptive Cluster Sampling, Sequential Adaptive Sampling, Network Sampling, Adaptif Küme Örneklemesi

Adaptive Cluster Sampling (ACS) is a probability-based survey design introduced by Steven K. Thompson in 1990 for estimating the abundance or total of rare, clustered populations. Starting from an initial random sample, the design adaptively adds neighboring units whenever a sampled unit satisfies a predefined condition—such as exceeding a count threshold—thereby concentrating sampling effort exactly where the population of interest occurs. It is most appropriate for ecologists, epidemiologists, and social scientists studying geographically or socially clustered rare phenomena.

ScholarGate
  1. Process / pipeline
  2. v1
  3. 1 Sources
  4. PUBLISHED
Cite this page →
Tools & resources
Download slides
Learn & explore

Read the full method

Members only

Sign in with a free account to read this section.

Sign in

Method map

The neighbourhood of related methods — select a node to explore.

Adaptive Sampling
Capture-RecaptureRespondent-Driven Sampli…Stratified SamplingAdaptive Cross-Sectional…

When to use it

Use Adaptive Cluster Sampling when the target population is rare relative to the sampling frame, highly clustered in space or social networks, and costly to visit units that yield zero observations. It requires a well-defined neighborhood structure and a meaningful threshold condition C. The design is less efficient than simple random sampling when the population is not clustered; if density is moderate and spread is uniform, stratified or systematic sampling may outperform it. The estimator assumes all networks can be fully traced, so open or unbounded networks can pose practical difficulties.

Strengths & limitations

Strengths
  • Concentrates sampling effort where the rare population occurs, improving precision for clustered rare populations compared with simple random sampling
  • Horvitz-Thompson estimator is provably unbiased regardless of how many adaptive additions are made
  • Probability-based design preserves inferential validity without subjective judgment about which clusters to expand
  • Flexible condition C allows the design to be tailored to ecological, epidemiological, or social science thresholds
Limitations
  • Final sample size is random and unpredictable, complicating logistical planning and budget management
  • When the target population is not clustered, adaptive expansion rarely triggers and the design offers no gain over simple random sampling
  • Accurate inclusion probability calculations require knowing the full network size, which may be difficult in real field settings
  • Edge units that do not satisfy C must still be visited and measured, increasing per-unit fieldwork costs at cluster boundaries

Frequently asked

Is the Adaptive Cluster Sampling estimator always unbiased?

Yes, provided inclusion probabilities αₖ are computed correctly based on the combinatorial probability that each network is intersected by the initial sample. Thompson (1990) proved that the Horvitz-Thompson estimator is unbiased under this design. Bias can arise in practice only if networks are mis-identified—for example, by stopping expansion prematurely or misassigning edge units.

How do I choose the condition C and neighborhood structure?

The threshold condition C should reflect a practically meaningful level of the study variable—often a non-zero count or a minimum abundance value. The neighborhood is typically defined by spatial adjacency (e.g., four-connected grid cells) but can reflect social ties in network surveys. Both choices should be justified by subject-matter knowledge before data collection, as post-hoc changes alter the design's probability structure.

When does Adaptive Cluster Sampling perform worse than simple random sampling?

When the population is spatially random rather than clustered, satisfying condition C in the initial sample is rare, so adaptive expansion seldom triggers. In this case ACS yields a sample close in size to the initial sample but with a more complex estimation procedure, offering no efficiency gain. A pilot survey or prior ecological knowledge should confirm clustering before committing to ACS.

Sources

  1. Thompson, S. K. (1990). Adaptive cluster sampling. Journal of the American Statistical Association, 85(412), 1050–1059. DOI: 10.1080/01621459.1990.10474975 ↗

How to cite this page

ScholarGate. (2026, June 2). Adaptive Cluster Sampling. ScholarGate. https://scholargate.app/en/survey-methodology/adaptive-sampling

Related methods

Capture-RecaptureRespondent-Driven SamplingStratified Sampling

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

  • Capture-RecaptureSurvey Methodology↔ compare
  • Respondent-Driven SamplingSurvey Methodology↔ compare
  • Stratified SamplingSurvey Methodology↔ compare
Compare side by side →

Referenced by

Adaptive Cross-Sectional Epidemiological Study

Similar methods

Adaptive Cluster SamplingAdaptive Simple Random SamplingAdaptive Stratified SamplingAdaptive Multistage SamplingAdaptive Weighted SamplingAdaptive Cross-Sectional Epidemiological StudyAdaptive Quota SamplingAdaptive Snowball Sampling

Related reference concepts

Survey Methods • Sampling MethodsSamplingGibbs SamplingImportance SamplingMarkov Chain Monte CarloCluster Analysis

Spotted an issue on this page? Report or suggest a fix →

ScholarGate — Adaptive Sampling (Adaptive Cluster Sampling). Retrieved 2026-07-21 from https://scholargate.app/en/survey-methodology/adaptive-sampling · Dataset: https://doi.org/10.5281/zenodo.20539026
Quick facts
Originator
Steven Thompson
Year
1990
Type
Probability-based adaptive design
Subfamily
Sampling design
Estimator
Horvitz-Thompson unbiased estimator
Trigger
Condition C on observed unit value
Related methods
Capture-RecaptureRespondent-Driven SamplingStratified Sampling
ScholarGate

A content-first reference library for research methods — what each one is, how it works, and where it comes from.

Open data (CC-BY)

Explore

  • Library
  • Search the library…
  • Browse by field
  • Fields
  • Journey
  • Compare
  • Which method?

Reference

  • Subjects
  • Atlas
  • Glossary
  • Methodology
  • Philosophy

Your tools

  • Bookshelf
  • Desk
  • Chat

Company

  • About
  • Pricing
  • Contact
  • Suggest a method

Entries are compiled from published sources for reference. Verifying the accuracy and suitability of any information for your own use remains your responsibility.

© 2026 ScholarGate · A research-method reference library
  • Privacy
  • Cookies
  • Terms
  • Delete account