Stratified Sampling
Stratified and Cluster Sampling Designs · Also known as: Proportional Stratified Sampling, Optimal Allocation Sampling, Stratum-Based Sampling, Tabakalı Örnekleme
Stratified sampling is a probability sampling design in which the target population is partitioned into non-overlapping, exhaustive subgroups called strata, and independent probability samples are drawn within each stratum. Formalized by William G. Cochran in Sampling Techniques (1977), the method exploits known population structure to reduce variance and guarantee representativeness of all major subgroups, making it a cornerstone of large-scale survey research and official statistics.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
+34 more
When to use it
Use stratified sampling when the population contains distinct subgroups with differing variances on the outcome of interest, when separate stratum-level estimates are required, or when certain small but important subgroups need guaranteed representation. Assumptions include a complete and accurate sampling frame, known or reliably estimated stratum sizes, and the ability to assign every unit to exactly one stratum prior to selection. It performs poorly when the stratifying variable is unrelated to the outcome or when stratum boundaries are defined post hoc.
Strengths & limitations
- Reduces sampling variance compared to simple random sampling by eliminating between-stratum variability from the error term.
- Guarantees representation of every defined subgroup, enabling reliable domain-level estimates.
- Allows flexible allocation strategies (proportional, optimal, equal) to balance precision and cost across strata.
- Design-unbiased estimator with well-understood variance properties facilitates honest uncertainty quantification.
- Requires a complete sampling frame with accurate stratum membership for all population units before sampling begins.
- Gains in precision are negligible if the stratifying variable is weakly correlated with the study outcome.
- Defining a large number of strata can fragment the sample, leading to unstable within-stratum estimates.
- Optimal allocation requires advance knowledge of within-stratum standard deviations, which may only be available from prior surveys or pilot studies.
Frequently asked
How does stratified sampling differ from cluster sampling?
In stratified sampling every stratum is sampled and the strata are internally homogeneous; in cluster sampling only a random subset of clusters is selected and clusters are internally heterogeneous. Stratified sampling typically produces lower variance, while cluster sampling reduces logistical costs by concentrating fieldwork in selected areas or groups.
When should I use proportional versus Neyman optimal allocation?
Proportional allocation is appropriate when stratum variances are roughly equal or unknown and when a single self-weighting estimate is desired. Neyman optimal allocation is preferred when stratum variances differ substantially and prior variance estimates are available, as it minimizes the overall estimator variance for a fixed total sample size.
Can stratified sampling be applied when some stratum sizes are unknown?
Stratum sizes N_h must be known to compute the weighted estimator. If exact counts are unavailable, researchers use auxiliary administrative data, census figures, or model-based estimates. When N_h are only approximately known, post-stratification can adjust estimates after data collection, though it introduces additional variance that must be accounted for.
Sources
- Cochran, W. G. (1977). Sampling Techniques (3rd ed.). Wiley. ISBN: 978-0-471-16240-7
How to cite this page
ScholarGate. (2026, June 2). Stratified and Cluster Sampling Designs. ScholarGate. https://scholargate.app/en/survey-methodology/stratified-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.
- Small Area EstimationSurvey Methodology↔ compare
- Survey WeightingSurvey Methodology↔ compare