Proportional Weighted Sampling
Also known as: proportional probability weighting, proportional weight sampling, probability proportional to size sampling, PPS sampling
Proportional weighted sampling is a probability-based survey design in which each subgroup (stratum or cluster) of the population is sampled and weighted in proportion to its true size in the population. By assigning sampling weights that mirror the actual composition of the population, the method ensures unbiased estimates without the need for post-hoc reweighting, and produces efficient estimates when variance within subgroups is relatively homogeneous.
Key highlights
- Produces self-weighting samples when strata are proportionally allocated, simplifying analysis.
- Unbiasedly represents the population composition without requiring post-hoc weight adjustment.
- Reduces variance relative to simple random sampling when the stratifying variable is correlated with the outcome.
- Transparent and straightforward to explain to funders, ethics boards, and non-specialist audiences.
- Compatible with all major survey analysis software and established variance estimation methods.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use proportional weighted sampling when the population has identifiable subgroups of known size, you want unbiased population-level estimates without post-stratification adjustment, and subgroup variances are broadly similar. It is particularly suitable for large-scale descriptive surveys — national health surveys, educational assessments, and labour force studies — where representativeness is paramount. Do not use it when subgroup variances differ markedly (disproportionate stratified sampling is more efficient in that case), when subgroup sizes are unknown, when the research goal is qualitative or exploratory rather than estimation, or when certain rare subgroups need oversampling to yield adequate precision.
Strengths & limitations
- Produces self-weighting samples when strata are proportionally allocated, simplifying analysis.
- Unbiasedly represents the population composition without requiring post-hoc weight adjustment.
- Reduces variance relative to simple random sampling when the stratifying variable is correlated with the outcome.
- Transparent and straightforward to explain to funders, ethics boards, and non-specialist audiences.
- Compatible with all major survey analysis software and established variance estimation methods.
- Requires a complete and accurate sampling frame with known subgroup sizes — often unavailable in practice.
- Less efficient than disproportionate stratified sampling when subgroup variances differ substantially.
- Small subgroups receive small absolute samples, leaving domain-specific estimates imprecise even if population-level estimates are sound.
- Proportional allocation does not guarantee adequate statistical power for subgroup comparisons.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
What is the difference between proportional weighted sampling and simple stratified sampling?
In proportional stratified sampling, the sample size in each stratum mirrors the stratum's share of the population (n_h = n × N_h/N). When design weights are explicitly calculated and applied in analysis — particularly in multi-stage or unequal-probability designs — the method is often called proportional weighted sampling. The two terms largely describe the same proportional-allocation logic; the 'weighted' label emphasises that analysis must account for the sampling design through explicit probability weights.
When should I use disproportionate allocation instead?
Use disproportionate (unequal) allocation when stratum variances differ substantially — sampling more from high-variance strata increases overall precision — or when you need adequate sample size within a small but important subgroup for separate analysis. You then compensate in analysis by applying design weights so that population-level estimates remain unbiased.
How do I handle subgroups whose actual size I don't know precisely?
If exact subgroup sizes are unavailable, you can use estimates from a recent census, administrative records, or a pre-survey screening study. Alternatively, post-stratification weighting adjusts a non-proportional sample after data collection using population benchmarks. If no reliable size estimate exists, proportional weighted sampling is not feasible and a different design should be considered.
Does proportional sampling guarantee representativeness?
It guarantees representativeness on the stratifying variables used to define subgroups. Units within each stratum are still selected randomly, so the sample is unbiased. However, characteristics not related to the strata are subject to ordinary sampling variability, and non-response can introduce bias regardless of the sampling design.
Can I use proportional weighted sampling in qualitative research?
The concept of proportional representation can inform purposive sampling in qualitative studies — for example, selecting interview participants in rough proportion to group sizes to ensure diversity. However, formal probability weights and design-based inference are meaningful only in quantitative probability sampling, where every unit has a known, nonzero selection probability.
Sources
- 1.Cochran, W. G. (1977). Sampling Techniques (3rd ed.). John Wiley & Sons.ISBN 978-0471162407
- 2.Kish, L. (1965). Survey Sampling. John Wiley & Sons.ISBN 978-0471489009
You have read it. What now?
Cite this page
ScholarGate. (2026, June 3). Proportional Weighted Sampling. ScholarGate. https://scholargate.app/survey-methodology/proportional-weighted-sampling