General population sampling for environmental valuation studies
Environmental valuation studies typically target the general public - but 'general public' needs to be defined precisely to produce policy-relevant WTP estimates.
This article explains how to define and recruit a general population sample for environmental valuation DCEs, including geographic scope, demographic quotas, and online vs offline modes.
Knowledge Base -> Respondent Sampling -> Environment
Ben White, 07.07.2026
Defining 'general public' for environmental valuation
When an environmental valuation study targets 'the general public', it is implicitly making claims about who bears the costs and who enjoys the benefits of the environmental good being valued. A water quality improvement in a specific river catchment is relevant to people who live in that catchment, use that river for recreation, or care about it for existence reasons - not to people in a different part of the country.
Getting the geographic scope of the sample wrong produces WTP estimates that are either too high (if the relevant affected population is smaller than the sample) or too low (if the sample includes many people who have no connection to the environmental good).
Designing the sampling frame for environmental studies
The sampling frame should match the policy-relevant population - the people whose preferences count in the policy decision. For a national air quality policy, the UK adult population is the relevant frame. For a local green space improvement, residents within a 2km catchment are more appropriate.
Demographic representation is also critical. Online panels over-represent younger, more educated, and more digitally engaged respondents. For environmental WTP estimates used in policy cost-benefit analysis, the sample should be representative of the relevant population in terms of age, income, and education, because these variables are correlated with WTP.
SurveyEngine's sampling team manages quota-based recruitment to ensure demographic representativeness across age, gender, region, and socioeconomic group. For studies where representativeness is critical, we also offer post-stratification weighting.
Setting up general population recruitment with SurveyEngine
Step 1: Define the geographic scope of the relevant population. Identify the geographic catchment that defines the affected population. Use postcode, region, or country to define eligibility in the screener.
Step 2: Set demographic quotas based on census data. Use national or regional census data to set quotas for age group, gender, and socioeconomic group. SurveyEngine manages interlocking quotas across multiple demographic dimensions.
Step 3: Screen for connection to the environmental good. Some studies benefit from a proximity or connection screen - for example, confirming that respondents have visited or are aware of the environmental resource being valued. This is a design choice that should be pre-specified.
Step 4: Plan for online and offline modes if needed. Online panels provide fast, cost-effective general population samples but miss non-internet users. For studies where representativeness is critical, supplement with telephone or postal components for older and lower-income respondents.
Step 5: Apply post-stratification weights in analysis. If the achieved sample composition differs from the target population despite quotas, apply post-stratification weights in the model estimation. SurveyEngine exports data in formats compatible with weighted analysis in R and Stata.
Worked example - national coastal WTP study
A national study of WTP for improved coastal water quality targets a quota-representative sample of 1,000 UK adults aged 18+. Quotas are set for: age (18–34, 35–54, 55+), gender (50% male), region (proportional to UK population), and socioeconomic group (ABC1/C2DE). Online panel recruitment delivers 940 completions within quota; the remaining 60 are achieved through telephone recruitment of over-65s who are underrepresented in online panels.
Post-stratification weighting is applied to the final sample to correct for residual imbalances in age and education. The weighted WTP estimate is £48/household/year, 12% lower than the unweighted estimate - reflecting the lower WTP of the older and less-educated respondents who were underrepresented in the unweighted online sample.
References
Bateman, I.J. et al. (2002). Economic Valuation with Stated Preference Techniques. Edward Elgar.
Planning an environmental valuation study? Contact SurveyEngine for sampling strategy and recruitment.
Or Contact us at support@surveyengine.com — we're glad to help.