Work

Whose voices are counted? survey sampling and public preferences for energy decarbonization

Year: 2026
Type: article
Study design: Observational Study
Abstract: Public acceptance of technologies used to decarbonize the electric grid can shape efforts to reduce climate risks. However, estimates of public preferences may depend on how survey respondents are recruited, raising questions about whether different sampling approaches yield comparable results. We compare five survey sampling approaches—three designed to produce nationally representative samples and two convenience samples—to assess demographic representativeness and preferences for grid-decarbonization technologies. The convenience samples were highly nonrepresentative and produced skewed estimates of public support that were not corrected by weighting. Among the nationally representative approaches, opt-in online sampling with quotas most closely resembled 2022 American Community Survey demographics. The three nationally representative approaches produced broadly similar energy-preference profiles in this study, although some technology-specific differences remained. Address-based sampling and random-digit dialing yielded similar preference profiles, while the opt-in sample showed statistically higher support for on-land wind and lower support for offshore wind and solar farms. Across these samples, rooftop solar ranked first, followed by solar farms, while nuclear and fossil fuels with carbon sequestration ranked last; the random-digit-dial sample differed by ranking offshore wind last. These findings demonstrate tradeoffs among survey sampling approaches: convenience sampling presented substantial representativeness challenges not remedied by weighting, while the three nationally representative approaches produced broadly similar relative rankings in this study despite some differences in demographic composition and technology-specific preferences. Researchers weighing trade-offs in survey sampling approaches should consider how sampling method may affect demographic representation and its relation to the variables they wish to measure.
Language: English
Open Access status: hybrid
Estimated APC paid: $3,390
OpenAlex ID: W4407717240
FWCI:
Cites: 42
Cited by:
Related to: 10