More support for hydrogen export than its domestic application in Australia

The study addresses whether Australians are more supportive of hydrogen for export or domestic applications within the broader context of accelerating green hydrogen development for decarbonisation. Australia has significant renewable resources and aspirations to become a major hydrogen exporter, while also exploring domestic uses such as blending hydrogen into gas networks and decarbonising heavy industry and transport. Social acceptance is crucial alongside infrastructure and environmental factors. Prior work has documented general support for renewable energy and hydrogen but also community resistance to local infrastructure (NIMBY/NIABY). Communication and information provision can shape acceptance. However, no prior research has directly compared public support for hydrogen used domestically versus for export. This study fills that gap by testing how brief, mixed-media information affects support for hydrogen in both contexts.

Research on social acceptance of renewable energy highlights generally positive attitudes toward technologies like solar PV and hydrogen, tempered by local opposition to siting (NIMBY/NIABY). Acceptance ranges from passive (no resistance) to active (engagement/participation) and is influenced by perceived justice, environmental consciousness, and communication processes. Studies have examined cultural and regional factors in Europe and Asia, links to sustainability strategies, and Australian attitudes showing strong support for hydrogen. Message framing and information dissemination (e.g., videos, visualisation) can influence perceptions and acceptance. Yet, comparative analyses of acceptance for hydrogen’s domestic versus export applications have been lacking; this study directly examines that distinction within Australia.

Design: A nationally representative online survey of Australian adults with quota controls (age, gender, geography) was conducted 29 Jan–20 Feb 2021. Of 11,089 who commenced, 3,405 failed eligibility, 3,670 did not complete, and 943 failed consistency checks; 3,020 completed the survey. Sampling: Non-probabilistic with quotas to approximate representativeness. Intervention: A pre/post design with a brief information package (“SV”) combining a 1:42 min explanatory video (ARENA: “What is renewable 'green' hydrogen gas?”) plus images and text. Respondents had to pass a video information check to proceed. Randomization: Participants were randomly assigned to one of two streams: Domestic (information on domestic applications) or Export (information on export routes/destinations). All respondents completed the pre- and post-survey within the same session. Measures: Baseline (pre) and post support for hydrogen in Australia on a 1–7 Likert scale (1=very unsupportive; 7=very supportive). Additional items captured subjective knowledge, demographics, climate change beliefs, environmental identity, trust, attitudes, energy use (e.g., gas), and innovativeness. Estimation: Primary analysis used OLS to estimate the effect of SV on support using a before–after setup with respondent fixed effects to control for unobserved time-invariant individual confounders: S2H_st = α + β SV + δ_r + ε. Ordered logit models complemented OLS given the ordinal outcome. To compare streams, models included stream-specific treatment indicators: S2H_st = α + β_Export + β_Domestic + δ_r + ε, followed by tests of equality of coefficients. Balance tests (difference-in-differences style and propensity score matching diagnostics) confirmed comparability of Domestic and Export cohorts at baseline. Robustness checks and summary statistics are reported in the Supplementary Information.

• Information increases acceptance: The SV increased mean support by 0.537 points on the 1–7 scale (OLS; SE=0.031), about a 10% rise from the baseline mean of 5.31 (SD=1.25). Ordered logit indicated respondents were 2.32 times more likely to report a higher support category after SV. OLS with respondent fixed effects yielded a similar effect (0.537; SE=0.020) and higher R-squared than OLS without fixed effects.
• Baseline equivalence: Domestic and Export cohorts had statistically indistinguishable baseline support (means 5.31 vs 5.30; balance tests non-significant).
• Stream-specific effects: Domestic stream support increased from 5.31 to 5.85 (~9% increase). Odds of higher support after SV were 2.1x (no FE) and 3.77x (with FE). Export stream support increased from 5.30 to 5.90 (~11% increase). Odds were 2.57x (no FE) and 5.3x (with FE).
• Export favored over domestic: The change in support was significantly larger for Export than Domestic by 0.10 points (≈2% of baseline mean); p=0.0059.
• Heterogeneity (Table 3): Effects were positive across subgroups but sizes varied. Differences (Export−Domestic) and p-values: Yes Gas: 0.118 (p=0.025); No Gas: 0.099 (p=0.1065); Metro: 0.122 (p=0.0079); Regional: 0.070 (p=0.4076); Yes PV: 0.183 (p=0.0078); No PV: 0.084 (p=0.0851); Innovator: 0.016 (p=0.8010); Follower: 0.163 (p=0.0011). Respondents with gas connections, those with solar PV, metro residents, and fast followers showed significantly larger preference shifts toward export. Over 80% of variance in support remains unexplained (R-squared across subgroup models 0.095–0.255).

The study directly addresses whether Australians support hydrogen more for export or domestic use and shows that brief, mixed-media information can increase support overall, with a larger effect for export applications. This suggests framing hydrogen within export and global decarbonisation narratives resonates strongly. Heterogeneous effects indicate context matters: households with gas or solar PV, metro residents, and fast followers are especially receptive to export-oriented messaging, possibly reflecting preferences for diversified energy options, existing clean energy investments, perceived national economic benefits, and lower personal switching costs. Regional residents showed relatively greater sensitivity to domestic applications, aligning with employment and industrial transition considerations in regions where energy production is concentrated. These findings underscore the importance of tailored communication and framing to different audiences and use-cases, and they validate the role of concise, multimedia information in shaping acceptance of emerging energy technologies.

The paper contributes by (1) demonstrating that a short, mixed-media information package can significantly raise public support for hydrogen, (2) showing that Australians exhibit greater support for hydrogen export than for domestic applications, and (3) revealing that socio-economic and geographic contexts shape responsiveness to information. Policy-wise, effective communication strategies can bolster societal acceptance, and emphasizing hydrogen’s export role may garner broader support, while development of a viable domestic industry remains crucial to underpin long-term export potential. Future research should unpack the substantial unexplained variance in support, explore longer-term persistence of information effects, test alternative framings and more granular application contexts, and examine community-level responses where infrastructure siting may trigger local concerns.

• No separate control group: The pre/post design without a pure control arm may not fully rule out testing or response biases, though the short within-survey interval mitigates external influences.
• Ordinal outcome: The 1–7 Likert scale is ordinal; OLS was complemented by ordered logit to address this concern.
• Potential confounding: While respondent fixed effects control for time-invariant unobservables, residual confounding may remain. Not all beliefs/attitudes are fully observable.
• Sampling: Non-probabilistic online sampling with quotas approximates national representativeness but may limit generalisability.
• Short-term effects: Measured immediate, within-survey changes; durability of attitude shifts over time is unknown.
• Unexplained variance: Substantial unexplained variation (>80%) indicates other factors influencing acceptance were not captured.