The energy crisis differentially impacted Swiss and German citizens’ energy literacy and efficiency preferences but not their support for climate policies

Russia’s invasion of Ukraine in 2022 triggered an energy crisis in Europe marked by supply shocks, price spikes, inflation, and heightened energy security concerns. Governments responded with emergency measures (e.g., price caps in Germany) and accelerated discussions on renewable energy and efficiency. While early cross-sectional evidence suggested stable or increased public support for renewables, support can be volatile as attention shifts to competing issues. Beyond overall support, it is unclear how citizens prioritize key policy attributes (e.g., energy independence, CO2 reductions, implementation timing, costs) over time. Concurrently, governments launched information campaigns to reduce demand and improve energy literacy, a known driver of conservation and energy-efficient purchasing, yet literacy is often low and may vary with numeracy. This study investigates how the crisis affected three climate-relevant domains across two critical time points (late 2022 and spring 2023) in Germany and Switzerland: (1) support for renewable energy policies and prioritization of their attributes, (2) energy literacy (accuracy of judgments of action energy savings), and (3) energy-efficient product choices. The study also tests how real-world, municipality-level electricity price changes in Switzerland moderated shifts in efficiency preferences.

Prior work indicates that public support for renewable energy policies was stable or even increased in mid-2022 after the invasion, though most evidence is cross-sectional and self-reported. Research on consumer behavior shows sensitivity to local energy prices in appliance purchases, with large price changes and high salience potentially intensifying attention to efficiency. Energy literacy is typically low but correlates with conservation behaviors and efficient purchases; simple interventions can improve literacy, and numeracy is positively associated with literacy. Policymaking during crises can alter the salience of energy independence and affordability, potentially reshaping how citizens weigh policy attributes. Gaps remain regarding longitudinal changes in policy support, attribute prioritization, literacy, and efficiency choices during the evolving crisis and under real price signals.

Design: Preregistered two-wave survey experiment in Germany and Switzerland with the same participants across waves: Wave 1 (Nov/Dec 2022) and Wave 2 (Mar/Apr 2023). N=1040 total (Germany n=552; Switzerland n=448 after minor exclusions). Samples were quota-based representative by gender, age, and income among adults 18–80. Main tasks each wave: (a) Renewable energy policy support via a 16-trial binary choice experiment. Each policy varied on four attributes: CO2 reduction potential (30% vs 15%), tax increase (6% vs 1%), implementation time (in 1 year vs in 7 years), and energy independence potential (+20% vs +10%). Participants accepted/rejected each package; attribute order randomized. (b) Energy literacy ("action energy literacy"): participants estimated annual kWh savings for 8 household actions (6 curtailment, 2 efficiency). Estimates were provided within a kWh range; actual values sourced from literature. Literacy index computed as mean absolute error using log10-transformed estimates vs actuals (higher scores = greater error = lower literacy; Cronbach’s alpha 0.83/0.82 in Waves 1/2). (c) Energy-efficient product choices: seven binary trade-off choices between two washing machines varying in purchase price (three levels) and annual kWh (46/58/70), constructed to avoid dominant options and to create high vs low efficiency differences; order of price vs efficiency information randomized. Natural experiment (Switzerland): Experimental choice data were merged with municipality-level electricity price changes (2023–2022 ct/kWh) from ElCom using postal codes. The Swiss non-liberalized market (fixed annual tariffs per municipality) enabled assessing how real-world price changes moderated within-person changes in efficiency preferences. Attention and context measures: Media/search attention trends tracked; self-reported emotions (fear, anger, hope, interest) about the crisis, energy expense concerns, and attention were assessed. Individual differences: Climate change concern (4 items), numeracy (Berlin Numeracy Test 1-item), political orientation, trust in government, demographics. Analyses: Mixed-effects logistic regressions for policy support (random slopes for attributes) and product choices (random intercepts by participant). Linear/mixed models for energy literacy (log-based errors). Key preregistered tests included stability of overall support and attribute weights across waves, effects of climate concern, changes in literacy (including frequency-of-use and confidence), relationship of literacy to product choices, and moderation of efficiency preferences by Swiss price changes. Deviations from preregistration (e.g., attention-check based exclusions only) and robustness checks are documented in Supplementary Tables.

Context shift: From Wave 1 to Wave 2, public/media attention to the energy crisis dropped substantially (Swiss media mentions of “energy crisis” down 71.6%). Emotions and concerns decreased: fear t(1039)=2.41, p=0.016; anger t(1039)=3.29, p=0.001; interest t(1039)=2.79, p=0.005; energy expense concerns t(1039)=2.21, p=0.028. Meanwhile, electricity prices increased in both countries. Policy support: All four attributes significantly affected support in both waves: taxes (OR=0.07, 95% CI [0.06; 0.08], p<0.001) had the strongest negative effect; longer implementation time (OR=0.29, [0.26; 0.34], p<0.001) reduced support; higher CO2 reduction (OR=1.46, [1.36; 1.56], p<0.001) and greater energy independence (OR=1.34, [1.25; 1.43], p<0.001) increased support. Overall support was stable across waves (OR=0.97, [0.92; 1.03], p=0.387), and the importance of attributes, including energy independence (OR=0.94, [0.83; 1.06], p=0.338), remained stable. Climate change concern predicted higher policy support (OR=2.02, [1.85; 2.20], p<0.001) and greater weight on CO2 reductions (b=0.11, [0.09; 0.12], p<0.001). Individuals with high concern particularly preferred immediate over delayed policies (interaction with implementation time OR=0.49, [0.44; 0.55], p<0.001). Effects were robust to demographic controls. Energy literacy (judgments): Participants systematically overestimated low-impact actions and underestimated high-impact actions in both waves. Low-impact actions were overestimated by 250% ([246%; 256%]) in Wave 1 and 259% ([255%; 264%]) in Wave 2; high-impact actions were underestimated by 29% ([26%; 31%]) and 23% ([20%; 25%]) respectively. Action impact strongly predicted bias (b=0.78, [0.74; 0.82], p<0.001). No overall improvement in judgment accuracy between waves (t(979)=0.64, 95% CI [-0.02; 0.05], p=0.523). Exploratory: Changes over time depended on numeracy (b=0.05, [0.00; 0.09], p=0.030); significant in Switzerland (b=0.09, [0.03; 0.15], p=0.004) but not Germany. In Switzerland, the low-numeracy group’s literacy declined (b=0.04, SE=0.02, p=0.008), while high-numeracy remained roughly stable (b=-0.05, SE=0.03, p=0.077). Appliance use frequency was unrelated to error (b=0.00, [−0.03; 0.03], p=0.873). Confidence showed a U-shaped relation to error (b=0.02, [0.01; 0.03], p<0.001). Product choices: Greater accuracy (higher literacy; i.e., lower error) predicted more frequent selection of the energy-efficient product (per unit increase in error: OR=0.84, [0.77; 0.92], p<0.001). Literacy interacted with efficiency differences (OR=1.41, [1.28; 1.55], p<0.001): those with better literacy were especially responsive when efficiency differences were large (e.g., 78% vs 42% efficient choices at high vs moderate kWh differences), whereas low-literacy participants were less sensitive (72% vs 50%). Climate change concern increased efficient choices (OR=1.51, [1.35; 1.68], p<0.001), with stronger effects for large efficiency differences. Temporal change in efficiency preferences and price moderation: Overall, energy-efficient choices declined from Wave 1 to Wave 2 (OR=0.79, [0.73; 0.86], p<0.001) in both countries, coinciding with reduced crisis salience. In Switzerland, real-world electricity price increases moderated this decline (χ2(1, N=409)=4.938, p=0.026): participants facing the largest price increases showed no significant decrease (slope at +1 SD price change: b=−0.18, SE=0.10, p=0.067), while those with modest increases showed a significant decline (slope at −1 SD: b=−0.29, SE=0.09, p=0.002). Results held controlling for 2022 price levels (χ2(1, N=409)=8.053, p=0.005).

The study shows that despite substantial shifts in public attention and emotions toward the energy crisis, overall support for renewable energy policies and the weighting of key policy attributes (costs, speed, CO2 reduction, energy independence) remained stable across winter 2022/23. Climate concern consistently bolstered support and preference for rapid, impactful policies, underscoring the importance of concern in shaping policy judgments. In contrast, consumer preferences for energy efficiency in product choices diminished as crisis salience waned, aligning with psychological evidence that attention and availability of information affect decision-making. Yet, economic rationality emerged: individuals exposed to strong real-world electricity price increases maintained efficiency preferences, indicating that price signals continued to guide choices even as salience fell. Action energy literacy was a key predictor of efficient purchases; however, literacy did not improve overall and gaps by numeracy persisted, widening in Switzerland where national campaigns were highly visible. Together, findings suggest that climate-relevant decisions depend on the interplay of contextual conditions (salience, price signals) and individual differences (concern, numeracy/literacy). Policy levers can act on both: sustaining salience through targeted campaigns and preserving or strategically targeting price signals can foster efficient choices without reducing support for renewables. Communications that accommodate lower numeracy—by using clearer, comparative, and absolute metrics relevant to household consumption—may reduce misperceptions and improve decision quality.

This two-wave cross-national natural experiment links experimental measures of policy support, energy literacy, and product choices with real-world price data, revealing robust stability in renewable policy support amidst shifting attention, but a decline in efficiency preferences unless reinforced by strong energy price signals. Energy literacy substantially shapes efficient choices, yet overall literacy did not improve and disparities by numeracy widened in Switzerland. Policymakers should consider maintaining the salience of efficiency (e.g., sustained, numeracy-sensitive campaigns) and designing targeted price relief that protects vulnerable households while preserving incentives for efficiency among others. Future research should: (1) extend to multi-wave longitudinal designs across more countries and policy contexts; (2) integrate survey experiments with actual retail purchase data; (3) test communication formats optimized for varying numeracy levels; and (4) explore distributional effects of pricing policies on efficiency behavior and policy support.

- Temporal scope: Only two assessment waves limit causal inference and may miss dynamic, non-linear changes. - External influences: Unmeasured factors (e.g., war-related anxiety, inflation, seasonal effects, concurrent policies) could confound observed changes. - Generalizability: Findings from Germany and Switzerland (with differing market structures and policy responses) may not generalize to other countries with distinct cultural, political, or market conditions. - Measurement constraints: Energy literacy reflects knowledge, an imperfect predictor of behavior; product choices were experimental rather than actual purchases. - Sample and exclusions: Deviations from preregistered exclusion criteria to preserve power; some participants excluded for inattentive responding in literacy tasks; attrition between waves. - Policy context: German price caps may have attenuated price signals broadly; design and timing of national campaigns differed across countries, affecting comparability.