Citizens can care about environmental protection yet disagree about whether the federal government should spend more money on it. This project studies a political judgment that may help explain that difference: whether people believe elections actually make government listen.
This is an adjusted difference between categories, not an invented absolute level of support. The association does not extend to the paper’s separate environmental-regulation item.
Beyond the abstract: the argument, study design, evidence, and limits of inference.
Perceived electoral responsiveness concerns whether elections make government listen to citizens. It is distinct from the respondent’s own sense of personal political influence. A preference for increased federal spending also differs from an assessment of environmental risks, because spending requires entrusting collective resources to public institutions.
The analysis combines American National Election Studies cross-sections from 2008 through 2024, with 19,829 respondents in the principal specification. Adjustments cover party identification, ideology, two government-trust items, demographics, and survey year. Respondents at the highest rather than lowest responsiveness category show about 4.6 percentage points more adjusted support for increased federal environmental spending. This is a probability contrast, not a causal treatment effect.
A joint comparison with personal external efficacy shows different conditional associations (Wald p = .042), but the distinction is modest and calls for replication. A separate 2020/2024 question about regulating business does not reproduce the result. That failure to generalize is substantive: endorsement of a spending program should not be equated with support for every environmental policy instrument.
Research basis: When Elections Matter unpublished ANES manuscript, 2008–2024.
These are standardized predicted probabilities from the party-adjusted ANES model, not raw percentages.
| Belief that elections make government listen | Predicted support for increased spending | Difference |
|---|---|---|
| Not much | 50.3% | Reference |
| A good deal | 54.9% | +4.65 percentage points |
The contrast is +4.65 points (95% observation-robust CI, +2.46 to +6.84); N = 19,829. The interval does not reproduce the unavailable full complex ANES design.
The ordered-logit coefficient is −0.260 (SE 0.063, p < .001) after party adjustment. The outcome is coded so higher categories indicate more opposition.
The separate external-efficacy coefficient is −0.045 (SE 0.071, p = .526) in a joint model; the coefficient contrast has Wald p = .042, which warrants cautious interpretation.
For environmental regulation in 2020 and 2024, electoral responsiveness has coefficient +0.071 (SE 0.088, p = .420). The spending result does not generalize automatically to other policy instruments.
When Elections Matter, current research manuscript; Table 2, Figure 1 and regulation comparison. Unpublished.
These are observational survey associations, not experimental proof that improving confidence in elections would change spending preferences. Different kinds of environmental policy cannot be assumed to share the same determinants. Manuscript in development.
Unpublished American National Election Studies manuscript, 2008–2024.