What is not known?

The common gap

Twelve articles of graded evidence converge on an absence: no study linking an on-pack label change to measured facility-scale contamination outcomes was found. The chain asserted across the field — better labels, better sorting, cleaner streams, better recycling — has not been demonstrated across its full length. Early links are evidenced: instructions improve simulated sorting; self-reported behaviour shifts in trials. The final link — the one implied by impact assessments accompanying statutes — has no located study behind it. Systematic reviews that searched for one report the same absence.

Stated precisely

The missing study is an evaluation of an on-pack label introduction or change — a scheme launch, a redesign, a mandated mark — with contamination or capture measured physically at facility scale (MRF audits, waste-composition sampling) in a before/after or controlled design. Nine structured searches were run across scheme evaluations (ARL, OPRL, How2Recycle, Triman, Nordic pictograms), MRF contamination studies, and the intervention literature. None returned a qualifying study.

BehaviourWorks Australia’s rapid review of recycling-contamination interventions — 1,306 citations screened, 137 papers fully reviewed — states: “Despite a comprehensive search… no reviews focusing specifically on the outcome of preventing contamination of recycling were found,” with “[l]ittle… known about effectiveness, and even less about cost effectiveness” (Kaufman et al., 2020). Sewak et al.’s systematic review of household waste-sorting interventions — 3,595 records screened, fourteen unique interventions — contains no on-pack label intervention with facility-scale outcomes (Sewak et al., 2021).

Nearby evidence

Cart-level feedback has the evidence that labels lack. The Recycling Partnership’s Atlanta “Feet on the Street” programme measured contamination physically: a 2017 pilot across 5,000 households produced a “57% reduction in overall contamination”; the 2019 citywide rollout across more than 98,000 households produced a 19 per cent decrease, measured by hand-sorting 125-pound samples from seventy single-family recycling truckloads into thirty categories at the MRF (The Recycling Partnership, n.d.). Evidential basis: moderate-to-strong — programme-published rather than peer-reviewed, but physically audited at facility scale under a defined protocol. The template exists. No application of it to a label change was found.

Label experiments exist with simulated or self-reported outcomes: Allison et al. (+11.15 points, n = 1,008, simulated bins); WRAP’s on-pack trial (91 versus 86 per cent, self-report); APCO’s ARL kerbside audits (238 households, 62 per cent conformance, cross-sectional, baseline-free). Between the simulated experiment and the cross-sectional audit lies the unbuilt bridge: the same audit protocol, run before and after a label change, with controls.

Other absences

Also not found: a peer-reviewed field RCT of an on-pack disposal label with objectively measured sorting; a before/after evaluation of any label scheme’s introduction on measured capture or contamination, for any scheme in any market; a government or peer-reviewed outcome evaluation of Korea’s mandatory grade label; an audited real-world packaging QR scan-rate study; a causal study isolating a labelling requirement’s effect on packaging design; a study sampling excluded populations — vision-impaired, low-literacy, non-native-speaking — on environmental labels specifically; a prevalence measurement of misleading on-pack claims on a defensible sampling frame. Measurement concentrates where it is cheap — recognition, stated intent, simulated choice. Mandates rest on behaviour, outcomes and populations that have not been measured.

The National Academies’ 2025 consensus report: “More regular collection and reporting of direct observations of household and commercial behavior related to recycling are needed to support recycling policy decision-making” (National Academies, 2025). Asked what is known, a consensus report of that standing answered substantially with what is needed.

What this theme has established

The gap analysis does not displace the findings already graded. Clear, action-specific disposal instructions improve sorting decisions in controlled settings; the largest measured effect is a worded negative instruction. Real-retail purchasing effects of sustainability labels are small, positive and significant — and far below stated willingness. Comprehension fails systematically toward “recyclable,” with error rates on technical vocabulary above ninety per cent, stable across a decade while recognition rises. Sustainability marks lose the on-pack attention contest. One mandatory grade label has survey evidence of perverse disposal effects. Compostable labelling’s contamination channel is real; the littering channel has failed a decade of looking. Measured “recycling” moves with the meter’s placement. The field is rich on decisions, comprehension and attitudes, and empty on outcomes.

Why the gap persists

Label changes roll out across thousands of SKUs over years, denying clean windows. Contamination is measured in expensive hand-sorts. Attribution through the waste stream is genuinely difficult. Schemes fund recognition surveys that market the label rather than audits that might indict it. Regulators legislate on impact assessments and then fund no ex post evaluation. Academics find simulated studies publishable and MRF logistics grant-hostile. Isolating the label’s contribution needs scheme cooperation at the point where schemes have least to gain. These conditions predict that the gap will not close except through a study commissioned for the purpose.

What to commission first

Three studies head the agenda. First, a facility-scale evaluation of a label change that is happening anyway: the EU’s harmonised sorting labels, with staggered implementation dates already set. Highest leverage; audit protocols exist. Second, systematic measurement of what consumers take “recyclable” to mean — capability, access or fate — because litigation and pending legislation will otherwise settle the definition on advocacy readings. Third, audited packaging scan rates from the platforms that already log them. That one costs publication rather than new fieldwork.

What closing it would take

The design is specifiable: before/after or stepped-wedge measurement of a mandated label change, with physical composition sampling at facilities, both capture and contamination recorded, and a comparison market or delayed-rollout group. The EU rollout is the natural experiment. If it is not instrumented, the largest labelling change in the field’s history will join the Korean grade label: a mandate whose behavioural and system effects remain unmeasured years after it is law.

References

APCO (Australian Packaging Covenant Organisation) (2023) ARL Impact Framework Report 2023. Available at: Open source (Accessed: 18 August 2026).

Kaufman, S., Meis-Harris, J., Spanno, M. and Downes, J. (2020) Reducing contamination of household recycling: a rapid evidence review for behavioural public policy (BehaviourWorks Australia, for the Department of Agriculture, Water and the Environment). Available at: Open source (Accessed: 18 August 2026).

National Academies of Sciences, Engineering, and Medicine (2025) Municipal Solid Waste Recycling in the United States: Analysis of Current and Alternative Approaches. Washington, DC: The National Academies Press. doi:10.17226/27978. Available at: Open source (Accessed: 18 August 2026).

Sewak, A., Kim, J., Rundle-Thiele, S. and Deshpande, S. (2021) 'Influencing household-level waste-sorting and composting behaviour: what works? A systematic review (1995–2020) of waste management interventions', Waste Management & Research, 39(7), pp. 892–909. doi:10.1177/0734242X20985608. Available at: Open source (Accessed: 18 August 2026).

The Recycling Partnership (n.d.) Improving curbside recycling in Atlanta (Feet on the Street programme results). Available at: Open source (Accessed: 18 August 2026).

Note on sources and verification

The flagship absence claim is stated in its narrowed, defensible form: no located study links an on-pack label change to measured facility-scale contamination outcomes — with cart-tagging feedback programmes expressly acknowledged as having such measurements, and the Atlanta figures (57 per cent pilot reduction across 5,000 households; 19 per cent citywide across 98,000+; seventy truckload hand-sorts) quoted from the programme's published results. Nine falsification searches were run for this article and their queries and outcomes are preserved in the research record; the claim is corroborated by the two reviews cited, whose quotations are verbatim from the retrieved documents (the BehaviourWorks screening and review counts as stated in the report; Sewak et al.'s counts from the paper's public record). National Academies recommendation language is verbatim from the readable chapter text; a specific numbered recommendation on label standardisation was not extracted and is not asserted here. The consolidated inventory of other absences restates claims documented, with their search bases, in the articles cross-referenced. Sewak et al.'s citation (authors Sewak, Kim, Rundle-Thiele and Deshpande; title, journal, volume and pages) is as given in the paper's SAGE record; an earlier draft of this reference conflated it with a separate 2021 paper and has been corrected. The proposed evaluation design is a synthesis of the located methodological templates prepared for this article, offered as specification rather than recommendation to any named actor.

Last verified: 18 August 2026.