What makes a label easier to understand?

What the design evidence actually supports

Other articles in this theme ask whether labels work. This one asks what makes them work better — the design factors schemes, regulators and producers control directly. Much of the located design research comes from adjacent fields: tobacco warnings, safety symbols, nutrition labels. Transfer to packaging sustainability labelling is by analogy, not demonstration.

Three claims were carried in strong form at the planning stage: that pictograms outperform text; that positional consistency outperforms every content change tested; and that negation costs more as an icon than as words. The located evidence supports a qualified version of the first, a modest version of the second, and does not support the third. The useful findings are operations — pairing, placement, wording — not a general preference between formats.

Pictures with words, not instead of them

Noar et al.’s meta-analysis of pictorial cigarette-pack warnings — 37 studies, 48 samples, 33,613 participants — found pictorial warnings “more effective than text-only warnings for 12 of 17 effectiveness outcomes,” spanning attention, reaction and intentions (Noar et al., 2016). Strength grade: strong for adding imagery to health warnings. The authors record the absence of measured behaviour outcomes. No equivalent meta-analysis for disposal or sustainability labels was found.

The packaging-specific evidence, where it exists, runs against a strong pictogram claim. In Allison et al.’s online disposal-label experiment, every effective condition was a worded instruction. Participants judged that “a logo alone would be insufficient as a disposal instruction strategy” and stressed “the importance of written disposal instructions to reduce any potential confusion” (Allison et al., 2024). The EU Joint Research Centre, synthesising behavioural literature for the harmonised sorting label, recommends “consistent use of text and graphical elements” — both, together (JRC, 2024). Action-oriented text-bearing labels have the highest recorded comprehension in located testing: Australian testing reported 82 per cent correct disposal answers (WRAP, 2020). Symbol-only formats populate the failure cases.

Pictorial elements register consistent gains in attention and emotional engagement (transfer evidence, strong). For disposal instruction, the evidenced optimum is pictogram plus text.

Positional consistency

Bialkova and van Trijp (2010), in visual-search experiments on nutrition labels, found a performance benefit “when the type of label and its location on the package did not change in two consecutive exposures,” with attention further enhanced by presence, doubled size and monochrome design. Strength grade: moderate — controlled experiments, transfer from nutrition, search-performance outcome. The JRC synthesis endorses “clarity benefitting from consistent placement” for waste-sorting labels (JRC, 2024).

The evidence supports consistency as a low-cost factor. It does not support a claim that consistency “beats every content change tested.” No located study ran positional consistency head-to-head against content manipulations. Of the located factors, consistency requires the least artwork change relative to the evidence behind it. That is a narrower proposition than dominance.

Negation in words and icons

The planning-stage claim that negation costs more as an icon than as words is not supported, and safety-symbol literature partly contradicts it. Hancock et al. (2004) tested 40 safety symbols on 52 younger and 52 older adults. Comprehension was generally below the 85 per cent criterion and worse for older adults. Prohibition symbols, however, recorded the highest comprehension of the symbol classes in the set.

Execution still matters. Murray et al. (1998), evaluating circle-slash variants across 16 pictorials, found over-slash and under-slash formats preferred over partial and translucent slashes. For some pictorials, “the slash’s intersection (overlap) with its critical features” made the underlying image harder to read — negation can occlude what it negates. Strength grade: moderate, transfer from safety signage, 1998. Allison et al.’s worded “do not recycle” remains the largest measured disposal-instruction effect in this theme. The two results are compatible: a clear prohibition works in both registers if the underlying image stays readable.

The instruction principle

Labels perform to the degree that they are instructions — imperative, behavioural, specific — and fail to the degree that they are evaluations, endorsements or technical vocabulary. That is the common element across the three findings, and it matches the comprehension gradient and the Korean grade-label misreading.

A worked application: compostable packs

The compostable package is the problem case with the largest documented evidence base here. Pair a destination pictogram with words that name the stream (“put with food waste”), not the material virtue (“compostable”). Place it consistently. If the pack does not belong in recycling, say so in a prohibition that does not occlude the image. Confine the format to applications where packaging and food waste travel together, as WRAP recommends. Design cannot repair a stream that cannot process the material. That limit is in the backfire article.

Limits of design

No located evidence shows that design optimisation overcomes the failure modes elsewhere in this theme. Design does not repair the attention budget on a crowded pack, the 54 per cent of shoppers who do not inspect product information, or a collection system that does not match the instruction. Design can make a usable instruction more usable. It cannot make an unused instruction effective.

Evidential basis

The record’s strength is convergence: tobacco warnings, safety symbols, nutrition search, disposal experiments and the JRC synthesis point the same way. Its weakness is transfer. Most of the strongest studies are not packaging sustainability labels. The one controlled packaging disposal experiment in this theme is online and simulated. The design claims that survive that filter are pairing, consistency, and worded instruction.

References

Allison, A.L., Băitanu, A., Purkiss, D., Lorencatto, F., Michie, S. and Miodownik, M. (2024) 'Enabling desired disposal of compostable plastic packaging: an evaluation of disposal instruction labels', Frontiers in Sustainability, 5, 1376519. Available at: Open source (Accessed: 18 August 2026).

Bialkova, S. and van Trijp, H.C.M. (2010) 'What determines consumer attention to nutrition labels?', Food Quality and Preference, 21(8), pp. 1042–1051. Available at: Open source (Accessed: 18 August 2026).

Hancock, H.E., Rogers, W.A., Schroeder, D. and Fisk, A.D. (2004) 'Safety symbol comprehension: effects of symbol type, familiarity, and age', Human Factors, 46(2), pp. 183–195. Available at: Open source (Accessed: 18 August 2026).

JRC (Joint Research Centre) (2024) Behavioural insights for waste-sorting labels in the European Union (Beaumais, O., Kirakozian, A., Lazaric, N., Bruns, H. and Dupoux, M.; JRC134206). Available at: Open source (Accessed: 18 August 2026).

Murray, L.A., Magurno, A.B., Glover, B.L. and Wogalter, M.S. (1998) 'Prohibitive pictorials: evaluations of different circle-slash negation symbols', International Journal of Industrial Ergonomics, 22, pp. 473–482. Author PDF available at: Open source (Accessed: 18 August 2026).

Noar, S.M., Hall, M.G., Francis, D.B., Ribisl, K.M., Pepper, J.K. and Brewer, N.T. (2016) 'Pictorial cigarette pack warnings: a meta-analysis of experimental studies', Tobacco Control, 25(3), pp. 341–354. Available at: Open source (Accessed: 18 August 2026).

WRAP (2020) On-pack labelling and citizen recycling behaviour. Report mirror available at: Open source (Accessed: 18 August 2026).

Note on sources and verification

Noar et al.'s meta-analytic findings, Hancock et al.'s symbol-class result, Murray et al.'s slash-variant preferences and occlusion finding, and Bialkova and van Trijp's consistency benefit are all quoted or characterised from the sources cited, and all four are transfer evidence — tobacco warnings, safety symbols and nutrition labels respectively — flagged as such wherever used. Noar et al.'s author list is as printed on the retrieved record. Allison et al.'s design-relevant findings, including the logo-insufficiency judgments, are from the open-access paper. JRC quotations are from the report and project pages retrieved; the project's own online-experiment results were not located, and the January 2026 technical proposal for the EU harmonised labels is referenced from its announcement without retrieval of its contents. The corrections to the three planning-stage design claims are documented in the body, with the direction of each correction stated. The instruction principle is a synthesis of the convergent evidence assembled here, not a finding of any single study.

Last verified: 18 August 2026.