The cost of open access is the barrier everyone names. Four more sit behind it.

The question

Why do researchers who say they support open access keep publishing behind paywalls?

The usual answer is money. Article processing charges are real, they are large, and they are the first thing anyone raises. A study of humanities and social sciences professors at Canada's fifteen research-intensive universities put the question directly to the people making the decision, and found that cost was the most frequently named barrier but not the only one, and arguably not the most stubborn.

Why it matters

Most of what we know about open access uptake comes from bibliometrics: counting what was published, where, and under which licence. That tells you the outcome. It does not tell you what the author was weighing when they chose the venue. Ayeni and Larivière note that studies collecting data directly from researchers are scarce, which leaves policy being designed against inferred motives rather than stated ones.

That gap matters because the interventions differ. If the barrier is price, waivers and transformative agreements are the answer. If the barrier is how careers are assessed, no amount of subsidy touches it.

What the researchers did

Twenty professors from humanities and social sciences disciplines took part in semi-structured interviews of roughly an hour each, conducted over Zoom between 8 March and 7 April 2022. Eight were early career researchers and twelve were mid-career. Participants came from the U15 group of Canadian research-intensive universities and were recruited by convenience sampling, drawn from people who had completed an earlier questionnaire and indicated willingness to be interviewed. Twenty-nine were invited and twenty took part.

Transcripts were returned to participants to check for accuracy, then coded in NVivo following Braun and Clarke's six-phase reflexive thematic analysis. Ethics approval came from McGill University's Research Ethics Board.

What they found

Five themes emerged as systemic and institutional barriers:

• Unaffordable article processing charges. The most frequently discussed barrier, and the one that most sharply affected early career researchers.
• Precarious career stage and tenure requirements. Participants described targeting high impact factor journals because that is what tenure and promotion committees reward, and those journals are typically paywalled or charge for the open option.
• Unequal privileges. Researchers whose work does not require large grants, and who therefore do not apply for them, described being cut off from the paid open access route entirely. Their methodology needs no funding, so they have no budget line to publish from.
• Gender. Women participants attributed lower engagement with open access to time: childcare, family responsibilities and service loads left less capacity to search out suitable open venues.
• Conflicting and unsupportive institutional policies. Participants reported receiving no institutional support, and described a disconnect between funder mandates and any mechanism to implement them. Some said they had been criticised for including publication charges in grant budgets.

The pattern worth noticing is that only the first theme is straightforwardly a price. The second and third describe an evaluation system that steers authors toward venues where open access is expensive or unavailable. The fourth describes time rather than money. The fifth describes institutions mandating a behaviour they do not resource.

What this does not establish

A good deal, and the study is reasonably clear about most of it.

Twenty interviews are appropriate for thematic saturation and cannot support any claim about prevalence. Nothing here licenses a statement about what proportion of researchers face a given barrier. The sample is Canadian, humanities and social sciences only, and drawn entirely from research-intensive universities, so it says nothing about STEM disciplines, teaching-focused institutions, or the Global South inequities the paper's own literature review discusses.

Sampling was by convenience, from people who had already completed a questionnaire on open access and volunteered for a follow-up. That is a group with more than average interest in the topic, which is useful for depth and unhelpful for representativeness.

The gender theme rests on participants' own explanations of their behaviour, not on any measurement of their publishing output. It reports a perceived cause. Establishing that caring responsibilities depress open access publishing would require a different design entirely.

Two limitations the paper does not foreground are worth stating. The cohort runs from early to mid-career, so the most secure group in the system, senior full professors, is absent from a study substantially about tenure precarity. And the interviews were conducted in early 2022 while the article appeared online in July 2025 and in a 2026 issue. Three to four years is a long time in open access policy. Read the findings as evidence about the mechanism rather than a current snapshot of Canadian practice.

Interpretation

The interesting claim in this paper is not that open access costs money. It is that the financial barrier and the assessment system are not two separate problems.

An author choosing a venue is choosing against a promotion criterion. When that criterion is journal prestige, and prestige correlates with paywalls or with high charges for the open option, then a waiver scheme removes the price while leaving the incentive pointing the same direction. The researcher without grant funding described in the third theme is not short of a discount. They are being asked to fund a venue choice their own method never generated money for.

The authors' recommendations follow from this: fund models that do not bill the author, diamond open access among them, and reform research assessment so that journal metrics carry less weight. The second of those is the harder one, and it has been argued for since DORA in 2013 without being resolved.

For a researcher reading this, the practical residue is smaller and still useful. Rights retention and repository deposit route around the price barrier without touching the venue decision, and the paper points at both. They remain the least discussed option in a conversation dominated by what publishers charge.

What remains unanswered

Whether these five barriers hold outside Canada, outside the humanities and social sciences, and outside well-resourced universities. Whether the gender finding survives a design that measures output rather than asking people to explain themselves. And the question the authors raise at the end, which is how indexing decisions by Scopus, Web of Science and DOAJ shape which open venues are considered legitimate in the first place. If discoverability and perceived legitimacy are set by the indexes, then the venue choice is constrained before cost enters the picture at all.

Source

Ayeni, P. and Larivière, V. (2026). Inequity, precarity, and disparity: Exploring systemic and institutional barriers in open access publishing. Journal of Librarianship and Information Science, 58(2), 1073 to 1090. DOI: 10.1177/09610006251353385. First published online 10 July 2025. Open access under CC BY-NC 4.0.

Journal quartile verified on SCImago: Q1, Library and Information Sciences, 2025 (SJR 0.720, H-index 45, ISSN 09610006). The journal has held Q1 in that category continuously since 2015. Category chosen because it is the journal's only SCImago subject category and matches the paper's subject.