Ecology claims to study all of animal life. A new audit of 1,383,803 scientific papers spanning nearly 150 years finds that it has been doing something far narrower — and that the gap may be widening.

Invertebrates — the insects, spiders, worms, and crustaceans that account for roughly 95% of all known animal species — are dramatically underrepresented in the ecological and evolutionary literature. Vertebrates, which make up fewer than 5% of described animal species, appear in nearly half of all ecology and evolution papers that focus on a single animal group. But a landmark study published in October 2026 in Ecology and Evolution by Pietro Pollo and Michael Kasumovic of the University of New South Wales reveals that the problem runs deeper than publication counts. Vertebrate papers published in the same journals as invertebrate papers draw, on average, 43.1% more online mentions across news outlets, blogs, and social media — and 45.1% higher Altmetric scores, a composite measure of digital reach. The bias does not live only in what gets published. It lives in what gets amplified.

That amplification matters because it shapes what scientists study next. A researcher scanning the literature finds vertebrates everywhere. The papers that fill the high-impact journals, gather the most citations, and trigger the most media follow-up are overwhelmingly about birds and mammals. The 95% of animal diversity represented by invertebrates is present, but quieter — channeled disproportionately into specialist entomology outlets with narrower readership, and generating far less follow-on coverage than studies about animals humans find instinctively appealing.

The study, the largest quantitative assessment of taxonomic bias ever conducted in ecology and evolutionary biology, has a name for this pattern: taxonomic chauvinism — a term coined in 2002 by Xavier Bonnet, Richard Shine, and Olivier Lourdais in Trends in Ecology and Evolution after surveying more than 1,000 papers and finding that research on ectothermic vertebrates such as reptiles was consistently framed more conceptually than research on the already-dominant birds and mammals, even within the non-mammal, non-bird vertebrate cohort. The 2026 study is the first to test whether 24 years of documented concern about this bias has had any measurable effect. Its verdict: essentially none, and in some areas the bias has grown.

How an Algorithm Read 148 Years of Science

Previous studies of taxonomic bias had a fundamental limitation: they relied on researchers manually reading and classifying papers, which capped sample sizes at a few hundred or a few thousand across one or two journals. The resulting snapshots were suggestive but not definitive — critics could argue that different journals, different eras, or different subfields might tell a different story.

Pollo and Kasumovic built a solution to that limitation: a keyword-based automated text classification algorithm trained on roughly 5,000 manually labeled papers, drawing on the GBIF Backbone Taxonomy and a custom vocabulary of taxonomic terms to identify the animal focus of each paper from its title, abstract, and keywords. When tested against 1,815 independently classified papers, the algorithm achieved at least 97.2% accuracy for every animal group, with an F1-score of 91.2% — robust enough to apply reliably across a corpus of unprecedented scale.

The result: 1,383,803 papers across 612 journals, from literature stretching from 1877 through 2025. Of the 991,894 papers that focused on animals, 45.1% concentrated exclusively on vertebrates and 46.1% on invertebrates. Those numbers might look nearly balanced until one recalls that vertebrates account for fewer than 5% of known animal species. Adjusted for biodiversity, insects emerge as the most dramatically underrepresented group, while mammals and fish are the most overrepresented relative to their actual species richness.

Vertebrate Papers Don’t Just Get Published More: They Travel Further

The study’s most consequential finding is not the publication ratio but the attention gap. Compared with invertebrate papers published in the same journal, vertebrate papers received more attention across every metric examined:

• 10.2% more academic citations
• 43.1% more online mentions (news, blogs, social media)
• 24.2% more readers on the reference manager Mendeley
• 45.1% higher Altmetric scores

The journal-level control is critical here. These are not comparisons between a high-profile mammal journal and a specialist entomology quarterly. They are comparisons between papers on vertebrates and papers on invertebrates published side-by-side in the same outlet, competing on an ostensibly level field. The vertebrate papers still pull significantly ahead.

Birds occupied a curious position in the data: they drew the greatest online buzz of any group but received comparatively fewer academic citations than some other vertebrate categories, illustrating that public appeal and scientific influence do not always align. The authors note that the attention analysis is correlational and cannot fully rule out confounding factors — subtle differences in research topic or methodology between vertebrate and invertebrate papers in the same journal may play a role. But the consistency of the advantage across four distinct metrics makes it difficult to attribute entirely to unmeasured confounders.

Why the Bias Has Survived 24 Years of Documented Concern

The question the study cannot escape is: why, after the scientific community was put on notice in 2002, has so little changed? Pollo and Kasumovic address this in a companion study — a preregistered survey of 868 researchers who study animal behavior published as a preprint in early 2026 — that proposes a three-axis framework for why bias persists.

Rational considerations — such as logistical ease of studying certain organisms and their perceived societal relevance — were associated with researchers’ organism choices within broad animal groups but not between them. In other words, practical reasoning explains why within mammals someone studies mice over elephants (they are easier to keep in a lab), but it rarely explains why someone chose mammals over beetles in the first place.

Contextual factors — such as the taxa that were prominent in a researcher’s peer network or that featured in early training — also shaped organism choice. Who supervised your doctoral research, and what animals appeared in your first field course, casts a shadow across an entire career.

Emotional factors — taxonomic affinities, which is to say simply liking certain kinds of animals — most closely mirrored the overall bias patterns in the literature. Researchers tend to study what they find compelling, and what humans find compelling is, by deep cultural habit, vertebrates.

This emotional dimension is significant because it suggests that awareness campaigns — publishing papers pointing out the bias — will not correct the bias. The pull toward familiar, furry, or feathered animals is not purely rational, and addressing it requires institutional interventions that reshape what organisms researchers are exposed to before their preferences calcify.

The publishing infrastructure reinforces this self-perpetuation. Nearly 40% of the 612 journals analyzed maintained explicit taxonomic restrictions, accepting papers only on specific animal groups. Of these specialist outlets, 42% focused on arthropods — including insects, spiders, and crustaceans. This means much of invertebrate research is effectively walled off in niche venues with narrower reach, making the vertebrate dominance in general ecology and evolution journals even more pronounced than the raw publication numbers suggest.

What Vertebrate-Centric Ecology Gets Wrong

Taxonomic chauvinism is not merely an equity problem within science. It carries an epistemological consequence the field is only beginning to reckon with: the theories ecology produces may be systematically wrong in ways that are not currently detectable.

The goal of ecology and evolutionary biology is to generate universal principles — laws of population dynamics, community assembly, and evolutionary change that hold across animal life. But if the empirical foundation for those principles is drawn overwhelmingly from the 5% of animal diversity that happens to appeal to human preferences, the universality of those principles is largely untested. Many of the most foundational findings in animal behavior, physiology, and developmental biology came from invertebrate model systems — the fruit fly Drosophila melanogaster, the nematode Caenorhabditis elegans, the sea slug Aplysia californica — precisely because their simplicity revealed mechanisms that vertebrate complexity conceals. The sustained neglect of invertebrates in ecology and evolution limits the reservoir from which the field can draw corrections to its own assumptions.

The biodiversity stakes compound the scientific stakes. Invertebrates underpin ecosystem services on which human food systems depend. Pollinators — primarily insects — are responsible for the reproduction of more than 75% of major food crops globally; crops that depend on insect pollination carry an economic value estimated at $235–577 billion annually by the United Nations science panel on biodiversity (IPBES). A 2025 meta-analysis found biomass declines exceeding 70% over 30 years in some regions. Yet the organisms sustaining that web are systematically monitored less rigorously, studied less thoroughly, and reported on in smaller venues than the animals that happen to feature on nature documentaries. Scientists cannot respond effectively to declines they cannot measure well. And they cannot measure well what the field has spent 148 years under-counting.

Temporal trends in the Pollo and Kasumovic dataset offer little optimism. Across the literature as a whole, the proportion of vertebrate papers held essentially steady over the 25-year window the study examined. In some subfields it increased. In aquatic journals, the share of papers on fish rose from 35.8% in 2001 to 49.7% in 2025. In journals covering behavior, conservation, and general ecology and evolution, birds and mammals remained the dominant subjects throughout. The only subfield where invertebrate research gained ground was zoology — largely driven by species-discovery work, which disproportionately involves invertebrates because the vast majority of undescribed species are invertebrates.

Malcolm Rosenthal and Maydianne Andrade’s Warning, Now Writ Larger

The 2026 study did not arrive without predecessors. In 2017, Malcolm Rosenthal and Maydianne Andrade published an analysis in The Scientist documenting that primates appeared far more often in Animal Behaviour than beetles did — roughly five publications per year on fewer than 500 primate species against a fraction of that for beetles, which comprise approximately 240,000 described species. They found that vertebrate papers received more citations on average than arthropod papers even after controlling for other variables, and that the bar for demonstrating relevance was measurably higher for papers on less charismatic taxa — invertebrate researchers had to frame their work more broadly just to clear the same editorial threshold.

Rosenthal and Andrade concluded that the evidence confirming the taxonomic skew was both clear and alarming. Their central question remains unanswered by the field: do papers that differ only in their study organisms get treated differently by reviewers and editors?

Pollo and Kasumovic’s analysis of 1.38 million papers effectively answers that question at field-wide scale: yes.

Proposed Remedies: Affirmative Action for Neglected Taxa

The study closes with what the authors describe as “urgent affirmative actions.” These include:

Mandatory taxonomic reporting. Journals and funding bodies should routinely collect and publicly disclose the taxonomic focus of submitted manuscripts and funded grants. Without consistent tracking, there is no accountability — and the available evidence suggests that awareness of the problem, in the absence of structural accountability, does not translate into behavioral change.

Taxonomic equity in evaluation criteria. Hiring committees, award panels, and editorial boards should incorporate taxonomic diversity into their formal criteria, analogous to how diversity initiatives in gender and institutional representation have gradually been embedded in scientific norms. The authors note that such initiatives were once dismissed as impractical and are now routine.

Targeted financial incentives. Dedicated grants, fellowships, and publication subsidies for research on understudied taxa could help correct the structural disadvantage that specialist invertebrate journals face in impact and reach relative to general ecology journals.

Educational reform. Curricula in ecology and evolutionary biology should deliberately expose students to a wide range of animal groups from early in their training, rather than defaulting to vertebrate examples. The companion survey data indicate that early exposure is one of the most durable predictors of which organisms researchers study throughout their careers.

Reviving descriptive biology. Taxonomy and natural history — the disciplines responsible for discovering and naming new species — have declined sharply in academic prestige and funding over the past half-century. Their revival is a prerequisite for any serious research agenda on the roughly 80% of invertebrate species that remain undescribed. A species that has not been named cannot be monitored, cannot be protected, and cannot be studied at all.

A Field Reckoning With Its Own Design Flaw

The study lands at a moment when the life sciences are grappling broadly with structural failures in their own knowledge production — from the replication crisis in psychology and medicine to growing recognition that genomics databases have been built predominantly on populations of European ancestry. The finding from Pollo and Kasumovic adds ecology and evolutionary biology to that list: the field has been conducting a 148-year experiment in which the organisms studied were selected not by scientific priority but by emotional preference, and the resulting theories have been treated as universal when they were built on a 5% sample.

The paper itself — an algorithmic audit of more than a million studies, deploying machine learning to quantify a bias hiding in plain sight — is in some ways a demonstration of the kind of rigorous meta-scientific examination the authors are calling for. A field that can produce a study this comprehensive about its own blind spots is at least capable of honest reckoning.

Whether that reckoning produces change will depend not on individual researchers deciding to study beetles instead of birds, but on whether the institutions that fund research, publish it, and train the next generation of ecologists build systems that make that choice less professionally costly than it has been for the past century and a half.

Frequently Asked Questions

Why do vertebrate research papers get more attention than insect research, even in the same journal?

The Pollo and Kasumovic study found that the attention gap — measured across citations, online mentions, reference-manager readership, and Altmetric composite scores — persists even when comparing papers published side-by-side in the same outlet, which rules out journal prestige as the explanation. The companion survey of 868 researchers suggests the strongest driver is emotional: scientists tend to study organisms they find appealing, and human preference for charismatic vertebrates (mammals, birds) over invertebrates (insects, worms, spiders) is deeply cultural rather than scientific. The attention gap then feeds back into itself: media outlets cover vertebrate findings more often, which makes vertebrate research seem more significant, which shapes what topics early-career researchers choose to pursue.

What is “taxonomic chauvinism” and why does it matter beyond academic fairness?

Taxonomic chauvinism — a term coined by ecologists Bonnet, Shine, and Lourdais in 2002 — describes the systematic preference for studying animals that happen to be culturally appealing rather than scientifically representative. It matters beyond fairness because ecology aspires to universal laws of population dynamics, community assembly, and evolutionary change. If those laws are built overwhelmingly on the ~5% of animal diversity that vertebrates represent, the universality of those laws is largely untested. Additionally, the 95% of animal life that invertebrates represent includes the pollinators, decomposers, and soil engineers that underpin global food production. A field that systematically understudies those organisms is less equipped to detect and respond to their declines.

Has any scientific field successfully corrected a structural research bias similar to this?

Yes, though slowly. Diversity initiatives in gender representation and geographic inclusion in clinical trials — both of which faced initial dismissiveness — have become standard requirements at major funding bodies in the US and Europe over a period of roughly two to three decades. The Pollo and Kasumovic study explicitly invokes this precedent to argue that extending a similar commitment to taxonomic diversity is achievable through the same mechanisms: mandatory reporting, formal evaluation criteria, and targeted financial incentives. The authors acknowledge that structural change in academia is slow, but they argue that the precedent shows it is possible when accountability measures are built into institutional norms rather than left to individual goodwill.

Could the bias in ecology research actually be making our theories about nature wrong?

The study does not test whether specific ecological or evolutionary theories are incorrect — but the concern is structurally serious. If the empirical base for claims about animal behavior, population dynamics, or evolutionary mechanisms has been drawn predominantly from a 5% sample of animal diversity, claims about the generality of those findings have not been rigorously tested against the other 95%. Several foundational results in biology that initially seemed general — drawn from vertebrate physiology or behavior — were later found to be exceptions when tested against invertebrate life. The invertebrate model organisms (Drosophila, C. elegans) that have generated some of biology’s most important mechanistic insights were productive precisely because the assumptions built on vertebrate study could be tested and, sometimes, corrected. A field that systematically produces fewer papers on invertebrates is running a smaller, less diverse test of its own hypotheses.