How We Rate Evidence: Animals

This page explains how the Animals & Veterinary section of CAB Direct judges the evidence behind the claims in its articles. The section covers pet health, dairy and beef production, animal behaviour and entomology, with topics that range from preventive veterinary checks, pet obesity and kidney disease in cats to dairy cow welfare, sheep vaccination programmes, antimicrobials in livestock, heat stress, animal cognition, bees and wasps, varroa mites and insect decline.

The approach described here is a working method for weighing sources claim by claim. It is not a formal academic grading system, and it does not give an article a single numeric score. The general, site-wide version of the approach is at /how-we-rate-evidence/; this page adapts it to questions about companion animals, livestock, animal behaviour and insects.

Articles in this section are general information about species, production systems and research findings. They are not veterinary advice for an individual animal, herd or flock, and they do not replace examination, diagnosis or treatment by a veterinary surgeon.

A claim-by-claim approach

An article about animals often combines several kinds of claim: a result from a controlled trial, a pattern seen in records from commercial farms or clinics, a figure reported by pet owners, a long-term trend in insect counts and a national statistic on disease or medicine use. Each claim is assessed against the evidence that supports it, rather than giving the whole article one label. A statement about how a vaccine performed in a controlled study and a statement about how many owners describe their dog as overweight carry different levels of certainty, even when they appear in the same article.

For every claim, the questions are the same: which species and population the evidence comes from, what exactly was measured and how, and whether the wording of the article stays within what the source shows.

What we assess

The points below guide the assessment of an individual claim. Not every point applies to every source: a national surveillance report raises different questions from a behavioural experiment or an owner questionnaire.

• Direct relevance: whether the study measured the outcome the article discusses, such as disease incidence, milk yield, lameness, lifespan, colony survival or a specific behaviour, rather than a proxy or a laboratory marker.
• Species and population: which species, breed, age group and production stage were studied, and how closely they match the animals the article discusses.
• Study design: whether the result comes from a randomised controlled trial, another experimental study, an observational study on farms or in clinics, an owner survey, a monitoring dataset or a promotional claim.
• Sample size and duration: the number of animals, herds, flocks, colonies or sites, and whether follow-up lasted long enough for outcomes such as lifespan, fertility or population trends.
• Housing and management: indoor or pasture-based systems, stocking density, diet, climate and husbandry, which can change results for farm and companion animals alike.
• Measurement: how outcomes were recorded, for example by clinical examination, laboratory tests, farm records, owner questionnaires, trap counts or behavioural scoring, and whether observers knew which group an animal belonged to.
• Consistency and uncertainty: whether other studies agree, the ranges or confidence intervals reported, and the limitations stated by the study authors.
• Source and interests: who published and funded the work where this is disclosed, and whether the source sells a feed, medicine, supplement or other product being assessed.

How different source types are used

Evidence about animals comes from several kinds of source, and each answers a different question. The section uses them as follows.

Controlled studies and clinical trials in animals

Controlled studies, including randomised clinical trials in the species concerned, are the main source for claims that a vaccine, medicine, diet, housing change or management practice causes a particular effect. We look at how animals were allocated to groups, whether a control group was used, whether assessment was blinded where that was possible, and which outcomes were measured. Results obtained in one species, or in laboratory animals, are reported as such and are not transferred to another species without saying so. Small or short trials are described with their size and duration.

Observational and on-farm studies

Observational studies follow animals in clinics, on commercial farms or in the wild without assigning a treatment. They show how outcomes such as lameness, mastitis, heat stress or antimicrobial use vary across real herds and flocks, and they often include far more animals than trials. Because farms, owners and animals differ in many ways at once, these studies show associations rather than proof of cause. Articles describe what was compared and which differences the researchers adjusted for, and avoid presenting an association as the effect of a single factor.

Owner-reported and survey data

Owner surveys and questionnaires are a common source for companion animal topics such as pet obesity, activity, diet and lifespan, and farmer surveys are used for practices such as vaccination or housing. They show what respondents report, which may differ from what an examination or a clinical record would show. Articles note how respondents were recruited, how many took part and whether the sample is likely to represent the wider population of owners or farms. Figures from surveys run by companies that sell pet food or insurance are identified by their source.

Long-term field datasets for insects and wildlife

Questions about insect decline, pollinator populations and the behaviour of wild animals often rely on long-term monitoring: repeated counts from traps, transects or fixed sites, and behavioural observation over many seasons. These datasets matter because short studies cannot separate a trend from year-to-year variation. We check where and when the monitoring took place, whether methods or sites changed during the series, and which species groups were counted. A decline recorded at monitored sites in one region is reported for that region and those groups, not as a figure for all insects everywhere.

Official surveillance statistics

Government and agency surveillance reports provide figures on notifiable diseases, outbreaks, veterinary antimicrobial sales or use, and animal numbers. They are the main source for national scale and trends. Articles state the reporting body, the year and the units, for example whether antimicrobial figures are sales by weight of active ingredient or estimates of use on farms, and note changes in reporting methods that affect comparisons between years or countries. Surveillance data record what was detected and reported, which can differ from the true level of disease.

From one study to a broader conclusion

A single study rarely settles a question about animal health, welfare or behaviour. A trial in one research herd, a study of colonies in one apiary over one season or an owner survey in one country is reported as a finding from that setting. Broader statements, such as whether a practice generally improves welfare or whether an insect group is declining, need support from several studies in different settings or from a well-conducted systematic review or meta-analysis.

Where studies disagree, the article describes the disagreement and its likely reasons, such as differences in species, breed, housing, climate or how outcomes were measured. Findings on animal cognition and play, where results can depend on how a behaviour is defined and tested, are presented together with the method used. A new study that contradicts earlier work is reported alongside that work, not as an automatic replacement for it.

Scope and applicability: species, breeds and housing systems

Evidence about animals often applies to a narrower group than a headline suggests. A result in housed Holstein cows may not hold for other dairy breeds kept at pasture. A finding on lifespan in one dog breed or size group may not apply to others. A study of honey bee colonies does not by itself describe bumblebees, solitary bees or wasps.

Articles therefore state the species, breed or type, age group, production system and country in which the evidence was collected. Housing and climate are named where they matter, for example for heat stress, where the same air temperature can affect animals differently depending on humidity, shade, ventilation and breed. Figures on medicine use, vaccination rules or disease status are given with their date, because they change over time. An average effect in a study group does not predict what will happen to a particular pet, cow or flock.

Matching language to evidence strength

The wording of an article should signal how strong the evidence behind each claim is. The section keeps a small set of distinctions:

• “Found” or “showed” is used for results measured in the study being described, with the species and setting stated.
• “Is associated with” is used for patterns in observational, on-farm and survey data where cause and effect have not been established.
• “Owners reported” or “farmers reported” is used for survey and questionnaire findings that were not checked against examinations or records.
• “Suggests” or “early evidence” is used when support comes from a few small or short studies, or from laboratory animals or another species.
• “Claimed by” or “according to”, with the company named, is used for figures from makers of feeds, medicines or equipment that have not been independently tested.

Articles avoid words such as “proven”, “cure” or “guaranteed”, and do not turn research findings into instructions for treating or feeding an individual animal. Where an article discusses signs of illness or treatment options, it presents them as general information and directs readers to a veterinary surgeon for decisions about their own animals.

Review, updates and corrections

Claims and figures in Animals articles are checked against traceable sources, and every number is given with its source and date. Each article names its author with a stated role. When a cited source is updated, corrected or withdrawn, or when newer evidence changes the picture, the affected wording or figure can be revised. Corrections are dated and remain visible in the article.

This page does not set a fixed interval for re-checking evidence. For the wider publishing rules, see /editorial-policy/; for how published fixes are recorded, see /corrections/. Readers who think a claim overstates or misreads its source can raise it through /corrections/ or /contacts/.

Who applies this approach

The approach is applied by the authors of the Animals & Veterinary section, each within their stated coverage. All five are listed with the role of Author.

Imogen Farrant covers companion animal health, preventive care, diet and welfare, where owner-reported data and clinic-based studies are common. Rhys Penwarden covers dairy and beef production, herd health, feeding, fertility and livestock welfare, where trial results need to be matched to breed and housing system. Lena Okonjo covers animal behaviour, insects, pollinators and pest insects, where the way a behaviour is defined and observed shapes the result. Gethin Morrow covers sheep vaccination programmes and clostridial diseases, antimicrobials in livestock, heat stress, poultry and pig systems, and preventive herd health, where surveillance statistics and on-farm studies are central. Rowan Siu covers bees and wasps, varroa and other mites, insect communication and insect decline datasets, where monitoring methods determine what a trend can show.

Author profiles and credentials are at /authors/.

Historical context

CAB Direct also has a separate history as a CABI research database platform. CABI traces its beginnings to an entomological committee in 1910 and records that, in 1973, the contents of its abstract journals were unified into the CAB Abstracts database. According to CABI, CAB Direct launched in 2003 as CABI’s own platform for CAB Abstracts. CABI’s help pages now describe searching the CABI Digital Library, in particular its main databases, CAB Abstracts and Global Health.

This page describes the evidence approach used by the current Animals & Veterinary editorial section. It should not be read as a claim that the historical CAB Direct database used this framework, or that CABI publishes, reviews or endorses this section.

CABI history: .

CABI Digital Library help: .

Related pages and contact

For background on the project, see /about/, and for the section hub, see /animals/. The site-wide evidence page is /how-we-rate-evidence/ and the editorial rules are at /editorial-policy/. Questions about how evidence is handled in a published Animals article can be sent through /contacts/.