Abstract

Research literacy is a critical skill for behaviour professionals, yet studies in domestic animal behaviour and training often privilege the suppression of overt behaviours over the consideration of emotional, cognitive, and relational impacts. This article outlines how methodological design, anthropocentric bias, and disciplinary priorities shape what research measures — and what it overlooks. Drawing on perspectives from ethology, anthrozoology, and critical animal studies, it examines key concepts for critical engagement, including operational definitions, validity, reliability, bias, and the distinction between statistical significance and practical relevance. The paper also addresses the influence of popular science and algorithm-driven media in distorting findings. A practical framework for reading research is provided to help practitioners move beyond uncritical reliance on “studies show…” toward nuanced, context-sensitive interpretations. The approach emphasises integrating efficacy with ethical integrity, fostering a more reflective and humane application of evidence in practice.

Why Research Literacy Matters

I read a paper on urban bird foraging years ago. It was something about altered patterns in response to human presence. The study was cleanly structured, the numbers lined up, and the analysis was sound. By most measures, it was a solid piece of research. But later, watching a small group of gulls along the shoreline near my house, I found myself thinking about everything that wasn’t in those tables. The way one gull would pause, then follow another’s lead. The shifting distances between them. The sudden, breathless lift-off when a dog passed too close. It wasn’t that the study was wrong, it captured something real. But it was narrow, as research is (and for good reason).

For someone who doesn’t work with birds, I think about them a lot. And for some reason, every time I observe the gulls’ squabble, it makes me think about how research, as powerful as it is, can’t help but reflect the questions we know how to ask. Methodology is always a filter. Some things make it through while others are edited out before the data sheet is even drawn up. This is why we need a plurality of approaches. While behavioural ecology might chart foraging routes or human-wildlife conflict in quantitative terms, other fields ask different, yet no less urgent, questions.

For example, in one of my master’s course classes, we read a paper that lodged itself in my thinking, titled How Pigeons Became Rats (Jerolmack, 2008). It traced how pigeons were transformed from symbols of prestige and aesthetic breeding to urban nuisances, revealing how cultural shifts, rather than biological changes, recast the meaning of an animal. I was particularly struck by the breeds whose curly feathers rendered them barely able to fly. It made me think differently not just about birds, but about the dogs I work and live with, about how we construct, devalue, and manipulate the lives of companion animals through systems of breeding, expectation, and control.

Similarly, I particularly remember a book by Peggs (2012) where she interrogates the commodification of nonhuman animals through the lens of aesthetic cruelty, revealing how selective breeding practices in nonhuman animals prioritise human desire over well-being. I remember that part really well, not because it was the only or best piece of work I have read on the subject (though I very much admire her work), but because it sparked a little flame that seemed to light the way for my brain to shift a little. Her work doesn’t aim to measure behaviour in the lab, but to expose the systemic violence encoded in the ordinary. This, too, is science. This, too, is an inquiry into lives shaped, controlled, (and sometimes diminished) by human design. These weren’t studies of training methods or behavioural outcomes that I had poured myself into in practice, but they were deeply relevant to behaviour. They helped me understand that our assumptions about dogs are always already shaped by culture. It reminds us that good research isn’t only about what is countable but also about what is accountable. This article, therefore, is about learning to read research with that in mind. Not with suspicion, but with curiosity and openness to what that research looks like. Not to dismiss its value, but to deepen it. Because when we learn to notice not only what is measured, but also what is missing, we begin to see research as a lens through which we understand the lives of others. You don’t need a PhD to read a research paper, but you do need a sharp eye, a healthy dose of curiosity, and a willingness to sit with complexity.

In the nonhuman animal behaviour and training industry, research is often cited as a stamp of authority: “The evidence says…” or “Studies show…” But what does that really mean? And how do we know if the study in question is relevant, ethical, or even well-designed? Reading research is about learning how to ask good questions. Whether you’re evaluating the claims behind a new training protocol, considering whether a medication trial applies to your client, just curious about the science behind enrichment, or diving into the research to write a well-supported journal article, research literacy helps you move from assumption to understanding. And that makes our work science-informed.

What Counts as Research?

When people hear the word “research,” they often picture lab coats, clipboards, and sterile environments. We tend to think in terms of control groups, sample sizes, and p-values. But research encompasses a vast and diverse set of approaches, many of which are deeply relevant to behaviour professionals. It’s not all about rats in Skinner boxes (though there’s enduring value there too). Understanding what counts as research requires us to look beyond traditional hierarchies of evidence and consider the richness and complexity of knowledge production and epistemological evolution.

Research Types and Methodologies

At a broad level, research is often divided into two overarching categories: primary and secondary. Primary research involves original data collection. This might take the form of experiments, field studies, surveys, or interviews. It is typically characterised by direct observation, measurement, or manipulation (Walliman, 2021). Secondary research, by contrast, involves the analysis or synthesis of existing studies. This includes literature reviews, meta-analyses, systematic reviews, and scoping reviews (Stewart & Kamins, 1993; Walliman, 2021). Both forms play vital roles in advancing knowledge, and each comes with its own epistemological commitments and limitations, which we will examine later.

Within both primary and secondary research, we encounter different methodological paradigms: quantitative, qualitative, and increasingly, mixed-methods designs (Creswell & Clark, 2017).

Quantitative research is grounded in the logic of measurement (Walliman, 2021). It seeks patterns, correlations, and causal relationships using numerical data and statistical analysis. It values replicability (the ability to reproduce a study and obtain similar results), generalisability (how applicable the findings are in broader populations), and objectivity (standardisation, control conditions, and statistical rigor). In the context of behaviour research, this might involve tracking the frequency and duration of a target behaviour across conditions, measuring physiological responses such as heart rate or cortisol, or calculating effect sizes to quantify the impact of an intervention. While quantitative methods can reveal consistent patterns and allow for predictions, they are often limited in capturing the lived experience, emotional tone, or contextual richness of behaviour (Creswell & Poth, 2016). As such, they are most powerful when paired with an understanding of what those numbers do and do not tell us.

Qualitative research, by contrast, prioritises depth over breadth (Walliman, 2021; Stewart & Kamins, 1993). Rather than aiming for large sample sizes or generalisable outcomes, it seeks to understand how meaning, experience, and behaviour are shaped by social, cultural, physical, and temporal contexts. Common methods include interviews, open-ended surveys, case studies, naturalistic observation, and textual or discourse analysis. These are analysed systematically using well-established approaches such as thematic analysis, grounded theory, or narrative analysis, depending on the research question and type of data (Tandon, 2021; Walliman, 2021). Instead of striving for universal truths, qualitative research aims to illuminate situated, relational, and subjective realities (Tandon, 2021). It explores how individuals make sense of their experiences, how relationships influence behaviour, and how meaning is co-constructed between human and nonhuman participants (Creswell & Poth, 2016; Tandon, 2021). In nonhuman animal behaviour research, this might involve examining how a caregiver interprets their dog’s responses during cooperative care, or how environmental stressors affect a cat’s interactions with veterinary staff. These studies offer rich, textured insights that may not be broadly generalisable but are crucial for informing compassionate, context-sensitive practice.

Mixed methods are exactly what they sound like. It is research that combines qualitative and quantitative approaches, acknowledging the strengths and limitations of each (Creswell & Clark, 2017).

Each methodology offers distinct insights, and none should be privileged in isolation. Behaviour professionals benefit most from being able to understand the affordances of different approaches so we can not only know what a study found, but how and why it was designed that way.

Common Research Designs in Behaviour Work

Primary

In applied animal behaviour, the research we encounter most often includes:

1. 1. Observational studies. Observational studies document behaviours in naturalistic or semi-naturalistic settings, often without experimental manipulation (Altmann, 1974. These are especially useful for understanding the context in which behaviour occurs and for identifying patterns that might otherwise go unnoticed (De Winkel et al., 2024). For example, an observational study might track the social interactions of shelter cats in a communal housing environment to examine affiliative or conflict behaviours. The strength of this design lies in ecological validity (studying wolves in the environments where behaviour naturally unfolds). However, because these studies do not involve controlled variables, it is difficult to establish causality with confidence (Brereton et al., 2022).
   2. Experimental designs. Experimental research, including randomised controlled trials (RCTs), attempts to establish causal relationships by manipulating an independent variable and measuring its impact on a dependent variable (Fraser, 2025; Lehner, 1987). For instance, an RCT might evaluate the effect of a specific training technique (independent variable) on reducing leash reactive-behaviours (dependent variable) by assigning participants to treatment (where their protocol is used) and control groups (where their protocol is not used). While such studies are valuable for isolating variables and testing hypotheses under controlled conditions, they often struggle with ecological validity (Gadagkar, 2021; Janmaat, 2019). Behaviour in a laboratory or simulated setting may not reflect behaviour in the lived environments of nonhuman animals and their caregivers. Small sample sizes and narrowly defined criteria can also limit generalisability (Gadagkar, 2021).
   3. Survey research. Surveys are commonly used to gather data on practitioner beliefs, client experiences, or broader population trends and can be either qualitative or quantitative in nature (King et al., 2018). For example, a survey might explore how behaviour professionals perceive the use of medications in training and behaviour work. While survey data can offer valuable insights and have a wide reach, their quality depends heavily on thoughtful design and careful interpretation (Farrar et al., 2021). Key considerations include whether the questions are clearly worded, unbiased, and appropriately scaled; whether the survey has been validated; and whether internal consistency has been assessed (for example, using Cronbach’s alpha to determine how reliably related items measure the same construct) (DeVellis, 2021; Farrar et al., 2021). It’s also important to ask whether the sample is both diverse and representative of the target population and recruitment of participants can be a limitation. Finally, interpretation should be cautious, keeping in mind the limitations of self-report data, potential biases, and incomplete responses (Farrar et al., 2021). It is important that the study lists how those are handled.
   4. Case studies. Case studies provide detailed, narrative accounts of specific individuals, dyads, or small groups (Yin, 2009). A well-documented case study might, for example, explore the behaviour change journey of a formerly fearful cat through a cooperative care protocol, contextualising each decision, challenge, and outcome. Though statistically ungeneralisable, case studies can highlight nuances and ethical tensions in real-world practice (Mills & Marchant-Forde, 2010). They are particularly powerful when addressing complex cases or underrepresented populations that would not fit neatly into controlled studies. So, if you have written a case study, guess what? You’ve done research.
   5. Ethnographic and autoethnographic research. Rooted in anthropology and sociology, ethnography involves immersive, long-term observation and engagement, often focusing on culture, identity, and power relations in human societies (Ellis et al., 2011). But increasingly, multi-species ethnographies have been utilised to examine the interconnectedness of humans and nonhumans (Alger & Alger, 1999; Koralesky et al., 2023; Wels, 2020). In applied animal behaviour, this might take the form of a researcher embedding themselves within a shelter or training community to understand shared practices and values (Alger & Alger, 1999). Or think of Jane Goodall and her work with the Chimpanzees in Gambe. Autoethnography, by contrast, centres the researcher’s own lived experience, as both observer and participant (Ellis et al., 2011). For instance, a behaviour consultant might reflect on their own emotional labour and ethical tensions in the consulting process. These methods, while uncommon in traditional behaviour literature, are increasingly recognised for their ability to surface unspoken norms, emotional dimensions, and structural influences.

Secondary

Secondary research plays a vital yet sometimes underappreciated role in evidence-informed practice. Secondary research includes systematic reviews of the existing literature, meta-analyses, narrative reviews, and scoping reviews. These forms of research do not generate new data but instead interpret, synthesise, or map existing studies. When done well, secondary research can offer powerful insights. A high-quality meta-analysis, for example, might aggregate data from dozens of studies to detect trends or effects that may not be visible in individual trials (Phillips, 2005). Systematic reviews aim to minimise bias by using transparent inclusion criteria and rigorous appraisal methods (Ineichen et al., 2024). These approaches can help identify where consensus exists, where findings diverge, and where gaps remain in the literature (Bahadoran et al., 2020).

However, secondary research is only as robust as the studies it draws from. If the included studies are methodologically weak, inconsistent, or overly heterogeneous, the conclusions of the review may be compromised (Bahadoran et al., 2020; Berger-Tal et al., 2019). There is also a risk that review methodologies — particularly in narrative or scoping formats — may introduce selective emphasis, depending on how studies are chosen, interpreted, or weighted (Ritskes-Hoitinga et al., 2022a/2022b). Furthermore, publication bias (the tendency for studies with positive or significant findings to be published more often) can distort what is available for synthesis in the first place (Bahadoran et al., 2020; Ritskes-Hoitinga et al., 2022a).

Despite these limitations, secondary research is invaluable for informing practice and policy, especially when direct evidence is limited or contested (Berger-Tal et al., 2019), such is largely the case in our industry. For practitioners, the key is to approach these syntheses with the same critical eye used for primary research, such as examining methodology, transparency, scope, and underlying assumptions. They are not definitive verdicts; rather they are evolving maps of a complex and dynamic evidence landscape. The most ethical and effective practitioners are those who can read across these differences, recognising the potential of each method while being attuned to its limitations (Ineichen et al., 2024). Importantly, they reflect different assumptions about what knowledge is, how it is acquired, and what it is for.

Interdisciplinarity and Epistemological Pluralism

Contemporary behaviour work is increasingly informed by research from adjacent and overlapping disciplines like ethology, veterinary medicine, psychology, education, anthrozoology, and even philosophy. This interdisciplinary influence enriches the field but also demands intellectual flexibility (Bekoff & Pierce, 2009; Marino & Frohoff, 2011). A study grounded in veterinary science may prioritise clinical outcomes while one grounded in education may prioritise learner experience (Weiss et al., 2015). An anthrozoological paper might attend to power, culture, ethics, or the human-animal bond in ways not common in empirical behavioural science (Marino & Frohoff, 2011). This diversity is a strength, but it also challenges us to think carefully about epistemological fit (Bekoff & Pierce, 2009; Biesta, 2015). What are the assumptions underpinning the research? How does the discipline define success, stress, learning, or welfare? How might those definitions shape both the questions asked and the conclusions drawn?

There is no one-size-fits-all approach to research literacy. It is not about memorising hierarchies of evidence or dismissing studies that don’t fit a particular mould. Instead, it is about cultivating the capacity to read widely, think critically, and appreciate the different ways knowledge is constructed (Denzin & Lincoln, 2011; Honey & Mumford, 1986). In behaviour consultancy, our scope of knowledge has to extend well beyond behaviour itself. We are not only interpreters of learning theory or ethograms; we are professionals working in a deeply dynamic context that touches veterinary science, ethics, sociology, education, welfare policy, and more. The lives of the nonhuman animals we work with are shaped by an interplay of health, environment, culture, and human decision-making. To navigate this terrain responsibly, we need fluency in more than one scientific language. This is why looking across disciplines — and across different types of research within them — is a professional necessity. It is about recognising that no single methodology captures the full complexity of behaviour, learning, and ethical practice, and that our role as professionals is not just to consume research, but to interpret it thoughtfully, contextually, and with care.

Evaluating Research: Core Concepts and Critical Engagement

Reading research critically requires more than recognising study design. It means interrogating how knowledge is produced, what is made visible through data, and what remains obscured. Each component of a research paper — each term, each measure, each omission — shapes how we come to understand behaviour and welfare. Below are some of the key lenses through which research can be critically examined as a reflective, context-aware practice.

Operational Definitions: What Counts as What?

A paper might begin with a well-formulated hypothesis and clearly defined variables. But how are those variables operationalised? When a study claims to measure “aggression,” what behaviours are included? Growling? Freezing? Lunging? Avoidance? The choice of what counts as a behaviour isn’t just semantic. It has profound implications for what is observed, coded, and concluded. Behaviour is complex and layered, and the process of reducing it to data points always involves value judgments.

Sample Size and Population

Context matters, and so does representation. A study involving 20 Labrador retrievers from a single shelter might yield interesting findings, but can those findings be extended to fearful mixed-breed dogs in urban homes, or working-line herding breeds with entirely different behavioural patterns? Sample size is only part of the picture. We must also ask who was studied, where, under what conditions, and whether that population bears any resemblance to the individuals whom practitioners are actually working with.

Validity: Is the Study Asking (and Answering) the Right Questions?

Validity refers to the soundness of a study’s design and conclusions. Internal validity considers whether the research accurately measures what it claims to measure, without confounding variables muddying the picture. External validity asks whether the findings are generalisable to other settings, species, or populations. A study might be methodologically airtight but still of limited use if the conditions are too far removed from practice.

Reliability: Can It Be Repeated?

Reliability is about consistency. If the same study were conducted again, or if the same video of behaviour were scored by multiple observers, would the results hold up? Inter-observer reliability is especially important in behaviour research, where interpretation is often subjective. Without strong reliability, data potentially becomes more of a reflection of the coder than of the behaviour itself.

Biases

Bias can arise at many stages. It might be in what questions are asked, in who is recruited, in how data is interpreted, and in what gets published.

1. 1. Confirmation bias can lead researchers to favour outcomes that support their hypothesis (Marsh & Hanlon, 2007). Like, if a practitioner believes that a certain tool reduces reactive behaviour, they might unintentionally focus on moments of quiet and overlook subtle signs of stress or avoidance that contradict their expectation.
   2. Funding bias may influence what kinds of research are conducted in the first place (Van der Schot & Phillips, 2013). Research often follows funding, trends, and dominant paradigms. Studies focused on pharmaceuticals, medicalised diagnoses, or quick-fix behavioural outcomes are more likely to receive support than those exploring relationships, cooperative care, or humanities-based questions. This doesn’t make them flawed, just partial. They reflect what is possible within current systems of priority and approval, which shapes not only what gets studied but how. Quantifiable, observable, and replicable metrics are privileged, and research outside these norms frequently lacks the institutional interest to flourish. As these dominant frameworks drive publication, citation, and replication, they create a sense of comprehensiveness while confining us to a narrow evidentiary terrain. Acknowledging this isn’t a critique of science, but a call to broaden what counts as evidence (and who gets to define it).
   3. Publication bias skews the literature toward positive findings. In other words, studies with positive or statistically significant results are more likely to be published than those with null or inconclusive findings (Higgins et al., 2023). Being aware of these structural influences helps us interpret research not as neutral, but as situated.

Statistical Significance Versus Practical Relevance

Statistical significance is not the same as real-world relevance. In research, a result is typically considered “statistically significant” if the probability that it occurred by chance is below a predetermined threshold, so often a p-value of 0.05 (AbdulRaheem, 2024). This means there is less than a 5% likelihood that the observed effect happened randomly. But a statistically significant result doesn’t automatically mean the effect is large, important, or noticeable in practice. For example, a two-second reduction in barking during a five-minute observation window might be statistically significant, especially in a tightly controlled environment with many repeated trials. But to the caregiver struggling with their dog’s distress at the front door, those two seconds may offer no meaningful change in day-to-day life. This is why it’s crucial to look beyond p-values. Effect size tells us how big or impactful a result actually is (Stewart & Kamins, 1993; Walliman, 2021). Confidence intervals show the range within which the true effect likely falls, offering a sense of the study’s precision (Stewart & Kamins, 1993; Walliman, 2021). And perhaps most importantly, readers should look for the researcher’s own reflection on whether the findings are likely to matter in real-world contexts. Not just in statistical terms, but in emotional, ethical, and practical ones.

Limitations

Often overlooked, the limitations section is where researchers disclose what didn’t go to plan, what might constrain their interpretations, and where uncertainty remains. These sections are not weaknesses. They are sites of integrity. A thoughtful limitations section suggests a reflective researcher who understands the boundaries of their own work. Reading it is essential to understanding what the study can (and cannot) tell us.

Reading in Context: From Scepticism to Synthesis

To engage critically with a study is not to seek flaws for the sake of discrediting it. It is to read with care, depth, and humility, and to be able to recognise what a study tells us, and what it doesn’t. Research lives in its context. It lives in the questions it asks, the assumptions it makes, the lives it seeks to capture, and the ones it overlooks. Critical reading means approaching research as a conversation partner, not a gospel. But reading in context requires both empathy and discernment. It means imagining the world behind the dataset. The caregivers trying to implement a protocol, the animal learners responding in ways the ethogram doesn’t capture, the practitioner wrestling with outcomes that don’t fit the prediction. It means asking who designed the study and what shaped their framing? Whose voices are included in the write-up and whose are absent? What was the study designed to explore, and was the question meaningful — and, not just to researchers, but to nonhuman animals, caregivers, and practitioners? Were the methods suited to that question? Who were the subjects, and how were they characterised or constrained? Were the findings interpreted with nuance, or were modest patterns inflated into general truths? Were the results acknowledged as provisional, shaped by setting and sampling, or were they positioned as universal?

Imagine a study examining the “efficacy” of e-collars. The data might show a statistically significant reduction in barking. But what else was happening? Did the study measure observable behaviours like tail carriage or head orientation? Or physiological responses like heart rate variability or recovery time post-intervention? Did it track generalisation across environments or only examine isolated sessions? Were comparisons made to other protocols like extinction or differential reinforcement? What counts as success and whose definition of success is being privileged? What if the study found no observable stress signals in video footage, but failed to consider that stress signals are themselves subject to interpretation, timing, and context? What if behaviours that indicate discomfort are missed by coders unfamiliar with that species, breed, or individual? Or dismissed as artefacts of the method? A flat ethogram may tidy up the data, but it can also erase its emotional resonance.

Even graphs and tables deserve critical scrutiny. Are they scaled in ways that visually exaggerate differences? Are averages presented without ranges or outliers? Is the variation in the data evident, or hidden behind summary scores that homogenise diverse experiences? Visual representation is powerful and sometimes unintentionally misleading. For example, a bar chart can mask individual variance and a line graph can suggest trends where none meaningfully exist.

And what of what is absent? Critical readers learn to notice silence, such as what wasn’t studied, what wasn’t measured, which populations were excluded, and which outcomes were deemed unimportant. Are fearful, anxious, old, or disabled individuals included in the study sample? Are marginalised caregivers or low-resource settings acknowledged as part of the training landscape? This isn’t about suspicion, it’s about curiosity. It’s about recognising that all research is bounded, and those boundaries matter. And it’s about noticing the moral implications of what’s omitted and not just what the study says, but what it declines to ask.

Reading critically is also about synthesis. It’s not enough to read one study in isolation and draw conclusions. How does this paper build on or depart from previous work? Does it replicate earlier findings or challenge them? Are the authors aware of competing interpretations? Do they engage with qualitative work, lived experience, or practitioner wisdom? The strength of a study often lies not in its control group but in its capacity to provoke thoughtful engagement. That’s why, for me, writing really helps me centre the available literature in practice. It allows me to dive deep into supporting and competing findings, and learn not from conclusions as much as the questions I end up with.

Ethics, Evidence, and the Limits of What Works

The phrase “evidence-based” has become a shorthand for credibility. It’s become a badge of authority that signals rigour, reliability, and rationality. But too often, this framing collapses a complex set of questions into a single one: “Does it work?” In doing so, it conflates efficacy with ethical validity. Yes, aversive methods can reduce behaviour, sometimes rapidly and with statistically measurable results. But what, exactly, are we measuring? And equally important, what are we not measuring? What behaviours are targeted for reduction, and why? What internal states, experiential dimensions, or long-term relational consequences are overlooked in the process? These omissions are not neutral, they reflect priorities, and those priorities reflect values. In this light, “evidence-based” practice is not an assurance of ethical integrity, but often a hollow credential. Powerful in rhetoric, but largely irrelevant to the actual moral quality of the work.

When evaluating research, particularly studies examining training methods in domestic dogs, cats, and horses, it’s important to look beyond the reported behavioural outcomes. Much of this research centres on the suppression or reduction of overt behaviours, often without assessing the emotional, cognitive, or relational costs involved. The “how” without the “why.” A paper might frame a reduction in barking, growling, scratching, or biting as a success, even when that “success” comes at the expense of the individual’s agency, emotional expression, or capacity to cope. Metrics may show compliance — something many clients say they want — but compliance alone tells us little about well-being. It does not capture the toll of coercion, the loss of predictability, or the flattening of affect that can accompany chronic stress.

An exclusive focus on observable outcomes risks sidelining what matters most in real-world relationships: trust, choice, curiosity, and connection. In practical terms, a method that “works” in a narrow behavioural sense may still result in relational fractures, eroded mutual understanding, or a diminished capacity for resilience in future encounters. For practitioners, this means that efficacy in training outcomes is not synonymous with ethical soundness. When we accept data at face value without interrogating what was measured, what was omitted, and whose interpretation we are relying on, we risk perpetuating harm under the guise of science.

The Hidden Human Bias in Research

A deeper issue lies in the human-centred design and interpretation of most animal behaviour research. Research questions, methodologies, and outcome measures are typically framed by human perspectives, often asking what nonhuman animals know about us, how they respond to us, or how they compare to humans and our constructs. In this framing, the nonhuman animal becomes a subject of human curiosity, rather than a knowing being in their own right. As Horowitz (2023) argues, it may be time to shift away from this anthropocentric default and reframe our inquiries to centre the nonhuman perspective.

Decentring the human in research means more than including nonhuman animals as data points. It means treating them as participants with their own perspectives, interests, and forms of meaning-making (Jones, 2022; Jones & Taylor, 2023). Rather than asking how a dog’s behaviour reflects human norms or expectations, we might instead ask what that behaviour means for the dog, within their own umwelt — that is, their unique perceptual world (Jones & Taylor, 2023; Maderson & Elsner-Adams, 2023). Yet this shift also reveals a methodological limitation, that our inability to fully access or comprehend that umwelt imposes epistemological boundaries on what we can measure (Scotto, 2024). Human cognitive frameworks, linguistic tools, and observational categories may never entirely capture the subjective experiences of nonhuman animals, no matter how well-intentioned the design (Bräuer et al., 2020; Colombino & Bruckner, 2023; Dacey, 2017; Scotto, 2024; Street et al., 2025; Wolfe, 2009).

This recognition should not lead to nihilism, but to humility. It demands a reconsideration of what counts as knowledge, and whose perspective defines its value (Street et al., 2025). It requires reimagining study design, interpretation, and reporting practices in ways that foreground the nonhuman animal’s experience, agency, and welfare (Colombino & Bruckner, 2023; Jones & Taylor, 2023). Traditionally, human participants are acknowledged — named, consent-giving, context-rich — while nonhuman animals are reduced to case numbers, variables, or anonymous subjects (Horowitz, 2023). Their perspectives are filtered through human interpretation, and their emotional and ethical relevance is frequently overlooked. Decentring the human challenges this double standard. It calls for nonhumans to be recognised not only as study subjects but as beings with lives that matter independently of their utility to human goals (Columbino & Bruckner, 2023; Kotzmann, 2023). Volsche et al. (2022) further advocate for research practices that explicitly honour the individuality and welfare of animal participants. This includes more ethical citation, description, and contextualisation, while refusing to subsume nonhuman experiences beneath human agendas and instead committing to a relational and inclusive science.

The Trouble with Pop Science

Isn’t it easier to just let someone else do the work for us? To report back in more palatable terms? Popular science articles, blogs, and social media posts play an important role in making research accessible. Obviously reading or even accessing the material to read from behind a paywall can pose issues. At its best, pop-science offers digestible summaries, sparks curiosity, and helps bridge the gap between academia and practice. But it also comes with risks, especially when nuance is lost in translation (Dempster et al., 2024; Erduran, 2025). Pop science often trades rigour for reach. Headlines may overstate findings, suggest causal relationships where none were tested, or extrapolate beyond the population or context studied. Correlation becomes causation. Preliminary findings become settled truths. And complex behaviours, especially in nonhuman animals, are reduced to catchy soundbites that oversimplify motivation, emotion, and context.

Sometimes, this distortion stems from the original press release. University media offices may amplify a study’s most marketable claim while glossing over limitations (Erduran, 2025). Other times, it’s the result of algorithms because attention-grabbing statements perform better than cautious ones (Metzler & Garcia, 2024). “Study shows dogs love praise more than food” will circulate more widely than “Under limited conditions, some dogs oriented more toward verbal markers than food rewards, possibly due to prior reinforcement history.” This simplification isn’t always malicious, but it does shape public understanding and, in our field, can subtly inform practitioner choices. When science is filtered through marketing, it becomes easy to miss key details like the population studied, the methods used, the degree of variance, or the assumptions embedded in the analysis (Erduran, 2025).

To read critically in the age of pop science, ask:

• • What did the original study actually measure? And, is the original study linked and accessible?
  • Are the claims in the article supported by the data, or do they stretch the implications?
  • What populations were studied, and are they comparable to the animals or clients you work with?
  • Are the limitations acknowledged — or are they conveniently absent?
  • Is the language definitive (“proves,” “shows,” “confirms”) or appropriately cautious?

Pop-science summaries may pique our interest, but for those of us in behaviour professions, it’s crucial to engage with the actual study. That means reading methods, questioning assumptions, and recognising that striking headlines often oversimplify nuanced science.

Your Research Reading Road Map

Don’t just take the “studies show…” at face value. Here’s how to dig in, decode the jargon, and decide what the research really means. This is how I tackle a paper:

Conclusion

Research is not infallible. It is shaped by human values, institutional structures, funding streams, and methodological limitations. But it remains a powerful tool for insight, discovery, and professional evolution. By learning to read research with discernment and care, behaviour professionals position themselves not only to apply evidence more effectively, but to contribute to the development of better, more ethical evidence. Critical reading is not a defensive posture. It is an active, generous practice of engagement — one that asks, always, how knowledge is produced, whose voices are heard, and what kind of world we are building when we treat evidence not as a command, but as a conversation.

References

AbdulRaheem, Y. (2024). Statistical significance versus clinical relevance: Key considerations in interpretation medical research data. Journal of Family and Community Medicine, 31(4), 217-226. https://www.doi.org/10.4103/ijcm.ijcm_601_23

Alger, J. M., & Alger, S. F. (1999). Cat Culture, Human Culture: An Ethnographic Study of a Cat Shelter. Society & Animals, 7(3), 199-218. https://doi.org/10.1163/156853099X00086

Altmann, J. (1974). Observational Study of Behavior: Sampling Methods. Behaviour, 49(3-4), 227-266. https://doi.org/10.1163/156853974X00534

Bahadoran, Z., Mirmiran, P., Kashfi, K., & Ghasemi, A. (2020). Importance of systematic reviews and meta-analyses of animal studies: Challenges for animal-to-human translation. Journal of the American Association for Laboratory Animal Science, 59(5), 469–477. https://doi.org/10.30802/AALAS-JAALAS-19-000139

Bekoff, M., & Pierce, J. (2009). Wild Justice: The Moral Lives of Animals. University of Chicago Press.

Berger-Tal, O., Greggor, A. L., Macura, B., Adams, C. A., Blumenthal, A., Bouskila, A., … & Blumstein, D. T. (2019). Systematic reviews and maps as tools for applying behavioral ecology to management and policy. Behavioral Ecology, 30(1), 1-8. https://doi.org/10.1093/beheco/ary130

Biesta, G. J. (2015). Good education in an age of measurement: Ethics, politics, democracy. Routledge.

Bräuer, J., Hanus, D., Pika, S., Gray, R., & Uomini, N. (2020). Old and New Approaches to Animal Cognition: There Is Not “One Cognition”. Journal of Intelligence, 8(3), 28. https://doi.org/10.3390/jintelligence8030028

Brereton, J. E., Tuke, J., & Fernandez, E. J. (2022). A simulated comparison of behavioural observation sampling methods. Scientific Reports, 12(1), 3096. https://doi.org/10.1038/s41598-022-07169-5

Colombino, A., & Bruckner, H. K. (2023). Hidden in plain sight: How (and why) to attend to the animal in human-animal relations. In Methods in Human-Animal Studies (pp. 1-29). Routledge.

Creswell, J. W., & Clark, V. L. P. (2017). Designing and conducting mixed methods research. Sage publications.

Creswell, J. W., & Poth, C. N. (2016). Qualitative inquiry and research design: Choosing among five approaches. Sage publications.

Dacey, M. (2017). Anthropomorphism as Cognitive Bias. Philosophy of Science, 84(5), 1152–1164. https://doi:10.1086/694039

Dempster, G., Sutherland, G. and Keogh, L. (2022). Scientific research in news media: a case study of misrepresentation, sensationalism and harmful recommendations JCOM 21(01), A06. https://doi.org/10.22323/2.21010206

Denzin, N. K., & Lincoln, Y. S. (Eds.). (2011). The Sage handbook of qualitative research. sage.

DeVellis, R. F., & Thorpe, C. T. (2021). Scale development: Theory and applications. Sage publications.

De Winkel, T., van der Steen, S., Enders-Slegers, M. J., Griffioen, R., Haverbeke, A., Groenewoud, D., & Hediger, K. (2024). Observational behaviors and emotions to assess welfare of dogs: A systematic review. Journal of Veterinary Behavior, 72, 1-17. https://doi.org/10.1016/j.jveb.2023.12.007

Ellis, C., Adams, T. E., & Bochner, A. P. (2011). Autoethnography: An Overview. Historical Social Research / Historische Sozialforschung, 36(4 (138)), 273–290. http://www.jstor.org/stable/23032294

Erduran, S. (2025). Beyond misalignment of science in the news and in schools. Science, 387(6729), eadu7468. https://doi.org/10.1126/science.adu7468

Farrar, B. G., Voudouris, K., & Clayton, N. S. (2021). Replications, comparisons, sampling and the problem of representativeness in animal cognition research. Animal behavior and cognition, 8(2), 273. https://doi.org/10.26451/abc.08.02.14.2021

Fraser, D. (2025). Two paradigms of research and their influence on the study of animal behaviour. Applied Animal Behaviour Science, 284, 106550. https://doi.org/10.1016/j.applanim.2025.106550

Gadagkar, R. (2021). How to design experiments in animal behaviour. Resonance, 26, 105–125. https://doi.org/10.1007/s12045-020-1108-6

Higgins, J. P. T., Thomas, J., Chandler, J., Cumpston, M., Li, T., & Page, M. J. (2023). Cochrane Handbook for Systematic Reviews of Interventions version 6.4 (updated August 2023) [Internet]. Cochrane. Available from: Available from www.training.cochrane.org/handbook.

Honey, M., & Mumford, A. (1986). The Manual of Learning Styles. Peter Honey Publications.

Horowitz, A. (2021). Considering the “dog” in dog–human interaction. Frontiers in Veterinary Science, 8, 642821. https://doi.org/10.3389/fvets.2021.642821

Ineichen, B. V., Held, U., Salanti, G., Macleod, M. R., & Wever, K. E. (2024). Systematic review and meta-analysis of preclinical studies. Nature Reviews Methods Primers, 4(1), 72. https://doi.org/10.1038/s43586-024-00347-x

Janmaat, K. R. (2019). What animals do not do or fail to find: A novel observational approach for studying cognition in the wild. Evolutionary Anthropology: Issues, News, and Reviews, 28(6), 303-320. https://doi.org/10.1002/evan.21794

Jerolmack, C. (2008). How pigeons became rats: The cultural-spatial logic of problem animals. Social problems, 55(1), 72-94. https://doi.org/10.1525/sp.2008.55.1.72

Jones, E. E. A. (2022). Silent Conversations: The Influence of Human Exceptionalism, Dominance and Power on Behavioural Expectations and Canine Consent in the Dog-human Relationship: a Thesis Submitted in Partial Fulfilment of the Requirements for the Degree of Doctor of Philosophy in Human-Animal Studies at the University of Canterbury (Doctoral dissertation, University of Canterbury).

Jones, E., & Taylor, N. (2023). Decentring humans in research methods: Visibilising other animal realities. In Methods in human-animal studies (pp. 33-49). Routledge.

King, N., Brooks, J., & Horrocks, C. (2018). Interviews in qualitative research. Sage Publication.

Koralesky, K. E., Rankin, J. M., & Fraser, D. (2023). Using institutional ethnography to analyse animal sheltering and protection II: Animal shelter work. Animal Welfare, 32, e67. doi:10.1017/awf.2023.83

Kotzmann, J. (2023). Sentience and intrinsic worth as a pluralist foundation for fundamental animal rights. Oxford Journal of Legal Studies, 43(2), 405–428. https://doi.org/10.1093/ojls/gqad003

Lehner, P. N. (1987). Design and execution of animal behavior research: an overview. Journal of Animal Science, 65(5), 1213-1219. https://doi.org/10.2527/jas1987.6551213x

Maderson, S., & Elsner-Adams, E. (2023). Two species ethnography: honey bees as a case study of an interdisciplinary “more-than-human” method. In Methods in Human-Animal Studies (pp. 87-107). Routledge.

Marsh, D. M., & Hanlon, T. J. (2007). Seeing what we want to see: Confirmation bias in animal behavior research. Ethology, 113(11), 1089-1098. https://doi.org/10.1111/j.1439-0310.2007.01406.x

Marino, L & Frohoff, T (2011). Towards a new paradigm of non-captive research on cetacean cognition. PLOS ONE 6(9): e24121. https://doi.org/10.1371/journal.pone.0024121

Metzler, H., & Garcia, D. (2024). Social drivers and algorithmic mechanisms on digital media. Perspectives on Psychological Science, 19(5), 735-748. http://doi:10.1177/17456916231185057

Weiss, E., Mohan-Gibbons, H., & Zawistowski, S. (Eds.). (2015). Animal behavior for shelter veterinarians and staff. John Wiley & Sons.

Mills, D., & Marchant-Forde, J. N. (Eds.). (2010). The Encyclopedia of Applied Animal Behaviour and Welfare. CABI.

Peggs, K. (2012). Animals and sociology. Springer.

Phillips, C. (2005). Meta-analysis — a systematic and quantitative review of animal experiments to maximise the information derived. Animal Welfare, 14(4), 333–338. https://doi.org/10.1017/S0962728600029663

Ritskes-Hoitinga, M., & Pound, P. (2022a). The role of systematic reviews in identifying the limitations of preclinical animal research, 2000–2022: part 1. Journal of the Royal Society of Medicine, 115(5), 186-192. https://doi.org/10.1177/01410768221093551

Ritskes-Hoitinga, M., & Pound, P. (2022b). The role of systematic reviews in identifying the limitations of preclinical animal research, 2000–2022: part 2. Journal of the Royal Society of Medicine, 115(6), 231-235. https://doi.org/10.1177/01410768221100970

Scotto, C. (2024). The Anthropocentric Bias in Animal Cognition. Artefactos. Revista de Estudios Filosóficos sobre Ciencia y Tecnología, 13(1), 85-116. https://doi.org/10.14201/art2024.31800

Stewart, D. W., & Kamins, M. A. (1993). Secondary research: Information sources and methods (Vol. 4). Sage.

Street, S. E., Hamilton, I., & Healy, S. D. (2025). Anthropocentric bias may explain research disparities between animal tool use and nest building. Animal Behaviour, 226, 123240. https://doi.org/10.1016/j.anbehav.2025.123240

Tandon, T. (2021). Qualitative research: An overview. International Journal of Indian Psychology, 9(3), 687–693. https://ijip.in/wp-content/uploads/2021/08/18.01.065.0210903.pdf

Van der Schot, A. A., & Phillips, C. (2013). Publication bias in animal welfare scientific literature. Journal of Agricultural and Environmental Ethics, 26(5), 945-958. https://doi.org/10.1007/s10806-012-9433-8

Walliman, N. (2021). Research Methods: The Basics, 3rd ed. Routledge.

Wels, H. (2020). Multi-species ethnography: methodological training in the field in South Africa. Journal of Organizational Ethnography, 9(3), 343-363. https://doi.org/10.1108/JOE-05-2020-0020

Wolfe, C. (2009). Human, All Too Human: “Animal Studies” and the Humanities. PMLA/Publications of the Modern Language Association of America, 124(2), 564–575. https://doi:10.1632/pmla.2009.124.2.564

Yin, R. K. (2009). Case study research: Design and methods (Vol. 5). Sage.