Testosterone and Aggression in Dogs - Separating Facts from Myths

1. Testosterone: A Brief Neuroendocrine Primer

1.1 Two Pathways of Action

Testosterone is a steroid hormone produced mainly in the Leydig cells of the testes, with smaller amounts from the adrenal glands in both sexes and the ovaries in females. It acts through two routes: a genomic pathway, binding androgen receptors to change gene expression over hours to days; and a non-genomic pathway, acting rapidly on cell membranes or — importantly — being converted to estradiol by the enzyme aromatase, altering neuronal excitability and neurotransmitter release (part of the wider neurochemistry shaping behavior). That aromatase route matters, because some of testosterone's behavioral effects may actually be estrogen effects.

1.2 Organizational versus Activational Effects

Testosterone rises during sexual maturation, but the timing, magnitude, and behavioral significance of that rise vary considerably across individuals, breeds, and developmental trajectories. A two-stage organizational–activational framework explains why castration cannot fully control sexually dimorphic behavior: prenatal testosterone has already organized certain brain structures, and removing circulating testosterone later cannot reverse that established wiring — it can only remove the ongoing activational signal (much of this laid down during early development).

1.3 Why Blood Levels Explain So Little

Circulating testosterone is the quantity that can be measured and the one least likely to predict behavior. Receptor density, aromatase activity, the timing of exposure and the individual's learning history all sit between the hormone and the behavior.

Two dogs with identical serum levels can behave entirely differently, which is the single most important fact in this article and the one least represented in how the hormone is discussed (why a physiological value is not a behavioral measure). It is also why measuring a dog's testosterone would not answer the question an owner is asking, which is occasionally proposed and would not help.

2. Testosterone and Aggression: A Modulatory, Not Causal, Role

2.1 A Weak Baseline Association

Studies comparing baseline testosterone in aggressive versus non-aggressive dogs yield mixed results. A cross-species review concluded that dominant or more aggressive animals tend to show higher plasma testosterone than submissive or less aggressive ones (Knol & Egberink-Alink, 1989), but that is a tendency across species and says little about an individual dog (and the "dominance" framing itself is scientifically unsound). Crucially, correlation is not causation: higher testosterone in aggressive dogs could be a consequence rather than a cause, since aggressive encounters can acutely raise testosterone — a feedback loop easily mistaken for a stable trait difference.

2.2 Testosterone as a Behavioral Amplifier

A more accurate framework treats testosterone as a modulator that can help bring about or escalate an aggressive state rather than cause it. On this account, compared with a neutered dog, an intact dog may react more readily, escalate faster, plateau at a higher level of reactivity, return to baseline more slowly, and possibly sit at a higher baseline overall (lowering the threshold of an already-aroused nervous system). That profile is proposed rather than measured; none of the canine studies cited here tracked those parameters by hormonal status. The framework reframes testosterone from "cause" to "amplifier" — a hormone that may lower the threshold for some forms of aggression without being necessary or sufficient for any of them.

2.3 Non-Genomic Mechanisms and the Amygdala

A dog-specific study examined androgen receptors in the basolateral amygdala of pathologically aggressive dogs, finding significantly more AR-positive neurons in aggressive dogs. The AR-negative fraction of neurons was increased as well, so the authors concluded that additional mechanisms are likely involved, and they judged most of the receptors to be unliganded, which argues against a major role for classic genomic androgen-receptor action in that region (Jacobs et al., 2006). One plausible alternative route is the conversion of testosterone to estradiol by aromatase, although that was not what this study measured. It is a single, small canine study, so its implications are suggestive rather than definitive — but it points away from a simple "testosterone → androgen receptor → aggression" story.

3. Acute versus Chronic Effects: Timing Matters

3.1 Acute Testosterone Fluctuations

In the comparative literature, testosterone can shift rapidly with social challenge: a male meeting a rival may show a transient rise that facilitates escalated responding in that moment. That pattern has been proposed as adaptive, preparing for conflict without chronically high levels, and it has not been characterized in dogs in the studies cited here. After GnRH-agonist implants such as deslorelin, testosterone can rise briefly before it falls; whether that short rise changes behavior has not been examined in the studies cited here.

3.2 Chronic Exposure

Chronic elevation may have different, more organizational effects, but in dogs the correlation between baseline testosterone and aggression is relatively weak — so chronic levels alone are poor predictors of aggressive behavior. Hormonal status may influence how a dog responds to frustration — a key trigger for many forms of aggression — more than it dictates aggression directly.

3.3 An Analogy Worth Being Careful With

Testosterone is often described as a volume control rather than a switch, and the metaphor is useful up to a point. It captures that the hormone changes intensity rather than creating behavior, and it understates how much depends on what is being amplified.

A dog with no propensity to defensive aggression does not acquire one when the volume rises. A dog that is already frightened of strangers may cross a threshold it would otherwise have stayed below (how an existing fear of strangers shapes the response).

3.4 What Would Falsify the Modulatory Account

If testosterone caused aggression, removing it should reliably remove the behavior, and the effect should not depend on the type of aggression or on the dog's history. Neither holds.

That the effect depends heavily on presentation and on individual history is the observation the modulatory account was built to explain, and it is why the account survives while the causal one does not.

4. Proactive versus Reactive Aggression: Differential Involvement

Not all aggression is alike. Behavioral neuroscience distinguishes proactive (instrumental, goal-directed, low-arousal) from reactive (affective, fear-driven, high-arousal) aggression, and these have different neurobiology and different relationships to testosterone. These categories are useful clinical models, but individual cases often overlap — a single dog can show fear, frustration and resource defense in the same episode.

4.1 Proactive Aggression: Testosterone-Associated

Proactive aggression — resource guarding, territorial defense, inter-male conflict — is the category usually described as testosterone-influenced, but the effect of castration on it is smaller and less consistent than often claimed, and for territorial aggression the older data report no change. In 57 male dogs castrated after two years of age, urine marking, mounting and roaming improved by at least 50% in at least 60% of dogs, whereas for the other problem behaviors, including the forms of aggression examined, an improvement of at least 50% occurred in fewer than 35% of dogs; effects on aggression toward unfamiliar people and on fear of inanimate stimuli were not significant (Neilson et al., 1997). An older study of 42 male dogs, based on unstructured telephone interviews an average of 27 months after surgery and without a control group, reported a 62% decrease in aggression toward other dogs, with territorial and fear-induced aggression unchanged (Hopkins et al., 1976, as described by Farhoody et al., 2018). Even here, testosterone is not the whole story: marking, mounting, and fighting are complex behaviors with a substantial learning component that castration will not undo.

4.2 Reactive Aggression: Minimal Testosterone Involvement

Reactive aggression — impulsive, emotional, triggered by perceived threat or frustration, and marked by high arousal, fear, and anxiety — appears far less testosterone-dependent, and may even increase after castration in some individuals (Farhoody et al., 2018) (the neurological picture of the reactive dog). Some evidence suggests neutering is associated with more fear-related behavior in certain dogs and contexts: in 6,235 male dogs, longer lifetime exposure to gonadal hormones was associated with reduced reporting of several fear- and aggression-related behaviors (McGreevy et al., 2018), and neutering was among the factors associated with fear of unfamiliar people and dogs (Puurunen et al., 2020).

4.3 Implications for Clinical Practice

The upshot is directional but weaker than usually presented. Older, uncontrolled studies reported decreases in aggression toward other dogs after castration (Hopkins et al., 1976, as described by Farhoody et al., 2018), while the largest dataset found no association between gonadectomy and aggression toward other dogs (Farhoody et al., 2018). That dataset measured aggression toward other dogs as a questionnaire category rather than conflict between intact males specifically, so a narrower effect on inter-male conflict is not ruled out. Castration is less likely to help — and in some dogs may be counterproductive — with reactive, fear-based aggression. For the various forms of aggression, the authors of the castration study concluded that fewer than a third of dogs can be expected to show marked improvement (Neilson et al., 1997).

5. What the Large Datasets Show

5.1 Why Sample Size Matters Here

Most of the classic castration studies involve dozens of dogs. Aggression is comparatively rare, varies by context and by target, and is affected by many variables at once — which is exactly the situation in which small samples produce unstable answers.

Two large questionnaire datasets have since addressed the question at a scale the earlier work could not.

5.2 The Gonadectomy Study

Using the Canine Behavioral Assessment Research Questionnaire, an analysis began with 15,370 surveys and, after excluding dogs gonadectomized before six weeks of age or to correct a behavior problem and those with incomplete data, retained 13,795 for aggression toward familiar people, 13,498 for aggression toward strangers and 13,237 for aggression toward other dogs (Farhoody et al., 2018).

Excluding dogs neutered to correct a behavior problem is the methodological decision that makes the study worth reading. Those dogs are the reason simpler analyses find that neutered dogs are more aggressive: they were aggressive first, which is why they were neutered.

5.3 The Result

The authors conclude that when the many factors affecting aggressive behavior are considered, there is no evidence that gonadectomy at any age alters aggressive behavior toward familiar people or dogs, and only a minimal increase in aggression toward strangers (Farhoody et al., 2018). One version of the abstract in circulation swaps strangers and dogs; the results tables are unambiguous that the only association concerned strangers.

That is a largely null result from more than thirteen thousand dogs, on the outcome for which the procedure is most often recommended. Null results of that size are rare in canine behavior research and are worth considerably more than a positive finding from fifty animals.

5.4 The One Exception

In the restricted model, there was a low but significant increase in the odds of moderate or severe aggression toward strangers among gonadectomized dogs, driven entirely by dogs gonadectomized between seven and twelve months of age, who were 26% more likely to show it. In the fully adjusted model the association was not significant, and the authors themselves consider that it may be a false-positive finding (Farhoody et al., 2018).

The age band matters. It corresponds to the period discussed in the developmental chapter below, and it is the one window in which the data show a possible effect — in the direction opposite to what owners are usually told. A 26% increase in odds is small in absolute terms, since the base rate of moderate or severe aggression toward strangers is low, and it is the opposite of the reason the procedure is recommended at that age.

5.5 What the Authors Say Plainly

Their own summary is that gonadectomy does not result in a predictable decrease in aggressive behavior across all male and female dogs, although it might be effective in some (Farhoody et al., 2018).

Both halves of that sentence are load-bearing. The procedure is not a treatment for aggression at population level, and individual dogs may still respond — which is why the honest answer to an owner is that it might help this dog and probably will not (where the wider neutering evidence is set out). That is an unsatisfying thing to be told and it is what the evidence supports.

5.6 A Second Dataset, a Different Outcome

A survey of around six thousand pet dogs examining social fearfulness found sterilization status among the variables associated with fear of unfamiliar people and unfamiliar dogs: intact dogs were less likely to be fearful than neutered dogs (Puurunen et al., 2020).

Fearfulness is not aggression, but much defensive aggression in dogs arises from it. A procedure that leaves aggression unchanged while being associated with more fearfulness is not neutral for behavior, even if it is neutral for the outcome that was measured. The association is one variable among several in that model and should not be read as an effect of the procedure; it is a reason to look rather than a finding to act on.

5.7 What Questionnaire Data Cannot Fix

Both datasets are cross-sectional and owner-reported. Neither can establish direction, and the sterilization association in particular is open to the same confound the gonadectomy study excluded by design: fearful dogs may be neutered more often.

The gonadectomy study handled that confound explicitly. That is what separates it from the many smaller analyses reaching louder conclusions (the wider evidence on aggression and its confounds).

6. The Behavioral Effects of Castration: A Complex Picture

A 2025 review concluded that the conclusions drawn from existing research on neutering, in both behavior and health, remain inconclusive and sometimes conflicting (Arroube & Pereira, 2025) — a fair summary of the whole field.

6.1 What Castration Consistently Reduces

Castration most consistently reduces behaviors tied to sexual dimorphism and male reproduction, but not in all males, and neither experience nor age at castration predicts which males will change (Hart & Eckstein, 1997). In the owner-interview study of 57 adult dogs, urine marking, mounting and roaming improved by at least 50% in at least 60% of dogs and by at least 90% in 25 to 40% (Neilson et al., 1997). Percentages often quoted from a 1976 study of 42 dogs rest on unstructured telephone interviews without a control group; this article cites them only as summarized in a later analysis (Hopkins et al., 1976, as described by Farhoody et al., 2018).

6.2 What Castration Does Not Consistently Change

Other behaviors respond inconsistently or minimally: aggression toward familiar humans (often no change or minimal improvement), fear-based aggression (may increase), learned aggressive behaviors (persisting through reinforcement history regardless of hormones), and aggression toward unfamiliar people, for which castration showed no significant effect (Neilson et al., 1997). In a survey of 491 Dutch owners, 58% of owners of castrated dogs named correcting unwanted behavior as a reason for castration, aggression was involved in half of those cases, and the most frequent report on aggression afterwards was no change (Roulaux et al., 2020).

6.3 Potential Negative Behavioral Effects

Most concerning for behaviorally challenged dogs, castration has been associated in some studies with increases in stress, fear, anxiety and certain types of aggression (Arroube & Pereira, 2025), with more fear- and aggression-related behavior at shorter lifetime hormone exposure (McGreevy et al., 2018), and with increased odds of fear of storms at every age of gonadectomy in Vizslas (Zink et al., 2014). For dogs with fear-based reactivity, removing testosterone may destabilize rather than settle the nervous system — which would fit a proposal, not directly tested in the studies cited here, that gonadal hormones also have stabilizing effects, so that too little testosterone may leave some dogs more anxious or insecure (and chronic stress compounds this). These findings challenge the assumption that castration is universally good for behavior.

6.4 What Owners Are Usually Promised

The behaviors castration most consistently reduces — roaming, urine marking, mounting, and some intermale conflict — are the testosterone-dependent ones and were described decades ago, in owner-interview studies without control groups (Hopkins et al., 1976, as described by Farhoody et al., 2018; Neilson, Eckstein & Hart, 1997). For intermale conflict the evidence is the weakest of the four.

Those are real effects and they are not what most owners are seeking when they neuter a difficult dog. The gap between what the procedure reliably does and what it is expected to do is where most disappointment in this area originates.

6.5 Effect Size Versus Reliability

Even for the behaviors that respond, the older literature reports substantial variation between individuals. A behavior reduced in a majority of dogs is not a behavior eliminated, and an owner told that castration stops marking has been told something that is true of most dogs and not of all.

Setting that expectation before the procedure rather than afterwards is the difference between an informed decision and a grievance. It also protects the professional giving the advice, since a prediction stated as a probability cannot be wrong in the way a promise can.

7. The Influence of Developmental Timing

The age at castration has emerged as a critical variable. A study of 6,235 male dogs castrated for reasons other than behavioral management found forty behaviors that differed between intact and castrated dogs; longer lifetime exposure to gonadal hormones was linked to reduced reporting of 26 mostly unwelcome behaviors, eight related to fear and seven to aggression, while indoor urine marking and howling when left alone were more likely with longer exposure (McGreevy et al., 2018).

7.1 Early Castration (Before 6 Months)

Early castration has been associated with increased behavioral risk: in a retrospective survey of 2,505 Vizslas, dogs gonadectomized at six months or younger had significantly increased odds of developing a behavioral disorder compared with intact dogs, and the younger the age at gonadectomy, the earlier such disorders were diagnosed (Zink et al., 2014).

7.2 Castration Between 7 and 12 Months

This window shows a mixed picture: dogs neutered at 7–12 months were slightly more likely to show moderate-to-severe aggression toward strangers in the restricted model of the large gonadectomy study, while the same study found no association between gonadectomy at any age and aggression toward familiar people or dogs (Farhoody et al., 2018).

7.3 Castration After 12 Months

The evidence on later castration is mixed. In the large gonadectomy study, gonadectomy at 13–18 months or after 18 months showed no association with aggression (Farhoody et al., 2018), whereas in Vizslas the odds of fear of storms were increased after gonadectomy at any age, including after twelve months (Zink et al., 2014).

7.4 The Two-Phase Testosterone Surge

The timing effects are often explained by two testosterone surges: a prenatal surge that organizes male-typical brain structures, and an adolescent surge that activates them. Castration before the second surge prevents activation but cannot reverse prenatal organization; castration after it removes circulating testosterone but cannot erase behaviors already learned and practiced. On this account, castration would be expected to work better before sexually dimorphic behaviors are fully expressed, yet would not fully eliminate established, rehearsed behaviors. The data do not bear that expectation out cleanly: neither experience nor age at castration predicted which males changed (Hart & Eckstein, 1997), and early gonadectomy was associated with more behavioral disorders, not fewer, in one breed (Zink et al., 2014).

7.5 Why the Seven-to-Twelve-Month Window Keeps Appearing

This period recurs in several places: adolescence, the two-phase testosterone surge described above, and now the one age band in which gonadectomy was associated with more aggression toward strangers (Farhoody et al., 2018).

Whether these are the same phenomenon is not established. Adolescence is a real developmental period on its own evidence; the association in the gonadectomy data is a single result, found in one of two models and possibly a false positive (Farhoody et al., 2018), and it cannot by itself show that the window matters for the outcome of the procedure. It is also the period during which most owners are advised to have the procedure done, which makes the convergence worth knowing about even while it remains unexplained.

8. Beyond Testosterone: The Multifactorial Nature of Aggression

The most important conclusion in the literature is that testosterone alone is a poor predictor of aggression, which is a complex, multifactorial phenotype.

8.1 Genetic Contributions

A genetic contribution to aggression-related traits in dogs is likely: in 13,715 Finnish dogs, breeds differed markedly in the prevalence of aggression and other anxiety-related traits, which the authors read as a genetic contribution (Salonen et al., 2020). Which genes carry that contribution is far less settled than the hormonal data above, and no single gene accounts for aggressive behavior (with experience shaping their expression).

8.2 Interactions with Other Hormones and Neurotransmitters

Testosterone does not act alone; the complex relationships between sex hormones and other hormones and neurotransmitters are one of the reasons this field is hard to study (Arroube & Pereira, 2025). One proposed route runs through serotonin: testosterone may modulate aggression partly by influencing serotonergic function rather than by acting on aggressive motivation directly. That route has not been examined in the canine studies cited here, so it describes how testosterone may interact with other neural and social systems rather than an established pathway in dogs.

8.3 The Role of Learning and Environment

Perhaps the strongest challenge to hormonal determinism is learning. Aggression, like any behavior, is shaped by reinforcement history: a dog that has successfully used aggression to make threats go away or secure resources has had that behavior reinforced. Removing testosterone may lower the threshold, but it does not erase the learned association — so behavioral modification is required regardless of hormonal status.

8.4 The Differential That Gets Skipped

Before any hormonal explanation is accepted for a change in a dog's behavior, the medical differential belongs first. Pain is a common and treatable cause of new irritability in an adult dog — a conservative estimate of around a third of referred behavior cases involve a painful condition (Mills et al., 2020) — and it produces exactly the presentation that gets attributed to testosterone in an intact male (how pain produces that presentation).

An intact male dog that has become defensive at three years old is not showing a hormonal problem simply because it is intact. The hormone was there last year too, and something else changed.

8.5 Learning History Does the Heavy Lifting

Of everything listed in this chapter, learning history is one of the most important modifiable contributors in an adult dog, alongside pain and the environment the dog lives in. A dog that has repeatedly succeeded in driving away something frightening has learned something that no endocrine intervention addresses (what graduated behavior work addresses instead).

That is the practical reason the hormonal question matters less than it appears to: whatever the answer, the work is the same. Management, graduated exposure below threshold and reinforcement of an alternative behavior are indicated for a fearful intact male and for a fearful neutered one alike.

9. Challenging Oversimplified Explanations

Three popular myths deserve rejection. Testosterone is not the sole driver: it modulates some aggression (especially proactive, inter-male) but is neither necessary nor sufficient, and neutered dogs can be more aggressive depending on type and individual. Castration is not a universal solution: by removing sexual hormones it creates potential for both benefit and harm, depending on the individual dog, and is no substitute for behavioral assessment, management, and training (Arroube & Pereira, 2025). Aggression is not purely hormonal: genetics, early experience, learning, pain, stress, fear, and environment all contribute, yet pain is easily overlooked in favor of hormonal explanations (because behavior is treated as a simple readout of a single cause).

10. Where the Myth Comes From

10.1 The Intuition Is Reasonable

The belief that testosterone causes aggression is not stupid. Male dogs were more often reported as aggressive in a large survey (Salonen et al., 2020), testosterone is the obvious biological difference, and removing it is a single intervention with a plausible mechanism. The inference is exactly the one a reasonable person would make.

It is also the inference the evidence does not support, and understanding why it persists is more useful than simply asserting that it is wrong.

10.2 Sex Differences Are Not Hormone Effects

Male dogs differ from female dogs in more than circulating testosterone. They differ in prenatal hormone exposure, in size in most breeds, in how they are acquired, kept, exercised and trained, and in what owners expect of them.

A higher rate of reported aggression in males is compatible with all of these. Attributing it to the hormone that can be removed is the explanation that suggests an action, which is part of why it wins.

10.3 The Confounded Comparison

The most persistent error in this area is comparing neutered and intact dogs without asking why each dog is in its group. Dogs are frequently neutered because they showed a behavior problem, which puts the difficult dogs in the neutered group before the procedure does anything.

Analyses that exclude those dogs find a different picture (Farhoody et al., 2018), which suggests that the confound, rather than the hormone, accounts for much of the association that simpler comparisons report. This is worth understanding rather than memorizing, because the same confound appears wherever a treatment is given selectively to the worst cases and then evaluated by comparing treated against untreated.

10.4 Confirmation in Practice

An owner who neuters a two-year-old dog for aggression is also, typically, working with a trainer, managing the environment more carefully, and watching the dog closely. Improvement over the following months has several candidate causes and one memorable one.

Where behavior does not improve, the usual conclusion is that the procedure was done too late rather than that it was the wrong intervention. The belief is therefore difficult to disconfirm from experience, which is the defining property of a claim that survives without evidence: every outcome confirms it.

10.5 Why Professionals Repeat It

Advice given to owners has been surveyed: castration advice came most often from veterinarian practitioners, and 58% of owners of castrated dogs named correcting unwanted behavior as a reason for the procedure (Roulaux, van Herwijnen & Beerda, 2020). The advice is not usually given cynically; it reflects a belief that was standard for decades and has not been revised as quickly as the evidence.

Naming that gently matters, because an owner who has already had the procedure done cannot undo it, and being told it was pointless helps nobody. The same applies to the professional who gave the advice: a recommendation that was standard practice a decade ago is not a failure of judgment, and treating it as one makes revision less likely rather than more.

11. Practical Recommendations

For inter-male aggression, older studies suggest castration helps some dogs, although marked improvement in aggression was seen in fewer than a third of dogs (Neilson et al., 1997) and the largest dataset found no association with aggression toward other dogs (Farhoody et al., 2018). For reactive, fear-based aggression, castration is less likely to help and may be detrimental in some dogs, so addressing the underlying fear through behavior modification and, when indicated, medication should take priority. On timing, if castration is elected, early castration (at six months or younger) was associated with increased behavioral risk in one breed study (Zink et al., 2014), while the evidence on later castration is mixed. Address learned behavior regardless: a dog that has practiced aggression needs systematic desensitization and counter-conditioning irrespective of hormonal status. Assess comprehensively first: rule out pain, fear, anxiety, frustration, and poor socialization before deciding, since for many dogs these matter more than hormones. And consider alternatives: vasectomy or hormone-sparing sterilization prevent reproduction while preserving gonadal hormones, and the authors of the large gonadectomy study call for these alternatives to be examined systematically (Farhoody et al., 2018).

11.1 What to Ask Before the Decision

Four questions do most of the work in a consultation. What exactly does the dog do, and toward whom? When did it start? What has been ruled out medically? And what is the household actually hoping the procedure will change?

The fourth is the one that most often goes unasked. An owner hoping for less marking and roaming has a reasonable expectation; an owner hoping the dog will stop lunging at strangers is expecting something the evidence does not support.

11.2 What to Say to an Owner Who Already Had It Done

Many owners arrive having neutered a dog on advice, for a behavior that has not changed. Telling them the procedure was pointless serves nothing and is not accurate either — it may have reduced marking or roaming, and it might have helped this dog.

The useful move is forward: the behavior that remains has a learning history and probably an emotional driver, and those are addressable now regardless of what the hormone was doing.

12. Testosterone Influence and Castration Effect by Aggression Type

The pattern below summarizes how strongly testosterone is implicated in each presentation and what castration can be expected to do. The categories are clinical models rather than cleanly separated biological classes, and individual dogs deviate from them in both directions.

Proactive / inter-male. Testosterone influence: moderate to high on the modulator account, with limited direct canine evidence. Effect of castration: improvement reported in some dogs; for aggression, marked improvement in fewer than a third (Neilson et al., 1997), and no association with aggression toward other dogs in the largest dataset (Farhoody et al., 2018).

Territorial / resource guarding. Testosterone influence: low to moderate. Effect of castration: territorial aggression reported unchanged in an older uncontrolled study (Hopkins et al., 1976, as described by Farhoody et al., 2018); strongly shaped by learning history.

Reactive / fear-based. Testosterone influence: low, and lower than for proactive forms. Effect of castration: often unchanged, and may increase in some individuals or contexts.

Predatory. Testosterone influence: very low. Effect of castration: no consistent effect.

Redirected / frustration-based. Testosterone influence: low. Effect of castration: unlikely to improve; the underlying frustration is what needs addressing.

Individual responses vary significantly. Castration is not a substitute for thorough behavioral assessment and targeted behavior modification.

13. Summary at a Glance

Gonadectomy did not change aggression toward familiar people or dogs — Across more than 13,000 dogs, no evidence that gonadectomy at any age alters aggressive behavior toward familiar people or other dogs, with only a minimal increase toward strangers (Farhoody et al., 2018).

One age band showed an effect in the opposite direction — Dogs gonadectomized between seven and twelve months were 26% more likely to show moderate or severe aggression toward strangers in the restricted model; the authors do not rule out a false positive (Farhoody et al., 2018).

The confound was excluded by design — Dogs neutered to correct a behavior problem were removed from the analysis, which is what separates this study from smaller comparisons (Farhoody et al., 2018).

The authors' own summary is even-handed — Gonadectomy does not produce a predictable decrease in aggression across dogs, although it might be effective in some (Farhoody et al., 2018).

Sterilization status appears in the fearfulness literature — It was among the variables associated with social fearfulness in around six thousand pet dogs (Puurunen et al., 2020).

Testosterone modulates rather than causes — On the best-supported account, the hormone lowers thresholds and amplifies existing tendencies; testosterone alone does not create aggression in a dog that has no propensity for it.

Reactive aggression is where the myth does most damage — Fear-based defensive aggression is a common presentation and the one least likely to respond to removing testosterone.

The medical case is separate — Nothing here bears on the medical reasons for or against neutering, which are a different question with a different evidence base.

14. Research Gaps and Critical Appraisal

The confidence attached to the overall picture is high; confidence in specific numbers is lower.

Most evidence is owner-questionnaire and correlational. Many studies lack control groups, are cross-sectional rather than longitudinal, rely on subjective owner reports, and do not distinguish aggression types (Arroube & Pereira, 2025).

Some figures rest on single or dated studies. The classic castration percentages come from small owner-interview studies without control groups (Hopkins et al., 1976, as described by Farhoody et al., 2018; Neilson et al., 1997), and the androgen-receptor findings from one small neuroanatomical study (Jacobs et al., 2006).

Timing evidence is uneven. Large datasets exist (McGreevy et al., 2018; Farhoody et al., 2018), but breed-specific findings rest on single breeds (Zink et al., 2014), and conclusions for females are limited.

Confounds abound. Owner-reported outcomes depend on owners' reasons and expectations (Roulaux et al., 2020), and breed, sex, age, and reason-for-neutering all interact (with breed itself a weak predictor of individual behavior).

The strongest evidence is questionnaire-based. The large gonadectomy analysis rests on owner-reported C-BARQ scores (Farhoody et al., 2018), which cannot establish causal direction even where the design excludes the most obvious confound.

Controlled trials are rare. Most comparisons in this area are observational, and the authors of the largest analysis call for randomized controlled studies (Farhoody et al., 2018).

The seven-to-twelve-month finding needs replication. A 26% increase in odds within one age band, in one dataset and in only one of two models, is a result worth following up rather than a settled developmental fact.

Serum testosterone is a poor proxy. Receptor density, aromatase activity and timing of exposure all intervene between circulating hormone and behavior, and none is routinely measured in behavioral work.

14.1 The Version Worth Remembering

On the modulator account, testosterone lowers thresholds and amplifies tendencies that are already present; for individual forms of aggression, the direct canine evidence for this is limited. Testosterone alone does not create aggression, and removing it does not reliably remove behavior that has been learned and rehearsed.

Where the hormone is genuinely doing the work — marking, roaming, some intermale conflict — castration helps. Where fear, pain or learning is doing the work, it does not, and those account for much of what owners actually present with.

15. Conclusion

Testosterone plays a role in canine aggression, but a far more nuanced one than popular discourse allows. It is best understood as a behavioral modulator that may lower the threshold for aggressive responding in specific contexts, and neither the sole cause nor a universal treatment target; for individual forms of aggression, the direct canine evidence for that account is limited. Castration most consistently reduces sexually dimorphic behaviors such as urine marking, mounting and roaming, although not in every dog (Neilson et al., 1997). Its effect on aggression is smaller than owners are usually told: marked improvement was seen in fewer than a third of dogs in the classic castration study, and in more than 13,000 dogs, with dogs neutered to correct a behavior problem excluded, gonadectomy at any age showed no association with aggression toward familiar people or other dogs and only a minimal, possibly false-positive increase toward strangers (Farhoody et al., 2018). Fear-based and reactive aggression is the least likely to respond, and neutering appears among the factors associated with more fear-related behavior and social fearfulness (McGreevy et al., 2018; Puurunen et al., 2020) — associations that cannot establish direction. Developmental timing shapes outcomes, though not in the simple way the two-surge account predicts (Hart & Eckstein, 1997; Zink et al., 2014). Clinical decisions should never rest on hormonal assumptions alone; they require full behavioral and medical assessment. Aggression cannot be reduced to a single hormone, but reflects the interaction of genetic predisposition, hormonal modulation, learning history, stress physiology, and environment. Managing it well means embracing that complexity rather than reaching for a convenient but inaccurate simplification — and, often, looking first for the pain, fear, or frustration that a scalpel cannot touch.

Key Insights (Takeaways)

• Testosterone modulates rather than causes aggression: on the best-supported account it can lower thresholds and amplify responses, but it is neither necessary nor sufficient, and the direct canine evidence for specific forms of aggression is limited. Across species, more aggressive animals tend to show higher testosterone (Knol & Egberink-Alink, 1989), and aggression can itself raise testosterone, so the correlation does not establish direction.

• In pathologically aggressive dogs, the basolateral amygdala contained more androgen-receptor-positive neurons, but also more receptor-negative ones, and most receptors appeared unliganded, which points to additional mechanisms (Jacobs et al., 2006). This rests on one small canine study.

• Castration most consistently reduces urine marking, mounting and roaming, which improved by at least 50% in at least 60% of 57 adult dogs, while marked improvement in aggression can be expected in fewer than a third (Neilson et al., 1997). In more than 13,000 dogs, gonadectomy at any age showed no association with aggression toward familiar people or dogs and only a minimal increase toward strangers (Farhoody et al., 2018).

• Timing matters, though not in a simple way: age at castration did not predict which males changed (Hart & Eckstein, 1997), gonadectomy at six months or younger was associated with increased odds of behavioral disorders in Vizslas (Zink et al., 2014), and longer lifetime hormone exposure was linked to fewer reported fear- and aggression-related behaviors (McGreevy et al., 2018). Learned aggression persists regardless of hormones.

• Aggression is multifactorial — genetics, other hormones and neurotransmitters, learning, pain, fear and environment all contribute (Arroube & Pereira, 2025). Undiagnosed pain is a frequent, often overlooked contributor (Mills et al., 2020). Decide on castration case by case, treat learned behavior directly, and rule out pain and fear first.

References

Arroube, A., & Pereira, A. F. (2025). Dog neuter, yes or no? A summary of the motivations, benefits, and harms, with special emphasis on the behavioral aspect. Animals, 15(7), 1063. https://doi.org/10.3390/ani15071063

Farhoody, P., Mallawaarachchi, I., Tarwater, P. M., Serpell, J. A., Duffy, D. L., & Zink, C. (2018). Aggression toward familiar people, strangers, and conspecifics in gonadectomized and intact dogs. Frontiers in Veterinary Science, 5, 18. https://doi.org/10.3389/fvets.2018.00018

Hart, B. L., & Eckstein, R. A. (1997). The role of gonadal hormones in the occurrence of objectionable behaviours in dogs and cats. Applied Animal Behaviour Science, 52(3–4), 331–344. https://doi.org/10.1016/S0168-1591(96)01133-1

Hopkins, S. G., Schubert, T. A., & Hart, B. L. (1976). Castration of adult male dogs: Effects on roaming, aggression, urine marking, and mounting. Journal of the American Veterinary Medical Association, 168(12), 1108–1110.

Jacobs, C., Van Den Broeck, W., & Simoens, P. (2006). Increased number of neurons expressing androgen receptor in the basolateral amygdala of pathologically aggressive dogs. Journal of Veterinary Medicine Series A, 53(7), 334–339. https://doi.org/10.1111/j.1439-0442.2006.00840.x

Knol, B. W., & Egberink-Alink, S. T. (1989). Androgens, progestagens and agonistic behaviour: A review. Veterinary Quarterly, 11(2), 94–101. https://doi.org/10.1080/01652176.1989.9694205

McGreevy, P. D., Wilson, B., Starling, M. J., & Serpell, J. A. (2018). Behavioural risks in male dogs with minimal lifetime exposure to gonadal hormones may complicate population-control benefits of desexing. PLoS ONE, 13(5), e0196284. https://doi.org/10.1371/journal.pone.0196284

Mills, D. S., Demontigny-Bédard, I., Gruen, M., Klinck, M. P., McPeake, K. J., Barcelos, A. M., Hewison, L., Van Haevermaet, H., Denenberg, S., Hauser, H., Koch, C., Ballantyne, K., Wilson, C., Mathkari, C. V., Pounder, J., Garcia, E., Darder, P., Fatjó, J., & Levine, E. (2020). Pain and problem behavior in cats and dogs. Animals, 10(2), 318. https://doi.org/10.3390/ani10020318

Neilson, J. C., Eckstein, R. A., & Hart, B. L. (1997). Effects of castration on problem behaviors in male dogs with reference to age and duration of behavior. Journal of the American Veterinary Medical Association, 211(2), 180–182. https://doi.org/10.2460/javma.1997.211.02.180

Puurunen, J., Hakanen, E., Salonen, M. K., Mikkola, S., Sulkama, S., Araujo, C., & Lohi, H. (2020). Inadequate socialisation, inactivity, and urban living environment are associated with social fearfulness in pet dogs. Scientific Reports, 10, 3527. https://doi.org/10.1038/s41598-020-60546-w

Roulaux, P. E. M., van Herwijnen, I. R., & Beerda, B. (2020). Self-reports of Dutch dog owners on received professional advice, their opinions on castration and behavioural reasons for castrating male dogs. PLoS ONE, 15(6), e0234917. https://doi.org/10.1371/journal.pone.0234917

Salonen, M., Sulkama, S., Mikkola, S., Puurunen, J., Hakanen, E., Tiira, K., Araujo, C., & Lohi, H. (2020). Prevalence, comorbidity, and breed differences in canine anxiety in 13,700 Finnish pet dogs. Scientific Reports, 10(1), 2962. https://doi.org/10.1038/s41598-020-59837-z

Zink, M. C., Farhoody, P., Elser, S. E., Ruffini, L. D., Gibbons, T. A., & Rieger, R. H. (2014). Evaluation of the risk and age of onset of cancer and behavioral disorders in gonadectomized Vizslas. Journal of the American Veterinary Medical Association, 244(3), 309–319. https://doi.org/10.2460/javma.244.3.309