Neutering and Behavior in Dogs: What the Evidence Shows

1. What the Question Actually Is

1.1 Two Different Decisions

Neutering for population control and neutering as a behavioral intervention are separate questions with separate evidence bases. The first has a clear public-health rationale. The second is a clinical claim about individual dogs, and it is the one examined here.

1.2 Why the Assumption Is Intuitive

Testosterone is associated with several behaviors that owners find difficult, and removing the gonads removes the primary source. The inference — less hormone, less behavior — is straightforward and, for a subset of behaviors, correct.

The problem is that the subset is narrower than the recommendation implies, and that the hormone modulates rather than drives most of what people want changed (a distinction the testosterone evidence sets out in detail).

1.3 Castration and Sterilization Are Not the Same Procedure

Two procedures are routinely conflated. Castration removes the gonads, so the hormone production goes with them; that is what every study in this article examined. Sterilization — vasectomy or tubal ligation — prevents reproduction while leaving the gonads and the hormones in place.

The distinction matters here because the entire behavioral literature concerns the first. Where the goal is preventing litters rather than changing behavior, the second achieves it without any of the effects discussed below, and it is rarely offered.

1.4 What Would Count as Evidence

A behavior that is hormone-dependent should decline after gonadectomy across populations, consistently, and independently of why the dog was neutered. That last condition matters more than it sounds, and section 9 returns to it.

1.5 Why the Recommendation Is So Widespread

Three reasons, and none is the evidence.

It is available. A behavioral problem that can be addressed by a surgical appointment requires nothing of the owner beyond attending it, unlike a training program that requires months of consistency.

It has a plausible mechanism. Testosterone is associated with several behaviors owners find difficult, and removing the source is an intuitive intervention regardless of whether the association is causal.

And it aligns with a recommendation already being made for other reasons. Where population control is the default advice, a behavioral rationale attaches to it easily.

1.6 How the Evidence Is Structured

Three lines run through this article and answer different questions. Comparative — do neutered and intact dogs differ, with the obvious confound removed. Dose-based — does the amount of lifetime hormone exposure relate to behavior. And health — what does timing do to joint and cancer risk.

They were designed independently and are read here together, which is why the convergences and divergences between them carry weight (with the general value of converging designs documented elsewhere).

2. What Neutering Reliably Changes

2.1 The Short List

Three behaviors show the most consistent response: roaming, urine marking, and mounting. These are the behaviors most directly linked to reproductive function, and the classic work on gonadal hormones and objectionable behavior in dogs and cats established them as the reliable targets (Hart & Eckstein, 1997). Roaming in particular is also driven by olfactory information that surgery does not remove (the channel dogs rely on most).

2.2 The Size of the Effect

Even here the effect is partial rather than absolute. A proportion of dogs show marked improvement, a proportion show some, and a proportion show none — and the behaviors in question also have a learned component that surgery does not touch. A dog that has marked indoors for two years has a habit as well as a hormone, and habits persist through their own mechanisms (which is why removing the driver rarely removes the behavior).

2.3 What Is Not on the List

Fear, anxiety, resource guarding, reactivity on leash, and most forms of aggression are not among the behaviors consistently shown to respond to neutering (which the reactivity literature would predict). This is where the recommendation and the evidence part company.

2.4 Why These Three and Not Others

The pattern is not arbitrary. Roaming in search of a mate, urine marking as a territorial and reproductive signal, and mounting are the behaviors most directly tied to reproductive function — which is what gonadectomy removes.

Everything else on the list of things owners want changed is downstream of emotion, learning or physical state rather than of reproductive hormones.

2.5 What "Partial" Means in Practice

The improvement is not uniform, and the variation has an identifiable source: how long the behavior has been performed.

A dog that has marked indoors for two years has both a hormonal driver and a well-established habit. Removing the first leaves the second, which is why the response rates reported for these behaviors are proportions rather than certainties.

2.6 What Mounting Actually Is

Worth separating, because it is frequently misread. Mounting occurs in reproductive contexts and also in play, in arousal, and as a displacement behavior under conflict.

Only the first is hormone-dependent. A dog that mounts when excited or overwhelmed is doing something gonadectomy will not address (with arousal having its own drivers).

3. The Aggression Data

3.1 The Design That Matters

The largest direct test used the Canine Behavioral Assessment and Research Questionnaire, a validated instrument, on an initial pool of 15,370 dogs. Critically, the analysis excluded dogs that had been neutered to correct a behavior problem — which removes the most serious confound in this literature, since dogs neutered because they were already aggressive would otherwise make neutering look causal (Farhoody et al., 2018).

After exclusions, 13,795 dogs were analyzed for aggression toward familiar people, 13,498 toward strangers, and 13,237 toward other dogs.

3.2 The Result

Neither gonadectomy nor age at gonadectomy showed an association with aggression toward familiar people or toward other dogs. For aggression toward strangers there was a low but statistically significant increase in the odds of moderate or severe aggression in neutered dogs compared with intact dogs (Farhoody et al., 2018).

3.3 An Earlier Finding in the Same Direction

The pattern is not new. A study of 1,842 dogs neutered and adopted between six and twelve months of age, surveyed two to three years later, found that aggression toward family members and toward strangers increased in males gonadectomized before five and a half months (Spain, Scarlett & Houpt, 2004).

That study is older, smaller and used a different design, which is precisely why the convergence is worth noting: two independent datasets, fifteen years apart, produced the same directional result on early neutering in males (with early development shaping much of what follows).

3.4 How to Read That

The honest reading is not "neutering causes aggression." The effect on strangers was small, the design is cross-sectional, and questionnaire data carry their own limitations (as measurement problems go generally).

The reading that does hold is narrower and still decisive for practice: in the largest dataset available, with the obvious confound removed, neutering was not associated with less aggression toward familiar people, strangers or other dogs. Those three categories are not the whole of canine aggression, but they cover what the recommendation is usually meant to address — and in none of them did the procedure perform as expected (with aggression being situation-specific rather than trait-like in any case).

3.5 What the C-BARQ Categories Cover

The instrument separates aggression by target and context, which is why the three categories are reported separately rather than pooled.

Aggression toward familiar people covers household members and regular visitors. Toward strangers covers unfamiliar people approaching or entering. Toward other dogs covers conspecific encounters. Each has its own item set and its own scoring.

3.6 What Falls Outside Them

Several forms do not appear in this analysis. Resource guarding is measured separately in C-BARQ and was not among the three categories reported. Aggression associated with pain, with handling, or directed at other animals sits outside the frame entirely.

That is the limitation stated in section 3.4, given in specifics: the study covers what the recommendation usually addresses and not the whole of canine aggression (with guarding having its own literature).

3.7 Why the Confound Exclusion Is the Study's Main Contribution

Without it, the analysis would compare a group containing dogs neutered because they were aggressive against a group that never gave anyone reason to. Any association found would be an artifact of that selection.

Removing those dogs is what turns a meaningless comparison into an interpretable one, and it is why this study carries more weight than the larger literature that preceded it.

4. The Hormone Exposure Studies

4.1 A Different Question

Rather than comparing neutered against intact, a second approach asked how much of a dog's life had been spent with gonadal hormones present. The measure — percentage lifetime exposure to gonadal hormones — treats the question as continuous rather than binary (McGreevy, Wilson, Starling & Serpell, 2018). It is a more informative framing than neutered-versus-intact, and it carries a statistical problem that section 9 sets out (of the kind that recurs wherever composite measures are used).

4.2 The Male Result

Across 6,235 male dogs castrated before ten years of age for reasons other than behavioral management, forty behaviors differed between intact and castrated dogs. Twenty-five were associated with hormone exposure and fourteen with age at castration.

Only two behaviors were more likely in dogs with longer hormone exposure: indoor urine marking and howling when left alone. Longer exposure was associated with significantly reduced reporting of twenty-six mostly unwelcome behaviors — of which eight related to fearfulness and seven to aggression (McGreevy et al., 2018).

4.3 What That Inverts

The expected pattern is that hormones drive unwanted behavior and removing them reduces it. The observed pattern was the reverse for nearly everything measured except the two behaviors that are directly reproductive.

The study is observational and does not establish that hormone exposure itself caused these differences — dogs neutered earlier may differ from dogs neutered later in ways the analysis cannot capture.

The authors framed the practical consequence as a paradox: castration may reduce the number of unwanted dogs while increasing the likelihood of problem behaviors that make individual dogs more likely to be surrendered (McGreevy et al., 2018).

4.4 How the Measure Works

Percentage lifetime gonadal hormone exposure is calculated as the proportion of the dog's life spent intact — a dog neutered at twelve months and surveyed at six years has spent about 17 percent of its life with gonadal hormones present.

The advantage over a binary comparison is that it captures timing and duration in one variable rather than treating a dog neutered at six months and one neutered at six years as the same case.

4.5 What Was Excluded and Why

Dogs neutered for behavioral management were removed from both analyses, as in the aggression study. Dogs neutered after ten years were also excluded, because a very high exposure percentage in an old dog carries a different meaning than the same figure in a young one.

4.6 The Two Behaviors That Went the Other Way

Indoor urine marking and howling when left alone were more likely in dogs with longer hormone exposure (McGreevy et al., 2018) — and both are reproductive or separation-related rather than emotional in the sense the other 26 measures capture.

Marking is directly hormone-dependent. Howling when alone sits closer to separation-related behavior, which has its own drivers (set out separately).

5. Where Males and Females Differ

5.1 The Parallel Study

The same method was applied to 8,981 female dogs spayed before ten years of age, again excluding behavioral indications. Twenty-three behaviors differed between intact and spayed dogs, twelve associated with hormone exposure and five with age at spay (Starling, Fawcett, Wilson, Serpell & McGreevy, 2019).

5.2 The Weaker Signal

Two behaviors were more likely with longer hormone exposure — chewing and howling. Ten mostly unwelcome behaviors were reported less often, of which one related to fearfulness and three to aggression (Starling et al., 2019).

Compare the male figures: eight fearfulness-related and seven aggression-related. The direction is the same; the magnitude is not.

5.3 Why This Matters Practically

Advice about neutering is routinely given without reference to sex, and these two studies used identical methods on the same instrument. The behavioral case against early neutering is substantially stronger for males than for females — which is not how the discussion is usually conducted (and individual variation complicates it further).

5.4 Why the Sex Difference Might Exist

Three accounts are available and the data separate none of them.

Hormonal. Testosterone and estrogen act on different receptor distributions, and removing them removes different things. The male effects concentrate on fear and aggression; the female findings do not.

Developmental timing. Females are typically neutered at a different point relative to skeletal and behavioral maturation than males in the same population, which confounds sex with timing.

Measurement. The behaviors measured — many of which concern aggression and fear — may simply be less frequent in females, leaving less variance for the analysis to detect.

5.5 What Follows Regardless

Whichever account holds, the practical implication is the same: a recommendation derived from the male data should not be transferred to a female, and the reverse.

That is not how neutering advice is usually given, and both studies used the same instrument on the same measures, which makes the comparison unusually direct.

6. Timing and Physical Health

6.1 The Shift in the Question

The health literature has moved from whether to when. Neutering in the first year became routine in the U.S. and much of Europe, and subsequent work indicates that for some breeds this is associated with increased risks of debilitating joint disorders and certain cancers (Hart, Hart, Thigpen & Willits, 2020a).

The joint disorders examined are hip dysplasia, cranial cruciate ligament tear or rupture, and elbow dysplasia. The cancers are lymphoma, mast cell tumor, hemangiosarcoma, and osteosarcoma.

6.2 The Breed-Specific Signal

The pattern first became visible in single-breed work. An evaluation of gonadectomized Vizslas examined both cancer risk and behavioral disorders and found elevated risk with earlier gonadectomy, with age of onset also affected (Zink et al., 2014).

That is one breed, and nothing in it transfers to another: single-breed studies of this kind cannot be generalized, and that limitation turned out to be the finding: the effect sizes differ enough between breeds that a single recommendation does not survive contact with the data.

6.3 Breed and Body Size

The findings across 35 breeds do not generalize into a single recommendation, which is the point of the paper: the risks differ substantially by breed. A parallel analysis of mixed-breed dogs across five weight categories found small breeds generally showing little vulnerability compared with larger-bodied dogs (Hart et al., 2020b).

Body size, not breed identity as such, carries much of the pattern (which is consistent with what breed explains generally).

6.4 Where Behavior and Health Converge

Both literatures point the same way on timing even though they measure different outcomes, and neither was designed with the other in mind. Early neutering is associated, in several large studies, with more unwanted behavior in males, and in larger breeds with elevated joint and cancer risk. Neither finding argues against neutering as such. Both argue against doing it early by default.

6.5 What the Breed Data Actually Contain

The 35-breed analysis is a lookup table rather than a conclusion, and using it as one requires reading it correctly.

For each breed it reports the incidence of joint disorders and cancers in intact dogs, and in dogs neutered within several age bands — under 6 months, 6 to 11 months, 1 year, 2 to 8 years. The recommendation for each breed follows from where those incidences diverge, not from a general principle applied to all of them (Hart et al., 2020a).

6.6 Why the Recommendations Differ So Much

Some breeds show substantial elevation in joint disorders with early neutering; others show none. Some show cancer effects in females and not males; others the reverse. A few show no significant effect at any age.

That is the finding, and it is why the paper exists: a single age recommendation across breeds would be wrong for most of them in one direction or the other.

6.7 The Mixed-Breed Analysis

For dogs without a known breed, the parallel analysis grouped by weight and found small breeds generally showing little vulnerability compared with larger-bodied dogs (Hart et al., 2020b).

That is the practical fallback where breed is unknown or uncertain, and it is a coarser instrument — body size is a proxy for growth trajectory rather than a cause in itself.

6.8 What the Design Cannot Deliver

Both papers draw on veterinary hospital records, which means the population is dogs presented for care rather than a random sample. Breeds are represented unevenly, some with a few hundred cases and some with far more.

And the analysis is retrospective: age at neutering was not assigned, so owners and veterinarians who chose early neutering may differ systematically from those who did not.

7. What This Means for a Decision

7.1 Behavior Is Not the Strongest Argument Either Way

Where a household is deciding, the behavioral evidence is the part least likely to settle it: the effects are modest, inconsistent in direction, and smaller than the individual variation around them.

7.2 The Question to Ask First

If the reason for neutering is a behavior problem, the first question is which behavior. Roaming, marking and mounting have a reasonable hormonal case. Fear-driven or defensive behavior does not, and the data suggest it may worsen (with fear learning proceeding on its own logic).

7.3 The Reversibility Problem

Surgical neutering is permanent, and the behavioral effects — in whichever direction — cannot be undone. Where the behavioral case is uncertain, temporary hormonal suppression allows the effect to be observed before the decision is made. This is a standard veterinary option and is underused.

7.4 What Neutering Cannot Do

It cannot resolve a learning history, and most behavior that owners want changed is maintained by consequences rather than by hormones (which is what makes the behavior persist). Where the behavior is anxiety-driven, the intervention that matters is the one addressing the anxiety (as the anxiety literature sets out), and where pain is involved, neither surgery nor training will substitute for a diagnosis (a route that is missed with striking regularity).

7.5 The Conversation Worth Having

Four questions structure it better than a yes-or-no framing.

What is the reason? Population control, a medical indication, or a behavior. The evidence differs completely between them.

If a behavior, which one? Roaming, marking and mounting have a defensible case. Fear-driven and defensive behavior do not.

What is the dog's breed and size? For the timing question, that is where the relevant data live (Hart et al., 2020a; 2020b).

Has a physical cause been ruled out? New or escalating behavior warrants examination before a surgical decision aimed at it.

7.6 What Cannot Be Undone

The asymmetry deserves stating plainly. If neutering helps, the benefit could also have been obtained later. If it does not help, or makes something worse, there is no route back.

That asymmetry is the argument for waiting where the case is uncertain — not an argument against the procedure.

8. What Neutering Is For

8.1 The Population Argument

The public-health case is separate from everything above and is not in dispute. Reducing unwanted litters reduces the number of dogs entering shelters, and where free-roaming populations exist the case is stronger still.

That argument concerns the population rather than the individual, and it is the reason blanket neutering policies exist.

8.2 The Tension the Data Create

The behavioral findings put the two in conflict. Castration may reduce the number of unwanted dogs while increasing the likelihood of problem behaviors that make individual dogs more likely to be surrendered (McGreevy et al., 2018).

The authors state this as a paradox rather than resolving it, and it is worth carrying as one: an intervention justified by relinquishment rates may raise a driver of relinquishment.

8.3 The Medical Indications

Two indications carry most of the weight, and both are better supported than the behavioral arguments discussed above.

Pyometra is the clearest. In Swedish insurance data covering more than 200,000 dogs, the crude twelve-month risk in intact bitches under ten years was around 2 percent, and survival estimates indicated that 23 to 24 percent of bitches would experience pyometra by ten years of age — with the proportion ranging between 10 and 54 percent depending on breed (Egenvall et al., 2001). Sweden is an unusually intact population, which is what makes the figures informative: they describe what happens when bitches are not routinely spayed. Ovariohysterectomy removes the risk entirely.

Benign prostatic hyperplasia is the most common prostatic disease in intact males and a major cause of canine infertility, with impaired semen quality, hematospermia and frequently concurrent chronic prostatitis (Domosławska-Wyderska, Zduńczyk & Rafalska, 2025). It is androgen-dependent, and castration resolves it. Medical alternatives exist — osaterone acetate and finasteride are used where breeding is to continue — which matters here because the condition does not force a surgical decision.

Testicular tumors are also cited, and that argument is weaker than it sounds: they are common in older intact males, largely benign, and removing both testicles to prevent a condition usually treated by removing both testicles is prophylaxis rather than therapy.

8.4 How These Sit Against the Behavioral Argument

They are individual medical decisions with their own evidential basis, and the behavioral literature addresses neither for nor against them. A dog with recurrent prostatic disease has a reason for castration that has nothing to do with how it behaves on a lead.

What follows is that a household should know which argument is being made. Where a medical indication exists it stands on its own; where one is asserted in support of a behavioral recommendation, two separate cases have been run together.

8.5 Where the Three Arguments Meet

Population control, individual health and behavior point in different directions for a given dog, and no framework resolves them into a single recommendation.

What the evidence supports is separating them in the conversation rather than allowing one to stand in for the others — which is what happens when a behavioral rationale is offered for a decision made on population grounds.

8.6 The Chemical Alternative

Where the behavioral case is uncertain, temporary suppression allows the effect to be observed before an irreversible decision.

Deslorelin implants and similar preparations suppress gonadal function for a defined period, after which function returns. If a behavior is hormone-dependent, the suppression period reveals it; if the behavior persists unchanged, surgical neutering will not address it either.

8.7 Why It Is Underused

It costs more per intervention, requires a return visit, and delivers an answer in months rather than immediately. It also produces an initial flare of hormone activity before suppression begins, which has to be explained to the owner.

None of those is a reason not to offer it where the indication is behavioral rather than medical or population-related — and the evidence in sections 3 to 5 makes that indication considerably weaker than it is usually presented.

8.8 What It Does Not Test

Suppression does not model the long-term physiological state of a neutered dog, and the joint and cancer findings in section 6 concern permanent absence of gonadal hormones during growth.

It answers the behavioral question and not the developmental one, which is a real limitation where the dog is still growing.

9. Where the Advice Comes From

9.1 Why Recommendations Differ by Country

Prevailing veterinary advice varies between regions for reasons that are partly historical and partly about population management rather than about the individual animal in front of anyone.

9.2 Why Practice and Evidence Diverged

The behavioral recommendation predates the evidence that tests it. Neutering to reduce aggression was standard advice long before the C-BARQ studies existed, and the studies arrived into a practice already established.

That sequence is familiar from other areas of this library, and it produces the same result: the finding has to displace a recommendation rather than inform one being formed (as puppy testing illustrates).

9.3 The Confounded Clinical Impression

Veterinarians and trainers observe that neutered dogs improve, and the observation is often accurate — because dogs neutered for aggression are the ones remembered, and any subsequent improvement is attributed to the surgery.

Regression to the mean does the rest: intervention typically happens when a problem is at its worst, so improvement is expected regardless. This is the exact confound the Farhoody exclusion removes, operating in everyday clinical impression rather than in a dataset.

9.4 What the Older Literature Said

Earlier work reported reductions in some forms of aggression after castration, and those studies did not exclude dogs neutered for behavioral reasons. The apparent contradiction with the newer findings largely dissolves once that difference is accounted for.

9.5 Why Correction Is Slow

A recommendation carried by two professions, aligned with a public-health message and supported by an intuitive mechanism does not change because three surveys reported null results.

Stating the evidence is what can be done. Expecting it to displace the practice quickly is not supported by how any of the other cases in this library have gone.

10. Summary at a Glance

Three behaviors respond reliably — Roaming, urine marking and mounting; the classic hormone-dependent set (Hart & Eckstein, 1997).

Aggression toward people or dogs is not among them — With behavioral indications excluded, neutering showed no association with aggression toward familiar people or other dogs, and a small significant increase toward strangers, across roughly 13,000–14,000 dogs per analysis (Farhoody et al., 2018).

Longer hormone exposure meant fewer problems, not more — In 6,235 males, longer exposure was associated with reduced reporting of 26 mostly unwelcome behaviors, including 8 fearfulness-related and 7 aggression-related (McGreevy et al., 2018).

Only two male behaviors went the expected way — Indoor urine marking and howling when left alone (McGreevy et al., 2018).

The female signal is weaker — In 8,981 females, 10 unwelcome behaviors were reduced with longer exposure, of which 1 related to fearfulness and 3 to aggression (Starling et al., 2019).

Breed recommendations differ substantially — Some breeds show marked elevation in joint disorders with early neutering, others none, and a few show no significant effect at any age (Hart et al., 2020a).

The health question is about timing — Early neutering is associated with elevated joint disorder and cancer risk in some breeds, with major breed differences (Hart et al., 2020a).

Body size carries much of it — Small breeds showed little vulnerability compared with larger-bodied dogs (Hart et al., 2020b).

Both literatures converge on timing — Neither argues against neutering. Both argue against neutering early by default.

11. Research Gaps and Critical Appraisal

All of this is owner-reported and cross-sectional. Every study cited here uses questionnaire data at a single time point. They establish associations across large populations and cannot establish that neutering caused any behavioral difference.

Observational designs carry confounders the analyses cannot remove. Nobody assigned the dogs to a group: owners and veterinarians decided who was neutered and when, and the people who choose early neutering differ systematically from those who do not. Shelter-origin dogs are routinely neutered before placement; breed, size and country shape the prevailing recommendation; urban and rural households differ in both the decision and the dog's daily life. Each of those is also plausibly related to behavior in its own right, which means a difference attributed to hormone exposure may belong to the circumstances that produced the exposure. Excluding behavioral indications removes the most damaging confounder (Farhoody et al., 2018; McGreevy et al., 2018) and does not remove this class of them.

The hormone exposure measure is confounded with age. Percentage lifetime exposure is calculated as a proportion of the dog's age at reporting. A dog neutered at six months and reported at ten years has a low value; the same dog reported at eighteen months has a high one. The measure therefore correlates inversely with current age at any fixed neutering age, and age has independent effects on behavior.

Confounding by indication is handled but not eliminated. Excluding dogs neutered for behavioral reasons is the right move (Farhoody et al., 2018; McGreevy et al., 2018), and it relies on owners correctly reporting why the procedure was done years earlier.

No randomized trial exists and none is likely. Randomizing dogs to neutering or not, at assigned ages, would answer the causal question and is not going to happen for practical and ethical reasons. This limitation is permanent, not temporary.

C-BARQ measures owner perception. It is a validated instrument, which means it measures consistently — not that it measures the dog's behavior independently of the owner's interpretation (a general constraint on this kind of data).

Breed-specific health data do not transfer. The joint and cancer findings are breed-specific by design, and the majority of breeds have not been examined. Applying a Golden Retriever result to an unstudied breed is not supported.

Effect sizes are frequently omitted in secondary coverage. The stranger-directed aggression increase was described by its own authors as low. Reporting that a study "found neutering increases aggression" without that qualifier misrepresents it.

11.1 What a Better Study Would Look Like

The design that would settle the behavioral question is describable and will not be run. Randomize dogs to neutering at assigned ages, follow them for several years, and measure behavior by direct observation rather than owner report.

The ethical objection is not to neutering but to the assignment: withholding or imposing a surgical decision for research purposes on companion animals owned by other people.

11.2 What Is Achievable

Prospective cohorts following dogs whose owners have already decided, with baseline measurement before the procedure. That removes the recall problem and the confound-by-indication in one step, and leaves selection intact.

Nothing of that kind has been published at the scale the question needs, which is why a literature this large still rests on cross-sectional survey data.

11.3 What the Owner-Report Basis Limits

Every behavioral finding here rests on C-BARQ, which measures owner perception of behavior. That instrument is validated in the sense that it measures consistently — not in the sense that it measures the dog independently of the person filling it in.

One direction of bias is worth naming: an owner who neutered a dog expecting improvement may report improvement. That would push the results toward finding a benefit, and the studies found none — which makes the null results more robust rather than less.

11.4 What Would Change the Picture

Three things, in order of feasibility. Prospective cohorts with pre-procedure baselines, which removes the recall problem. Direct behavioral observation rather than owner report, at least in a subsample. And breed-specific behavioral data, since the health findings show breed matters enormously and the behavioral analyses pool across breeds entirely.

The last is the most striking omission: the same literature that established large breed differences in physical outcomes has not asked whether behavioral outcomes differ by breed at all.

12. Conclusion

The recommendation that neutering will calm a dog down has an unusual property: it is widespread, it is offered by professionals, and the largest datasets addressing behavioral outcomes do not support neutering as a general calming intervention. With dogs neutered for behavioral reasons excluded, neutering showed no association with reduced aggression in any category tested, and a small increase toward strangers. In more than six thousand males, longer exposure to gonadal hormones was associated with fewer unwanted behaviors across twenty-six measures, including eight relating to fear and seven to aggression — the reverse of the expected pattern, with the only two exceptions being the directly reproductive ones. This does not demonstrate that hormones are protective: the studies are observational and cannot separate a hormonal effect from differences between dogs that were neutered earlier or later. The female data run in the same direction and considerably weaker, which is itself worth knowing given how rarely the advice distinguishes by sex. None of this makes neutering wrong, and none of it is causal evidence: these are questionnaire studies, the hormone exposure measure is entangled with age, and the randomized trial that would settle it will never be run. What the evidence does support is a narrower and more useful position than either camp usually takes. Neutering reliably addresses roaming, marking and mounting. It does not reliably address fear, anxiety or aggression, and longer hormone exposure was associated with fewer fear-related and aggression-related behaviors in males — which means early neutering may be associated with more of them. The question worth asking is therefore not whether but which behavior, and when — with the timing question now carrying weight from two independent literatures that were not designed to agree.

Key Insights (Takeaways)

• Neutering was not associated with less aggression in the three categories the largest dataset examined. With dogs neutered to correct behavior problems excluded, gonadectomy showed no association with aggression toward familiar people or other dogs, and a low but significant increase toward strangers, across roughly 13,000 to 14,000 dogs per analysis (Farhoody et al., 2018). The exclusion matters: without it, dogs neutered because they were aggressive would make the procedure look causal.

• Longer hormone exposure was associated with fewer unwanted behaviors, not more. Across 6,235 male dogs, longer lifetime exposure to gonadal hormones was associated with reduced reporting of 26 mostly unwelcome behaviors — 8 relating to fearfulness and 7 to aggression — while only indoor urine marking and howling when alone went the other way (McGreevy et al., 2018). This runs counter to the expectation behind the standard recommendation — though the study establishes association, not cause.

• The effect is substantially weaker in females. The same method applied to 8,981 spayed females found 10 unwelcome behaviors reduced with longer hormone exposure, of which one related to fearfulness and three to aggression (Starling et al., 2019), against eight and seven in males. Advice that does not distinguish by sex is not tracking the evidence.

• Three behaviors do respond reliably, and they are the reproductive ones. Roaming, urine marking and mounting remain the defensible targets (Hart & Eckstein, 1997) — partially, and with the learned component untouched by surgery.

• Population control, individual health and behavior are three separate arguments. They point in different directions for a given dog, and the behavioral literature does not address the medical indications or the public-health case. Allowing one to stand in for the others is what happens when a behavioral rationale is offered for a decision made on population grounds.

• The timing question now carries more weight than the yes-or-no question. Early neutering is associated with increased joint disorder and cancer risk in some breeds, with small breeds showing little vulnerability compared with larger-bodied dogs (Hart et al., 2020a; 2020b). Behavior and physical health point in the same direction on timing, from two literatures that were not designed to agree.

References

Domosławska-Wyderska, A., Zduńczyk, S., & Rafalska, A. (2025). Negative impact of benign prostatic hyperplasia on fertility in dogs – a mini-review. Frontiers in Veterinary Science, 12, 1582705. https://doi.org/10.3389/fvets.2025.1582705

Egenvall, A., Hagman, R., Bonnett, B. N., Hedhammar, Å., Olson, P., & Lagerstedt, A.-S. (2001). Breed risk of pyometra in insured dogs in Sweden. Journal of Veterinary Internal Medicine, 15(6), 530–538. https://doi.org/10.1111/j.1939-1676.2001.tb01587.x

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)01126-4

Hart, B. L., Hart, L. A., Thigpen, A. P., & Willits, N. H. (2020a). Assisting decision-making on age of neutering for 35 breeds of dogs: Associated joint disorders, cancers, and urinary incontinence. Frontiers in Veterinary Science, 7, 388. https://doi.org/10.3389/fvets.2020.00388

Hart, B. L., Hart, L. A., Thigpen, A. P., & Willits, N. H. (2020b). Assisting decision-making on age of neutering for mixed breed dogs of five weight categories: Associated joint disorders and cancers. Frontiers in Veterinary Science, 7, 472. https://doi.org/10.3389/fvets.2020.00472

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

Spain, C. V., Scarlett, J. M., & Houpt, K. A. (2004). Long-term risks and benefits of early-age gonadectomy in dogs. Journal of the American Veterinary Medical Association, 224(3), 380–387. https://doi.org/10.2460/javma.2004.224.380

Starling, M., Fawcett, A., Wilson, B., Serpell, J., & McGreevy, P. (2019). Behavioural risks in female dogs with minimal lifetime exposure to gonadal hormones. PLoS ONE, 14(12), e0223709. https://doi.org/10.1371/journal.pone.0223709

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