Guide · updated from the protocol library

Separation anxiety in dogs: what actually works, and what only sounds like it does.

Everything below is rendered from the same sourced library our paid plans are built from — the protocols with their real attributions, the mistakes with the reason they backfire, and the places where the published sources contradict each other. Where the field has no data, this page says so instead of filling the gap.

• During active treatment, the dog cannot be left alone beyond its current trained threshold — real-world absences must be covered (sitter, daycare, work, family) rather than left to happen solo.non-negotiable

  ASPCA, stated plainly: 'During treatment for separation anxiety, your dog cannot be left alone except during your desensitization sessions.' Mechanism (Polite Paws, height-phobia analogy): forcing exposure beyond the dog's threshold reinforces panic rather than resolving it — 'they associate your departure with panic, and anxiety escalates even before reaching their previous threshold.' A single over-threshold absence is framed as a real setback: 'that single breach of trust can cause setbacks.' ASPCA independently lists 'progressing too quickly' among the specific things that worsen SA.

• Never scold, punish, or use aversive tools (shock/prong/choke collars, anti-bark collars, leash corrections) for anxiety-related damage, accidents, or vocalization.non-negotiable

  AVSAB position statement: aversive methods 'should not be used under any circumstances'; adverse effects include 'inhibition of learning, increased fear-related and aggressive behaviors, and injury.' Because fear/anxiety often underlie the behavior, 'aversive techniques often make the behaviour issue even worse.' ASPCA, direct: 'Do not scold or punish your dog. Anxious behaviors are not the result of disobedience or spite.' Cross-study convergence: the IAABC literature review found 'eliminate punishment' the MOST frequently recommended intervention across 14 reviewed studies (cited in 7), ahead of desensitization itself (5/14).

• Arrange temporary coverage during the training period — work, a sitter, family/friends, daycare (preferably small home-based operations), or bringing the dog along — rather than leaving the dog alone.

  This is how mgmt.suspend-absences gets implemented in real life. ASPCA and VCA both list the same menu (work, sitter, family, daycare, car in moderate weather only/never heat). Polite Paws frames it explicitly as temporary: 'a means to success during the learning phase,' not a permanent lifestyle change. No source gives a rigorous head-to-head comparison of these options' tradeoffs — treat any ranking as this product's own judgment call, not sourced.

• Keep departures and arrivals low-key: no excited greetings, no big emotional goodbyes.

  VCA: 'Both punishment and excited greetings must be avoided' at homecoming — recommended practice is to ignore the dog until it settles (10-15 minutes), then give attention. High-arousal departures/arrivals reinforce that leaving/returning are big emotional events, working against the goal of making them unremarkable (Patricia McConnell's framing, blog-level source only — her book itself was not accessible).

• Choose crate vs. room confinement vs. free-roam based on which setting the individual dog is calmest in (per camera assessment) — not a blanket rule. Switch away from a crate immediately if it produces heavy panting, salivation, frantic escape attempts, or persistent vocalizing.

  ASPCA: crating 'can be helpful for some dogs if they learn that the crate is their safe place,' but switch to a baby-gated single room if escape/distress signs appear. VCA is more cautionary: crates 'should be used with caution because they can promote intense escape attempts and may result in fairly serious injuries.' Confinement anxiety (a distinct differential, see shard 01) means some dogs get WORSE, not better, in a crate regardless of SA status — this is the deciding diagnostic question, not a fixed preference.

• Scold or punish the dog for destruction, house-soiling, or vocalization that happened during their absence.

  These are anxiety symptoms, not disobedience — AVSAB: aversive methods increase 'inhibition of learning, increased fear-related and aggressive behaviors.' Cited in 7 of 14 studies reviewed by IAABC as the most consistently recommended thing to eliminate.

• Use an anti-bark or shock collar to stop distress vocalization during absences.

  Barking may stop short-term out of fear while the underlying anxiety keeps building, later resurfacing as other behaviors (destroying doors, breaking windows, house-training accidents) once the vocal outlet is suppressed. Not independently verified against a named peer-reviewed study — tracks the AVSAB punishment position but is search-aggregated, multi-source synthesis.

• Leave the dog alone for a full-length absence, hoping the fear will 'burn itself out' (flooding / 'cry it out').

  'A dog who is being flooded is usually over threshold, which means the situation is too intense for them to think clearly, learn effectively, or respond calmly' — rather than extinguishing fear, it typically intensifies it. This is functionally the exact over-threshold absence that mgmt.suspend-absences exists to prevent. No single named peer-reviewed study on flooding specifically for canine SA was located; this is general dog-behavior-science consensus reporting.

• Get a second dog, hoping company will fix the first dog's separation anxiety.

  True SA is about the absence of the attachment figure specifically, not aloneness in general — 'the presence of another dog is not likely to be a replacement for the human company that your dog seeks.' Anxiety more commonly spreads dog-to-dog than confidence does, and the two dogs can become interdependent, creating a second SA case. IAABC literature review independently confirms: 'second dog presence' is 'generally not associated with improvement.'

• Give big, excited hellos and emotional goodbyes.

  VCA recommends ignoring the dog until it settles (10-15 min) instead — high-arousal comings/goings reinforce that departures and arrivals are significant emotional events, working against the goal of making them boring/unremarkable.

• Blame the dog's SA on being 'spoiled' or overly attached — too much affection, sleeping in the bed, being carried — and consider withdrawing affection to fix it.

  Directly contradicted by a controlled study: Parthasarathy & Crowell-Davis (2006) formal attachment test, 32 SA dogs vs. 43 controls, found 'no significant difference in the type of proximity-seeking behaviors' between groups; dogs with SA 'spent no more time in contact with or proximity to their owners' than dogs without. IAABC's literature review independently corroborates this by name.

• Exercise the dog heavily right before leaving, expecting it to sleep through the absence.

  Separation anxiety is a panic/fear response, not an excess-energy problem — 'the anxious dog is not behaving this way because they're bored or from lack of exercise, but because they're having a panic attack when they're left alone.' No controlled study was located testing exercise as sufficient treatment for diagnosed SA. A commonly repeated figure ('physically fit dogs often have 20-30% lower anxiety rates') could not be traced to any study — flagged as folklore.

• Rely on a food-stuffed Kong/puzzle toy at departure, expecting it to resolve the anxiety.

  Overall, directly: 'Food toys will not "fix" a dog with separation anxiety, but they are good monitors of improvement in the dog's distress. Distressed dogs cannot eat.' For moderate-to-severe anxiety the toy is 'usually ignored or add[s] frustration'; a refused toy is itself a distress signal, not a failed enrichment attempt.

• Assume a crate automatically makes the dog feel 'safe,' like a den, without checking how the dog actually responds in it.

  Overall, directly: crating 'has less to do with making the dog feel "safe" — something often stated but never proved — than it does with minimizing damage and protecting the dog.' VCA separately warns crates 'can promote intense escape attempts and may result in fairly serious injuries' for SA dogs, and a distinct differential (confinement anxiety) means some dogs get WORSE, not calmer, when crated.

• Leave music or the TV on, expecting it alone to calm a genuinely panicking dog.

  Deborah Wells (Queen's University Belfast) found that for dogs already in acute separation distress, classical music produced 'little to no measurable benefit, and silence was sometimes better.' Most background-noise research was done with radio/music, not TV, so TV-specific claims are unproven. Overall herself recommends noise as environmental normalcy, not as anxiolytic treatment.

• Plan around a supposed 48-72 hour cortisol-recovery window — e.g., spacing training sessions or stressful outings far apart believing the dog needs days to physiologically 'reset.'

  Traced and debunked: the IAABC Foundation Journal article's own author, trying to cite the claim, 'was unable to find any evidence that this information is based on scientific findings.' The actual evidence: 'cortisol... has been shown to leave a dog's system 30 to 60 minutes following a stressor' — roughly 50-100x faster than the popularly repeated figure. A related claim ('4-6 weeks to detox from chronic stress') is flagged by the same author as equally unevidenced.

• Give an L-theanine supplement (e.g. Anxitane), expecting it to address separation-specific anxiety, on the strength of broad 'calming supplement' marketing.

  skeptvet's review: 'the research evidence in dogs is limited and weak, and it is not possible to say with confidence whether l-theanine has a meaningful benefit for dogs with anxiety.' One source explicitly states Anxitane 'can be useful for certain anxiety issues but not for anxiety issues related to separation' — i.e. this specific indication is excluded by at least one source, despite the general marketing framing.

What success actually meansNo universal fixed-duration target across sources. Naismith (FAQ): success = the owner's real-life target duration — 'leave for normal amounts of time — work, errands, social events — without any distress' — not a fixed number. A separate, narrower reported figure defines success as one hour or more of alone time for a specific cohort (families training 3x/week for 6 months); this is flagged as an inconsistency with her general FAQ definition, not resolved into one number. Overall (the strictest, most authoritative primary source) explicitly rejects 'absence of destructive behavior' as sufficient: 'It is not sufficient that your dog does not bark or destroy something when left alone' — the bar is an internal calm state (happy looks, lowered heart rate, slowed respiration). No source claims universal full 4-8h solo tolerance as the goal. Treatment 'may be lifelong,' and relapse can follow household changes (a death, a move, a family member leaving) — Overall recommends proactively planning around any known upcoming change.

• “The '72-hour cortisol vacation': after a stressful event, it takes 24-72 hours (sometimes stated as 48-72h) for a dog's cortisol/stress hormones to return to baseline, so give several days of total quiet after any bad walk or trigger exposure.”

  The IAABC Foundation Journal author (Luisa Depta) who had herself previously advocated 'cortisol vacations' using this exact figure went looking to cite the underlying research and 'was unable to find any evidence that this information is based on scientific findings.' She published a correction in her own professional journal, also flagging the related claim that 'chronically stressed dogs require four to six weeks to detox from stress' as equally unsupported.

  What is true insteadCortisol has been shown to leave a dog's system roughly 30-60 minutes following a stressor (peaking ~15-30 min post-event, normalizing within about 60 min per Beerda et al.'s data); estimated cortisol half-life is ~66 minutes — a timescale roughly 50-100x shorter than the popular claim. Slower-recovering systems (broader HPA-axis regulation, sleep, general arousal baseline, behavioral carryover) may genuinely take longer, which is a defensible, hedged reason to still give a dog a calm day after a bad outing — but that is a different claim from 'cortisol takes 3 days.'

• “Separation anxiety is caused by the dog being 'spoiled,' overly clingy, or hyperattached to the owner — too much affection, sleeping in the bed, being carried, and similar.”

  Parthasarathy & Crowell-Davis (2006) ran a formal attachment test on 32 dogs with owner-reported SA vs. 43 without, followed by 30-minute videotaped alone-time, and found 'no significant difference in the type of proximity-seeking behaviors exhibited by dogs with and without separation anxiety' — dogs with SA 'spent no more time in contact with or proximity to their owners during the attachment test than dogs without separation anxiety.' The IAABC Foundation Journal literature review independently corroborates this, citing the same study: 'hyperattachment behaviors were no more likely in SA dogs when compared to a control group.'

  What is true insteadSA is not based on degree of attachment ('hyperattachment'/spoiling); the study authors suggest a different attachment STYLE, not degree, may be a factor. Blaming affection or bed-sharing for SA is unsupported and risks shaming owners for behaviors that aren't actually causal.

• “Tire your dog out with a big walk or run right before you leave and it will sleep through your absence instead of panicking — physically fit dogs supposedly have '20-30% lower anxiety rates.'”

  No controlled study was located testing exercise as an intervention for diagnosed separation anxiety specifically. The '20-30% lower anxiety rates' figure could not be traced to any study in this research pass. SA is a panic/fear response to the owner's absence, not an excess-energy problem: 'the anxious dog is not behaving this way because they're bored or from lack of exercise, but because they're having a panic attack when they're left alone.'

  What is true insteadExercise is a plausible general-welfare practice, not shown to be sufficient treatment for SA on its own. No source in this corpus treats it as a fix.

• “Punish, scold, or correct a dog for growling at a person or another dog — 'nip the growling in the bud' before it 'becomes' aggression.”

  AVSAB: 'Aversive methods... should not be used under any circumstances' and 'There is no evidence that aversive methods are more effective than reward-based methods in any context.' The growl-specific mechanism: 'If you punish a dog for growling, the dog may stop growling, but they will still feel uncomfortable or scared. Without the growl as a warning, the dog may skip straight to biting when they feel threatened.' ASPCA corroborates: 'The ASPCA is opposed to any training equipment that causes a pet to experience physical discomfort or undue anxiety.'

  What is true insteadA growl is a warning signal, not misbehavior to eliminate — it sits on Kendal Shepherd's Ladder of Aggression below snapping/biting. Suppressing the growl removes the warning, not the underlying stress, and raises bite risk. Management (increase distance, remove the dog from the situation) plus counter-conditioning to the trigger is the sourced alternative.

• “A reactive dog just needs more exposure/socialization — take it to the dog park, let it 'work it out' by meeting more dogs, and it will grow out of the reactivity.”

  Cornell Riney Canine Health Center, direct: 'Immersing your dog in situations where they are not comfortable will increase their fear and may make their behavior worse,' and explicitly advises avoiding the dog park during the reactive-dog training period. AVSAB names flooding explicitly alongside pain- and intimidation-based tools as techniques an appropriate trainer should avoid. This is flooding — forced over-threshold exposure — not desensitization; the whole DS/CC and BAT literature depends on staying under threshold, which more exposure at a busy dog park guarantees will not happen.

  What is true insteadReactivity is treated by staying under threshold and building calm associations gradually (DS/CC, LAT, BAT 2.0, Engage-Disengage), not by more unstructured exposure. 'Your dog does not have to be friends with all people or every dog.'

• SA-like signs (house-soiling, vocalizing, pacing, destruction when alone) alongside any indicator of a possible medical driver: recent onset with no clear trigger, GI or urinary signs, restlessness that could be pain, vague/nonspecific symptoms, or comorbid panic-type signs.

• SA-like or aggressive behavior appearing suddenly, with no apparent trigger, in an adult or senior dog who tolerated being alone (or was otherwise stable) before.

• Escape attempts causing injury (broken teeth/claws, injured paws), self-mutilation (chewing own tail), or excessive self-licking/chewing beyond ordinary destructive chewing.

• Repeated or severe attempts to escape through windows, doors, or crates during owner absence.

• Dog consistently refuses food, treats, or a food-toy offered at the start of an owner's departure.

• A new or changed elimination pattern during absences (previously house-trained dog begins soiling, or the pattern of accidents changes) in the context of a suspected SA case.

• New anxiety, disorientation (getting lost in familiar places, going to the wrong side of a door), altered sleep-wake cycle (daytime sleeping, nighttime pacing), house-soiling, or activity-level change in a senior dog — the DISHAA pattern.

• Aggression when approached while lying down/resting, reluctance to move or exercise, or an owner describing the dog as "Jekyll and Hyde" with sudden unpredictable aggression bursts.

• New house-soiling, appetite change, or lethargy alongside a behavior complaint.

• House-soiling the owner describes as "behavioral," straining to urinate, or increased urination frequency.