How to Pick a Service Dog Candidate: Temperament, Health, and Age -- What Actually Predicts Success
Ask ten people what makes a good service dog and nine will start naming breeds. It's the wrong first question. I've spent 11 years in behavioral science -- the last six as a Registered Behavior Technician -- and if there's one thing that field teaches you, it's that the learner's individual history matters more than the category they belong to. Dogs are no different. The right question isn't "what breed?" It's "what temperament, what health picture, and what age am I actually starting with?"
I want to be honest with you up front, because nobody else seems to be: even experienced service dog programs with decades of breeding behind them expect a substantial share of candidates to wash out. Choosing well doesn't guarantee success -- it shifts the odds. I trained my own service dog, Calipso, from the ground up, and I chose her the way this article describes. Here's the full framework, so you can apply it before you fall in love with a puppy you saw on the internet.
Temperament beats breed -- every time
Breed can tell you about size, energy, and some genetic health risks. It cannot tell you whether the individual dog in front of you can stay calm in a crowded grocery store. That answer lives in temperament, and it shows up early if you know how to look.
Confidence: not fearlessness -- fast recovery
The trait I care about most is not a dog who never startles. Startling is normal; it's ancient survival wiring. What predicts a working career is what happens in the next three seconds. A good candidate startles at the dropped metal bowl, recovers, and then walks over to investigate it. A poor candidate bolts, hides, and stays tense for the rest of the session. I've watched this one sequence predict more outcomes than any pedigree.
When evaluating any dog or puppy, deliberately create a small, harmless surprise (dropped keys, a bag tipping over) and watch the recovery, not the reaction.
Motivation: a dog who wants something is a dog you can teach
Every behavior is built on consequences -- that's the core of behavioral science, and it doesn't care about species. A service dog candidate needs a reinforcer bank you can actually draw from: strong food drive, toy drive, or genuine social motivation. If the dog is indifferent to everything you can offer, training turns into arm-wrestling. That's not a character flaw, but it is a harder road.
Run the "pocket test." Show food, a toy, and yourself (play and attention). A candidate who engages with at least two of the three gives you room to build almost any task.
Neutrality: the most underrated trait in service work
Service dogs spend their working lives ignoring things: other dogs, strangers who want to pet them, food on the floor, kids shrieking two aisles over. A dog who needs to greet everyone, or who fixates on other dogs, will fight that job every day. Neutrality isn't coldness -- it's the ability to notice the world without needing to engage with it.
Watch the dog around everyday chaos -- other animals, crowds, traffic. You're looking for relaxed observation, not longing stares, whining, or barking.
Resilient, not fragile
Public access training involves mistakes, awkward moments, and gradual exposure to stress. A resilient dog shrugs off a clumsy session and tries again. A fragile dog folds, and folding doesn't fix itself in public settings. Note the distinction: resilient does not mean "tough enough for harsh handling." It means the dog bounces back from ordinary training friction.
Pick the dog that keeps coming back for more after a confusing moment -- not the one that melts or the one that explodes.
Red flags I never talk myself out of
Over the years I've developed a rule: when I see one of these in a candidate, I stop rationalizing. Listing them here so you can borrow the rule.
A dog that startles hard and stays frightened -- no recovery within seconds, and the fear carries into the next interaction.
Fear-based aggression toward dogs or people in a young dog. Management exists, but you'd be building a career on a fault line.
Intense early resource guarding -- guarding food, space, or people fiercely before any training has occurred.
Shutting down in novel environments: freezing, refusing food, refusing to move. A dog like this can improve, but public access will be a long fight.
Noise sensitivity that doesn't soften with gentle, repeated exposure. Fire alarms and traffic are part of a working life.
Obsessive chasing of moving things -- cars, bikes, squirrels. Proofing against that drive is possible but slow, and it fails at the worst moments.
None of these makes a dog worthless. They just make that dog the wrong bet for service work specifically.
Health: the working-life math
A service dog isn't a pet who jogs occasionally. It's an athlete whose career depends on a body that holds up for eight to ten years of working. Health screening isn't paranoia -- it's due diligence on the asset doing the work.
Orthopedics first. Hips and elbows carry the load, literally. For larger breeds, ask for OFA or PennHIP results from the breeder. For rescues, at minimum get a hands-on orthopedic exam and watch the gait at a walk and a trot.
Vision and hearing. Many tasks rely on both, and some congenital issues show up gradually. A basic ophthalmic and hearing check is cheap insurance.
Sound structure. Watch the dog move. An efficient, symmetrical gait at this age predicts easier decades. Lopsided movement, bunny-hopping, or persistent lameness are genuine concerns, not quirks.
Full veterinary workup. Bloodwork, heart, skin, teeth. For rescue dogs especially -- you're building the future on unknowns, so shrink the unknowns early.
Spay/neuter timing. For large breeds this affects joint development; discuss timing with your vet rather than defaulting to an early date.
Age: what you're really buying at each stage
Age doesn't determine temperament, but it changes what you know and what you're betting on.
Puppy (8 weeks to 6 months)
You control the entire history -- that's the superpower. The socialization window between roughly 3 and 21 weeks is where confidence gets installed, and you can't get that time back later. The trade: you won't see the finished temperament for a year or more, and real public access work waits until social maturity around 18-24 months. Puppies are the long road.
Adolescent (6 to 18 months)
The stretch nobody warns you about. Hormones arrive, focus wobbles, and yesterday's perfect recall evaporates. It's survivable with structure, but temperament here is still developing -- you're betting on a moving target.
Young adult (18 months to 4 years) -- the sweet spot
Temperament is now visible and reasonably stable. Health history is checkable. Training can start immediately instead of waiting out puppyhood. This is why adult rescues with foster histories are quietly one of the best-kept secrets in owner-trained service work: someone has already observed this dog in a real home for months.
Senior (7+)
Older dogs can be wonderful companions and can learn, but a first service career starting late in life fights both physiology and time. I don't recommend it as a first candidate.
The 10-minute candidate snapshot
This isn't a certification and it isn't science's final word -- it's a filter you can run before investing months. Five checks, five minutes each if you're being generous.
Startle and recovery: drop something metal at a distance. Watch the seconds after, not the flinch itself.
Novel object: place a strange item (umbrella, bag) on the floor. Curious approach = good. Persistent avoidance = note it.
Motivation check: food, toy, social play. You want the dog to care about at least two.
Handling: touch paws, ears, tail briefly. Mild squirming is fine; panic or hard protesting is a flag.
Environment: a short walk somewhere new. Looking around calmly, checking in with you, sniffing -- good signs.
Run this on several candidates. Patterns across dogs teach you faster than any article can -- including this one.
Breeder or rescue?
Both paths produce service dogs. They just sell different things: the breeder sells predictability of health and lineage; the rescue sells an adult whose temperament is already observable. I chose to owner-train, and the path you pick matters less than doing the assessments above with clear eyes. Programs are a third route -- excellent dogs, long waitlists, significant cost, and no guarantee the match is right either.
The honest takeaway
No candidate is perfect. The goal isn't a flawless dog -- it's good odds plus a training method that matches how animals actually learn. Fast recovery, real motivation, genuine neutrality, a sound body, and an age where temperament is knowable: stack those, and the wash-rate math starts tilting your way.
Calipso passes every check in this article -- not because she was destined for the job, but because the job was built for her, step by step, using the same science I use every day as an RBT. How that building actually happens is the next post in this series: task analysis, and how one fluid service-dog task is really thirty small behaviors in a trench coat.
Next in the series: how a service-dog task is actually built, from a one-sentence definition to a dog who performs it anywhere.