The German Shepherd was built for a job and is now mostly kept as a pet. Almost every argument about the breed sits in that gap. This page sticks to what has been measured in German Shepherds, and says so when a number had to be borrowed from other dogs.
How big they get
Every figure in the panel above was taken off live young adult dogs on a scale and a measure, so it describes pets rather than a ring ideal.
Height is essentially finished by about 18 months. Weight keeps filling in until roughly two years, as the chest deepens and muscle builds. A tall, narrow 16-month-old is normal and will not stay that shape.
The sex gap, measured two ways
Most breed pages treat the male-female difference as a footnote. Here it is the main event, and it has been measured better than in any other breed we know of.
Start with the written standard. The American standard asks for males of 24 to 26 inches and females of 22 to 24, measured “at the top of the highest point of the shoulder blade.” Those two ranges touch at exactly one point. A clean two-inch offset between the sexes is rare in a breed standard.
Now the skeleton. A 2020 bone study measured seven lengths and widths on each of four limb bones (the humerus and radius in the front leg, the femur and tibia behind) in 25 German Shepherds, 12 male and 13 female. Males were bigger on average for every single measurement.
The paper reports a mean sexual dimorphism ratio of 1.07. Sexual dimorphism just means the size gap between the sexes. The difference was statistically solid for 92.8% of the measurements.
The striking bit is what one bone could do. Working from bone measurements alone, the authors identified a dog’s sex correctly between 82% of the time using the radius and 93% using the femur, the thigh bone. That is the highest score reported for sexing a dog from any one skeletal element.
Now put the two kinds of evidence side by side, which we can do because we hold the bodyweight data. A ratio of 1.07 is a linear ratio: it compares lengths. Mass goes roughly with the cube of a length, and 1.07 cubed is about 1.22. The male and female midpoints in our own dataset give a ratio of 1.22.
Twenty-five skeletons and 115,000 live dogs, measured by completely different methods, agree to two decimal places. Take the sample size caveat seriously: 25 dogs, in a study built for archaeology rather than health research. But the match is real, and almost no other breed lets you run that check.
What actually predicts your puppy’s adult size
Guide-dog charities have a hard reason to care about adult size: a dog outside the workable range cannot be placed. One programme kept full records from 1979 to 1997 and analysed them.
Across 1,333 German Shepherd offspring, the heritability of mature bodyweight was 0.57. For height it was 0.35. Heritability is the share of the differences between dogs that is explained by their genes.
So more than half the spread in adult weight came down to breeding. Look at the parents. If you cannot, our calculator blends the breed average with your puppy’s own growth curve, and leans harder on the curve as the dog gets older.
One detail for context. That programme aimed at 18 to 32 kg, which is smaller than the FCI standard’s 30 to 40 kg for males. A guide-dog charity selects for a size that works in harness, not one that wins in a ring.
The three-year myth
“German Shepherds grow until they’re three” appears on page after page. We went looking for where it comes from and found nothing.
Here is what the primary research says instead:
- The largest growth study in dogs, built on more than six million young dogs, says larger breeds need up to 24 months to reach adult bodyweight. German Shepherds sit inside its validated size band, so those curves really do apply here.
- The guide-dog programme took mature weight as the last weight recorded between 300 and 600 days, about 10 to 20 months, and measured mature height at around 14 months.
- The 115,000-dog population study counts dogs of 12 to 24 months as young adults.
Three datasets, three methods, all landing near 18 to 24 months. Every source we found offering “three years” was a content farm or a breeder page, and none cited anything.
The likely explanation is a mix-up. Bone growth finishes in that 18 to 24 month window. Chest depth, muscle and behavioural maturity keep going to two or three years, which owners genuinely do notice. Calling that “growing” is understandable, but it is not what the charts measure, and it leaves people waiting for height that will never arrive.
What living with one is actually like
Start with the finding that should change how you read any breed page, this one included.
A 2022 study surveyed 18,385 dogs and genotyped 2,155 of them. Breed explained about 9% of the behaviour differences between one dog and another. Two German Shepherds can differ from each other nearly as much as two dogs of different breeds do.
It did pick up something breed-shaped here. A German Shepherd was less likely than a randomly chosen dog to score in the top quarter for human sociability, by a factor of 0.78. Labradors and Golden Retrievers went the other way. Being less outgoing with people is a real tendency in this breed: a tendency, not a rule.
What breed-specific work adds
The best behaviour dataset in this breed comes from 1,041 German Shepherds in the UK and Sweden, scored on the C-BARQ, a long owner questionnaire that turns everyday behaviour into numbers. The heritability of the traits it measured was low to moderate: 0.23 for playfulness with people, 0.18 for fear of non-social things like noise, 0.18 for interest in strangers, and lower or unclear for the rest.
Read that plainly. Even inside one breed, and even for the traits genes touch most, most of the difference between two German Shepherds is not written in their genes. How a puppy is raised is doing a lot of the work.
The aggression question, answered with numbers
This gets asked about German Shepherds more than anything else, so here is what has been counted.
In a large owner survey using the same questionnaire, 292 German Shepherds were scored. 16.4% hit the top of the scale for aggression toward other dogs, the level the authors define as biting or trying to bite. Toward unfamiliar people the figure was 4.5%, and toward their own household 2.1%. So the breed’s issue, where there is one, is mostly with other dogs rather than with people.
UK vet records agree on the shape of it. Among 12,146 German Shepherds seen in primary-care practice, aggression was recorded as a problem in 4.76%, which put it among the five most common things vets wrote down. It showed up in 6.75% of males against 2.78% of females.
Then there is the guide-dog record, which is the least flattering evidence and also the most carefully collected. Across 7,770 guide dogs followed for twenty years, German Shepherds were seven times more likely than Labradors to be withdrawn for fear or aggression, rising to 7.5 times for males. The same dogs were 70% less likely to be withdrawn for lacking willingness or confidence.
Three limits on that, all of which matter. Guide work is a specific and demanding job, and being wrong for it is not the same as being a problem at home. Only 286 of those dogs were German Shepherds. And a programme that already expects trouble from a breed may look harder for it.
Take a direction from all of this, not a forecast. A German Shepherd is more likely than a Labrador to have an opinion about the dog across the street. Whether yours does is decided mostly outside the breed label.
Working lines and show lines
This split is real, it matters in a pet home, and almost everything written about it is untested. Here is the line between the two.
What has been measured
Both written standards call for a straight back. The American one says the back is “straight, very strongly developed without sag or roach.” The FCI one describes a straight back running to a croup that “should be long and slightly sloping (approx 23° to the horizontal).”
Neither standard authorises the extreme sloped or roached topline. It is a show-ring reading of the standard, not a requirement of it.
A study of 60 healthy German Shepherds using 3D motion capture and a pressure walkway found the difference is real and measurable. Dogs with sloped backs put more vertical force through the front legs, bent more through the mid-back, flexed the stifle and hock more, and moved less evenly as the slope increased.
It also tied conformation to the ring directly: 68% of the sloped-back group were shown, against 17% of the level-back group.
What it did not find was a size difference. The paper reports no difference in body mass or height between the two groups. And on the question everyone wants answered, the authors decline to guess: “further studies are required to determine whether the variation in movement, posture and conformation within the breed relates to an increased susceptibility to musculoskeletal disorders.”
So there are proven effects on posture and movement, and no proven effect on disease. Anyone telling you the sloped back causes hip dysplasia has gone further than the people who measured it.
What has not been measured
American show, West German show, West German working, Czech and DDR lines are described everywhere as differing in size, bone, drive and health. We found no peer-reviewed study and no kennel-club document putting a number on any of it. The type difference is documented. The size difference is not, and neither is the temperament difference. The behaviour study above came closest (its UK dogs were pets and its Swedish dogs came from an armed forces breeding programme), but it reported no head-to-head comparison.
The practical advice survives the missing evidence, because it never depended on it. Ask the breeder what the parents were bred to do, then go and meet them. A line selected hard for drive is not the easy starting point for a quiet house, whatever the paperwork says.
Growing up: adolescence, joints and neutering
Puberty, and the sit that stops working
Somewhere around eight months, a lot of German Shepherds start acting as though they have never met you.
This has been measured, and this breed is in half of it. Guide-dog carers rated 285 dogs, German Shepherds among them, for trainability at 5, 8 and 12 months, and 8 months came out worst of the three. The dip was personal. Trainers who knew the same dogs less well rated them higher at that age. The hands-on sit test alongside it used a different 93 dogs at 5 and 8 months, and those were retrievers only, so that half is borrowed.
A separate UK cohort of over a thousand puppies backs up the timing. Owner reports of unwanted behaviour peaked at 12 months, when 42.1% of owners had something to report, then fell away. Barking, jumping up, pulling on the lead and ignoring recall were the usual complaints.
That pattern is documented, it is aimed at you specifically, and it ends. It is not evidence that you have trained the dog badly.
Keeping the growth boring
Keep the dog lean and let the growth take its time. In this breed the elbow is the reason. One suggested mechanism for elbow dysplasia is a growth mismatch: the radius lengthening faster than the ulna, so the joint surfaces stop meeting cleanly while the dog is still building.
On neutering, take the figures to your vet rather than deciding alone, because the trade-off is real in both directions. Among 1,170 German Shepherds, joint problems turned up in about 7% of intact males against 21% of males neutered before a year. In females it was 5% against 16%.
A larger follow-up covering 1,257 German Shepherds found the same pattern and put a shape on it: 6% in intact males, 19% neutered under six months, 18% at six to eleven months, and 9% at one to two years. Urinary incontinence reached 9% in females spayed under a year. The authors’ own line for this breed is that “the suggested guideline for males, given the risks of joint disorders, is delaying neutering until over 2 years of age.”
Across a lifetime
What follows is a list of things to mention to a vet, not a list of things to decide at home. Vets record these more often in German Shepherds than in dogs generally.
In UK primary-care records, median lifespan was 10.3 years. Females reached 11.1 years against 9.7 for males, which is a wide gap for a single breed.
The hind end is the breed’s defining problem. Two of the four leading causes of death were losing the ability to stand (14.9%) and spinal cord disease (13.6%). Musculoskeletal disease led at 16.3%, with cancer at 14.5%. Very few breeds have a death list that leans this hard on movement.
Day to day, ears and joints dominate. Otitis externa (inflammation of the outer ear canal) was the single most common diagnosis at 7.89% in one year. Osteoarthritis followed at 5.54%, then diarrhoea at 5.24% and being overweight at 5.18%.
Degenerative myelopathy is the one owners fear. It is a slow, painless wasting of the spinal cord. It starts with a swaying back end and ends in paralysis, usually from 8 or 9 years old. Most owners choose euthanasia within a year of the first signs, though some dogs take over three. About 2% of German Shepherds seen at US veterinary teaching hospitals were affected. A DNA test exists, for a change in a gene called SOD1, and a dog needs two copies of it to be at risk. But this is “an autosomal recessive disorder with incomplete penetrance”, meaning “not all the animals identified with a pair of the abnormal genes have developed the disease.” A two-copy result is a risk, not a sentence.
Exocrine pancreatic insufficiency is oddly this breed’s own. The pancreas stops making enough digestive enzymes, so food passes through barely used. German Shepherds make up roughly two-thirds of all cases in dogs. Signs usually start between six months and six years, at a median of 36 months. The giveaway is a dog losing weight while eating ravenously, with large volumes of loose, foul-smelling stool. It is treatable for life with enzyme powder and vitamin B12, a good reason to ask for the test rather than keep changing the food.
A word on hip scores
You will see a hip dysplasia rate quoted for this breed. Which one you see depends entirely on whose database somebody opened.
- The German breeding registry’s own screening data, covering 48,367 German Shepherds born 2001–07, found 8.5% affected hips.
- The North American OFA database, current to August 2026, has 146,557 German Shepherds evaluated, of which 29,951 were dysplastic, 20%. That places the breed 50th of the 219 breeds OFA ranks. Its elbow figure is harsher: 19% of 68,041, which is 24th.
- An Italian study of 18,696 dogs found 26.6%.
- A Swedish study of 402 dogs found 49%.
That is a six-fold spread, and German Shepherds do not differ that much between countries. The databases count different dogs.
Screening for the German registry is effectively compulsory for breeding registration, so nearly every dog gets submitted, good hips or bad. OFA submission is voluntary, and an owner whose vet has just shown them a bad X-ray often does not send it in. Put three different scoring systems on top and the numbers stop being comparable at all.
That last point is not a guess. Researchers X-rayed 200 Golden Retrievers and 140 Rottweilers and then watched which films went to the registry. Normal hips were 8.2 times as likely to be submitted. Counting every dog X-rayed instead of every dog submitted pushed dysplasia up to 53–73% and 41–69% in those two breeds. Nobody has run that study in German Shepherds, so we cannot give you the corrected number for this breed. We can tell you which direction the error runs.
The lesson travels well beyond this breed. A prevalence figure is only as meaningful as its denominator, the group of dogs it was drawn from. Any page quoting one hip dysplasia rate as though it were the answer has not checked where it came from.
The same registry data gives heritabilities of 0.25 for hips and 0.31 for elbows. Both are moderate, which is why decades of screening have produced real but slow improvement. Reported elbow dysplasia runs 12 to 21%.
What we could not verify
Size or health differences between lines. Covered above. No peer-reviewed study, no kennel-club document, and every source making the claim was a breeder or enthusiast site.
Growth-rate control protecting German Shepherd hips. Raising a large-breed puppy lean and slow is good advice, and we give it. But the controlled feeding trial behind it used 48 Labrador Retrievers. No equivalent trial in German Shepherds appears to exist. The mechanism makes sense and the effect in Labradors was large, so we would still follow it, while saying out loud that it is an extrapolation.
How common the SOD1 risk variant is in German Shepherds. Breed-wide figures are quoted a lot. We could not open the paper they trace back to, and a German Shepherd study we could open, from Brazil, found no two-copy dogs at all in 95 tested. So we are giving you the disease rate and the inheritance instead of a carrier percentage we cannot stand behind.
A German Shepherd growth chart from real data. No published breed-specific growth curve exists. The table below comes from our own model using this breed’s measured adult weights, which makes it reproducible and checkable, not a record of what German Shepherd puppies actually weighed.