The Published Push-Up Table by Decade and Sex, the Test Behind It, and What Its Numbers Can and Cannot Tell a Reader Over 40
Credit: Daily Vitality
A push-up count is one of the few fitness numbers anyone can produce on a living room floor with no equipment at all. That simplicity is exactly why "push-ups by age" charts are everywhere. It's also why so many of them disagree. For a man in his 40s, one chart calls 25 push-ups excellent, while another saves that label for 35 or more.
The answer to "how does a count compare?" depends on which table is being used, how its numbers were collected and whose bodies they describe. There is one published table with a documented history behind it, and almost every credible chart online is a copy of it, or a copy of a copy. It's older than most people expect, and it asks men and women to do two different exercises.
This guide sets out that table for ages 40 to 69, with younger decades for context. It walks through the exact test protocol behind the numbers, explains what the dataset does and doesn't represent, and shows how to retest so that two counts months apart can be fairly compared. It's a performance benchmark and nothing more: a way to place a number, not a verdict on health.
The push-up test is a muscular endurance test. It measures how many times the chest, shoulders and the backs of the arms can lift a fixed share of body weight before fatigue or a breakdown in form ends the set. That makes it different from a strength test like a one-rep bench press, which asks how much can be lifted once.
The body has to stay rigid from shoulders to hips (or to knees in the modified version) for every rep. So a push-up also asks something of the trunk. Researchers who have tested push-ups against bench-press benchmarks point out that any push-up protocol assesses a component of trunk strength and endurance as well as arm and shoulder endurance. That helps explain why the correlation between push-up counts and bench-press results is real but only moderate.
A benchmark is only useful if it gives roughly the same answer twice. On that front the push-up holds up reasonably well. Across published studies, test-retest reliability ranges from moderate to high depending on the protocol and the group tested. On average, two sessions have differed by only about one and a half to two repetitions.
A Canadian graduate thesis that reviewed the test for use in a clinical study summarized the figures behind the two versions. The bent-knee push-up in women showed test-retest reliability of 0.83 and inter-rater reliability of 0.997. The toe push-up in men showed test-retest values of 0.91 to 0.93. These come from college-age samples, so they describe how consistent the test is as a measurement, not how people in midlife perform.
Reliability is about the measurement, not the person. A change of one or two reps between sessions sits inside the test's normal wobble, and that's worth keeping in mind before reading much into it.
A count lines up with the table only if it comes from the same test. The version reproduced alongside the Canadian norms by the American Council on Exercise lays out the standard elements. Hands sit shoulder-width apart, directly under the shoulders, with hips and shoulders aligned in a rigid trunk. Each rep finishes with the elbows fully straight at the top. The reps are steady, with no resting between them, and there's no time limit.
The Canadian manual's own wording for the stopping point is slightly different in tone but the same in spirit. A graduate study that followed the Canadian protocol stopped the test when participants were seen straining forcibly or using compensating movements. Sagging hips, a piked-up rear end or half-depth reps all count as that kind of compensation.
This is the detail most online charts drop, and it changes how the whole table reads. In the Canadian protocol, and in the American College of Sports Medicine's reproduction of it, men perform the standard push-up from the toes while women perform the modified push-up from the knees. Those are two different exercises. A knee push-up moves less body weight, so the counts run higher.
That means the women's column describes knee push-ups, and the men's column describes toe push-ups. A woman who tests from her toes has done a harder test than the one her column was built on.
The women's numbers in the table are knee push-up counts. A toe push-up count from a woman is not comparable to that column, and the men's column is not a substitute, since it describes a different population.
For another benchmark where the version of the test changes the numbers, Daily Vitality's Single-Leg Balance by Age guide runs into a similar problem with eyes-open and eyes-closed versions.
Here is the Canadian push-up table as reproduced in the CSEP Physical Activity Training for Health resource manual (2019) and in the American Council on Exercise's assessment sheet, which cites it. The 40s, 50s and 60s are the focus. The 20s and 30s are included so the shape of the change across decades is visible. Men's counts are toe push-ups. Women's counts are knee push-ups.
The table stops at 69. There is no published row for 70 and older in this dataset, so any chart that extends it has borrowed or estimated those numbers from somewhere else.
A few patterns stand out once the rows sit side by side. Counts fall with every decade for both sexes. The drop is steep between the 20s and 40s and flatter after that. For men, the entry point to "Excellent" goes from 36 in the 20s to 25 in the 40s and 18 in the 60s. The "Good" band for men in their 50s spans just three reps, from 10 to 12. That's narrow enough that a single rep can move a count across a label.
The women's column tells a slightly different story. By the 50s, "Excellent" (21 or more) sits at the same number as the men's, but it's measuring knee push-ups against toe push-ups, so the two rows can't be read as equal performance. In the 60s, the women's "Good" band (5 to 11) is wider than the "Very good" band above it. That's reproduced exactly as published and reflects how the original categories were cut, not a typo here.
Daily Vitality's Walking Speed by Age guide shows a similar decade-by-decade shape for a very different kind of measurement.
The usual promise of a benchmark chart is a percentile: a number that says what share of people scored lower. This table doesn't offer that. It's published as five named bands per decade and sex, and the source data for the individual percentiles isn't part of what was published. So the honest reading of a count is "this lands in the Good band for men aged 50 to 59." It isn't "this is the 60th percentile."
Some sites attach tidy percentiles to the same decades anyway. Those figures are derived by someone, somewhere, and can't be traced back to the published table. There's more on that below.
How far apart are the two versions? One small study of 72 college women had every participant do both. They averaged 9 toe push-ups and 17.5 knee push-ups, close to a two-to-one ratio on average. The spread was wide, though. Eight women couldn't complete a single toe push-up, while seven did more than 20.
That's a young sample, and the authors say more research is needed to validate a toe-based scale for women. So no conversion factor is offered here. The study's useful lesson is narrower: the two counts are far enough apart that mixing them produces a misleading placement.
Many sites credit these push-up norms to a 2000 paper by Payne and colleagues, "Canadian Musculoskeletal Fitness Norms." That paper did test push-ups. According to its own abstract, though, its new norms covered partial curl-ups, vertical jump and leg power, the measurements "for which Canadian norms are not currently available." Push-ups, grip strength and sit-and-reach were measured in its 571 self-reportedly healthy participants (312 women and 259 men, aged 15 to 69). The team used them to confirm that the sample matched the 1981 Canada Fitness Survey and a 1988 national survey.
So the push-up norms already existed by 2000. A 2022 study of the women's protocol traces the standards to the Canadian Association of Sport Sciences, the predecessor of today's Canadian Society for Exercise Physiology (CSEP), in 1987. Those authors note that the ratings are more than 30 years old and are still used today.
A 2016 national study in Norway, which tested 726 adults aged 20 to 65, opened with a blunt line: up-to-date research on this kind of fitness testing is lacking. It listed the ACSM's Canadian-based values among the latest field-test norms available. The Norwegian team built its own reference values, but it used a different push-up variant, so those numbers can't be laid over the Canadian table. It did confirm the same broad pattern: men outscored women on push-ups, and scores fell with age.
For grip strength, US reference values exist from large national projects. For push-ups, no comparable US population table turned up in this research. No adult push-up component turned up in the CDC's national health examination survey materials, and the CDC's 2012 youth fitness survey used a modified pull-up instead.
The closest American alternative comes from the Cooper Institute, and it's a different test. Cooper's norms use a one-minute timed push-up: in the institute's own example, a man aged 30 to 39 who does 34 push-ups in one minute ranks at the 70th percentile. A timed test rewards speed, while the Canadian test runs to fatigue with no clock. Counts from the two can't be swapped.
The Cooper Institute also makes a distinction that explains a lot of the confusion online. A norm shows where someone ranks compared with others of the same sex and age. A standard is the minimum needed to do a physical job, such as police or military work. Job standards are pass-or-fail thresholds set for a purpose. They aren't descriptions of the general population, and a chart that blends the two is neither.
Readers following the benchmark series can find the same norm-first approach in Average Daily Steps by Age and Sex.
A search for push-ups by age turns up three recurring problems.
Unknown sources. One widely used push-up calculator rates men aged 40 to 49 as "Excellent" only above 34 push-ups, and states that the original source for its data is unknown. The Canadian table puts that line at 25. A man doing 28 push-ups in his 40s is "Excellent" by the documented table and merely average-to-good by the undocumented one.
Misattributed percentiles. Another reference site publishes neat 5th-to-95th percentile values for each decade, credited to Payne 2000 and its 571 participants. As covered above, that paper didn't generate push-up norms, and the Canadian table isn't published as percentiles. Those percentile figures don't appear in the published table.
Blended standards. A popular running site builds its charts, covering the 40s through 70s and beyond, by combining Army and Marine Corps fitness standards with YMCA norms and school testing data, with no source for any single row. Mixing job standards, timed tests and untimed tests in one chart produces numbers no single group of people actually recorded.
A chart that can't answer those questions can still be interesting. It just can't be used to place a count.
A single count is a snapshot. The more useful picture comes from the same person over time, and that only works when the test conditions stay put. The measurement research points to a handful of things that tend to move a count without any real change in capacity: a different depth marker, a different surface, a different time of day, or a warm-up one day and none the next.
Published push-up studies have found average differences of about 1.5 to 2 repetitions between two test sessions with nothing else changing. A retest that comes in one rep higher or lower is inside that range. A shift of several reps, or a move across a band boundary that holds on a second retest, is a clearer signal of real change in the measurement.
The table's bands are also narrow in places. Some are only three reps wide for people in their 50s and 60s, so a label can change on a small shift. The count itself is the more stable number to track; the label is a rough guide to where it sits.
The 5 Simple At-Home Fitness Tests roundup covers other self-tests that can sit alongside a push-up count in the same session. The sit-to-rise test is one of them, and you can see how sit-to-rise scores compare by age.
About the Author
Angela Nightingale is a Senior Editor at Daily Vitality with over two decades of experience in digital publishing and health and wellness content. She specializes in turning complex, often-confusing health topics into clear, calm, and practical guidance that respects the reader's intelligence. Her work focuses on helping people feel informed and confident — never overwhelmed or alarmed — as they make everyday decisions about how they eat, move, rest, and age.
In the Canadian table, 13 to 16 toe push-ups falls in the "Good" band for men aged 40 to 49, 17 to 24 is "Very good," and 25 or more is "Excellent." Those are category bands from 1980s Canadian data, not percentiles.
For women aged 50 to 59, the table's "Good" band is 7 to 10 knee push-ups, "Very good" is 11 to 20, and "Excellent" is 21 or more. Those numbers describe push-ups from the knees.
Not directly. The women's column was collected with knee push-ups, and in one study of young women the knee count averaged close to double the toe count. There's no validated conversion, so a toe count doesn't place cleanly against either column.
Not in this dataset. The published Canadian table ends at 69. Charts that show 70-plus bands have drawn those numbers from elsewhere or estimated them.
The Canadian version isn't. It runs at a steady pace with no rest until form breaks down or no more full reps are possible. Some American norms, such as the Cooper Institute's, use a one-minute limit instead, and the two sets of counts aren't interchangeable.
The credible ones trace back to Canadian norms reproduced by the Canadian Society for Exercise Physiology and the American College of Sports Medicine. A 2022 study dates those standards to 1987. Many other charts don't name a source at all, or blend in military and job standards.
Mostly because the sites use different tests (timed or untimed, knees or toes), different populations, or pass-or-fail job standards presented as if they were population norms. A 10-rep gap for the same age group is common.
This table is a performance benchmark: it places a count among Canadian adults of the same decade and sex from the era the data were collected. It isn't built to assess health, and a band label isn't a health rating.
The research doesn't set an interval for personal benchmarking. What makes retests meaningful is keeping the conditions identical. Counts that differ by one or two reps usually fall within normal session-to-session variation.
Daily Vitality's Body Fat Percentage by Age and Sex guide is another entry in the same series.
