The Sit and Rise Test: What Your Score Actually Predicts (And What It Doesn't)

A low sit and rise test score predicts earlier death in middle-aged and older adults, and that is the whole of what it does — it predicts, it does not cause. The 2025 study everyone is quoting (published online in June 2025) followed 4,282 adults aged 46 to 75 for a median of 12.3 years, recorded 665 deaths, and found that people scoring 0 to 4 out of 10 were 3.84 times as likely to die of natural causes as people scoring a perfect 10. Here is the part nobody quotes: in that same study, every score of 8 and above was not statistically distinguishable from a 10. So the real line is 8, not perfection, and the viral "anything under a 10 means trouble" framing is not in the paper it claims to come from.

How do you score the sit and rise test?

You get five points for sitting down and five for standing back up, and you lose one full point every single time a hand, forearm, knee or the side of a leg touches down — plus another half point if you wobble.

The published protocol is stricter than the version circulating on social media, and the details change the score:

• Barefoot, on a non-slip flat surface, in a space of at least 2m x 2m, in unrestrictive clothing.
• The evaluator reads one standardised instruction, word for word: "Without worrying about the speed of the movement, try to sit and then to rise from the floor, using the least support that you believe is needed."
• Each support costs one point — each hand, forearm, knee, or side of the leg. A hand placed on your own knee counts. People miss that one and score themselves two points high.
• Half a point comes off for unsteady execution — a partial loss of balance on the way down or the way up.
• Crossing the legs is allowed, as long as the sides of your feet are not used for support.
• Take several attempts and keep the best — the evaluator is allowed to give advice between tries, and only your best sitting score and best rising score count. The composite runs 0 to 10 in half-point steps, so there are 21 possible scores.

Score it honestly or don't score it. A test this cheap is only worth anything if you count the hand on the knee.

What is a good sit and rise test score?

Eight is the honest line — in the larger of the two mortality studies on this test, every score of 8 or above carried a mortality hazard ratio whose confidence interval crossed 1.0, meaning the risk difference was not statistically significant, while everything at 7.5 and below was.

Here are the 2025 numbers band by band, adjusted for age, sex, body mass index and clinical variables, with a perfect 10 as the reference. For natural-cause death: 0 to 4 carried a hazard ratio of 3.84 (95% CI 2.25-6.97), 4.5 to 7.5 carried 2.23 (95% CI 1.37-3.95), 8 carried 1.51 (95% CI 0.92-2.66), and 8.5 to 9.5 carried 1.25 (95% CI 0.74-2.24).

For cardiovascular death the pattern repeats, larger and noisier: 6.05 (95% CI 2.30-20.94) at 0 to 4, 3.25 (95% CI 1.30-10.91) at 4.5 to 7.5, 2.17 (95% CI 0.87-7.26) at 8, and 2.01 (95% CI 0.78-6.80) at 8.5 to 9.5.

Read the intervals, not the headline numbers. A confidence interval that crosses 1.0 means the study could not rule out no difference at all — the estimate leans one way, but the data are consistent with zero effect. At 8 and above, on both causes of death, that is exactly what happens. At 7.5 and below, on both, it does not. So "a 9 means you're in trouble" isn't a cautious reading of this study, it's a misreading of it — the same failure as the strength benchmarks everyone quotes, where the answer to is 135 a good overhead press flips depending on whose dataset somebody quietly picked for you.

Does the sit and rise test really predict how long you'll live?

It predicts — it does not cause, and that distinction decides whether the test is useful to you or just scary.

The association is real and it has been found twice. The original 2014 paper followed 2,002 adults aged 51 to 80 for a median of 6.3 years, recorded 159 deaths, and found adjusted hazard ratios of 5.44 (95% CI 3.1-9.5) in the lowest band against the 8-10 reference, then 3.44 (95% CI 2.0-5.9) and 1.84 (95% CI 1.1-3.0) as scores climbed, p < 0.001. Its most-quoted line: "Each unit increase in SRT score conferred a 21% improvement in survival." What the 2025 follow-up adds is the finer split at the top — the first look at whether a 9 differs from a 10, and it could not show that one does.

Now the limits, because they're large. Both cohorts came from the same single private exercise-medicine clinic in Rio de Janeiro (CLINIMEX). Participants were self-referred or sent by their own physician — people who show up at an exercise clinic are not a random slice of anyone, and the 2025 paper notes none of them had physical or clinical limits that would interfere with fitness testing. Men were 67.5% of the 2025 sample and 68% of the 2014 one, and the 2025 sample ran 46 to 75. These are observational findings on a selected clinical population: not a randomised trial, not a general-population sample, and no evidence at all that raising your score lowers your risk. Wikipedia's entry puts the media problem plainly — headlines suggest the test predicts mortality broadly across age groups, and "that broad of a claim has not been suggested in, nor is it supported by, existing literature."

Is there a real sit and rise test chart by age?

There is exactly one published age-reference chart for this test, and it is almost certainly not the one you found — the real one is a colour-coded quartile figure built from 6,141 adults, and it does not give you a single target number for your age.

That paper is Araújo and colleagues, 2020, in the European Journal of Preventive Cardiology: 6,141 adults (4,101 men, 2,040 women) aged 16 to 98, tested at the same Rio clinic between July 1998 and February 2018. It reports 15 consecutive five-year age bands, from 16 through over 85, split by sex, as quartile colour bands — Q1 covering percentiles 1 to 25 through Q4 covering percentiles 76 to 99 — separately for sitting, rising and the composite. It tells you which quartile a score falls in for your age and sex, and the target it sets is a band rather than a number — the authors say aim for the green or blue band, that is, "equal to or above the median (P50) for his/her sex and age range." You can read the median off the chart as the line where yellow turns green, but the paper never prints a single "your age equals this score" figure, because it was not built to hand you one. It does report one number worth having before you panic about a 9: fewer than 8% of men and women over 55 scored a perfect 10.

So the crisp "age 45 = this number" tables all over page one are not in the published literature. Somebody made them up and everybody copied them — the same thing that happened with grip strength and the dead hang age charts, where invented tables outrank the real data because a clean number is easier to publish than an honest one. One detail from the 2020 paper carries into the next section: athletes, children and young adolescents were deliberately excluded, along with anyone with locomotor or clinical limitations.

Should a healthy lifter in their 30s even care about this test?

If you train, you're under 45, and you scored a 10, this test just told you nothing you didn't already know — its published evidence comes from adults aged 46 to 80, and the reference sample deliberately excluded athletes.

It earns its 30 seconds as an instrument. It costs nothing and reads four things in one movement — the authors list them as muscular strength and power, flexibility, balance and body composition. In gym terms: single-leg strength through a deep range, hip and ankle mobility, balance, and what you weigh. That combination catches a specific guy — strong in every loaded pattern he trains, unable to get off the floor without a hand down. A 315 squat tells you nothing about whether you can do that. It's a bilateral, mid-range, supported lift with a bar on your back. The floor asks a different question.

The other input it reads is bodyweight, and that one I know from the inside. I went from 285 lbs to about 200 and I hold it year-round, lifting four days a week. Eighty-five pounds came off every single movement where my own body is the load — and getting off the floor on one leg is exactly that movement. Nobody files fat loss under "mobility work," but for a lot of guys carrying extra weight it's the biggest lever on this test they've got.

How do you improve your sit and rise test score?

You don't train the test — you train what it measures: single-leg strength through a deep range, hip and ankle mobility, balance, and bodyweight relative to that strength.

• Single-leg strength through a deep range. Split squats, step-downs and cossack squats, 2-3 hard sets each. The rise out of a cross-legged position is a one-leg effort from a dead stop with no stretch reflex, which is why strong bilateral squatters still fail it. Balance comes along for free here — you cannot stand up off one leg without it. For a separate read on balance alone, see how long you should be able to stand on one leg.
• Ankle dorsiflexion. If your heel lifts in a deep squat, the ankle is your limiter. Loaded calf stretches and knee-over-toe work, a few minutes most days, move it faster than anything passive.
• Hip flexion and rotation. Deep goblet squat holds, 90/90 transitions, and sitting on the floor in the evening instead of the couch.
• Bodyweight. Same lever as everything else you do against your own mass — every pound off is a pound your legs no longer have to stand back up.

Then keep the honesty straight. Rehearse the sit and rise itself for two weeks and your score will go up while almost nothing the score was detecting has changed — the same trap as squeezing a gripper to beat a grip-strength test. Train those inputs and the score moves because you moved. Take a reading, change something real, take the reading again.

If the inputs are the problem and you have no structure to train them in, that's what the Momentum Method is — the four-day split, the daily minimum standard, and the one-screen scorecard that tells you whether any of it is working. Retest every few months. Thirty seconds on the floor is a cheap reading to take, and worth taking. It is not a verdict on how long you've got.