The sitting-rising test: how to score it and what your score means

The test comes from a research group at an exercise-medicine clinic in Rio de Janeiro, and their 2020 reference paper sets out the protocol. You do it barefoot, on a non-slip floor, with at least 2 m by 2 m of clear space. The instruction is to sit down on the floor and then stand back up "using the least support that you believe is needed", without worrying about speed.

Crossing your legs is allowed, as long as you do not use the sides of your feet for support. You may try more than once: the best score for each action counts, whatever the number of attempts.

Each action, sitting down and standing up, starts at 5 points, for a possible total of 10. Then:

• Take off 1 point for each hand used for support.
• Take off 1 point for each forearm, knee or side of the leg used.
• Take off 1 point for a hand placed on the knee.
• Take off 0.5 if the evaluator sees an unsteady execution.

So sitting down with one hand on the floor and standing up with no help scores 4 + 5 = 9. Using a hand and a knee to stand up scores 3 for that action.

The half-point deduction is one judgment by the person watching, made once for the action, not a penalty for every wobble. That makes scores near 10 depend partly on who is rating. If you test yourself, film it, or use the same helper each time.

The first large study looked at 2,002 adults aged 51–80 at the clinic, 68% of them men, followed for a median of 6.3 years; 159 died. Scores were grouped into four bands, and each band was compared with the highest, 8–10. Hazard ratios, adjusted for age, sex and BMI:

• 0–3: 5.44 (95% CI 3.1–9.5)
• 3.5–5.5: 3.44 (2.0–5.9)
• 6–7.5: 1.84 (1.1–3.0)
• 8–10: the reference group

A hazard ratio of 5.44 means that, at any point during follow-up, people scoring 0–3 were dying at about five and a half times the rate of people scoring 8–10, once age, sex and BMI were accounted for. Each extra point on the test went with 21% better survival.

The press release at the time gave the lowest band as 0–2.5. The paper’s abstract says 0–3, and that is the figure used here.

A larger, prospective cohort from the same research group followed 4,282 adults aged 46–75 for a median of 12.3 years, with 665 deaths from natural causes (COVID-19 and external causes excluded). This time the scoring included the half-point deduction for unsteadiness, and the groups were finer. Raw death rates over follow-up, lowest score group to highest:

• 0–4: 42.1%
• 4.5–7.5: 20.4%
• 8: 11.1%
• 8.5–9.5: 7.0%
• 10: 3.7%

Against a perfect 10, the 0–4 group had a hazard ratio of 3.84 for natural-cause death and 6.05 for cardiovascular death: at any point during follow-up, they were dying of natural causes at nearly four times the rate. The cardiovascular figure has a very wide interval (2.29–20.94) because few people in the top group died of heart disease. The 0–4 group was the lowest 10% of participants.

The detail worth noticing is the score of 8. In the 2012 bands, 8 sat in the lowest-risk group. In 2025, a score of 8 still had three times the death rate of a 10. Getting up with one hand is good; it is not the top of the scale.

Reference scores exist for 6,141 non-athlete adults aged 16 to 98, from clinic data collected between 1998 and 2018. The age- and sex-specific percentiles are in that paper’s figure and supplement, which we could not read for this review. We therefore do not quote any band value.

Two things limit that reference set even once its figures are read. It comes from one clinic, two-thirds men, who chose to attend. And it is cross-sectional: it shows how scores differ between age groups, not how one person’s score changes over the years.

Why the floor matters at all: among uninjured fallers aged 72 and over in New Haven, 47% could not get up unaided after at least one fall. Among people over 90 in Cambridge who fell, 80% could not get up at least once and 30% lay on the floor for an hour or more. Those long lies were strongly linked to serious injury, hospital admission and moves into long-term care. These studies support getting up from the floor as a goal; they do not show that practising the test prevents long lies.

Do the test barefoot on a clear, non-slip floor, with someone nearby the first time. Skip it if you have a recent knee or hip injury. A score is a benchmark, not a diagnosis, and none of this is medical advice.

Longstanding includes the floor sit-and-rise as one of its ten tests. Two of its sixteen goals for age 85 use it: getting up off the floor with one hand (a score of 8) and getting down to play and back up unaided (a score of 6). Those targets sit at the lower edges of the 2012 study’s 8–10 and 6–7.5 bands.

Two gaps, stated plainly. First, the app’s percentile bands for this test are our estimates. They have not been checked against the 2020 reference percentiles, because we could not read them, so this site labels them unverified. Second, the app’s scoring differs from the published protocol: it takes off 0.5 for each loss of balance rather than once, does not mention a hand on the knee, and records one attempt rather than the best of several. Both are known gaps, listed on the test page.

The score is coarse, and one rating can move by half a point. Re-tested on the same floor every quarter, it still shows whether getting down and up is becoming easier or harder.