Tango and Parkinson's Research: How a Small Striking Finding Becomes a Global Claim

The claim that tango does something specific for Parkinson’s disease is repeated far more often than it is sourced, and this page quotes no figure for it, because the original study, its sample size and its design have not been traced to source here. What can be documented is the pattern by which claims like it travel, and that pattern is worth more to a reader than one more unsourced percentage would be.

The reason this page exists is a month in which the same claim arrived three times from three unconnected people. A social prescribing link worker mentioned it on the phone. A relative of somebody in one of my care home groups asked me whether it was true. And a funder’s project officer used it in a sentence beginning with the words research shows. Nobody offered a source and nobody had one when I asked, which they all did politely and none of them found awkward. That is not a story about three careless people. It is a story about a claim that has become ambient, which is a much harder thing to correct than a claim somebody is defending.

Where a claim about effect appears below it carries a tier (Tier 1 supported, Tier 2 promising but limited, Tier 3 practice and anecdote), explained on reading arts in health research.

What this page can and cannot say

It can describe the mechanism of propagation in detail, because that has been traced. It cannot describe the study, because that has not.

This site’s rule is that a figure is quoted only after somebody has opened the document it came from and recorded its design, its sample size, its outcome measure, its effect size, its interval, its certainty rating and its year. For the tango finding, that work has not been done here, so no number appears. Naming the gap makes for a weaker page and an honest one, and it is a considerably better outcome than the alternative, which is a number I could not defend if somebody asked me where it came from.

What would fill the gap, stated plainly: a citable trial record, its design, how many people were in each arm, what the comparator was, which outcome was primary, and the effect size with a confidence interval. If a claim of superiority over other dance styles is being made, that additionally requires a head to head trial of one style against another, powered to detect a difference between two active interventions. That is a rarer and much more demanding design than a trial of dance against no treatment, and the difference between the two is the single most useful thing to hold on to when reading anything in this area. The wider picture of what the Parkinson’s dance literature does and does not contain is on dance for Parkinson’s.

The five stages, and how to spot each one

Claims in this field decay along a predictable route, and each stage strips out something a reader needed.

Stage one: a small study with a striking result. This is a normal and healthy thing for a field to produce. Small early studies are how research begins, and a large apparent effect in a small sample is a reason to fund a bigger trial. Nothing has gone wrong yet.

Stage two: the comparator disappears. The paper compared the intervention against something, and the something is the first casualty. If the comparator was no treatment or a waiting list, dropping it turns a modest claim, people who did something did better than people who did nothing, into a much larger one about the specific activity. If the comparator was another active intervention that performed just as well, dropping it hides the most important result in the paper.

Stage three: the number loses its denominator. A percentage or an effect size survives, and the sample size, the interval and the timepoint do not. This is where a result stops being checkable, because a figure with no denominator cannot be interrogated, only accepted or refused.

Stage four: the label drifts. The outcome that was measured becomes a bigger outcome. A result about walking speed becomes a result about mobility, then about independence, then about quality of life. Every step is a small stretch and the cumulative distance is enormous.

Stage five: it becomes a fact. It appears in a funding bid, a service description, a conference slide, a newspaper piece. At this point it is being quoted by people who have never seen the paper, from sources who never saw it either, and its frequency becomes the thing that makes it credible.

Four claims that were traced, and what each one lost

The tango study has not been traced here. Four other claims on this site have, and they show the stages above running in real documents.

The arts on prescription savings set. Three figures, all from one uncontrolled evaluation of one scheme, written by a GP in December 2011, which describes itself as a simple observational study and states in its own conclusion that it “does not imply causality”¹. The commonly quoted 37% fall in GP consultations is the months 7 to 12 figure; the report’s own full year figure, in the same section, is 24%. The 27% is a reduction in overall NHS spend, from £157,473 to £115,050, and is quoted almost everywhere as a reduction in admissions, which appear separately as 54 before and 33 after with no percentage and no significance test attached. And the per patient saving divides £42,423 by the 90 patients analysed to get £471, adds £105 of GP time, and sets the total against £180,000 divided by the 500 patients referred. Numerator and denominator describe different populations. That is stages two, three and five in one document, and the full trail is on does arts on prescription save money.

Ulrich’s 1984 window study. Still cited in hospital design guidance four decades later. It is 46 patients in 23 matched pairs, one Pennsylvania hospital, one operation, one season, retrospective records from 1972 to 1981, with a length of stay of 7.96 days for tree view patients against 8.70 days for wall view². It contains no artwork of any kind. A finding about a window has carried the case for paintings in corridors for forty years, which is stage four at industrial scale: see does hospital art speed recovery.

The singing and postnatal depression trial. Across the whole sample there was no significant effect (p=0.16). The result people quote is a moderate to severe subgroup at week 6, which was no longer significant by week 10, and singing did not beat the trial’s active comparator of creative play³. Stage two and stage three, in a trial that was itself well designed and reported honestly. The honest summary is faster recovery among the more severely affected rather than a better endpoint.

The American social prescribing pilot. The figure of more than 300 cultural organisations is real and belongs to a statewide programme for FY24 to FY25⁴, not to the pilot it is usually attached to, which involved 8 organisations and then 12. The independent peer reviewed evaluation of the pilot reports 12 cultural organisations, 20 healthcare providers, 414 referrals, 363 participants and participant data on 84⁵. Using the larger number as the pilot’s size overstates it about twenty five times over. That is stage three and stage five, with an evaluation sitting there the whole time saying something smaller. More on the international picture at arts on prescription around the world.

The closest relative of the tango claim

The mislabelled dance and depression figure is the same failure in the same corner of the field, and it has been traced.

The number that circulates is -7.33, and it circulates as an effect size. What the Cochrane review of dance movement therapy for depression actually reports across its whole sample is SMD -0.67 (95% CI -1.40 to 0.05) at very low quality, from 3 studies and 147 participants, with the authors stating that the evidence “does not allow any firm conclusions to be drawn”⁶. The -7.33 is an adults only subgroup figure from two studies and 107 participants, printed in the review’s own abstract under a standardised mean difference label and immediately described there as failing to meet clinical significance. A standardised mean difference of -7.33 would be more than seven standard deviations, which is not a possible effect size for a psychological intervention on a symptom scale. It is a raw mean difference on a questionnaire subscale, wearing the wrong label.

That is the tango problem in a case where the paperwork survived: a striking number, detached from its subgroup, its sample and its own author’s warning, repeated by people acting in good faith. The full account is on dance for depression.

What a checkable claim looks like

For calibration, here is the one movement claim in this field that survives every stage of the journey intact.

Rhythmic auditory stimulation after stroke improved gait velocity by 11.34 m/min (95% CI 8.40 to 14.28), from 9 trials and 268 participants, at moderate quality⁷. Tier 1. Notice what is attached to it: the outcome named exactly, the effect size, the interval, the certainty rating, the trial count, the sample size, and a review anybody can open. Notice also how narrow it is. Gait velocity. Not mobility, not independence, not quality of life, and not mood. A claim that arrives with all seven of those things attached is a claim you can argue with, which is the point. See rhythmic auditory stimulation and, for the contrast this site draws deliberately, dance and health.

The five questions

Before repeating any dance and Parkinson’s claim, including anything on this site:

1. What design? Randomised, controlled but not randomised, or before and after with nothing to compare against.
2. How many people, and in which arm? A total is not enough; a difference between two arms of fifteen people each is a different animal from a total of thirty.
3. What was the comparator? Another activity, or nothing. This is the question that changes the meaning of the result most and is asked least.
4. Which outcome, exactly? Named, and as the study named it.
5. What is the effect size and its interval? Not whether it was significant.

If whoever is making the claim cannot answer those, the claim has been forwarded rather than checked. That is not a reason to be rude to them; three sincere people asked me the same question in a month and none of them had done anything wrong except trust the person before them.

What none of this argues

It does not argue that tango classes are a waste of anybody’s time, and reading it that way would be the same error running the other way.

An unsourced claim is a statement about a citation, not about an activity. People go to these classes, enjoy them, keep going, and get movement, music and company out of them. A Tier 3 label on this site means the research question is open rather than answered against, and most of what happens in community halls sits there, including the work I do myself. What is ruled out is a promise about a clinical outcome, and what is not ruled out is anything at all about whether the hour is worth having. Nothing on this page is a reason to change, delay or decline any part of anybody’s treatment, and no dance class substitutes for one.