Research reference
Published study figures, collected in one place for homeowners, gym owners, wellness facility teams, and writers who need a clear source to cite.
Read the numbers correctly
These figures describe published research and survey samples. They do not establish a medical benefit, a universal session plan, or a substitute for professional guidance.
For a purchase decision, the equipment still has to fit the room, the intended user, and the manufacturer requirements. For a personal health decision, a qualified clinician is the right source of advice.
Cold-water immersion
The review evaluated randomized studies of healthy adults using cold showers, ice baths, or plunges. Its numbers show the scope and variation of the evidence base.
11
The review included 11 studies of cold-water immersion in healthy adults. It assessed what the published research measured, rather than establishing a universal use plan.
Source: Cain et al., PLOS ONE, January 20253,177
Those 11 studies represented 3,177 participants. The authors noted that the available evidence is still limited by study size and participant diversity.
Source: Cain et al., PLOS ONE, January 20257-15 C
The reviewed cold-water interventions ranged from 7 to 15 degrees Celsius, or about 45 to 59 degrees Fahrenheit. This describes the studies, not a ColdCraft recommendation.
Source: Cain et al., PLOS ONE, January 202530 sec-2 hr
The reported sessions ranged from 30 seconds to 2 hours. That wide spread is one reason a study result should not be treated as a single protocol for every person or setting.
Source: Cain et al., PLOS ONE, January 202510:1
Ten of the included studies used baths and one used a shower. That matters when comparing a cold-plunge setup with other forms of cold exposure.
Source: Cain et al., PLOS ONE, January 20257 / 4
The review rated seven included studies as moderate quality and four as high quality using the PEDro scale. That rating describes study methods, not a guarantee that every outcome is settled.
Source: Cain et al., PLOS ONE, January 202555.4%
Across the 3,177 participants, 55.4% were female. One study supplied the female-participant data, while the other included studies enrolled only men, an important limit on how broadly the findings can be applied.
Source: Cain et al., PLOS ONE, January 202510
The authors searched 10 research databases through January 17, 2024 before selecting eligible trials. This shows the review’s search scope, not the number of studies that ultimately met its criteria.
Source: Cain et al., PLOS ONE, January 20256 / 5
Six included studies used randomized crossover designs and five used randomized parallel designs. The different designs are another reason individual study results need to be read in their original context.
Source: Cain et al., PLOS ONE, January 2025Sauna bathing
The Global Sauna Survey documented the habits and experiences reported by regular sauna users. It was an online, self-selected sample, so it should not be read as a census of all sauna owners.
482
The 2019 online Global Sauna Survey produced 482 valid responses from regular sauna bathers. It was a self-reported survey, not a population study.
Source: Hussain, Greaves, and Cohen, Complementary Therapies in Medicine, June 20191-2x
Survey respondents reported using a sauna approximately one to two times per week. This describes the surveyed group, not a frequency recommendation.
Source: Hussain, Greaves, and Cohen, Complementary Therapies in Medicine, June 201951.3% / 48.7%
The valid responses were 51.3% men and 48.7% women. The balance is useful context, while the online sample remains self-selected.
Source: Hussain, Greaves, and Cohen, Complementary Therapies in Medicine, June 2019Methodology and sources
ColdCraft selected figures that are directly reported in the cited publications, then added only enough context to explain what each number does and does not mean. No figures were estimated or projected.
ColdCraft Recovery. “Sauna and Cold Plunge Research Statistics.” Updated October 20, 2026. https://coldcraftrecovery.com/recovery-research-statistics
