Method
Participants
We recruited 60 people with a sauna habit (attending sauna at least once a week; regular sauna condition), 60 with almost no sauna habits (not attending sauna but had the opportunity to attend in the last 6 months; irregular sauna condition), and 60 with no sauna habit at all (had no opportunity to go to saunas in the last 6 months; no sauna condition). The sample size was calculated assuming analysis of covariance (ANCOVA) as the main analysis method (moderate effect size f = .25, power = .80, significance level = .05), with the three conditions among these participants as independent variables and measures such as well-being as dependent variables, controlling for SES and health orientation. Calculations using G*Power (3.1.9.7; Faul et al., 2009) yielded a total required sample size of 158 participants. Considering that a 10% missingness rate is possible in an online survey (Hoerger, 2010), 180 participants (60 per condition) were included in the study. After excluding participants who failed the attention check, 57 had a regular sauna habit (Mage = 41.72, SDage = 9.35, 36 men, 21 women), 60 had an irregular sauna habit (Mage = 41.92, SDage = 8.89, 29 men, 31 women), and 60 had no sauna habit (Mage = 41.40, SDage = 8.98, 31 men, 29 women).
Materials
Twelve scales were utilized for measurements. The first three variables measured were more general variables, as examined in previous studies; variables four and five were related to bodily sensation; variables six to 10 were related to higher-order emotional experiences, including self-transcendent emotion and spirituality; and variables 11 and 12 were controls.
Satisfaction with Life Scale. The Satisfaction with Life Scale (SWLS; Diener et al., 1985) is a five-item instrument that is answered using a 7-point Likert scale and includes items such as “In most ways, my life is close to my ideal.” We used the Japanese version of this scale (Sumino, 1994) as an indicator of well-being.
State-Trait Anxiety Inventory. We used only the characteristic scales of the Japanese version of the State-Trait Anxiety Inventory (STAI; Nakazato & Mizuguchi, 1982; Spielberger et al., 1971). The participants rated 20 items, such as “I feel good,” on a 4-point scale from “1 = Almost never” to “4 = Often.”
Ten-Item Personality Inventory. We used the Japanese version of this Big Five A scale, which consists of 10 items on five subscales that measure personality (Gosling et al., 2003; Oshio et al., 2012). For example, extraversion is rated using a 7-point scale ranging from “1 = I think not at all” to “7 = I strongly agree.”
Five Facet Mindfulness Questionnaire, Short Version. The Japanese version of the Five Facet Mindfulness Questionnaire, short version (FFMQ; Baer et al., 2006; Takahashi et al., 2022) comprises 24 items that measure trait mindfulness across five subscales (nonreactivity, observing, acting with awareness, describing, and nonjudging). For example, observation, measured using items like “I pay attention to sensations such as the wind in my hair or the sun on my face,” is rated on a 5-point scale from “1 = Never or very rarely true” to “5 = Very often or always true.”
Plymouth Sensory Imagery Questionnaire, Short Version. The Plymouth Sensory Imagery Questionnaire (short version) (Psi-Q; Andrade et al., 2014) utilizes seven subscales to measure multisensory mental imagery ability. The Japanese version, used in this study (Hitsuwari & Nomura, 2023), does not include “emotional feeling” and has 21 items across six subscales: vision, sound, smell, taste, touch, and bodily sensation. For example, for vision, items such as “Imagine the sound of …” followed by the word “bonfire” are rated using a 7-point scale ranging from “1 = No image at all” to “7 = As vivid as real life.”
Dispositional Positive Emotion Scale. We used the Japanese version of the Dispositional Positive Emotion Scale (DPES; Shiota et al., 2006; Sugawara et al., 2020), a 38-item instrument that measures the tendency to feel positive emotions across seven subscales (joy, contentment, pride, love, compassion, amusement, and awe). For example, for joy, respondents answer the item “I often feel bursts of joy” on a 7-point scale ranging from “1 = I totally disagree” to “7 = I strongly agree.”
Gratitude Questionnaire-Six Item Form. The Gratitude Questionnaire-Six Item Form (McCullough et al., 2002) is a six-item scale measuring gratitude propensity. The Japanese version, used in this study (Shiraki & Igarashi, 2014), contains five items because one item was dropped during the reliability and validity validation stages. For example, items such as “I have so much in life to be thankful for” are answered on a 7-point scale ranging from “1 = Strongly disagree” to “7 = Strongly agree.”
Aesthetic Evaluation of Art (Haiku Poetry). The haiku, the world’s shortest poem, was used to measure the aesthetic perception of art. A haiku can be appreciated in a short time, and evaluation tasks can be easily carried out on a computer (Hitsuwari & Nomura, 2022). Haikus are associated with self-transcendent emotions (Kato & Hitsuwari, 2024). Five haikus were evaluated using a 7-point scale ranging from “1 = Not beautiful at all” to “7 = Very beautiful.”
Japanese Youth Spirituality Rating Scale. The Japanese Youth Spirituality Rating Scale (Nigorikawa et al., 2016) uses five subscales to measure Japanese spirituality (harmonizing with nature, Ikigai [the Japanese word for one’s sense of purpose in life], awe toward the invisible, connecting with one’s ancestors, and self-determination) on a 27-item scale. For example, for items such as “harmonizing with nature,” the scale includes statements like “Being in nature gives me energy” and “I feel that I have a lot of energy.” Items are rated on a 7-point scale ranging from “1 = Not at all true” to “7 = Very true.”
Time Perception. A four-item scale measuring time perception was used (Rudd et al., 2012). Items such as “I have lots of time in which I can get things done” were answered using a 7-point scale ranging from “1 = Strongly disagree” to “7 = Strongly agree.” The higher the score, the more time the participant perceived that they had available.
Socioeconomic Status. Several questions were used to measure SES (Shimizu et al., 2021; Yanagisawa et al., 2013). First, participants reported their approximate annual household income using a 12-point system, ranging from “less than 12 million yen” to “more than 12 million yen” (SES income). Second, the participants answered the question, “If the whole of current Japanese society were to be divided into five strata, where would you place your household income?”, using a five-factor system: “high,” “above middle,” “below middle,” “above low,” and “below low” (SES ranking). Third, participants answered three questions regarding their subjective current and childhood SES. For example, for subjective childhood SES, the item “My family usually had enough money for things when I was growing up” was rated using a 7-point scale ranging from “1 = Strongly disagree” to “7 = Strongly agree.”
Health Orientation. We used a 10-item scale measuring health awareness and health-related behavioral practices (Takahashi & Muto, 1992). An example item is, “Do you think you are a person who takes care of your health?” The other items were answered using a 4-point scale ranging from “1 = Not at all true” to “4 = Very true.”
Procedure
Participants were recruited using CrowdWorks (https://crowdworks.jp/), a Japanese online crowdsourcing company, and completed a questionnaire created on Qualtrics (https://www.qualtrics.com/). After signing an informed consent form, they answered the abovementioned questionnaire and questions related to sauna use (e.g., frequency of sauna attendance, sauna history, motivation, sauna bathing methods, psychological and physical effects), their haiku and art experiences and interests, and their demographic information.
Results and Discussion
Descriptive statistics for the three conditions and the results of the one-factor analysis of variance between participants are presented in Table 1. The analysis of variance (ANOVA) was conducted using the anovakun package (ver. 4.8.6.; Iseki, 2021) in R (ver. 4.2.3; R Core Team, 2023), and multiple comparisons were conducted using the Shaffer method.
[Table 1 near here]
The results in Table 1 show that there were significant differences in the three-group comparisons in the Psi-Q’s somatosensory imagery ability (F(2, 174) = 3.38, p = .04, η2 = .04), evaluations of the beauty of the haikus (F(2, 174) = 3.88, p = .02, η2 = .04), and subjective current SES (F(2, 174) = 3.87, p = .02, η2 = .04). Next, we conducted an ANCOVA, controlling for SES and health orientation. Several measures of SES were employed in this research; however, in the ANCOVA, we used the participants’ subjective current SES, which was significantly different in the ANOVA. Using the lmres function within the R pequod package (Mirisola & Seta, 2016), bodily sensory imagery ability and the aesthetic evaluation of the haikus were entered as dependent variables, the experimental condition of sauna bathing frequency was entered as the independent variable, and SES and health orientation were entered as control variables. The effects of bodily sensation imagery ability (a subscale of the Psi-Q; Figure 2a) and the aesthetic evaluation of haiku poetry (Figure 2b) remained significant in an ANCOVA that controlled for health orientation and subjective SES (Supplemental Material, Table S1).
[Figure 2 near here]
The irregular sauna and no sauna conditions were grouped together and compared post hoc according to regular sauna attendance. The t-tests showed a significant effect of the sauna condition on the aesthetic evaluation of haiku poetry and subjective current SES. The t-test also showed significant trends of the sauna condition in nonreactivity, Psi-Q, taste imagery ability, contentment, pride, amusement, awe, harmonizing with nature, and Ikigai (Figure 3; see also Supplemental Material, Table S2 for statistical values). The aesthetic evaluation of haikus remained significant when controlling for subjective current SES and health orientation (b = -.31, SE = .13, t = -2.42, β = -.18, p = .02), whereas other variables were not significant.
[Figure 3 near here]
The results of Study 1 showed an association between the frequency of sauna bathing and some psychological variables, particularly bodily sensation imagery ability and the aesthetic evaluation of haikus. These variables have been suggested to be related at the observational level; however, using statistical analysis, this study empirically revealed an association for the first time. With its repetition of hot and cold exposure, sauna can draw attention to the body and cause its imagery to become clearer, even if an individual is not in the sauna room. The association with the aesthetic evaluation of haikus was also consistent with the aim of the above-mentioned exhibition (teamLab, 2021), which sought to enhance the art experience by clearing the mind in the sauna. However, Study 1 did not reveal causal relationships, such as whether sauna bathing (a habit) affected these physical, emotional, and psychological tendencies, or whether individuals with a higher propensity toward such tendencies were more likely to take sauna baths. In Study 2, we examined the psychological effects of sauna bathing through a field experiment at a bathing facility with saunas.