Laughter in everyday life: an event-based experience sampling method study using wrist-worn wearables

The study investigates how often people laugh in daily life and how laughter frequency relates to individual differences (Big Five personality traits, happiness, life satisfaction, gelotophobia, and cheerfulness). Motivated by the scarcity of accurate in-situ assessments of laughter—given its fleeting nature—the authors employ an event-based experience sampling method using wearables to minimize recall bias and burden. Two research questions guided the work: (RQ1) whether happiness is higher during laughing than at non-laughing reference times; and (RQ2) whether laughter frequency is associated with personality traits, well-being (happiness, life satisfaction), cheerfulness, and gelotophobia. The study aims to validate a novel measurement approach combining a one-button wrist-worn wearable and a physical analogue scale (PAS) and to replicate and extend prior findings using a longer assessment window and modern statistical methods.

Prior research on daily laughter includes early event-contingent diary and ESM designs. Mannell and McMahon (1982) studied humorous experiences but not explicit laughter. Kuiper and Martin (1998) and Martin and Kuiper (1999) used paper diaries (3 days) and reported about 18 laughs/day (range 0–89), with frequency peaking in the evening. Observational and audio-recording studies (e.g., Vettin and Todt, 2004; Smoski and Bachorowski, 2003) suggested people may underreport laughter in self-reports. Valeri (2006) found similar frequencies (mean 19/day). Other work examined laughter in interviews, social interactions, or naturalistic contexts (Provine and Fischer, 1989; Provine, 1993; Reay, 2015; Kashdan et al., 2014). Limitations of prior studies include short assessment windows, reliance on paper-and-pencil with compliance uncertainty and recall/postponement issues, frequent occurrence and social context making in-the-moment logging difficult, and correlates often assessed retrospectively rather than concurrently. Advances in ESM using smartphones and wearables allow low-burden, sensor-augmented, in-situ measurement (e.g., PAS), enabling better capture of frequent, fleeting states and behaviors. Based on prior associations, the authors expect higher happiness during laughter, and links between laughter frequency and extraversion (via Type A links), cheerfulness, and well-being, while the association with gelotophobia remains unclear.

Design: Event-contingent ESM with supplemental time-based sampling over 4 consecutive weeks. Participants logged laughter events in situ using a one-button wrist-worn wearable and reported happiness at each event via a Physical Analogue Scale (PAS). Additionally, three pseudo-random time-based prompts per day (9–12, 12–15, 15–18) assessed baseline happiness. Laughter events were categorized as belly laugh (single press) or fit of laughter (double press); time-based baseline was indicated by triple press. Participants and ethics: Initially, 140 participants in a methodological study (smartphone vs wearable). Due to higher burden and missed logs in the smartphone group, only wearable users were retained. Of 65 wearable users, 13 excluded (technical issues n=6; procedural misunderstandings n=7), yielding N=52 (k=9,261 valid assessments after removing 32 invalid button counts >3). Mean age 29.2 years (SD=12.61; range 18–74); 67.3% women; wearable predominantly on left hand (69.0%); position unchanged during study (98.1%). Conducted per ethical standards; anonymous, voluntary; informed consent obtained; institutional waiver; raffle incentives. Measures: - Wearable/PAS (field): Happiness assessed via PAS using wrist/forearm angle 0–90° (rescaled 0–100). Laughter type via button presses (1=belly, 2=fit, 3=baseline prompt). - Online final survey: Big Five (BFI-44; German version; reliabilities: Extrav α=0.87, Agree α=0.87, Consc α=0.86, Neuro α=0.87, Open α=0.78); Satisfaction with Life Scale (SWLS; α=0.88); Gelotophobia Scale (15 items; α=0.92); State-Trait Cheerfulness Inventory – trait version (STCI-T; 20 items each): Cheerfulness α=0.93, Seriousness α=0.75, Bad mood α=0.96. Procedure: Face-to-face onboarding; 4-week field phase with 3 daily prompts for baseline happiness and ad libitum logging of laughter events; post-study online questionnaire; device return and optional personal feedback. Data collected July 2019–January 2022. Data handling and statistical analysis: PAS angles had 8.4% negative values (likely due to posture); sensitivity analysis with winsorized measure (negatives set to 0) did not change outcomes. Multi-level modeling in R (lme4): random-intercept, fixed-slope models with level-1 predictors (baseline vs belly vs fit) and level-2 predictors (gender, age, Big Five, gelotophobia, life satisfaction, STCI cheerfulness, seriousness, bad mood). Predictors grand-mean centered except gender. Baseline model used to compute ICC. Final model omitted cross-level interactions to avoid overfitting. Reported R²GLMM (marginal/conditional). For laughter frequency (RQ2), recurrent event regression using reReg with a Cox-type proportional rate model for pooled belly+fit events. Time of day initially continuous/cyclic; due to non-proportionality, modeled categorically in 3 bins (1–9, 9–17, 17–1).

Descriptives: Across 4 weeks, per participant mean belly laughs = 155 (Median 69; SD=125.98; range 6–556), i.e., ~2.5–5.5 per day (mean vs median). Fits of laughter were rarer: mean 16 (Median 7.5; SD=33.30; range 0–228), i.e., ~0.27–0.57/day (~every 2nd–4th day). Time-based baseline responses: mean 46 (SD=27.7) out of 56 prompts. Laughter increased over the day, peaking in evening; higher Friday–Sunday than Monday–Thursday. RQ1 (Happiness during laughter): Baseline happiness mean ≈ 38.5 (PAS 0–100). Multi-level model showed significant within-person increases: belly laugh +8.6 PAS points (SE=0.66; t=13.01; p<0.001), fit of laughter +12.9 points (SE=1.08; t=11.91; p<0.001) vs baseline. No level-2 predictor reached significance in this model. Model: R²≈43% (conditional); ICC=36%. RQ2 (Predictors of laughter frequency; recurrent event regression): - Positive associations with frequency: happiness (B=0.25, SE=0.02, Z=13.21, p<0.001), openness (B=0.23, SE=0.05, Z=4.46, p<0.001), conscientiousness (B=0.15, SE=0.06, Z=2.59, p=0.009), life satisfaction (B=0.18, SE=0.07, Z=2.46, p=0.014), bad mood (trait) (B=0.29, SE=0.12, Z=2.37, p=0.017; pattern suggests non-linearity), gender main effect (women > men; B=0.76, SE=0.23, Z=3.35, p<0.001). - Negative association: seriousness (B=-0.22, SE=0.08, Z=-2.62, p=0.008). - Age: main effect negative (B=-1.75, SE=0.24, Z=-7.16, p<0.001), with a strong gender×age interaction (B=1.07, SE=0.13, Z=8.18, p<0.001): women showed higher frequency with increasing age, whereas men showed decreasing frequency with age. No significant gender×time-of-day interactions (p>0.23). Time-of-day categorical effects non-significant vs night reference (9–17: B=0.35, p=0.19; 17–1: B=-0.13, p=0.67). - Non-significant predictors: extraversion, agreeableness, neuroticism, gelotophobia (trend negative, especially among women), cheerfulness overall (but significant in men in gender-stratified analyses).

Findings validate the wearable plus PAS approach: happiness was significantly higher during laughter events than at random baseline moments, with larger increases for fits of laughter than belly laughs, directly addressing RQ1. Laughter occurred frequently in everyday life and followed known diurnal and weekly patterns, replicating earlier diary-based research while benefiting from in-situ logging over a longer window. Regarding RQ2, laughter frequency related positively to contemporaneous happiness and life satisfaction, consistent with prior links between laughter and positive affect. Personality associations emerged for openness and conscientiousness, aligning with the idea that more open and organized individuals may experience more social engagement or contexts conducive to laughter. Seriousness related negatively, and trait bad mood showed a positive but likely non-linear relation, suggesting complex dynamics where social norms and context (e.g., affiliative laughter, etiquette) may decouple laughter from trait cheerfulness. Gender differences indicated higher overall frequency in women and a gender-by-age interaction (women’s laughter frequency increasing with age; men decreasing), broadly consonant with earlier descriptive trends. Gelotophobia showed no significant overall association but trended negative, especially among women, indicating a potential avenue for finer-grained state-level assessments in future work. Overall, results replicate and extend prior literature with improved ecological validity, demonstrating the utility of recurrent event models for laughter data.

This study demonstrates the feasibility and validity of assessing laughter in daily life using wrist-worn one-button wearables combined with a physical analogue scale for momentary happiness. Over a four-week ESM, participants reported frequent belly laughs and occasional fits of laughter; happiness increased markedly during laughter events, and laughter frequency showed meaningful associations with personality (openness, conscientiousness), well-being (happiness, life satisfaction), and traits like seriousness, as well as gender and age patterns. The work replicates key findings from prior diary and observational studies while leveraging modern sensing and statistical approaches (recurrent event regression) over a longer timeframe. Future research should enrich contextual information (e.g., social presence, laughter source) via wearables with displays or brief smartphone prompts; examine state-level gelotophobia and situational moderators; model potential non-linearities (e.g., bad mood) more explicitly; and recruit more diverse and balanced samples to test interaction effects with greater power.

- One-button wearable limited per-event information (e.g., could not capture laughter source or social context); devices with screens could allow multi-item in-situ assessments. - Only two laughter types were assessed (belly laugh, fit of laughter); subtler forms (e.g., social smiling) may be under-captured in self-paced event logging. - Sample skewed toward students and convenience recruitment; generalizability may be limited. - Data collection spanned the COVID-19 pandemic; contextual constraints may affect representativeness. - Modest sample size relative to number of predictors limits power for interactions and subgroup analyses; some gender-specific findings should be interpreted cautiously.