A three-month study at a private-sector corporate research institute in Japan reported higher average self-reported PERMA scores alongside use of a daily reflection application, according to a preprint. Average overall PERMA rose from 5.90 at baseline to 6.28 immediately after the intervention. The mean difference was 0.39, with a reported 95% confidence interval of 0.18 to 0.60.
That change is an association in this study, not a causal finding. The field study had no randomized control condition, so the reported PERMA, word-diversity and workshop patterns were exploratory rather than causal.
The tool and its measures
Participants were instructed to use the application daily after work and, after reflecting on positive experiences from that day, select up to three action-oriented well-being words. Subjective well-being was assessed with the 23-item PERMA Profiler, a self-report questionnaire.
Ninety participants took part. Two with missing PERMA data were excluded from the primary analysis, leaving 88 people in that analysis. During the formal intervention period, participants generated 3,480 input records containing 10,104 selected action words and used all 72 words available in the tool.
The researchers summarized PERMA scores and selected-word diversity using means and standard deviations. They used paired-sample t-tests for pre-post comparisons, with Wilcoxon signed-rank tests as a sensitivity check, and used Welch's t-test and Mann-Whitney U tests for workshop-group comparisons. Selected-word diversity was calculated with Shannon entropy from the proportions of selected words in specified time windows.
Where the score changes appeared
The clearest domain-level changes were in positive emotion and accomplishment. Positive emotion rose from 5.67 at baseline, with a standard deviation of 1.53, to 6.22 after the intervention, with a standard deviation of 1.57. Accomplishment rose from 5.73, with a standard deviation of 1.48, to 6.31, with a standard deviation of 1.40. Engagement did not reach the conventional p < .05 threshold.
Among the 74 people who supplied follow-up data, overall PERMA averaged 6.32 immediately after the intervention and 6.33 one month later. The mean difference was 0.01, with p = .911. The available-case follow-up therefore showed no detectable overall change during that month.
Workshop comparisons produced mixed associations
Researchers compared 44 workshop participants with 44 non-workshop participants. Overall PERMA changed by a mean of 0.57, with a standard deviation of 0.93, among the workshop participants, compared with 0.20, with a standard deviation of 1.03, among the non-workshop participants. The numerical difference was not clearly supported at the conventional p < .05 threshold: Welch's t was 1.74, p was .085, and Cohen's d was 0.37.
Cumulative selected-word diversity increased by 0.11, with a standard deviation of 0.13, in the workshop group and by 0.05, with a standard deviation of 0.09, in the non-workshop group. The primary comparison gave p = .028 and Cohen's d = 0.48, while a Mann-Whitney sensitivity analysis gave p = .053.
Among workshop participants, lower pre-workshop alpha diversity tended to be associated with larger subsequent increases. Pearson's correlation was -0.34, with p = .025, and Spearman's correlation was -0.36, with p = .017. These correlations describe an association within the workshop group.
Word choices also varied by department. Hope-related words accounted for 22.5% of selections in research-oriented departments, compared with 12.4% in administrative departments. Contact-related words accounted for 18.6% in administrative departments and 15.9% in research-oriented departments; acknowledgment-related words accounted for 14.5% and 10.0%, respectively. The departmental patterns were exploratory and may reflect department composition, input frequency, workshop participation or organizational culture.
The signal remains tentative
The study's setting limits how broadly the results can be read. It was one three-month field study at a single private-sector corporate research institute in Japan, with outcomes based on self-reported PERMA scores and recalled, self-selected words. The application logs recorded participants' reflections but did not objectively verify that corresponding actions took place. Perceived burden, session duration and completion time were not directly measured.
Several analytic limitations add uncertainty. The analyses were exploratory and were not formally corrected for multiple comparisons. Follow-up used available cases, and those who responded had more input days and selected more words during the intervention. The action-word set was not independently validated as a psychometric instrument, while the internally translated Japanese PERMA Profiler was not formally back-translated or independently validated. Shannon entropy is also sensitive to the number of observations in a time window.
The workshop combined several components, so the study could not isolate their individual contributions. The design leaves the relationship between workshop participation, well-being and cumulative word diversity as exploratory, not causal. The study also did not show that more varied selected words represented better well-being or more diverse real-world behavior.
The next test is controlled
The authors describe the findings as exploratory and hypothesis-generating. They say controlled comparisons are needed to test the tool against alternative, blank-logging or delayed-intervention conditions, while larger studies across languages, cultures, industries and occupations would be needed to assess whether the findings extend beyond this workplace.
Toyota Central R&D Laboratories, Inc. provided financial support for the research and publication, and the authors stated that funding did not influence the study design or reporting. All authors are employees of the company; patent applications were filed for the technology, and all authors are listed as inventors. The manuscript is a preprint submitted for peer review.
Paper data and sources
Original title: The Well-Being Palette: An Action-Word Selection Tool Designed for Low-Burden Reflection on Workplace Well-Being
Authors: Nobuhiko Muramoto, Takayuki Nagaya, Tomoko Tanaka et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text