Assessing research trends and scientific advances in well-being studies

Abstract

Well-being is becoming increasingly important as an innovation goal for products and services, in addition to traditional value axes such as function, eficiency, quality, price, and size. The value axes of well-being include more individualized mental preferences compared to traditional ones. Therefore, the measurement of well-being incorporating the changes of people's perceptions and academic attention is needed. Using large scale citation data of academic papers, we investigated the elements of well-being and how they have evolved with the research trends in related disciplines. The analysis result of this study uncovered the elements of well-being that cannot be measured only by economic indicators and the necessity for creating new values for future society.

Keywords

• eol>well-being
• bibliometrics
• natural language processing
• information retrieval
• linguistic inquiry

1. Introduction

1.1. Expanding dimensions by two axes

Well-being research is an emerging cross-disciplinary The notion of well-being has repeatedly been updated by ifeld, which attracts many attentions from researchers, referring to the two axes of contrasting concepts, thus policy makers and practitioners recently [ 1, 2 ]. It in- expanding its dimensions. A classic work was Bradburn cludes wide range of disciplines such as economics, psy- and Caplovitz’s psychological well-being, which involved chology, health studies, social sciences, and biomedicine two unrelated dimensions composed of positive and neg[ 3 ]. Responding to the COVID-19 pandemic that triggerd ative feelings, concluding that happiness was determined a renewed awareness of the wealth of well-being more by the balance of the two feelings [ 7 ]. Reviewing this idea than money, World Health Organization reported that with other studies of avowed happiness, Wilson extracted the notion of well-being goes beyond individual percep- the characteristics of happy person as a "young, healthy, tions to societal level that unites the health, economic, well-educated, well-paid, extroverted, optimistic, worrysocial and environmental dimensions of the Sustainable free, religious, married person with high self-esteem, Development Goals by its comprehensive nature [ 4 ]. high job morale, modest aspirations" [ 8 ], which was up

Researchers have revealed the multi-dimensional as- dated by Diener’s works of subjective well-being (SWB). pects of well-being and the necessity of including multi- He explained the trend in SWB theories of understanditem scales for more reliable measures [ 5 ]. For example, ing the processes that underlie happiness with its stress Ryf operationalized six prominent aspects of psycho- on people’s goals, coping eforts, and dispositions, in logical well-being based on the litarature survey, and contrast to classical happiness correlated with the demorevealed that the aspects of positive relations with others, graphic characteristics [ 9 ]. autonomy, purpose in life, and personal growth were Social well-being is one of such counter notions connot strongly captured by previous indexes [ 6 ]. Recently, trasted to individual’s subjective well-being that can have researchers from diferent ethnic backgrounds gathered positive efects towards neighbors and community [ 10 ]. and proposed a global understanding of well-being that Similar viewpoints are seen in the recent concepts of intakes into account cultural relevances for more inclusive terdependent happiness that evaluate happiness based on well-being measures [ 2 ]. relationships over individual happiness [ 11 ], and family well-being over personal well-being [ 12 ].

Other examples that refer to the two axes of well-being AAAI 2023 Spring Symposia, Socially Responsible AI for Well-being, includes new dimensions proposed by representative paMarch 27–29, 2023, USA * Corresponding author. pers in each field detected in this study; such as the Greek $ [e-mail] (H. Yamano) word "eudaimonia" that gives meaning and direction to € https://ifi.u-tokyo.ac.jp/en/people/category/core-faculty/ life contrasted with "hedonia" [ 6 ], youth life satisfaction (H. Yamano) to adults’ ones [ 13 ], positive intervention on depressive 0000-0001-9027-8062 (H. Yamano)

© 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License symptoms [ 14 ], and mental well-being vs physical wellCPWrEooUrckReshdoinpgs IhStpN:/c1e6u1r3-w-0s.o7r3g ACttEribUutRion W4.0oInrtekrnsahtioonpal (PCCroBYce4.0e).dings (CEUR-WS.org) being [ 15 ], which we describe in the result section.

1.2. Meta-analyses of well-being studies

the clusters, using psychometric properties of the word use. It also considered disciplinary shares within clusters and their diefrences based on the journal classifications.

Finally, it aimed to discuss implications of the results for Socially Responsible AI for Well-being and outline future research directions.

There are many meta-analyses extracting the essential features and summarize the complexity of well-being studies with diferent scopes of diciplinary foundation and wide ranged dimensions. These studies include systematic reviews summarizing relationship between outcomes of the states of illness and well-being that have both common and distinct factors on the latent mental 2. Methods health. Based on the meta-analysis, the authors showed the impact of diferent psychological interventions on We took four main steps listed as the following; (1) permental well-being, revealing the eficacy of mindfulness- form citation network clustering, (2) extract represenbased and multi-component positive psychological in- tative words and key features from each cluster and its terventions in both clinical and non-clinical populations sub-clusters, (3) classify sub-clusters by their published [ 16 ]. Recent systematic reviews focused on the specific years and psychometric properties. (4) classify clusters well-being subjects such as life satisfaction of older adults by their published years and academic fields. [ 17 ], entrepreneurs’ well-being [ 18 ], and importance of First, we explored papers that used the terms "well bebiodiversity of greenspace [ 19 ]. Some bibliometric analy- ing" or "well-being". In total, 99,282 papers were retrieved ses investigated the happiness studies in general, using from the Web of Science database. We built a citation netmultiple queries for extracting papers of corresponding work using direct citation that links the articles with their ifelds from the database, and showed the development references. We focused on 67,464 papers in the largest and trends of the fields such as subjective well-being, connected components of the citation network to discard social indicators, and positive psychology, revealing their unrelated papers that had no citation links among them. separate origins and multi-discipnalities [ 1, 20 ]. To obtain research clusters from the citation network,

Many meta-analysis and systematic reviews extract we applied the modularity maximization algorithm of papers based on the existing guidance such as PRISMA Louvain method [ 23 ]. [ 21 ] to avoid the bias in the selection process of papers. Second, to investigate the characteristics of well-being However, because it relies on the human reviewing of clusters, we extracted representative words of clusters, every selected papers, these studies have limitations in calculated by the values of terms in each abstract of the the amount investigated from dozens to hundreds of pa- retrieved papers using tf-icf [ 24 ], that is the product of pers at best. In addition, because they adopt the authors’ the term frequency (tf) and the inverse cluster frequency judgement in discriminating relevant and irelevant pa- (icf) defined as the following equation, pers, it might not exclude bias within their concerns. On tbheer ootfhepraphearnsd,annedt wthoerkwainsdaloymsesoifnccroorwpodrsa;tejuldarggeemneunmts- tf-icf(, ) = , · log , (1) by many researchers reflected in the citations and word use they adopted. Among these approaches, there are attempts to elucidate a conceptual framework to find the linkages between distinct fields such as employee well-being and innovativeness, based on the semantic similarities of the clusters in the network [ 22 ].

This study aimed to build a comprehensive overview of the components of diferent types of research topics of well-being studies created by the citation among papers, irrespective of their existing classifications of the disciplines, and capture the diferential development of these fields being updated by researchers. The primary outcome was the main constructs of well-being studies in general, with a secondary aim to explore diferences in their growth and scientific advances of the fields represented by the clusters in the citation network with their constituent sub-clusters. It further considered the content relationships between the well-being clusters based on their word use and research fields. In detail, it considered the weight of physical and mental content of where and represent the term and the cluster, , is the occurrence of in , is the total number of terms in , and is the number of clusters that include the term , is the total number of clusters.

Third, to investigate the psychometric properties of word use in the clusters of well-being, we investigated the linguistic features of the words, i.e., tens of thousands of words in the title, abstract, or keywords of the papers in each cluster. We employed the LIWC2015 (Linguistic Inquiry and Word Count) dictionary [ 25 ] composed of almost 6,400 English words and 73 categories, including psychological constructs with two-tiered categories such as Afective process (Positive emotion, Negative emotion, Anxiety, Anger, Sadness), Personal concerns (Work, Leisure, Home, Money, Religion, Death), and Biological processes (Body, Health, Sexual, Ingestion). In this study, we used 46 psychological categories and sub-categories in total, divided by 21 mental categories and 25 physical categories. By using tf-icf values of every word in each sub-cluster with their psychological categories, we clas- clusters with less than 3000 papers, because our focus sified sub-clusters of well-being research. We assumed a is the investigation of main constructs in the well-being sub-cluster belongs to mental category, if weighted sum ifeld. As the total number of papers of top 12 clusters of the values of its words in mental categories exceeded were 53,211 which is about 79% of 67,464 papers in the those in physical categories. largest connected components of the citation network,

Fourth, to investigate disciplinary diferences among it covered most papers in the data. We noticed that clusthe clusters, we used the journal categories assigned ters that related to physical health (#11) and family relato each paper, which was provided by Web of Science tions (#10) tended to have older average published year, database. We focused on 12 main categories and excluded whereas those about maintaining mental health (#9, #12) detailed sub-category descriptions followed by ’;’ in each had newer publications, which were indicated by *. category name.

3. Results

3.1. Transition of the cluster shares and their psychometric properties

The component of well-being research have changed dramatically over time. Figure 2 shows that the share of publications had changed diferently among 12 clusters though years from 1995 to 2022. Although well-being research is a growing field with a rapid increase of papers in total, we found many clusters with decreasing shares. Clusters that related to physical health or economic and social conditions (#1, #6, #10, #11) had decreasing share tendencies, compared to those about nature and mental health (#7, #9, #12). The largest increasing trend was seen in the yellow colored cluster #9 related to mindfulness.

We further investigated top 5 sub-clusters of each of the well-being 12 clusters, as shown in Figure 3. Many clusters had both mental and physical sub-clusters with broad distribution of the average publication years. However, the newest two clusters (#9, #12) had only mental sub-clusters with accumulated published years after 2016.

3.2. Trends in the growing clusters

ecosystem service, ecosystem, biodiversity, service, land use, human nature, nature connectedness, forest, green, nature relatedness, biophilic green, urban, green space, biodiversity, ecosystem service, urban green space fishery, conservation, biodiversity, fishing, ecosystem, marine, forest lighting, ofice, noise, occupant, indoor, building, ieq, open plan, employee physician, resident, burnout, medical, wellness, satisfaction, professional mindfulness, meditation, mindfulness based, stress, mbsr, dispositional compassion, self compassion, mindfulness, self, compassionate mindfulness, teacher, stress, mindfulness based, meditation, ibsr, workplace student, medical student, medical, medical school, yoga, stress, pass/fail positive psychology, positive psychology intervention, psychology intervention flourishing, perma, student, spane, teacher, adolescent, flourishing scale gratitude, gratitude intervention, trait gratitude, counting blessing, grateful character strength, strength, character, strength use, positive psychology emotion, emotion regulation, regulation, reappraisal, valuing happiness environment (#7-5) to ecosystem and biodiversity (#7-1, #7-3, #7-4). The newest and large interest in this cluster was seen in the topic related to nature connectedness (#7-2). In Mindfulness cluster, we noticed that newer sub-clusters tended to linked with the wellness of professionals such as physicians and teachres in the workplaces (#9-1, #9-2). In Gratitude cluster, we found a new notion of "flourishing" proposed by Seligman [ 26], which is a model for well-being based on positive psychology (#122). We also found a measure named Scale of Positive and Negative Experience (SPANE), which was proposed by Diener with the scale of flourishing [ 27] in the same sub-cluster.

The distributions of the academic fields and their average publish year within each cluster difered considerably among clusters. For example, Economics (blue) was dominantly seen in Income clurster (#3), whereas Familystudies (green) in Youth cluster (#5) and Family cluster (#10). On the other hand, Psychology-Multidisciplinary (grey) had large shares in many clusters (#1, #2, #3, #4, #5, #6, #12). As for the overall age of the fields, PsychologySocial (orange) had comparativery earlier publication years in most clusters, whereas Environmental-sciences (pink) was the latest in every cluster. We also noticed the origin of the fields in the newest three clusters, and found that of the 12 clusters, Environmental-sciences was orig3.3. Disciplinary diferences and the inated in Nature cluster (#7), whereas Mindfulness (#9) origins of the clusters and Gratitude (#12) cluster had no oldest academic fields within them. In other words, every fields in these two Figure 4 represents the yearly distribution by journal cat- clusters tended to have newer average published years egories, i.e., academic fields of the well-being 12 clusters. than those in the other clusters.

4. Discussion

scientific attention to "connectedness". For example, the latest sub-cluster Nature / connectedness (#7-2) indicated As we have seen in this study, well-being research is a the importance of nature experience and relatedness to multi-disciplinary field with growing dimensions. There green spaces such as forests for human well-being. Anhave been many measures and eforts for updating these other latest sub-cluster was Gratitude / fourishing (#12-2), measures to fit the scales to the latest understanding of where positive relationship is the key recent finding in its human well-being [ 5 ]. However, because of its complex- five metrics named PERMA (Positive emotion, Engageity and growing nature, it might be hard to continue ment, Relationships, Meaning, Achivement). Predicting updating the measures with only human reviewing. To research trend of well-being studies would not be easy. understand the main constructs of well-being study, in- Overall, however, the results of the analysis indicate that cluding recent research findings, we used citation net- the mental aspects are becoming more and more imporwork analysis and text mining to extract basic features of tant, and furthermore, scientific interests are extending the major clusters. We noticed a clear distinction of the to connections with nature and others beyond ones’ own contents and the main fields of study varied considerably happiness or individuals self-interest. among clusters, suggesting that the multi-disiplinarity Based on the results of the meta-analysis, we believe of the well-being research stems from the distinct ori- that inclusiveness, i.e., awareness of oneself and others gins and developmental processes. Finding linkages to would be a key issue for Socially Responsible AI for Wellifll the gap between the clusters / disciplines for more being. Therefore, it is important to ensure that AI not comprehensive understanding of well-being will be the only generates value and convenience, but also does not challenge for the future research. have a negative impact on the mental well-being of citi

The annual transitions of the clusters showed that zens. Specifically, it is important to curb social discriminaclusters about Mindfullness and Gratitude had growing tion, social fragmentation, and the spread of misinformashares, which were in line with the result of the psycho- tion and hate speech that incite anxiety among citizens. metric properties of the words used in their sub-clusters, In bridging the gap between social and economic values, suggesting recent trend of mental well-being research. policy intervention is considered essential in achieving However, these new clusters were consist of relatively old these goals. Specific measures to promote the above inacademic fields, indicating they are growing by connect- clude: encouraging companies to disclose information ing the new notions with established fields. On the other on the development and use of AI; third-party audits of hand, the third newest cluster of Nature had an origin of AI development and use; government R&D support for the academic field; Environmental sciences which spilled the development of AI that enhances wellbeing; and enover to the other clusters. Investigating the contents of hancing AI ethics education at educational institutions. its sub-clusters and diciplinary distributions, we found There are several caveats in our study. As papers withthe Nature cluster started from ofice environmental and out the words of "well-being" or "well being" in any of occupational health reseach, to biophilic nature reseach, their title, abstract, keywords were excluded from the which might reflect people’s growing awareness of finite dataset, we did not include papers that referred to wellnature and its relevance to well-being under the crisis of being in diferent wording or in the body of the paper. climate change. For the prediction and understanding of the evolution A notable new trend seen in these new clusters was the of well-being research, detailed observatory analysis is still needed. Detailed feature design that consider proper information granularity for predicting the future trend of well-being studies is remained for future study. [26] M. E. Seligman, Flourish: A visionary new understanding of happiness and well-being, Simon and

Schuster, 2012. [27] E. Diener, D. Wirtz, W. Tov, C. Kim-Prieto, D.-w.

Choi, S. Oishi, R. Biswas-Diener, New well-being measures: Short scales to assess flourishing and positive and negative feelings, Social indicators research 97 (2010) 143–156.

References

[1] C. Kullenberg , G. Nelhans, The happiness turn? mapping the emergence of “happiness studies” using cited references , Scientometrics 103 ( 2015 ) 615 - 630 .

[2] L. Lambert , T. Lomas , M. P. van de Weijer , H. A. Passmore , M. Joshanloo , J. Harter , Y. Ishikawa , A. Lai , T. Kitagawa , D. Chen , et al., Towards a greater global understanding of wellbeing: A proposal for a more inclusive measure , International Journal of Wellbeing 10 ( 2020 ).

[3] A. C. de Chavez , K. Backett-Milburn , O. Parry , S. Platt , Understanding and researching wellbeing: Its usage in diferent disciplines and potential for health research and health promotion , Health Education Journal 64 ( 2005 ) 70 - 87 .

[4] World Health Organization , Towards developing WHO's agenda on well-being , Technical Report , World Health Organization, 2021 .

[5] M.-J. Linton , P. Dieppe , A. Medina-Lara , Review of 99 self-report measures for assessing well-being in adults: exploring dimensions of well-being and developments over time , BMJ open 6 ( 2016 ) e010641 .

[6] C. D. Ryf , Happiness is everything, or is it? explorations on the meaning of psychological wellbeing ., Journal of personality and social psychology 57 ( 1989 ) 1069 .

[7] N. M. Bradburn , D. Caplovitz , Reports on happiness: A pilot study of behavior related to mental health, 3 , Chicago, Aldine Publishing Company, 1965 .

[8] W. R. Wilson , Correlates of avowed happiness., Psychological bulletin 67 ( 1967 ) 294 .

[9] E. Diener , E. M. Suh , R. E. Lucas , H. L. Smith , Subjective well-being: Three decades of progress ., Psychological bulletin 125 ( 1999 ) 276 .

[10] E. Diener , Guidelines for national indicators of subjective well-being and ill-being ., in: Journal of Happiness Studies: An Interdisciplinary Forum on Subjective Well-Being , Springer, 2006 .

[11] H. Hitokoto , Y. Uchida , Interdependent happiness: Theoretical importance and measurement validity , Journal of Happiness Studies 16 ( 2015 ) 211 - 239 .

[12] K. Krys , C. A. Capaldi , J. M. Zelenski , J. Park , M. Nader , A. Kocimska-Zych , A. Kwiatkowska , P. Michalski, Y. Uchida, Family well-being is valued more than personal well-being: A four-country study , Current Psychology 40 ( 2021 ) 3332 - 3343 .

[13] C. L. Proctor , P. A. Linley , J. Maltby , Youth life satisfaction: A review of the literature , Journal of happiness studies 10 ( 2009 ) 583 - 630 .

[14] N. L. Sin , S. Lyubomirsky , Enhancing well-being and alleviating depressive symptoms with positive psychology interventions: A practice-friendly metaanalysis , Journal of clinical psychology 65 ( 2009 ) 467 - 487 .

[15] C. L. Keyes , The mental health continuum: From languishing to flourishing in life , Journal of health and social behavior 43 ( 2002 ) 207 - 222 .

[16] J. van Agteren , M. Iasiello , L. Lo , J. Bartholomaeus , Z. Kopsaftis , M. Carey , M. Kyrios , A systematic review and meta-analysis of psychological interventions to improve mental wellbeing , Nature Human Behaviour 5 ( 2021 ) 631 - 652 .

[17] S. Khodabakhsh , Factors afecting life satisfaction of older adults in asia: A systematic review , Journal of Happiness Studies 23 ( 2022 ) 1289 - 1304 .

[18] J. C. Sánchez-García , G. Vargas-Morúa , B. R. Hernández-Sánchez , Entrepreneurs' well-being: a bibliometric review, Frontiers in Psychology 9 ( 2018 ) 1696 .

[19] V. Houlden , A. Jani , A. Hong , Is biodiversity of greenspace important for human health and wellbeing? a bibliometric analysis and systematic literature review , Urban Forestry & Urban Greening 66 ( 2021 ) 127385 .

[20] M. Dominko , M. Verbič , The economics of subjective well-being: A bibliometric analysis , Journal of Happiness Studies 20 ( 2019 ) 1973 - 1994 .

[21] A. Liberati , D. G. Altman , J. Tetzlaf , C. Mulrow , P. C. Gøtzsche , J. P. Ioannidis , M. Clarke , P. J. Devereaux , J. Kleijnen , D. Moher , The prisma statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration , Annals of internal medicine 151 ( 2009 ) W-65.

[22] Y. Elsamani , C. Mejia , Y. Kajikawa , Employee wellbeing and innovativeness: A multi-level conceptual framework based on citation network analysis and data mining techniques , Plos one 18 ( 2023 ) e0280005 .

[23] V. D. Blondel , J.-L. Guillaume , R. Lambiotte , E. Lefebvre, Fast unfolding of communities in large networks , Journal of statistical mechanics: theory and experiment 2008 ( 2008 ) P10008 .

[24] J. W. Reed , Y. Jiao , T. E. Potok , B. A. Klump , M. T. Elmore , A. R. Hurson , Tf-icf: A new term weighting scheme for clustering dynamic data streams , in: 5th International Conference on Machine Learning and Applications (ICMLA'06) , IEEE, 2006 , pp. 258 - 263 .

[25] J. W. Pennebaker , , R. J. Booth , R. L. Boyd , M. E. Francis , Linguistic inquiry and word count: Liwc2015, Austin, TX: Pennebaker Conglomerates (www . LIWC.net) . ( 2015 ).