Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Gratitude, Kindness, and Strengths-Based Practices

Abstract. Gratitude, kindness, and strengths-based practices are intentional psychological activities that target beneficial appraisal, prosocial conduct, and deliberate use of valued capacities. This fundamental research monograph examines their historical experimental emergence, conceptual boundaries, mechanisms, randomized studies, cross-cultural evidence, replication, meta-analyses, delivery conditions, methodological limitations, and ethical implications. It distinguishes effects on immediate affect, life evaluation, sustained functioning, social relationships, and recipients' welfare. The synthesis finds small average improvements for several well-being outcomes, especially in gratitude research, coupled with substantial heterogeneity and sensitivity to comparators and measurement. Evidence for acts of kindness is heterogeneous and often actor-centered; strengths and virtues interventions cover unevenly studied constructs. Future research requires active controls, mechanism identification, long-term follow-up, cross-cultural psychometrics, recipient-centered outcomes, and monitoring of unwanted effects. The article is a critical narrative synthesis of published research, without a claim of new original empirical data or a newly conducted systematic review.

Keywords: gratitude interventions; gratitude journaling; expressed gratitude; acts of kindness; prosocial behavior; strengths-based practices; signature strengths; savoring; positive psychology interventions; randomized controlled trials; active comparators; well-being; positive affect; person–activity fit; intervention dosage; cross-cultural validity; meta-analysis; replication; Artificial Era

Scientific Scope and Research Questions

Gratitude, kindness, and strengths-based practices form a family of intentionally organized psychological activities designed to alter attention, appraisal, interpersonal conduct, and the use of existing capacities. Gratitude activities direct attention toward a received benefit and its significance; kindness activities organize chosen actions intended to benefit other people; strengths-based activities identify characteristic capacities and arrange opportunities for their deliberate use. These activities can be administered as brief experimental tasks, self-directed routines, group programs, or components of longer interventions. Their defining feature is an observable practice with a specified exposure, comparator, and outcome, rather than an aspirational statement about positive personality. The relevant research questions concern what participants actually do, which outcomes change, for whom, relative to what alternative, for how long, and through which processes. Emmons and McCullough (2003) Seligman et al. (2005) Lyubomirsky and Layous (2013)

The present monograph critically integrates foundational randomized experiments, theory, subsequent trials, research syntheses, and methodological debate. Its scientific object is the intervention pathway from assigned gratitude, kindness, and strengths activities to measurable psychological and interpersonal consequences. Trait classification, comprehensive models of flourishing, and general prosociality serve as explanatory background but have independent research identities. Savoring is examined when it operates as an attentional component of an intervention. The inquiry distinguishes an effect on the participant's immediate affect from durable change in life evaluation, psychological functioning, relationship quality, or another person's welfare. It is a critical narrative synthesis of the cited literature: no new empirical observations, exhaustive systematic search, prospectively registered review protocol, or original meta-analysis are claimed.

Four levels of explanation organize the evidence. At the task level, the experimenter assigns a prompt or behavior. At the psychological level, participants interpret benefits, imagine recipients, experience emotion, or rehearse a capability. At the relational level, their conduct enters social exchanges whose consequences may depend on trust and prior obligations. At the population level, opportunities, cultural norms, material resources, and institutional constraints govern who can participate and what outcomes are feasible. These levels should be connected through explicit measurements rather than inferred from a positive group difference. The central estimand in a randomized trial is typically the mean consequence of assignment compared with an alternative condition; a mechanism claim requires evidence about the intervening process. A policy question about who benefits under realistic delivery demands a different estimand and often an implementation study. Lyubomirsky and Layous (2013) van Agteren et al. (2021)

Conceptual Boundaries of Gratitude, Kindness, and Strengths

Gratitude denotes an appraisal or disposition involving recognition of favorable circumstances or benefits and, in interpersonal instances, appreciation of a benefactor. It can appear as a transitory emotion, a recurrent orientation, a communicated expression, or a structured writing exercise. The distinctions matter because a person can appreciate nature, luck, or an event without identifying a human benefactor; can feel privately grateful without expressing thanks; or can express politeness without experiencing corresponding gratitude. Wood et al. (2010) proposed an integrative understanding that encompasses recognition of interpersonal help and habitual attention to positive aspects of life. An intervention may therefore move several psychological processes simultaneously, requiring separate outcome measures rather than treating all gratitude indicators as interchangeable.

Kindness refers here to concrete action undertaken with the intention of benefiting another person, including practical assistance, prosocial spending, considerate communication, and other acts that have an identifiable recipient or beneficiary. The welfare of the actor and that of the recipient constitute distinct endpoints. A pleasant feeling after helping establishes a possible actor benefit while leaving recipient effects to be studied. A character strength is a relatively enduring, positively valued capacity, while a strengths-based practice specifies a behavior in which that capacity is deliberately recognized, applied, or adapted. The scientific question concerns the consequences of the practice and its fit with tasks and contexts. The broader validity of virtues and strength taxonomies requires a different measurement program. Curry et al. (2018) Seligman et al. (2005) Cebolla et al. (2025)

The distinction between an intervention and an evaluative label also protects construct validity. Gratitude can be encouraged through listing past events, expressing appreciation, or attending to ordinary benefits, but these operations involve different social exposures. Kindness may be spontaneous, assigned, anonymous, reciprocal, or publicly visible; their effects need not have the same psychological meaning. A strength may be reported as self-knowledge, inferred from behavior, or assigned by an instrument, with different measurement assumptions. Even closely related constructs should not be collapsed because they share a favorable moral connotation. A cumulative science would specify the observable activity, the intended latent construct, evidence for their connection, and the outcomes likely to change. It would then ask whether an intervention has incremental explanatory and practical value relative to equally credible activities that do not require the proposed label. Wood et al. (2010) Cebolla et al. (2025)

Origins of Experimental Positive-Activity Research

Research on prosocial conduct, beneficial appraisal, motivation, and emotional experience long preceded the contemporary intervention label. The early twenty-first-century positive psychology program brought short, repeatable practices into an explicitly experimental framework with randomized comparison conditions and prospective assessment. Emmons and McCullough (2003) compared gratitude listings with accounts of hassles and neutral or social-comparison writing, observing several favorable outcomes across three experiments, with positive affect the clearest recurring finding. Seligman et al. (2005) then compared a set of Internet-delivered positive exercises, including identifying good things, conveying gratitude, and using signature strengths in a new way. These investigations translated theoretically interesting psychological operations into assignable treatments and generated questions about expectancy, participant selection, practice persistence, and the durability of measured change.

The experiments were historically consequential because they showed that specific patterns of attention and action could be manipulated prospectively. Their experimental framing also made alternative explanations more visible. Completing a weekly journal requires time and reflection; following any assigned routine may produce mastery, perceived agency, or social meaning. A participant who knows that an exercise is designed to increase happiness may report improvement because of demand or personal expectations. The initial findings therefore establish activity effects within particular designs and populations, with their explanatory specificity left open to independent tests. The subsequent history of the field is partly the history of developing more credible controls, broader population samples, longer follow-up, and cumulative reanalysis. Davis et al. (2016) White et al. (2019)

The historical trajectory also illustrates a recurrent research cycle: a compelling laboratory demonstration stimulates dissemination, broad claims attract diverse applications, and subsequent replication and meta-analysis refine the original inference. The gratitude comparisons in Emmons and McCullough (2003) established a manipulable pattern of reflective attention. The multiple-exercise comparison in Seligman et al. (2005) encouraged differentiation among activities with potentially different time courses. Later literature brought more focused quantitative assessment of gratitude Davis et al. (2016) Kirca et al. (2023) and kindness Curry et al. (2018), while methodological reappraisal White et al. (2019) exposed the importance of selection and small-study bias. The field's scientific identity now resides as much in comparative validity and cumulative evidence as in the earliest demonstrations. Chronology should therefore be interpreted as a sequence of changing experimental questions, not a simple record of increasingly decisive confirmation.

Theoretical Models of Intentional Positive Activity

The sustainable-change account developed by Lyubomirsky et al. (2005) proposed that intentional activity is one route through which psychological well-being may be altered over time. Its significance lies in placing behavioral repetition, adaptation, and individual choice into the causal story, even though the proportions attributed to different determinants of happiness in early presentations should not be read as fixed biological constants. Lyubomirsky and Layous (2013) articulated a positive-activity model in which activity characteristics, motivation, effort, and person–activity fit affect outcomes through positive emotions, positive thoughts, positive behaviors, and need satisfaction. The framework produces testable moderation and mediation questions, including whether an intervention succeeds because it changes relationships, perceived agency, attention, or how an individual interprets ordinary events.

Several adjacent theories offer complementary explanations. Fredrickson (2001) argued that some positive emotions broaden immediate cognition and support the acquisition of enduring resources; this is a plausible sequence for gratitude and prosocial engagement but requires direct measurement of intervening steps. Ryan and Deci (2000) linked psychological functioning to autonomy, competence, and relatedness, suggesting that a freely chosen act of kindness may operate differently from a socially compelled task. These models describe hypothesized mechanisms rather than evidence that each gratitude or strengths exercise engages every proposed process. A rigorous synthesis separates theory-based plausibility from experimental demonstration, tests mediators on their own timescales, and recognizes that the same outward behavior may arise from different motives.

The hypothesized pathways can be arranged as competing and complementary causal models. An attentional model predicts that gratitude writing first modifies what events are retrieved and noticed; a relational model predicts subsequent changes in appreciation expression and social response; an agency model predicts that strengths use or kindness increases perceived effective action; an emotional model anticipates a proximal affective shift that may or may not outlast the task. Each model predicts a different temporal ordering of mediators and outcomes. A factorial experiment that separately manipulates attention, communication, and enacted conduct could determine whether their effects are additive, substitutable, or dependent on context. Theoretical flexibility becomes a scientific advantage only when mechanisms yield distinct, falsifiable predictions. The broad accounts of positive activity Lyubomirsky and Layous (2013), positive emotion Fredrickson (2001), and basic needs Ryan and Deci (2000) provide variables for such tests, rather than definitive labels for every observed outcome.

Gratitude Journaling and Benefit Recognition

Gratitude journaling typically asks participants to identify events, people, or circumstances that they value and record them with some regularity. Its experimental attraction is that assignment, frequency, and reported completion can be specified in advance. In Emmons and McCullough (2003), participants were randomized across gratitude, hassles, and alternative writing conditions and assessed repeatedly. Some psychological and behavioral indicators favored the gratitude conditions across the three studies, with positive affect particularly robust. The contrasts identify what happened when one kind of sustained writing was substituted for another, rather than isolating a single biological gratitude mechanism. The studies also demonstrate why a response on a particular well-being scale must be kept distinct from a global claim about flourishing.

An attentive journaling task can increase retrieval of positive memories, alter the salience of benefactors, encourage reframing, and provide a valued daily ritual. Those components may reinforce one another but are also possible alternative mechanisms. The relevant control could therefore be neutral event recording, writing about pleasant experiences, recording helpful actions, or an equally credible reflective practice. Their choice changes the causal quantity being estimated. In meta-analytic comparisons, effects tend to be smaller when controls are psychologically active than when they require minimal engagement. Davis et al. (2016) The methodological implication is that researchers should record the exact prompt, interval, total exposure, and comparison activity, and should refrain from pooling markedly different gratitude tasks as if they were one standardized treatment.

A useful experimental contrast would randomize participants to gratitude listing, neutral episodic recall, and positively valenced recall without beneficiary attribution. The three conditions permit separate estimation of the effect of recalling good events, the effect of positive valence, and the additional effect of framing those events as benefits worthy of gratitude. Researchers could sample experience during daily life, test memory accessibility, and independently measure interpersonal behavior. Compliance should be assessed through task completion without coercing disclosure of private details. Analysis of missing journal entries is essential: people who continue journaling may be those who enjoy or benefit from it, producing biased per-protocol estimates. Such a design would move beyond asking whether one journal is better than no journal and toward identifying a specific causal operation. It remains a proposed refinement of the established work by Emmons and McCullough (2003) and the comparative caution of Davis et al. (2016).

Gratitude Letters, Visits, and Expressed Appreciation

Expressed-gratitude exercises invite participants to articulate appreciation toward another person, sometimes by writing and delivering a letter or making a personal visit. They add communication and social exposure to the cognitive task of benefit recognition. Seligman et al. (2005) reported a favorable but differently timed response for a gratitude visit within a program that also tested other positive activities. In a separate synthesis, Kirca et al. (2023) analyzed 25 randomized controlled studies with 6,745 participants and found a small average effect of expressed gratitude on psychological well-being. These studies support the possibility of benefit in some settings while demonstrating that a letter, a visit, and a private inventory cannot be regarded as identical psychological exposures.

The social consequences of gratitude expression may include recognition, reciprocal warmth, embarrassment, relief, perceived indebtedness, or unexpected relational strain. The recipient's prior relationship with the writer, the context in which gratitude is voiced, cultural expectations, and whether the communication is welcome should be considered before generalizing from self-report improvement in a volunteer sample. Consent is relevant when an exercise requires contacting another person: a grateful participant's objective may differ from the recipient's wishes. Studies that measure only the writer's happiness cannot establish improvements in the relationship or recipient well-being. A stronger design incorporates relational endpoints, assesses the beneficiary when feasible, and distinguishes expression itself from beneficial consequences. Wood et al. (2010) Kirca et al. (2023)

Expressed-gratitude research benefits from treating the communication event as an interpersonal intervention rather than a solitary emotion exercise. The participant may prepare a message, anticipate its reception, transmit it, and then interpret the reply or lack of reply. This sequence allows separate measurements of expectation, anticipatory affect, actual contact, response quality, and later relational functioning. A no-contact writing comparator can distinguish the effect of expression from the effect of composing a grateful narrative, while a non-gratitude supportive message may isolate the role of gratitude-specific content. Participants should be allowed to choose not to contact a recipient and to record discomfort. Studies also need to recognize that recipients may experience unequal obligation or pressure. These questions follow from the intervention forms evaluated by Seligman et al. (2005) and Kirca et al. (2023), but direct recipient-centered causal evidence is needed.

Gratitude as Emotion, Appraisal, and Social Information

A grateful response often begins with perceiving something favorable, assigning value to it, and sometimes inferring that another person acted with beneficial intent. Emotional intensity may depend on how unexpected, costly, or personally meaningful the benefit appears, while the interpretation of another person's intention can alter the experience. Repeated gratitude exercises may strengthen accessibility of positive information without changing the objective frequency of benefits received. This suggests a distinction between a change in environment and a change in attention or appraisal, both of which may affect self-report outcomes. Wood et al. (2010) placed these components within a broader model connecting dispositional gratitude, relationships, health, and well-being, while its review cannot by itself identify which component is causally necessary.

Mechanism studies should capture affect close to the gratitude event, benefactor recognition at the relevant social moment, and subsequent relational action over a longer interval. The mediator should be measured after the intervention and before the outcome where the proposed sequence requires that order. A postintervention correlation between feeling grateful and reporting satisfaction is compatible with shared method, general positive response style, or reverse causal influence. Experimental variation in whether participants recall benefactors, express gratitude privately, or communicate with recipients would provide more discriminating evidence. Attributional, attentional, social, and motivational accounts remain compatible with portions of existing results; their relative explanatory contributions require direct comparisons. Lyubomirsky and Layous (2013) Davis et al. (2016)

A persistent analytical challenge is the distinction between feeling gratitude and reporting gratitude. Reflective tasks make culturally valued emotions especially accessible in language, and participants may infer that favorable evaluations are expected. Triangulation could combine immediate affect ratings with privately recorded event descriptions, behavioral expressions of appreciation, and reports from people involved in the relevant exchange. Yet no behavioral indicator is an infallible measure of internal appraisal: thanking someone may reflect politeness or institutional obligation. Longitudinal mediation also requires separating stable individual differences from within-person changes, because people who are more grateful on average may differ systematically in social resources. The most convincing causal evidence would show that manipulating a proposed appraisal component changes subsequent outcomes while keeping alternative components as comparable as feasible. Wood et al. (2010) Lyubomirsky and Layous (2013)

Savoring and Attentional Reappraisal Within Gratitude Practice

Savoring is the deliberate attention paid to positive experience, its anticipation, present enjoyment, or recollection. Within this dossier, savoring is relevant as a possible component of gratitude activity: recalling a helpful encounter may prolong positive feeling, and recording valued experiences may promote reflective enjoyment. The intervention logic differs from expressing thanks to a benefactor because the primary manipulation can be directed toward sensory, autobiographical, or social attention. Treating savoring as the hidden active ingredient of gratitude would require experiments that hold writing time and positive content constant while manipulating grateful appraisal separately. The positive-activity model offers a reason to expect positive thoughts and emotions to contribute, but does not resolve which operation is distinctive. Lyubomirsky and Layous (2013)

Measurement should distinguish the remembered quality of a pleasant event from the emotion felt while recalling it, and should assess whether repeated recollection changes the emotional meaning of the event. When the outcome itself asks participants to list good things, the apparent treatment effect may partly reflect practice at the assessment task. Ecological momentary assessment and independent observation can reduce overlap between assigned writing and evaluation. Researchers can also examine whether familiarization leads to diminishing responsiveness or whether variety preserves engagement. The appropriate theoretical question is whether deliberate attention to beneficial experience changes everyday experience beyond the duration of an exercise and relative to a comparably attentive activity. Lyubomirsky et al. (2005) Fredrickson (2001)

Savoring may produce benefits through repeated contact with positive memory, yet repetition can create adaptation or make an emotionally significant event feel routine. Experiments can test whether participants respond differently when asked to vary examples, attend to sensory details, identify beneficiaries, or imagine future opportunities for appreciation. Randomized micro-interventions embedded within daily life could help estimate immediate affective responses without relying only on retrospective global satisfaction. The duration of benefit should also be measured: a prompt that elevates mood for a few minutes can be useful for a specific purpose, even if it does not alter long-run life satisfaction. Conversely, a slow relationship effect might be missed by an immediate mood measure. The research objective is to map these time scales empirically and evaluate how experiential attention fits within the broader positive-activity model. Lyubomirsky and Layous (2013) Fredrickson (2001)

Acts of Kindness as an Experimental Intervention

Acts-of-kindness protocols ordinarily instruct participants to perform a certain number of helpful actions over a defined period. The actual behaviors can differ radically: offering assistance, sending support, performing a household task, volunteering, or quietly helping a stranger. Nelson et al. (2016) compared kindness directed toward other people or the world with kindness directed toward oneself and a neutral activity over six weeks. The study provides experimental evidence that the destination and social meaning of a positive act can matter for the participant's psychological flourishing. Yet the range of participant-selected behaviors complicates attribution to a single mechanism, and self-reported compliance may differ from observed helpful conduct.

In a synthesis of 27 experiments totaling 4,045 participants, Curry et al. (2018) reported a small-to-medium average benefit to the actor's well-being, estimated at δ = 0.28. This estimate aggregates heterogeneous tasks and settings, and it measures changes in the helper rather than direct improvements for beneficiaries. A rigorous causal description should identify the assigned behavior, how the recipient was selected, what the participant genuinely did, and what rival activities might have produced similar benefits. Kindness is a family of behaviors embedded in social opportunity structures; the effectiveness of assigning kindness cannot be separated entirely from whether participants have meaningful, safe and voluntary opportunities to help.

To interpret acts-of-kindness trials, researchers need a taxonomy of what was enacted, who benefited, and how much autonomy the actor retained. A participant might provide companionship, perform a burdensome service, share resources, or offer information; these actions can have different costs and different effects on belonging or competence. Digital self-reports can register intention without verifying enactment, and social praise can alter reporting. Behavioral confirmation, recipient consent, and measures of perceived cost can strengthen assessment, although full surveillance would be ethically intrusive. The intervention's dose should reflect the burden and meaning of each activity rather than simply counting acts. The empirical results of Nelson et al. (2016) and Curry et al. (2018) justify examining actor well-being, while stronger claims about social welfare require direct measurement of recipients and consequences beyond the actor's immediate satisfaction.

Prosocial Spending, Replication, and the Importance of Context

Spending money on others provides a relatively standardized way of manipulating one form of prosocial behavior. Dunn et al. (2008) reported that expenditure directed toward others was associated with or experimentally related to happiness in its particular designs. Later preregistered work tested the claim at greater scale. Aknin et al. (2020) reported that initial replication experiments did not consistently reproduce the original contrast, whereas a later experiment involving 5,199 participants found a small advantage for recalling prosocial rather than personal expenditure. The authors emphasized methodological features and engagement as determinants of apparent effects. These results illustrate the scientific value of replication: a directionally plausible intervention may have a smaller, more context-dependent effect than a highly cited initial demonstration suggests.

The meaning of giving money depends on purchasing power, interpersonal obligation, the beneficiary's needs, participants' socioeconomic position, and whether a spending decision is discretionary. A small gift chosen freely may create connection or agency; a demanded transfer can produce financial strain or resentment. Consequently, donation amount alone does not supply a psychologically adequate treatment definition. Future trials should measure material opportunity, perceived autonomy, recipient benefit, and exposure to gratitude or social approval as separate variables. Large preregistered replications supply more credible precision for the tested conditions, while their conclusions still apply to the specific form of prosocial expenditure rather than every act of kindness. Aknin et al. (2020) Ryan and Deci (2000)

The registered replication literature helps disaggregate study credibility from the social appeal of a hypothesis. An initially positive expenditure effect may vary with whether people physically spend money, recall prior spending, report predicted pleasure, or complete a hypothetical task. Each procedure measures a different process, and participant engagement may interact with the manipulation. Aknin et al. (2020) is particularly informative because it reports both null and small positive findings rather than forcing the evidence into a single dichotomy. A useful next step is a multicountry, preregistered field trial that varies recipient choice, transaction magnitude and anonymity while assessing the welfare of both parties. Such a study would also stratify by economic security and treat opportunity cost as an outcome, allowing the analysis to distinguish an affordable voluntary gift from a financially consequential demand.

Strengths Identification and Deliberate New Use

A strengths-based intervention frequently begins by identifying capacities that participants regard as characteristic, valuable, and available for purposeful application. A common experimental instruction is to use a recognized signature strength in a new way. Seligman et al. (2005) included such an activity and reported favorable follow-up outcomes in its Internet-based intervention comparison. Its practical appeal is that participants can adapt the operation to work, learning, family relationships, and self-directed goals. The same flexibility makes replication challenging. A respondent who reports using curiosity, perseverance, or kindness may perform very different behaviors from another participant with the same self-description.

There are at least four distinguishable causal components: receiving an affirmative self-description, attending to previously successful behavior, planning a novel application, and enacting that application under real constraints. Positive feedback alone may influence self-efficacy; novelty may make the exercise engaging; deliberate enactment may alter competence or relationships. To distinguish them, investigators can compare a strengths-identification-only group with a strengths-application group, match novelty and planning in active controls, and assess whether the claimed strength is actually used in observable activity. The distinction matters because success in a broad strengths program cannot verify each proposed character classification or imply that every individual should intensify every positively labeled disposition. Seligman et al. (2005) Cebolla et al. (2025)

Strengths interventions should assess the difference between naming a capacity and deploying it. A person may endorse curiosity yet have little practical latitude to explore; another may value perseverance while facing conditions in which additional persistence is harmful. Novel strength use can thus be tested as an interaction between capacity, opportunity and task demands. Structured diaries and observable goal completion can complement an inventory score, while assessments of autonomy, fatigue and role conflict reveal costs. Experimental groups might receive strengths feedback without an application plan, feedback plus planning, or planning for a neutral skill, thereby testing the specific contribution of the strengths frame. This structure also prevents a favorable self-description from being mistaken for behavioral change and clarifies which parts of the early intervention evidence Seligman et al. (2005) can be replicated across real-world settings.

Strengths and Virtues Intervention Evidence

A contemporary synthesis by Cebolla et al. (2025) examined randomized trials of interventions fostering strengths and virtues through 21 systematic reviews. It included 162 randomized trials and approximately 33,032 adult participants, covering a limited subset of the wider virtues and strengths universe. The evidence base is therefore extensive in aggregate yet uneven across particular constructs and delivery formats. Some programs target gratitude or kindness and overlap with the practices examined earlier in this monograph; others focus on distinct capacities or multi-component interventions. Their coexistence in an umbrella synthesis does not establish that every virtue has the same evidence base or that a uniform mechanism explains improvements across them.

Study independence is critical when several reviews include the same original trials. Counting reviews, intervention labels, or meta-analytic outcomes as if they represent distinct sets of participants inflates an impression of evidentiary breadth. Evaluation should return to the underlying trials for the intervention–outcome combination of interest, inspect risk of bias, sample characteristics, active comparators, follow-up and measure validity, and specify whether outcomes represent a strength, a general well-being scale, or functional performance. Broad literature synthesis is useful for identifying candidates for further research and documenting gaps. Claims that specific strengths can be trained, remain durable, or transfer to unrelated domains require stronger targeted designs than are provided by general intervention-level averages. Cebolla et al. (2025) White et al. (2019)

Umbrella reviews require special attention to the unit of evidence. The 21 reviews assembled by Cebolla et al. (2025) do not amount to 21 independent interventions, and many study reports contribute multiple outcomes. The strongest inference about a particular virtue should return to its independently randomized trials and consider intervention fidelity, risk of bias, the definition of that virtue, and the quality of measurement. This review also demonstrates a selection problem: more intensively studied virtues tend to be those for which exercises are easier to operationalize, while difficult-to-measure strengths remain underrepresented. A high proportion of positive comparisons within an intervention category can coexist with weak evidence for long-term functioning if outcomes are short-term and self-reported. A transparent evidence map should therefore display both where trials exist and where they are absent, and resist ranking untested constructs by the reputation of a broader taxonomy.

Operational Definitions, Measurement, and Outcome Hierarchies

An intervention can change a narrow self-report item without producing a corresponding change in a person's broader life. In gratitude research, possible endpoints include momentary gratitude, grateful disposition, positive and negative affect, life satisfaction, relationship quality, and behavioral expression of appreciation. In kindness trials, the possible outcomes include number and character of prosocial acts, the actor's happiness, perceived social connection, recipient welfare, and longer-term relationship functioning. Strengths studies may measure self-reported strength use, personal efficacy, engagement, well-being, or performance in the activity's target domain. These endpoints should be prespecified and interpreted independently. A finding about increased positive affect cannot simply be translated into a conclusion about reduced clinical symptoms or durable functioning. Emmons and McCullough (2003) Davis et al. (2016) Nelson et al. (2016)

Psychometric quality requires reliability, validity, sensitivity to change, and comparison of measurement properties across groups and cultures. A journal-based intervention may increase the readiness to describe positive experiences and thereby affect later self-report. Social desirability may be especially salient when participants are assigned gratitude or kindness. Behavioral logs, recipient reports, repeated experience sampling, blinded coding where feasible, and established scales offer converging but imperfect measurements. To evaluate causal relevance, the analysis must also distinguish adherence to instructions from the quality of a completed exercise and the opportunity to perform it. A participant's inability to find a safe recipient for a kindness task should be represented as part of feasibility, not silently excluded from effectiveness estimates.

A coherent outcome hierarchy can distinguish immediate proximal effects, intermediate behavioral processes and distal valued consequences. For gratitude, proximal outcomes might be attention to benefits and momentary appreciation, intermediate outcomes might be voluntary expressions of thanks, and distal outcomes might be independently assessed relationship functioning. For kindness, actual behavior and recipient experience should precede claims about community trust or durable flourishing. For strengths use, task-specific engagement and competence should be observed before claims about personality development. Researchers can prespecify a primary outcome within the hierarchy and label exploratory findings accordingly. In multi-outcome trials, correction for multiplicity and reporting of null results are necessary to avoid selecting whichever favorable measure happens to reach significance. The measurement issues raised by Davis et al. (2016) and White et al. (2019) become concrete once each claimed benefit is tied to its proposed timescale and observational method.

Randomized Trials, Comparators, and Expectancy Effects

Random allocation helps identify the average causal effect of being assigned to a practice when study conduct preserves the intended comparison. Its strength does not eliminate participants' awareness of receiving an attractive intervention, expectancy-induced reporting, nonadherence, attrition, or outcome selection. A waitlist comparison identifies a different quantity from an activity-matched control involving a credible alternative reflective exercise. Davis et al. (2016) found that gratitude interventions outperformed activity-matched comparators only modestly, while differences from psychologically active comparators were negligible in the comparisons it synthesized. The practical implication is that an intervention may have a benefit compared with inactivity while offering little advantage over another engaging form of attention or self-reflection.

For a definitive trial, investigators should preregister one or a few primary outcomes, follow-up intervals, eligibility rules, planned covariates, and rules for handling missing data. Intention-to-treat estimates preserve the randomized comparison, while per-protocol analyses answer narrower questions and risk selection bias. Control tasks should be comparable in time, credibility, participant attention and assessment schedule when the question concerns a practice-specific mechanism. Researchers should report both the effect of assignment and actual participation without conflating them. If the intervention requires social contact and the control does not, an observed difference may partly represent increased social exposure. These are design choices central to the interpretation of gratitude, kindness, and strengths practices, not administrative details. White et al. (2019) Lyubomirsky and Layous (2013)

Blinding is unusually difficult when participants must knowingly undertake gratitude or kindness exercises, but investigator masking during coding and automated outcome collection are often possible. Designs can also assess the credibility of comparison activities, measure pretrial expectations and record whether participants guessed the preferred hypothesis. When active treatment is defined by a participant-selected action, heterogeneity of actual exposure complicates the interpretation of assignment effects. Randomization should be maintained for the primary analysis, with secondary adherence analyses clearly separated. Cluster randomization may be required when kindness behavior affects shared classrooms or teams and risks spillover across assigned participants. A multisite protocol should document deviations, missing outcomes, adverse reactions and whether control groups received social interaction of comparable intensity. Evidence synthesis is especially sensitive to these details because nominally similar trials can estimate different causal contrasts. Bolier et al. (2013) White et al. (2019)

Meta-Analytic Evidence for Gratitude Practices

Across the broader gratitude literature, reported benefits are typically small, although direction and magnitude vary with design and outcome. Davis et al. (2016) found a modest effect against activity-matched controls for psychological well-being, d = 0.14, whereas comparison with psychologically active controls yielded d = −0.03 with an interval spanning zero. Kirca et al. (2023) synthesized 25 randomized controlled studies with 6,745 participants evaluating expressed gratitude and reported a small positive average effect. These results should be read as estimates for particular sets of trials and comparators; the values are neither interchangeable nor direct measures of an individual's expected improvement. Gratitude interventions include heterogeneous writing, expression, reflection and combined protocols, making design-level subgroup analysis necessary.

A preregistered cross-cultural meta-analysis by Choi et al. (2025) gathered 145 papers, 163 samples, 727 effect sizes, and 24,804 participants from 28 countries. Its pooled effect on well-being was Hedges' g = 0.19 with a 95% confidence interval of 0.15 to 0.22. The authors reported substantial heterogeneity and statistically detectable between-country variation, while evaluated moderators did not establish an explanation for that cultural variation. Importantly, the prediction interval encompassed zero. The pooled result therefore supports a small average effect across studied contexts and leaves room for negligible or different effects in particular populations. More positive results for certain outcome types or intervention combinations do not establish a universally effective practice dosage. Choi et al. (2025)

The cross-cultural synthesis also makes visible how an effect size's confidence interval differs from its expected variability across future settings. The estimate g = 0.19 in Choi et al. (2025) has a narrow confidence interval for the pooled mean, but the reported 95% prediction interval extends approximately from −0.21 to 0.59. A narrow interval around the mean therefore does not establish that every new sample will benefit. The model incorporated dependent outcomes across studies and examined moderators, yet heterogeneity remains relevant to any recommendation. Researchers should communicate the reference population, intervention types and outcome definitions when translating a pooled effect into expected practical value. Decision makers also need baseline distributions and minimally important differences, which standardized effect estimates alone cannot determine. This distinction is central to a scientifically responsible statement that gratitude practices have modest average efficacy and uncertain context-specific magnitude.

General Positive-Psychology Reviews and Bias Correction

Evidence from the broader positive-intervention literature gives useful context while combining methods beyond this dossier's three core practices. Bolier et al. (2013) synthesized 39 randomized studies representing 6,139 participants and found average standardized effects of approximately 0.34 for subjective well-being and 0.20 for psychological well-being, alongside heterogeneity and indications of publication bias. Such estimates generated optimism about brief interventions, yet they are conditional on study selection, outcome aggregation, analytic choices, and the quality of included trials. They cannot be treated as the causal effect of gratitude alone or of a specific strengths application. Differences between pooled estimates across reviews may arise from both genuine intervention diversity and methodological handling of the same literature.

White et al. (2019) revisited meta-analyses of positive psychology interventions and argued that corrected effects were considerably smaller than commonly presented. van Agteren et al. (2021) broadened the assessment of psychological interventions aimed at improving mental well-being, and Wilkie et al. (2026) synthesized randomized trials of well-being-focused programs using network meta-analysis. Together, these reviews reinforce the importance of precise program definitions and active-control distinctions. Their populations, intervention nodes, endpoint definitions and methods differ, so a single ranking of gratitude, kindness and strengths exercises cannot be extracted without preserving what each comparison actually measures. A scientific summary should prefer appropriately calibrated intervals, transparent sensitivity analyses, and transparent recognition of small average effects over claims of universal transformation.

Network meta-analysis extends beyond pairwise comparisons by combining direct and indirect evidence under assumptions of transitivity and consistency. These assumptions become demanding when intervention nodes combine distinct practices and when control conditions differ substantially in intensity. Wilkie et al. (2026) examined 183 randomized trials including 22,811 adults without diagnosed conditions and reported improvements in several intervention categories compared with inactive controls; the analysis also identified considerable risk-of-bias concerns and potential publication bias. These findings illuminate broader program options but do not isolate gratitude from kindness or strengths use within combined programs. van Agteren et al. (2021) reviewed 419 randomized trials in clinical and nonclinical populations, finding effect sizes moderate at most and generally low-to-moderate evidence quality. The appropriate comparative inference is thus conditional on inclusion rules and designs, rather than a universal hierarchy of named practices.

Dose, Repetition, Adherence, and Durability

The behavioral dose of a positive activity includes its frequency, duration, intensity, content, and total time in use. Practice quality and social context may matter more than the number of repeated tasks. Frequent gratitude listing can become habitual and cognitively economical, but it can also become rote or burdensome. Concentrated acts of kindness may differ from actions distributed across the week; writing an appreciation letter once differs from a sustained interpersonal exchange. Lyubomirsky and Layous (2013) identified dosage, variety, motivation, effort, and person–activity fit as potential moderators. These are empirically testable claims: an optimal dose cannot be inferred from a simple finding that one assigned schedule performed better than one control schedule.

Long-term benefit requires distinguishing active practice from residual consequences after practice stops. When participants voluntarily continue an activity, follow-up comparisons combine initial assignment with subsequent adherence and self-selection. Some early exercises appeared to retain effects for months in Seligman et al. (2005), but attrition, continuing engagement and the original online recruitment conditions must be considered. Future studies should record practice trajectories, examine randomized variations in spacing and reminders, report outcomes after discontinuation, and use models accommodating repeated measurement and person-specific change. A sustained improvement in a well-being questionnaire is informative, while the stronger claim of an enduring change in psychological organization requires additional evidence about functioning across contexts.

Dosage trials should distinguish frequency from cumulative time, and scheduled exposure from actual participation. One group could complete a short gratitude exercise daily while another spends equal total time on one weekly session; a third could vary the content while holding exposure constant. This structure identifies whether temporal spacing or novelty matters. Measuring the persistence of benefit requires planned assessments after reminders and task assignments stop. Statistical models must account for participants who disengage because the task is difficult, boring or distressing; otherwise the apparent long-term effect among adherent users may overstate the effect of offering the program. Results can be reported as immediate gain, sustained gain during practice, and residual gain after discontinuation. Such distinctions operationalize the important but broad sustainable-change claims in Lyubomirsky et al. (2005) and the activity characteristics proposed by Lyubomirsky and Layous (2013).

Who Benefits and the Limits of Generalization

An average treatment effect can conceal substantial individual variability. Baseline gratitude, social resources, depressive symptoms, cultural context, developmental stage, material security, and willingness to engage may influence both exposure and outcome. Volunteers recruited because they wish to become happier may be unusually motivated or have different expectations from people participating through schools, employers, or health services. Choi et al. (2025) found variation across countries in gratitude intervention outcomes without identifying a decisive explanatory moderator. This result warrants direct cross-cultural comparison and measurement validation. It does not establish that a single national or cultural trait causes variation in responsiveness.

Cross-population transfer should be treated as an empirical problem. Henrich et al. (2010) described the disproportionate representation of populations from Western, educated, industrialized, rich, and democratic settings in behavioral research. Positive activities may be particularly sensitive to local norms of obligation, self-presentation, gratitude expression, economic giving, and appropriate praise. Researchers should distinguish cultural adaptation of an activity from changes in the construct being measured, ensure linguistic and psychometric comparability, and avoid assuming that social willingness is independent of safety and opportunity. Testing in underrepresented populations requires collaboration with local communities and reporting of socioeconomic and relational conditions. Heterogeneity is scientifically meaningful when it improves causal specification rather than becoming an unexplained label.

Cultural variation should not be reduced to national labels. Within every country, people differ by family obligation, socioeconomic opportunity, institutional trust, language, migration history, and access to safe social relationships. An act praised as generosity in one situation may be treated as an expected duty in another. A gratitude exercise that prompts recognition of a supportive relationship in one context may highlight exclusion in another. Multilevel studies should therefore measure individual values, relationship circumstances and structural opportunities alongside sampling location. Measurement invariance should be evaluated before interpreting differences on happiness or gratitude scales as direct differences in psychological effects. Since Choi et al. (2025) identified significant country variation without identifying an explanatory moderator, research priorities include better culturally adapted manipulations, broader non-Western samples, and prospectively specified contextual variables rather than retrospective interpretation of national averages.

Ethics, Adverse Experiences, and Social Obligation

Activities framed as beneficial can create demands on participants. A gratitude assignment may be unwelcome when a relationship involves coercion, neglect, or conflicting obligations; exhortations to appreciate a benefactor can obscure legitimate anger or the absence of fair treatment. Kindness programs can increase unpaid caring burdens or pressure participants to interact with people who are unsafe. Strengths labeling can reward socially valued behaviors while overlooking structural barriers or turning personal capacities into managerial expectations. These possibilities support an ethical requirement for voluntary participation, context-sensitive prompts, and clearly defined boundaries. Positive intervention research needs explicit assessment of nonbenefit, negative affect, relational discomfort, opportunity cost, and adverse social consequences rather than measuring improvement alone.

The welfare of others is a separate ethical and empirical concern. A participant may gain satisfaction from helping while imposing unwanted assistance on a recipient. A gratitude letter may shift pressure onto a person whose involvement was never sought. Workplace strengths programs can create reputational expectations that distort consent. Responsible research allows participants to decline, substitute safe tasks, and report aversive responses; where appropriate it asks recipients about experience rather than presuming benefit. Clinical applicability requires further qualified assessment: a positive activity tested in volunteers should not be treated as a replacement for established diagnosis or indicated care. The literature supports targeted investigation of these issues, while the absence of systematically collected harm data prevents strong estimates of their frequency. Curry et al. (2018) Davis et al. (2016)

Ethical assessment also requires distinguishing voluntary positive activity from externally imposed emotional labor. In workplaces, a manager's demand for gratitude can implicitly discourage criticism; in dependent relationships, prescribed thanks may reinforce power differences; in schools, mandated kindness can assign unpaid social responsibility to already burdened students. These are possible pathways of harm that should be evaluated rather than assumed to be widespread. Trial protocols can include privacy-preserving negative-effect reporting, participant refusal options, alternative tasks, and follow-up about relational consequences. These safeguards support both better ethics and better science because individuals with unwanted responses are informative observations, not inconvenient deviations. A practice's moral vocabulary does not establish its safety, and an average positive effect does not eliminate the need to consider the distribution of burdens and benefits.

Applied Settings and Comparative Program Design

In educational, organizational, community and counseling settings, positive activities can be delivered individually, in groups, through online applications, or as elements of multi-component programs. Their implementation differs in intensity of supervision, cultural fit, privacy expectations, and opportunities to translate reflection into action. The strongest practical questions concern what a specific setting can support, whether participants understand and endorse the task, how much time it takes away from other activities, and what meaningful outcome would justify continued use. van Agteren et al. (2021) and Wilkie et al. (2026) provide broader comparative evidence on well-being-focused psychological interventions, but comparisons between heterogeneous program categories require caution about transitivity and the limited range of available head-to-head trials.

An applied evaluation should compare the practice with a realistic alternative of similar cost and credibility, measure feasibility and unwanted effects, and report well-being at time points meaningful to the recipient population. In a school, recipient and peer reports may matter more than an immediate mood rating; at work, autonomy and workload should be measured alongside enthusiasm; in counseling, the activity's role must be situated within the individual's broader needs. An intervention that participants enjoy may have implementation value without demonstrating a unique therapeutic mechanism. Cost-effectiveness and equity require their own research designs, including who can take part, who receives the material or social benefit, and whether an apparently small mean effect conceals unequal distributions.

Implementation research should distinguish effectiveness under ideal adherence from effectiveness when a program is offered through an ordinary institution. Randomized pragmatic trials could compare a voluntary digital gratitude module with an equally engaging reflective module, assess reach and completion, and record changes in participant functioning over an academically or organizationally meaningful period. Cost includes facilitator time, participant attention, data stewardship, and potential displacement of other services. Implementation can also alter the intervention itself: a private diary becomes a different practice when shared with supervisors, while a kindness task changes when participants know their conduct is being monitored. A realistic comparative evaluation therefore documents the social arrangement that delivers the exercise. Broad intervention syntheses such as Wilkie et al. (2026) are useful for motivating these comparisons, but their pooled outcomes do not replace direct cost, feasibility and safety data within particular institutions.

Human–Artificial Environments and New Research Questions

Digital systems can prompt gratitude journaling, suggest opportunities for kind acts, and scaffold the use of recognized human strengths. In the Artificial Era, these functions create novel intervention environments: an artificial assistant can adapt reminders, preserve a user's records, propose context-specific tasks, and mediate how social recognition is expressed. The psychological objects established by existing human studies are the user's attention, appraisal, conduct, perceived relationship, and outcomes. A system's production of grateful language should be analyzed as an observable communicative output, with claims about artificial subjective feeling requiring independent operationalization. The first research question is therefore how system-mediated practices alter human experience and action compared with equally credible nonpersonalized and human-supported methods.

Promising designs would randomize prompt timing, personalization, social reciprocity cues, and the availability of human follow-up while preserving data protection and participant autonomy. Outcomes could include enacted kindness toward human recipients, retention of gratitude activities, unassisted strengths use after the tool is removed, relational opportunity costs, and bias in the system's suggestions. Repeated measures would help separate convenience from lasting functional change. Such studies could also test whether artificial assistance redistributes emotional labor or becomes a gatekeeper of what is recognized as a positive event. These are research proposals within the horizon From Homo to Artificial and should be clearly distinguished from established effects of human-only trials.

Automated personalization raises separate validity and governance questions. An artificial system might infer a user's circumstances from prior text and propose appreciation or kindness exercises; the same inference may be mistaken, privacy-sensitive, or culturally inappropriate. Researchers should test the accuracy of the system's context detection, the user's ability to override suggestions, and how often prompts displace real-world human contact. Randomized comparison against simple, nonpersonalized reminders would establish whether complex adaptation offers incremental benefit. A further experiment could compare assisted practice while a tool is available with independent transfer after its removal. The resulting endpoints would concern human cognition, action and relationships; any claim about artificial experience would require a separate conceptual and empirical investigation. This distinction preserves the scientific scope of the current dossier while opening a precise program for Artificial Era research.

Critical Integration and Unresolved Scientific Problems

Across the three practice families, the most secure scientific conclusion is that deliberately structured positive activities can produce modest average changes in selected self-reported well-being outcomes under some experimental conditions. Evidence is comparatively abundant for gratitude practices, smaller and behaviorally heterogeneous for direct kindness experiments, and unevenly distributed across distinct strengths and virtues. Choi et al. (2025) provides a recent quantitative benchmark for gratitude, while Curry et al. (2018) synthesizes experimental kindness and Cebolla et al. (2025) maps the uneven interventions literature on strengths and virtues. The underlying studies do not converge on a single indispensable mechanism. Positive affect, attention, perceived competence, relatedness, social opportunity, expectancies, and reflection remain partially overlapping explanatory pathways.

The largest unresolved questions concern component-specific causal identification, recipient-centered benefits, the long-term trajectory of change, valid cross-cultural translation, individual and institutional moderators, and adverse experiences. The field needs factorial experiments separating content from time and expectancy, active-control trials of specific activities, preregistered longitudinal follow-up, and sufficiently powered designs for credible effect modification. Multi-informant outcomes are particularly important for kindness and interpersonal gratitude. Aggregated effect sizes should be interpreted alongside practical significance, ecological feasibility, baseline variation and the consequences of asking vulnerable participants to perform these activities. A scientific account remains informative precisely when it states both what randomized comparisons support and where measurement and research design leave the explanation open. White et al. (2019) Lyubomirsky and Layous (2013)

A mature evidence program should develop a public, outcome-specific matrix connecting each practice to its experimental comparator, duration, participant population, direct behavioral endpoint, subjective well-being endpoint, recipient assessment, and adverse-effect monitoring. It should mark whether the inference comes from an original randomized trial, a registered replication, a dependent meta-analytic estimate, or a theory paper. Such a matrix could reveal where repeatedly cited experiments supply much of the literature and where independent replications are missing. It could also inform choice of control conditions and prioritize trials with meaningful uncertainty about practice mechanisms. The current evidence supports small average well-being benefits for some assigned activities, and it provides a clear methodological agenda for understanding when those effects persist or carry beyond self-reported experience. Choi et al. (2025) Curry et al. (2018) Cebolla et al. (2025)

Conclusion

Gratitude, kindness, and strengths-based practices constitute a scientifically tractable set of intentional activities that engage appraisal, attention, behavior, and the organization of relationships. Experimental investigations from the early 2000s established feasible manipulations of benefit recognition, interpersonal generosity, and deliberate strengths use. Later trials, replications, quantitative reviews and methodological critiques have produced a more discriminating evidence base. The central finding is a small, context-sensitive average improvement in some well-being outcomes, with substantial variation across practices, comparisons, measurements and participants. Effects in a questionnaire, durable changes in everyday functioning, and benefits for other people must be treated as separate empirical claims. Emmons and McCullough (2003) Seligman et al. (2005) Choi et al. (2025) Curry et al. (2018)

The next stage of research requires better explanation rather than merely additional positive activities: carefully matched controls, component isolation, transparent measurement, long-term observation, cross-cultural validation, recipient-centered evaluation and assessment of unwanted effects. The most useful scientific understanding arises when a practice is specified through its observable operations, its opportunity structure, its proposed mechanisms, its outcome measures and its causal contrast. This framework makes the evidence applicable to educational, relational, organizational and emerging human–artificial environments while preserving the distinct status of observed psychological results and future theoretical possibilities.

The epistemic significance of these interventions lies in their testability. A journal entry, prosocial action, or planned application of a strength can be defined, assigned and compared with credible alternatives. Its consequences can then be evaluated across immediate emotion, sustained functioning, interpersonal consequences, and context-specific costs. Research becomes more reliable as it separates these outcomes and uses independent replications to determine which effects survive stronger designs. The weight of current evidence favors calibrated claims about moderate-to-small intervention effects in the studied samples, rather than uniform effects across all people or all practices. A well-specified intervention is therefore best understood as a context-dependent arrangement of attention, action and opportunity whose value can be assessed against measured alternatives. Davis et al. (2016) White et al. (2019) Choi et al. (2025)

References

  • Aknin, L. B., Dunn, E. W., Proulx, J., Lok, I., & Norton, M. I. (2020). Does spending money on others promote happiness? A registered replication report. Journal of Personality and Social Psychology, 119(2), e15–e26. https://doi.org/10.1037/pspa0000191
  • Bolier, L., Haverman, M., Westerhof, G. J., Riper, H., Smit, F., & Bohlmeijer, E. (2013). Positive psychology interventions: A meta-analysis of randomized controlled studies. BMC Public Health, 13, Article 119. https://doi.org/10.1186/1471-2458-13-119
  • Cebolla, A., Navarro-Siurana, J., Galante, J., Sarrión, P., Vidal, J., Fernández-Castilla, B., Lecuona, O., Vázquez, C., & Baños, R. M. (2025). A synthesis of RCTs on psychological interventions fostering strengths and virtues: Evidence from 21 systematic reviews. Applied Psychology: Health and Well-Being, 17(4), Article e70069. https://doi.org/10.1111/aphw.70069
  • Choi, H., Cha, Y., McCullough, M. E., Coles, N. A., & Oishi, S. (2025). A meta-analysis of the effectiveness of gratitude interventions on well-being across cultures. Proceedings of the National Academy of Sciences, 122(28), Article e2425193122. https://doi.org/10.1073/pnas.2425193122
  • Curry, O. S., Rowland, L. A., Van Lissa, C. J., Zlotowitz, S., McAlaney, J., & Whitehouse, H. (2018). Happy to help? A systematic review and meta-analysis of the effects of performing acts of kindness on the well-being of the actor. Journal of Experimental Social Psychology, 76, 320–329. https://doi.org/10.1016/j.jesp.2018.02.014
  • Davis, D. E., Choe, E., Meyers, J., Wade, N., Varjas, K., Gifford, A., Quinn, A., Hook, J. N., Van Tongeren, D. R., Griffin, B. J., & Worthington, E. L., Jr. (2016). Thankful for the little things: A meta-analysis of gratitude interventions. Journal of Counseling Psychology, 63(1), 20–31. https://doi.org/10.1037/cou0000107
  • Dunn, E. W., Aknin, L. B., & Norton, M. I. (2008). Spending money on others promotes happiness. Science, 319(5870), 1687–1688. https://doi.org/10.1126/science.1150952
  • Emmons, R. A., & McCullough, M. E. (2003). Counting blessings versus burdens: An experimental investigation of gratitude and subjective well-being in daily life. Journal of Personality and Social Psychology, 84(2), 377–389. https://doi.org/10.1037/0022-3514.84.2.377
  • Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist, 56(3), 218–226. https://doi.org/10.1037/0003-066X.56.3.218
  • Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2–3), 61–83. https://doi.org/10.1017/S0140525X0999152X
  • Kirca, A., Malouff, J. M., & Meynadier, J. (2023). The effect of expressed gratitude interventions on psychological wellbeing: A meta-analysis of randomised controlled studies. International Journal of Applied Positive Psychology, 8, 63–86. https://doi.org/10.1007/s41042-023-00086-6
  • Lyubomirsky, S., & Layous, K. (2013). How do simple positive activities increase well-being? Current Directions in Psychological Science, 22(1), 57–62. https://doi.org/10.1177/0963721412469809
  • Lyubomirsky, S., Sheldon, K. M., & Schkade, D. (2005). Pursuing happiness: The architecture of sustainable change. Review of General Psychology, 9(2), 111–131. https://doi.org/10.1037/1089-2680.9.2.111
  • Nelson, S. K., Layous, K., Cole, S. W., & Lyubomirsky, S. (2016). Do unto others or treat yourself? The effects of prosocial and self-focused behavior on psychological flourishing. Emotion, 16(6), 850–861. https://doi.org/10.1037/emo0000178
  • Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68–78. https://doi.org/10.1037/0003-066X.55.1.68
  • Seligman, M. E. P., Steen, T. A., Park, N., & Peterson, C. (2005). Positive psychology progress: Empirical validation of interventions. American Psychologist, 60(5), 410–421. https://doi.org/10.1037/0003-066X.60.5.410
  • van Agteren, J., Iasiello, M., Lo, L., Bartholomaeus, J., Kopsaftis, Z., Carey, M., & Kyrios, M. (2021). A systematic review and meta-analysis of psychological interventions to improve mental wellbeing. Nature Human Behaviour, 5, 631–652. https://doi.org/10.1038/s41562-021-01093-w
  • White, C. A., Uttl, B., & Holder, M. D. (2019). Meta-analyses of positive psychology interventions: The effects are much smaller than previously reported. PLOS ONE, 14(5), Article e0216588. https://doi.org/10.1371/journal.pone.0216588
  • Wilkie, L., Fisher, Z., Geidel, A., Goodall, I., Kamil, S., Davies, E., & Kemp, A. H. (2026). A systematic review and network meta-analysis of randomized controlled trials of well-being-focused interventions. Nature Human Behaviour, 10, 715–726. https://doi.org/10.1038/s41562-025-02369-1
  • Wood, A. M., Froh, J. J., & Geraghty, A. W. A. (2010). Gratitude and well-being: A review and theoretical integration. Clinical Psychology Review, 30(7), 890–905. https://doi.org/10.1016/j.cpr.2010.03.005