Bringing Past Kindness Into the Present: Memories of Acts of Kindness Are Vivid, Memorable, and Feel Good

Do memories of doing good to others make us feel better? And what characterizes these memories? In recent years, there has emerged an interest in the beneficial effects of recollecting memories of acts of kindness dedicated to others (Ko et al., 2021; Titova & Sheldon, 2021). This line of research has yielded promising findings, showing evidence to suggest that recalling past acts of kindness may increase well-being to a similar extent as performing acts of kindness (Ko et al., 2021). Memories of doing good to others (i.e., other-focused kindness) also seem to benefit the remembering person more than memories of doing good to oneself (i.e., self-focused kindness; Titova & Sheldon, 2021). Besides that, bringing these memories to mind might stimulate more acts of kindness toward others (L. Aknin et al., 2011), which may have both intra-individual (e.g., well-being) and interinterindividual benefits (e.g., promoting future acts of kindness). However, although this line of research is thriving these years, it is too early to draw definitive conclusions regarding the benefits of recollecting memories of acts of kindness toward others. While a wide array of studies on the benefits of prosocial (monetary) spending to the actor has emerged, considerably fewer studies have examined the implications of acts of kindness toward others, even though conceptual differences render generalizations across these concepts challenging. The current contribution addresses this research gap.

A second gap in this line of research is how memories of other-focused kindness contribute to well-being after the event took place. While earlier research has primarily looked into the role of need-satisfaction of autonomy, competence, and relatedness to prosocial behaviors (e.g., Gherghel et al., 2019), this approach does not sufficiently explain how past experiences shape well-being through memory recollections. A larger focus on the phenomenological experiences of memories can be helpful in that endeavor. When individuals remember past acts of kindness, a mental re-experiencing of the event takes place (Schie et al., 2019). Yet the features of this phenomenological experience remain overlooked, even though the phenomenological experiences of memories may shape their emotional implications (Schie et al., 2019). While it is commonly assumed that the recollection of past experiences entails a re-experiencing of the emotions felt during the event, these emotions can also influence the vividness and memorability of our memories (Conway, 2005; Conway & Pleydell-Pearce, 2000). In turn, the more vivid and memorable these memories are, the more likely they are to amplify our emotional response when remembering them at later points (Schie et al., 2019). In this sense, investigating the phenomenological experiences of memories of acts of kindness could shed further light on how such memories influence well-being upon later recollection.

The current preregistered research had two main objectives. First, to examine the well-being benefits of recollecting memories of acts of kindness toward others. Second, to investigate the phenomenological features of such memories, and how they relate to well-being. By doing so, this research aims to contribute theoretically and methodologically to the ongoing scientific efforts to understand the emotional implications of acts of kindness.

Previous research suggests that both eudaimonic and hedonic conceptions of well-being characterize a healthy, well-functioning life (C. Keyes, 2002). Hedonic well-being, sometimes also called happiness or subjective well-being, is commonly defined as the presence of positive emotions and life satisfaction (Oishi & Westgate, 2022). This approach focuses on individuals’ subjective evaluation of their life rather than postulating that certain features characterize a good life (Diener et al., 1998). However, proponents of eudaimonic well-being contend that a fulfilling life refers to realizing one’s potential in terms of, for instance, feeling a sense of relatedness and meaningfulness (Deci & Ryan, 2000; Huta & Waterman, 2014). While various approaches have touched upon what eudaimonic wellbeing encompasses, Ryff’s (1989) multidimensional framework is perhaps the most influential approach. According to Ryff (1989), eudaimonic well-being encompasses six distinct facets: selfacceptance, positive relationships with others, autonomy, environmental mastery, purpose in life, and personal growth.

As such, the controversy centers on the question of what constitutes a fulfilling, healthy life. On the one hand, proponents of the eudaimonic conception of well-being have argued that hedonic well-being is an unreliable marker of healthy living in the long term, as emotional states are believed to be highly dependent on external circumstances and situations (C. Ryff & Singer, 1998). On the other hand, proponents of the hedonic conception contend that it provides a more accurate measure of healthy living because it reflects the extent to which individuals live in alignment with their own desires and preferences (Diener et al., 1998). While several researchers have discussed which concept most accurately describes healthy living, others highlight that eudaimonic and hedonic signify different levels of analysis, where eudaimonic well-being constitutes an antecedent of hedonic wellbeing (Martela & Sheldon, 2019).

Despite these controversies, eudaimonic and hedonic well-being may both be critical aspects of a healthy life. Results from cross-sectional and longitudinal studies suggest that people scoring high on both aspects of well-being tend to function better overall. This superior functioning is indicated by fewer instances of mental disorders and better physical health indices (C. Keyes, 2002; C. L. Keyes et al., 2010). When people experience high levels of both eudaimonic and hedonic well-being, they also demonstrate fewer symptoms of depression and anxiety (C. Keyes, 2002; C. L. Keyes et al., 2010), perform better in high school (C. L. Keyes et al., 2012), and have the lowest all-cause mortality risk (C. L. M. Keyes & Simoes, 2012) compared to people experiencing low levels of either eudaimonic or hedonic wellbeing. When seeking to promote overall mental health and functioning, it thus seems crucial to target both aspects of well-being. Numerous interventions have been shown to alter eudaimonic and hedonic well-being, including positive thinking, visualization, gratitude, journaling, mindfulness, engaging in acts of kindness (Bolier et al., 2013) , and, potentially, recollecting memories of past acts of kindness.

Acts of kindness toward others (or other-focused kindness) denote actions performed to benefit others (Nelson-Coffey et al., 2016), such as donating to charity, complimenting someone, visiting a lonely family member or friend, or volunteering. Other-focused kindness augments not only the happiness of the recipient but also the well-being of the actor, highlighting the bidirectional benefits of doing good to others (Chancellor et al., 2018; Curry et al., 2018). As an illustration, engaging in weekly acts of other-focused kindness has been found to increase eudaimonic and hedonic well-being more than weekly acts of self-focused kindness, such as self-care, watching a favorite movie, or having a favorite meal (Nelson-Coffey et al., 2016).

The basic psychological needs framework can be useful for understanding how doing good to others seemingly benefits the actor (Deci & Ryan, 2000). This framework conceptualizes three intrinsic psychological necessities that are considered universal nutrients for optimal function and well-being: autonomy, which refers to the sensation of self-determination; competence, denoting the perception of capability; and relatedness, which signifies the sensation of interpersonal connection (Deci & Ryan, 2000). Empirical work shows evidence to suggest that prosocial behaviors, including other-focused kindness and prosocial spending, contribute to the satisfaction of these psychological necessities (Gherghel et al., 2021; Lok & Dunn, 2020; Titova & Sheldon, 2021). Theoretically speaking, other-focused kindness may foster feelings of autonomy if people experience that they act on their own accord. Additionally, it seems likely that people can feel competent when they experience how their actions benefit others. Finally, considering the prosocial nature of other-focused kindness, engaging in such actions could improve relationships with others and thereby enact a sense of relatedness (L. B. Aknin & Whillans, 2021).

In broad terms, the field of experimental research on prosocial behavior and well-being encompasses two overarching approaches. While a large body of studies on prosocial spending has emerged in the last decade (Curry et al., 2018), other-focused kindness has received notably less attention. Despite similarities, conceptual differences challenge the generalizations of findings across these lines of research. Whereas prosocial spending is being operationalized by instructing individuals to spend money on others (or recollect such memories; e.g., L. B. Aknin & Whillans, 2021), other-focused kindness includes acts of spending money and time on others (Zhang et al., 2021). When taking a closer look at each dimension (i.e., money and time spending) from a basic psychological needs approach, it becomes apparent that their conceptual differences may impact their potential for influencing the actor’s well-being. Time spending has been found to be characterized by a stronger connection to the recipient than money spending (Zhang et al., 2021). In addition, time spending has been shown to increase happiness significantly more than money spending, possibly because time spending more often is directed toward individuals with whom people have a stronger emotional connection (Zhang et al., 2021). In contrast, a common example of money spending is donating to charity, where an immediate and close social interaction with the recipient is unlikely (L. B. Aknin & Whillans, 2021). As such, it cannot necessarily be expected that the benefits of prosocial spending can be generalized to predict the benefits of other-focused kindness, and vice versa.

Although the basic psychological needs approach is commonly employed to explain the connection between prosociality and well-being, additional processes could contribute to this relationship, especially longitudinally. One crucial process might be the mental re-experiencing of past acts of other-focused kindness when recollecting these memories at later points. By revisiting instances where one engaged in other-focused kindness, there is a potential for re-experiencing the satisfaction of basic psychological needs, namely autonomy, competence, and relatedness (Lekes et al., 2014). For instance, the re-experiencing of actions that cultivated meaningful relationships might serve as a reminder of one’s relatedness to others, possibly enhancing the sense of well-being. The richness of the phenomenological re-experiencing of past events can be characterized by several interrelated features, including vividness (i.e., how clearly the memories are experienced), memorability (i.e., how easily the memories are recollected), and their emotionality, which can vary in terms of valence and intensity (Alea & Bluck, 2007; Conway & Pleydell-Pearce, 2000; Lekes et al., 2014; Wilson et al., 2018).

Such phenomenological memory features may play a pivotal role in promoting the emotional implications of memories, such as memories of acts of kindness. When individuals easily and clearly re-experience past events, the emotional response to them can be intensified (Ford et al., 2012). Previous research has shown that vividness is linked to a deeper level of autonoetic consciousness, which refers to a state of remembering past events from a first-person perspective and indicates a richer re-experiencing of emotions during recollection (Schie et al., 2019; Tulving, 1985). In other words, the possibility that we re-experience memories of past other-focused kindness in a more immersive manner than memories of less need-satisfying events could shape our emotional responses to them.

Consistently, previous research has found associations between need-satisfying memories, phenomenological memory features, and well-being (Austin & Costabile, 2021; Lekes et al., 2014; F. Philippe et al., 2010; F. L. Philippe & Bernard-Desrosiers, 2017; Wilson et al., 2018). Wilson et al. (2018) found that vividness in a mental imagery task significantly predicted the observed positive emotional response. Lekes et al. (2014) found that a composite measure of vividness, memorability, and emotionality of recollected memories of intrinsically valuable events predicted life satisfaction. Moreover, Philippe and Bernard-Desrosiers (2017) prompted participants with words from their previously described need-satisfying memories, which increased well-being significantly more than word prompts from others’ need-satisfying memories. Taken together, research points to the notion that other-focused kindness aligns with basic psychological needs, potentially rendering the memories of such events more vivid, memorable, and positive than memories of less need-satisfying events. The present research investigated the phenomenological memory features of past acts of kindness (other; self) to yield further insights into the memory processes potentially contributing to their implications to well-being.

The main objective of the first preregistered study was to examine the well-being benefits and phenomenological experiences of memories of acts of kindness toward others. To pursue this aim, this research utilized two comparison conditions that were instructed to recollect memories of self-focused kindness or trivial, everyday experiences.

To the best of our knowledge, only a few similar studies exist. In a sample of 532 college students, Ko et al. (2021) found that recalling acts of kindness increased hedonic well-being to a similar extent as performing acts of kindness and significantly more than no intervention. Furthermore, Titova and Sheldon (2021) compared memories of acts of kindness toward others vs. oneself. They found that recollecting memories of other-focused kindness increased hedonic wellbeing significantly more than recalling self-focused kindness memories. The internal and external validity of these studies are problematic for the following reasons. First, Titova and Sheldon (2021) applied a within-condition design, which comes with a risk of a carry-over effect (Meltzoff & Cooper, 2018). As mentioned by the authors, it cannot be ruled out that beliefs about how one ought to feel after acting kindly to others compared to oneself influenced participants’ ratings (Titova & Sheldon, 2021). The current study employed a between-condition design to rule out this possible confounder. Second, earlier studies’ internal validity is threatened by the possibility that journaling one’s experiences may affect well-being by itself. Indeed, Baikie et al. (2011) observed that journaling neutral experiences from one’s daily life improved mental and physical health at levels comparable to expressive and positive writing. The current study therefore included a memory comparison condition in which participants were instructed to recall a trivial memory from their daily lives (i.e., shopping for groceries). Finally, previous studies’ external validity is challenged by their reliance on student samples. According to a second-order meta-analysis including data from more than one million participants, student responses are significantly more homogenous than in non-student samples (Peterson, 2001). Additionally, both the magnitude and directionality of outcomes can differ between student and non-student samples (Peterson, 2001). Such findings highlight the importance of replication in non-student samples. To address this, the current study recruited participants through the crowdsourcing platform Prolific, as earlier research has shown that Prolific samples are more heterogeneous in terms of age, education, and socioeconomic status than student samples (Douglas et al., 2023). Besides addressing these methodological limitations, the current research contributes to our understanding of the role of phenomenological experiences of memories of past acts of kindness.

Our secondary aim was to explore the phenomenological memory features of two main categories of acts of kindness. Specifically, this study investigated the phenomenological memory features of acts of spending time and money on others by qualitatively coding participants’ descriptions of other-focused kindness memories. By doing so, this study extends upon earlier findings showing that spending time on others positively shapes well-being when compared to spending money on others (Zhang et al., 2021). This dichotomous categorization was also chosen to increase the statistical power of the analyses (Wilcox, 2001).

This study was conducted in line with the ethical guidelines of the American Psychological Association. Preregistration and Supplementary Materials are accessible at the Open Science Framework (https://osf.io/6gfdt) We explicitly mention where we deviated from our preregistration.

Participants were recruited via the crowdsourcing platform Prolific in exchange for 0.50$. Within our given budget, we could collect data from about 500 participants (Lakens, 2022). A sensitivity power analysis revealed that a sample size of 500 participants would yield sufficient statistical power for detecting between-condition differences of d > 0.31 (β = .80, α = .05, two-tailed). Of the 610 participants starting the study, 75 ended prematurely, 9 failed attention checks, and 5 misinterpreted the assigned recall exercise. The final sample included 521 participants (M_age = 40.49, SD_age = 14.25, range_age = 18 – 78, 50.0% women, 98.8% from North America).

Data collection took place through the PsyToolkit survey platform (Stoet, 2010, 2017). Participants first signed an informed consent form. Participants were then randomly allocated to one of three conditions: 1) describe a past act of kindness towards others, 2) describe a past act of kindness towards self, or 3) describe the last time they were shopping for groceries. Condition three (i.e., recall and describe a memory of grocery shopping) was included to isolate the effect of recalling acts of kindness from the potential effect of journaling one’s experiences. Participants were required to write at least 150 characters to enhance the cognitive and emotional reexperiencing of retrieved memories (L. Aknin et al., 2020). They were also required to label their memory descriptions with a headline to ease the coding of memory descriptions (see Supplementary Materials for full instructions).

After the recall exercise, two sequences of assessments were completed. In the first sequence, memory features (memorability, vividness, emotional intensity, emotional valence, and personal importance) were measured in counterbalanced order. Memorability (“How memorable was this event for you?”), vividness (“How vivid (or clear in your mind) is the memory you have for this event?”), and emotional intensity (“Overall, how emotional was this event for you?”), and personal importance (“How significant or important is this memory to you?”) was measured with the Personal Significance subscale of the Memory Qualities Questionnaires (Alea & Bluck, 2007). Each item was rated on a 5point Likert scale ranging from “Not at all” to ”Extremely”. Emotional valence was assessed with a single item (“How did you feel during the event?”) rated on a 7-point slider scale (-3 to +3) ranging from ”Negative” to “Positive”.

In the second sequence of assessments, eudaimonic well-being and hedonic well-being were measured in counterbalanced order. Eudaimonic well-being was assessed with the Flourishing Scale, which has demonstrated satisfying structural and convergent validity (Diener et al., 2010). It consists of 8 items that were rated on a 5-point Likert scale (α > .88). Examples of items include “My social relationships are supportive and rewarding” and”I lead a purposeful and meaningful life” (Diener et al., 2010, p. 263). Hedonic well-being was assessed with a single item (“How do you feel at the moment?”) that was rated on a 7-point slider scale (-3 to +3) ranging from “Unhappy” to “Happy”. In the end, demographic information (age, gender, geographic location) was collected.

Finally, the first author and a student assistant independently coded all memory descriptions in the other-focused kindness condition with respect to the following categories: Spending money included all acts in which the main cost of the actor was spending something with a monetary value on others; spending time included all acts where time was the main cost of the actor (Zhang et al., 2021). When there was uncertainty as to whether the descriptions referred to acts of spending time or money on others, they were not assigned to any category. Inter-rater reliability was acceptable (IRR = .89). Disagreements were resolved by discussion.

The results are displayed in Table 1 and Table 2. To minimize the risk of Type 1 error, Bonferroni correction was used to adjust for family-wise errors in all planned contrasts, setting the significance level at p < .017 (of note, this is a deviation from the pre-registration). Overall, no significant differences in well-being were observed between the other-focused and self-focused kindness conditions (see Table 2). On the other hand, analyses showed that the self-focused kindness condition reported significantly higher hedonic well-being than the control condition.

Vividness, memorability, emotional intensity, and emotional valence were rated significantly higher in the kindness conditions than in the control condition, yet no differences were observed between the kindness conditions. Non-preregistered Pearson’s correlation analyses showed several statistically significant relationships between the dependent variables (see Table 1). Notably, eudaimonic and hedonic well-being were significantly correlated with each of the phenomenological memory features.

To assess the between-condition differences in well-being, we preregistered ANCOVA tests with time spent on the recall task as a covariate, as prior findings suggest a relationship between time spent on recall tasks and memory accuracy (Chang, 2011). Because no between-condition differences in recall time was detected (F(2,518) = 0.30, p = .738, n² = 0.00), we deviated from the preregistration here by running ANOVA tests instead (see Table 2). Surprisingly, no significant overall difference was detected in eudaimonic well-being. Thus, no further analyses were performed. On the other hand, a significant overall difference in hedonic well-being was found. Planned contrasts revealed that participants in the other-focused kindness condition (n = 177) did not report significantly more hedonic well-being than participants in the control condition (n = 173, t(518) = 1.92, p = .056, d = 0.21, 95% CI = [-0.01, 0.42]. Analyses also showed that participants in the self-focused kindness condition (n = 171) reported significantly higher hedonic well-being than participants in the control condition, (t(518) = 3.08, p = .002, d = 0.33 95% CI = [0.12, 0.54]. No significant difference in hedonic well-being was found between other-focused kindness and self-focused kindness memories, (t(518) = -1.19, p = .237, d = -0.13, 95% CI = [-0.34, 0.08]. The means for the conditions can be found in Table 2.

A series of one-way ANOVA tests were performed to assess between-condition differences in memory features (see Table 2). A significant overall difference was found in emotional valence. Other-focused kindness memories were rated as significantly more positive than control memories (t(518) = 10.21, p < .001, d = 1.09, 95% CI = [0.87, 1.31]. Self-focused kindness memories were also rated as significantly more positive than control memories (t(518) = 10.10, p < .001, d = 1.09, 95% CI [0.87, 1.31]. However, no significant difference in emotional valence was found between other-focused and self-focused kindness memories (t(518) = 0.02, p = .988, d = 0.00, 95% CI = [-0.26, 0.26].

We also found a significant overall difference in emotional intensity. Other-focused kindness (t(518) = 10.35, p < .001, d = 1.11, 95% CI = [0.89, 1.33] and self-focused kindness memories (t(518) = 9.43, p < .001, d = 1.02, 95% CI = [0.80, 1.24] were rated as significantly more emotionally intense than the control memories. No difference between other-focused kindness and self-focused kindness memories was found in emotional intensity (t(518) =0.84, p = .401, d = 0.09, 95% CI = [-0.12, 0.30].

Likewise, an overall difference was found in vividness. Other-focused kindness (t(518) = 4.88, p < .001, d = 0.52, 95% CI = [0.31, 0.73]. and self-focused kindness memories (t(518) = 5.44, p < .001, d = 0.59, 95% CI = [0.37, 0.80] were rated as significantly more vivid than the control memories. No significant difference between other-focused kindness and self-focused kindness memories was found in vividness (t(518) = -0.60, p = .547, d = -0.06, 95% CI = [-0.28, 0.15].

We also found a significant overall effect on memorability. Planned contrasts showed that other-focused kindness (t(518) = 12.39, p < .001, d = 1.33, 95% CI = [1.10, 1.55] and self-focused kindness memories (t(518) = 11.73, p < .001, d = 1.27, 95% CI = [1.04, 1.49] were rated as significantly more memorable than control memories. No significant difference was observed between other-focused kindness and self-focused kindness memories (t(518) = 0.56, p = .576, d = 0.06, 95% CI = [-0.15, 0.27].

Finally, we found a significant overall difference in personal importance. Planned contrasts showed that other-focused kindness memories were rated as significantly more personally important than control memories (t(518) = 14.48, p < .001, d = 1.55, 95% CI = [1.32, 1.78]. Self-focused kindness memories were also rated as significantly more personally important than control memories (t(518) = 14.24, p < .001, d = 1.54, 95% CI = [1.30, 1.77]. No difference was observed between other-focused kindness memories and self-focused kindness memories in personal importance (t(518) = 0.11 p = .910, d = 0.12, 95% CI = [-0.20, 0.22].

Based on the qualitative coding of the memory descriptions in the other-focused kindness condition, we found that 109 participants described acts of spending time on others, whereas 60 participants described acts of spending money on others. Four memory descriptions were not assigned to any category due to uncertainty on how to classify the described act of kindness. In terms of inferential analyses, we deviate from the preregistration here. As above, we preregistered ANCOVA tests with time spent on the recall task as a covariate to assess differences between giving time and money to others. Yet no significant differences was observed in time spent on the recall task (t(167) = -0.99, p = .32, d = -.16). Hence, we ran independent t-tests instead to examine differences in wellbeing and phenomenological features between memories of giving time (n = 109) and giving money (n = 60) to others (see Table 3). No significant difference was observed in eudaimonic well-being. Analyses further revealed that participants who recalled memories of giving time to others did not report significantly more hedonic well-being than participants who recalled memories of giving money. We found no significant differences in memorability, vividness, emotional valence, and personal importance. However, memories of giving money were rated as significantly more emotionally intense.

One of our objectives was to examine the well-being benefits of other-focused kindness memories compared to self-focused kindness memories and everyday control memories. We found no significant differences in well-being between other-focused kindness memories and self-focused kindness memories. However, other-focused kindness memories had a small, almost significant positive effect on hedonic well-being compared to everyday control memories, yet no difference was observed in eudaimonic well-being. Another objective was to explore the phenomenological features of memories of kindness as the first study to date. Other-focused kindness and self-focused kindness memories were significantly more memorable, vivid, positive, intense, and personally important than the control memories.

At first glance, our findings challenge the notion that recalling other-focused kindness memories can alter the sense of eudaimonic well-being, suggesting that the benefits of recalling otherfocused kindness memories are limited to the subjective, hedonic sense of well-being. However, it is necessary to rule out the possibility that the null findings reflect measurements on asymmetrical levels. While hedonic well-being was measured at the state level, eudaimonic well-being was measured at the trait level. Traits are characterized by being relatively stable over time, whereas states are more sensitive to situational influences (Steyer et al., 1999). Theoretically, trait-level change requires stable changes in people’s lives, such as new habits (Bleidorn et al., 2021). State levels, on the other hand, fluctuate depending on situation-specific triggers yet may rapidly return to a baseline level after the trigger has disappeared (Bleidorn et al., 2021), e.g., a retrieved memory. The Flourishing Scale, such as the item “I am a good person and live a good life” (Diener et al., 2010, p. 263), measures eudaimonic well-being on the trait level by prompting participants to reflect on their life in general. In contrast, the single-item tool for measuring hedonic well-being taps into the state level by asking participants how they feel in the current situation. Hence, we conducted another study with an instrument for measuring state eudaimonic well-being to investigate whether the different outcomes reflect measurements on asymmetrical levels.

The second preregistered study had two aims. The first aim was to replicate the results concerning hedonic well-being, as the results contradicted similar research in this field (e.g., L. Aknin et al., 2020; Titova & Sheldon, 2021). The second aim was to investigate whether recollecting memories of other-focused kindness increases eudaimonic well-being when measured with a state-sensitive tool. To this end, an ad-hoc self-report instrument was developed for assessing state eudaimonic wellbeing. It operationalizes the main features commonly associated with eudaimonic well-being, including personal growth, autonomy, relatedness, competence, and meaning/contribution to a greater whole (Huta & Waterman, 2014). These features are also measured in other instruments, such as the Psychological Well-Being Scale (C. D. Ryff, 1989), the Flourishing Scale (Diener et al., 2010), the Mental Health Continuum (C. Keyes, 2002), and the Questionnaire for Eudaimonic Well-Being Scale (Waterman et al., 2010). Unlike these instruments, the items in the present instrument were formulated to be more sensitive to situational influences by asking participants to reflect on their current experience of eudaimonic well-being.

This study was conducted in line with the ethical guidelines of the American Psychological Association. Preregistration and Supplementary Materials are accessible at the Open Science Framework (https://osf.io/6gfdt).

As in Study 1, based on the available financial budget, we aimed to recruit about 500 participants. Participants were recruited through the crowdsourcing platform Prolific in exchange for 0.50$ Of the 631 participants starting the study, 93 ended prematurely, 21 failed attention checks, 2 did not submit a valid Prolific ID, and 2 misinterpreted the recall exercise. The final sample thus included 513 participants (M_age = 41.00, SD_age = 13.35, 50% women, 98.8% from North America). A sensitivity power analysis revealed that with this sample size, we can detect between-condition differences of d > 0.30 (β = .80, α = .05, two-tailed).

The research design, conditions, and survey flow were similar to Study 1, with the exception that memory features were not assessed. Content coding of other-focused kindness memory descriptions was done using the same guidelines as in Study 1. Inter-rater reliability was satisfying (IRR = .93), and disagreements were resolved by discussion.

Hedonic well-being was assessed using the same instrument as in Study 1. State eudaimonic well-being was assessed with a 5-item self-report instrument (range = 5 – 35). Each item operationalized a core feature of eudaimonic well-being as described earlier (Huta & Waterman, 2014). The items were rated on a 7-point Likert scale (from “Strongly disagree” to “Strongly agree”) as findings suggest that the ability to differentiate between responses increases up to this point (Preston & Colman, 2000). Participants received the following instruction: ”Please take a moment to notice how you feel at this very moment when rating the statements* below”. Below the instructions, the following items were outlined: 1) “Right now, I feel connected to other people.” (relatedness), 2)”Right now, I feel I am contributing something good to my surroundings.”* (autonomy), 3) “Right now, I feel I am contributing something good to my surroundings.” (meaning/contribution to a greater whole), 4) “Right now, I feel competent.” (competence), and 5) “Right now, I feel I am growing as a person.” (personal growth). We tested the assumption of unidimensionality through principal axis factor analysis of participants’ ratings of the five items. A factor with an eigenvalue of 3.25 explaining 65.07% of the total variance was extracted. Moreover, all factor loadings were above .70.¹ In other words, a single latent variable is measured when combining the items. Internal reliability was also satisfying (α > .86).

To minimize the risk of Type 1 error, Bonferroni correction was used to adjust for family-wise errors in all planned contrasts, setting the significance level at p < .017 (of note, this is a deviation from the pre-registration). As preregistered, we performed one-way ANOVA tests for comparing between-condition differences in well-being. An overall difference was found in state eudaimonic well-being (see Table 2).² Planned contrasts showed that participants in the other-focused kindness condition (n = 163) reported significantly more state eudaimonic well-being than participants in the control condition (n = 177, t(510) = 3.42, p < .001, d = 0.37, 95% CI = [0.16, 0.59]. Analyses also revealed this pattern between self-focused kindness memories (n = 173) and control memories (t(510) = 3.92, p < .001, d = 0.42, 95% CI = [0.21, 0.63]. No difference was observed between other-focused kindness memories and self-focused kindness memories in state eudaimonic well-being (t(510) = -0.45 p = .656, d = -0.05, 95% CI = [-0.26, 0.17].

An overall between-condition difference was also found in hedonic well-being (see Table 2). Participants in the other-focused kindness condition reported significantly more hedonic well-being than participants in the control condition (t(510) = 3.37, p < .001, d = 0.37, 95% CI = [0.15, 0.58]. Participants in the self-focused kindness condition also reported significantly higher hedonic wellbeing than participants in the control condition t(510) = 3.97, p < .001, d = 0.42, 95% CI = [0.21, 0.64]. No difference was found between other-focused kindness memories and self-focused kindness memories in hedonic well-being (t(510) = -0.58, p = .593, d = -0.06, 95% CI = [-0.27, 0.16].

Finally, we ran non-preregistered independent t-tests to examine differences between recalling memories of giving time (n = 90) and money (n = 62) to others in well-being (see Table 3). There were no significant differences in state eudaimonic well-being and in hedonic well-being. Overall, the study showed that participants recalling acts of kindness reported more state eudaimonic and hedonic well-being than participants recalling everyday memories, while no differences were observed between the kindness conditions.

The current research aimed to assess the well-being benefits and phenomenological features of memories of acts of kindness. For this purpose, we conducted two preregistered experimental studies using an experimental recollection paradigm. Study 2 aimed to replicate the findings concerning hedonic well-being, as well as to utilize a more sensitive tool than in Study 1 to measure eudaimonic well-being at the state-level.

Overall, we found that recalling memories of acts of other-focused kindness improves the sense of state well-being compared to recalling everyday memories, although not in comparison to self-focused kindness memories. Specifically, we generally found higher state hedonic well-being after recalling memories of acts of other- and self-focused kindness than in the memory control condition, likely reflecting the heightened immersiveness and positive emotional valence associated with kindness memories. Similar findings have been observed in a meta-analysis of 27 prospective experimental studies that investigated the emotional implications of engaging in various acts of other-focused kindness in comparison to a wide variety of conditions (δ = 0.28), such as self-focused kindness/personal spending, tracking daily activities, and measurement only (Curry et al., 2018). In line with Ko et al. (2021), our findings therefore suggest that performing and recalling other-focused kindness yield similar improvements in state hedonic well-being. Our findings also indicate that recalling memories of other-focused kindness differentially influences various levels of eudaimonic well-being. Whereas recalling other-focused kindness seemingly increased state eudaimonic wellbeing, no effects were observed in trait eudaimonic well-being. As such, the current research highlights that highly sensitive outcome measures are necessary when examining the well-being benefits of short memory exercises.

We further found that memories of spending time vs. money on others did not differentially affect well-being or the associated phenomenological experiences. These null findings in well-being are in conflict with previous research showing evidence to suggest that spending time on others more strongly influence happiness than spending money on others (Zhang et al., 2021). However, it is challenging to explain these conflicting findings. Possible explanations range from research artifacts (e.g., unreliable assessment tools, and low statistical power) to population characteristics. Indeed, Zhang et al. (2021) participants were Chinese college students, as opposed to North American crowdsourcing participants in the current studies, implying population differences with respect to age, education, and culture. Moreover, natural (statistical) variability across studies should not be neglected as a possible explanation, pointing to the need for more studies on this topic. Future research should also consider the possibility that a more fine-grained conceptualization of prosocial behaviors (e.g., conceptualized in terms of the purpose, reward, and/or recipient of prosocial behaviors) is necessary to delineate robust and replicable differential influences on well-being (Pfattheicher et al., 2021).

Intriguingly, the current research showed evidence to suggest that memories of acts of kindness toward others and oneself do not differentially influence well-being. There might be several possible explanations for this finding. One possibility is that recalling memories of self-focused kindness and other-focused kindness yields similar improvements in well-being by satisfying different important goals (Carver, 1997; Conway & Pleydell-Pearce, 2000). For instance, positive feelings embedded in self-focused kindness memories may be reactivated because these memories tap into innate motives for seeking happiness and pleasure. (Christensen, 2017). Moreover, prior findings suggest that self-focused kindness memories satisfy the need for autonomy significantly more than other-focused kindness memories (Titova & Sheldon, 2021). Various memories of kindness might therefore achieve similar benefits through different pathways.

However, this explanation is unsatisfactory, as it fails to take contradictory findings into account. An alternative explanation is that population characteristics have a moderating role in the well-being benefits of prosociality, as studies conducted with college students yield conflicting results. Specifically, these studies have shown that performing or recalling prosocial behaviors (other-focused kindness/prosocial spending) significantly enhances hedonic well-being compared to self-focused kindness or personal spending (Dunn et al., 2008; Nelson-Coffey et al., 2016; O’Brien & Kassirer, 2018; Titova & Sheldon, 2021; Whillans et al., 2019). In contrast, we included crowdsourced participants and did not replicate the results from studies including college students. To the best of our knowledge, only a few similarly large studies (> 100 participants per condition) also relied on crowdsourced participants. These studies have yielded mixed findings, including insignificant results (Miles et al., 2022), very small effect sizes (d ≤ 0.06) (L. Aknin et al., 2020), and inconsistent results across a series of experimental studies (Varma et al., 2022). As such, these findings indicate a discrepancy in the outcomes observed in studies involving college students and those involving crowdsourced participants.

Further research is necessary to determine if population characteristics account for the contradictory findings in this field. One important area to explore is whether differences in values across populations alter the benefits of prosocial behaviors. As mentioned, it has been suggested that the level of coherence between individuals’ current goals and experiences affects their emotional states, with people feeling better when their actions align with their values (Carver & Scheier, 1990). In line with this notion, cross-sectional research has showed a link between self-reported monthly prosocial spending and happiness specifically among individuals scoring high in self-transcending values, but not among those with lower self-transcendence (Hill & Howell, 2014). Systematic variations in values across different populations may therefore be a promising avenue to explore for whom various acts of kindness are beneficial.

The present paper presented the first exploration of the phenomenological memory features associated with past acts of kindness. Our aim was to assess the vividness, memorability, and emotionality of memories of prosociality, and how these features relate to well-being. We found evidence to suggest that memories of acts of kindness (other; self) are more easily recollected, richer in sensory and contextual detail, and personally more significant than memories of a trivial activity (i.e., grocery shopping). Likewise, we observed relationships between well-being and all of the phenomenological memory features, suggesting an intricate link between how people experience these memories and how they feel. The more immersive, emotional, and positive the memories are reexperienced, the greater the sense of well-being people seemingly felt.

Our findings can have significant implications for understanding how acts of kindness benefit the actor. While it is highly acknowledged that feeling a sense of autonomy, competence, and relatedness mediates the benefits of other-focused kindness (Gherghel et al., 2021; Lok & Dunn, 2022), the question of how such actions continue to influence well-being after the act remained unanswered. Our findings point to the notion that the immersiveness and positivity of kindness memories could promote their beneficial effects. The recollections of such memories can, for example, occur when individuals think about them or share them with others. Indeed, earlier research has shown that people are prone to rehearse positive memories more than negative memories (e.g., Walker et al., 2009). In such circumstances, the recollection of kindness memories could create an opportunity for re-experiencing the event’s need-satisfying properties in an immersive manner, possibly contributing to well-being.

Of note, other-focused and self-focused kindness may give rise to the observed phenomenological memory features through different pathways. While earlier research has shown that other-focused kindness enhanced the sense of relatedness and competence more than self-focused kindness, self-focused kindness was found to increase the sense of autonomy significantly more than other-focused kindness (Titova & Sheldon, 2021). In this line, the observed similarities in the phenomenological experiences of other-focused and self-focused kindness could reflect different pathways to similar outcomes. Recollecting memories of past other-focused kindness could specifically entail the re-experiencing of relatedness and competence, whereas the recollection of self-focused kindness may specifically entail a re-experiencing of autonomy. However, it is important to note that the idea that phenomenological memory features mediate the benefits of past acts of kindness should be interpreted with caution. Indeed, this account remains speculative considering the correlational, non-time-lagged nature of our analyses. Alternatively, the phenomenological memory features could be an epiphenomenon. The current research cannot reject that possibility. Neither does the current research shed light on whether each phenomenological memory feature differentially contribute to well-being.

This contribution has strengths associated with its between-condition experimental design and the inclusion of a memory comparison condition. By incorporating these study features, we were able to delineate the impact of recalling memories of prosociality on well-being from any potential influence that might arise from solely journaling one’s experiences. Nevertheless, the current work also has limitations. We already mentioned the non-time-lagged nature of our analyses above. We want to acknowledge the following additional limitations:

First, outcome assessment was not performed prior to the recall exercises. This decision was based on an assumption that the relatively large sample sizes of our studies would balance the wellbeing levels across conditions. This implies that the observed well-being differences most likely can be attributed to the assigned recall exercises. The fact that Study 2 replicated Study 1’s findings concerning hedonic well-being supports this assumption.

Second, a weakness of this research is the single-item assessment tools. Besides being more difficult to capture fine-grained variance using single-item scales, it is also complicated to assess multifaceted constructs (Robinson, 2018). Thus, it cannot be ruled out that measurement artifacts are responsible for the fact that we observed no difference in state well-being between recalling other-focused kindness and self-focused kindness, which, as previously noted, is in conflict with earlier findings. However, if measurement artifacts explain the contradicting results, it implies that the wellbeing benefits of prosocial behaviors are highly contingent on the manner they are assessed. While replicating prior findings when using different assessment tools increases the confidence in the results, finding variability across tools that assess a similar construct suggests that the outcomes are more complex than initially assumed. In other words, using a variety of assessment tools with different properties within a research field may reveal the subtle aspects of the phenomena being studied, ultimately providing a more comprehensive understanding of the topic.

Third, it is important to highlight the limitations associated with using an ad-hoc tool for measuring state eudaimonic well-being. While its structural validity and internal reliability were supported, the lack of analyses of its convergent and discriminant validity raises concerns about the construct validity of our ad-hoc tool (Flake et al., 2017; Flake & Fried, 2020). To increase construct validity, the assessed features (i.e., autonomy, competence, relatedness, meaningfulness/contribution to a greater whole, and personal growth) were selected to capture the most common features of eudaimonic well-being as conceptualized across various theoretical perspectives (Huta & Waterman, 2014). While it remains to be empirically established that our ad-hoc tool accurately assesses state eudaimonic well-being, the conceptual similarities between our tool and validated tools support this notion.

Finally, we want to acknowledge an alternative explanation for the observed well-being improvements that cannot be ruled out due to the experimental design. As the findings showed, the experimental conditions differed not only in terms of memory content but also in terms of phenomenological memory features. On the one hand, as we have proposed, this finding can be interpreted as suggesting that kindness memories give rise to improved state well-being because of the immersiveness and positivity associated with these memories. On the other hand, the experimental design does not allow for disentangling a relative contribution of various aspects of these memories, as there was no comparison condition to delineate the impact of kindness specifically. For example, an act of kindness toward others may also involve a positive social interaction that by itself could be associated with positive sentiments and an immersive phenomenological experience. Thus, while this study provides preliminary insights into the possible link between acts of kindness, phenomenological memory experiences, and well-being, more research is necessary to further substantiate this suggested link.

To conclude, this contribution examined the well-being benefits and phenomenological experiences of memories of acts of kindness. We showed that methods of memory assessment within the research field of prosociality may contribute to understanding the processes responsible for the benefits of acts of kindness after the event took place. Our findings also support the notion that acts of kindness toward others are beneficial to the actor yet, in contrast to previous research, other-focused kindness was not more beneficial than self-focused kindness. We propose that researchers should put efforts into understanding the conflicting findings in this field by looking further into what works for whom rather than whether acts of kindness benefit the actor.

We express our gratitude to Sophia Agatha Seibt Rijkhoff for her assistance with data coding.

Contributed to conception and design: ABH, AB & SP

Contributed to acquisition of data: ABH & SP

Contributed to analysis and interpretation of data: ABH, AB & SP

Drafted and/or revised the article: ABH, AB & SP

Approved the submitted version for publication: ABH, AB & SP

None.

Supplemental material can be found online at the Open Science Framework: (https://osf.io/6gfdt).

This study was preregistered. Preregistrations, materials, and data are available at the Open Science Framework: (https://osf.io/6gfdt).