God of the good gaps: prevalence, eliciting situations, and demonstrations of gratitude to God as compared to interpersonal gratitude

ABSTRACT

Gratitude for another person’s actions has received exponential attention from the scientific community for its many benefits. Yet, this research is virtually silent on one key target of gratitude – god – despite billions of people believing in a personal, intervening, and benevolent god. In a large multi-method study, we sampled U.S. Hindus, Muslims, Jews, and Christians (N = 1270). First, we document the prevalence of spontaneous mentions of god as the target of a gratitude expression following a personal success. Only 16% of our religious participants did mention god but priming god increased this number to 29%. Second, we document a wider array of eliciting situations of gratitude to god (GTG) compared to gratitude to another person (GTO) and particularly for broad good things in life that don’t have a clear agent. Finally, we document ways that GTG vs. GTO is demonstrated, suggesting that GTG sustains religious practice and builds morality.

KEYWORDS:

• Cross-cultural
• religion
• Islam
• Hinduism
• Judaism
• Christianity

The past two decades have produced exponential growth in scientific understanding of a fundamental human experience (Baumgarten-Tramer, 1938; Trivers, 1971) gratitude for another person’s actions (Algoe et al., 2008, 2020; Algoe, 2012; Allen, 2018; Bartlett & DeSteno, 2006; McCullough et al., 2001; Tsang, 2006a, 2006b; Watkins et al., 2006). In the past decade, building on theory about the social functions of emotion (Keltner & Haidt, 1999) and from relationship science (e.g. Reis & Shaver, 1988) researchers have recognized that some of the key benefits from the experience of gratitude, such as relationship development and personal wellbeing, are driven through its expression. As such, researchers have begun to rigorously study how people express, or demonstrate, their gratitude to others (GTO; Algoe & Haidt, 2009; Algoe & Zhaoyang, 2016; Algoe et al., 2016, 2020; Chang & Algoe, 2020; Watkins et al., 2006). The extensive body of empirical work that now looks at gratitude expressed between romantic partners, friends, co-workers, and strangers continues to grow but is virtually silent on one key target of an individual’s gratitude: god¹.

This gap in knowledge becomes all the more surprising when one considers several facts: (1) gratitude for another’s kind actions helps to build relationships (cf. Algoe et al., 2008; Algoe, 2012), (2) most people identify as religious and/or believe in a god (Pew Research Center, 2013; Stark, 2008), (3) god is an entity with whom people are said to have a relationship (Granqvist et al., 2010), and (4) given god’s ubiquitous presence and most world religions’ focus on regular engagement with god through prayer and worship (Yaden et al., 2020), god could be considered an important and close relationship. Further, across world religions, god not only is believed have originally provided all that an individual has, but also to regularly intercede on behalf of an individual’s well-being (Pew Research Center, 2018). Such generous actions are foundational to the experience and expression of gratitude (Algoe et al., 2008; Algoe, 2012; Koenig et al., 2001). Indeed, important religious texts refer to people experiencing and expressing gratitude to god (GTG). Islam: Quran 2:152, ‘So remember Me; I will remember you. Be thankful to Me, and never ungrateful’. Christianity: Thessalonians 5:18, ‘in everything give thanks; for this is the will of God in Christ Jesus for you’. Judaism: Psalm 100:4, ‘Enter into His gates with thanksgiving, And into His courts with praise. Be thankful to Him, and bless His name’. Hinduism: The Mahabharata of Krishna-Dwaipayana Vyasa, Book 3, ‘Whoever with a heart full of gratefulness and free from malice strives to do what is good, attains wealth, virtue, happiness and heaven (hereafter)’.

Yet god is most definitely not the same as another human. There are important theoretical differences in the ‘relationship’ one could have with god versus humans, and very real constraints on one’s ability to directly engage in that relationship through grateful behavior. Here, we merge theory from affective science, relationship science, and the science of religion to delineate important areas of potential similarities and differences in GTG vs. GTO, then begin to address those open research questions with a large-scale multi-method study.

GTG Happens

Most people are religious. A survey by Pew Research Center indicates that in 2010, 83.6% of the world’s population was estimated to be affiliated with a religion, a number projected to grow, not decline by 2050 (Pew Research Center, 2015). Focusing on Americans, a striking 96% report believing in a god or higher power (Stark, 2008). Multiple studies have documented that most theists (including Christians, Jews, and Hindus, but there is not much data among Muslims) represent god as a benevolent agent, as someone who is compassionate, merciful, healing, and helping (see for a review, Johnson, 2021). Moreover, in most world religions, god or a higher power is attributed as the cause of good outcomes in one’s life and often, life itself. For example, Pew Research reports that ‘Nearly eight-in-ten U.S. adults think God or a higher power has protected them, and two-thirds of Americans say they have been rewarded by the Almighty’ (Pew Research Center, 2018).

Critically, the emotional response of gratitude is often triggered by just such attributions (Algoe et al., 2008; Koenig et al., 2001; Ortony et al., 1988). Gratitude is a short-term positive feeling one may experience upon noticing that another person has created a positive outcome for the self, out of care or concern for one’s welfare (Algoe & Stanton, 2012; Algoe et al., 2008; Watkins et al., 2006). As such, psychological scientists have typically defined the emotional response of gratitude as inherently interpersonal.

Given the centrality of god in so many people’s lives and that god is thought to be a benevolent agent causing a wide range of positive outcomes for believers, it stands to reason that religious individuals likely experience GTG. Cross-cultural and cross-religious research has already documented that religious people want to feel more gratitude in general and do indeed experience more gratitude in general (Emmons & Kneetzel, 2005; McCullough et al., 2002; Vishkin et al., 2019). Anecdotally, public figures regularly thank their god following a major success (cf. e.g. ‘to God all the glory’). A very small body of empirical research provides preliminary evidence that religious people experience GTG specifically. In an early and small study among nuns and priests, GTG was the second most experienced emotion after love (Samuels & Lester, 1985). In a cross-cultural study among Black, White, Latinx, and Korean members of Pentecostal and Presbyterian churches, linguistic analyses of narrative accounts of closeness to God and spiritual transformations also showed much reference to gratitude (Abernethy et al., 2016). Therefore, one aim of the present study is to document the prevalence of spontaneous acknowledgement of god’s contributions to positive outcomes in one’s life that would underlie religious individuals’ experiences of gratitude in everyday life.

Relatedly, another aim is to better understand similarities and differences in what people are grateful for when they are grateful to god vs. humans. This is related to the prevalence question because it is possible that people have more reasons to be grateful to god and therefore experience GTG often. In addition to being more numerous, reasons for GTG may also be more diverse. Indeed, given that god is perceived as directly intervening in people’s lives, as creator of the universe, and as having infinite resources, virtually everything can be construed as a gift from god. This includes events that human beings could not typically be responsible for like being spared from a natural catastrophe. Careful documentation of how laypersons use the term ‘gratitude’ shows that people use it to refer to a more general appreciation of positive outcomes in one’s life. This latter form has been referred to as ‘generalized’ gratitude (Lambert et al., 2009), and may include gratitude for the blue sky, one’s health, and even for life itself. In contrast with the more typical form of gratitude, generalized gratitude is not triggered by a specific benefit received and attributed to another person’s actions on one’s behalf. However, religious people have a more readily available agent to attribute such positive outcomes to: their god. God can act as the ultimate source of events and explanation. We predict that GTG may also stem from events that would fall under this umbrella of generalized gratitude.

Expressions of gratitude as a key behavior

In humans, research from dozens of studies supports the theory that the emotional response of gratitude to another for their kind actions helps to build a high-quality, communal relationship with that benefactor (Algoe et al., 2008; Algoe, 2012). This evidence comes from people interacting with strangers as well as with people they already know and love, leading to the conclusions that gratitude is a signal that helps to find new people or remind of current people in one’s life who – by virtue of having demonstrated care and concern for one’s welfare – are good potential relationship partners; and the evidence comes from prospective studies with data from both members of the dyad, which facilitates the conclusion that gratitude promotes the relationship (or ‘binds’ the two together), over time (Algoe, 2012). Critically, the find, remind, and bind theory of gratitude builds on both affective and relationship science to postulate and test hypotheses about the social functions of this emotion.

Functionalist approaches to emotion (cf. Keltner & Haidt, 1999) reveal the value of understanding how emotions are demonstrated. Because emotional responses are theorized to be short-term responses to (real or perceived) events in the environment, involving a coordinated response of mind, body, and behavior to address that event, one take-away is that the behavioral response should – on average – further the aim of ‘addressing that event’. That is, the specific behavior should be relevant to the situation at hand (Baumeister et al., 2007). Another is that, because most events involve other people, expressive behavior often coordinates social interactions. In the case of gratitude, evidence supports the theory that the expressive behavior facilitates the relationship between the grateful person and his or her generous benefactor (Algoe et al., 2016; Chang & Algoe, 2020; Williams & Bartlett, 2015).

Critically, although early research on emotion expression focused on the face (Ekman & Friesen, 1986 but see for full body expression of GTG, Van Cappellen & Edwards, 2021), modern research has identified a wide range of modalities and time-scales in which behavior purported to demonstrate a particular emotional response systematically emerge (e.g. review in Keltner et al., 2019). In the case of gratitude, although the verbal phrase ‘thank you’ is an iconic expression of interpersonal gratitude, in fact, research going back to one of the earliest publications (Baumgarten-Tramer, 1938), updated in a unique modern investigation (Watkins et al., 2006), as well as a very recent investigation (Chang & Algoe, 2020), among others, all reveal that people who feel grateful to another person will demonstrate that sentiment through a wide variety of actions that meet situationally-determined opportunities to promote the relationship with the benefactor, including hugs, trying to improve themselves, wanting to spend more time with the person, or finding a way to repay the person, in addition to simply saying ‘thanks’ to that person. (self-improvement; Chang & Algoe, 2020)All of this raises the fundamental question: what do we know about god and religion that may shape demonstration of GTG in ways that are not obvious in gratitude toward humans?

Specificity of god as the target of the gratitude expression

On one hand, across religions, god is often represented as a social agent with many of the same psychological properties as humans (e.g. think and feel) and with whom one can have a personal relationship (Barrett & Keil, 1996; Granqvist et al., 2010; Heiphetz et al., 2016; Shtulman & Lindeman, 2016). For example, thinking about god in prayer engages neural regions of social cognition suggesting that this experience is comparable to a ‘normal’ interaction with another person (Schjoedt et al., 2009). Therefore, there should be many aspects of the demonstrations of gratitude that may serve the same purpose as or look the same as what is seen in ‘interpersonal gratitude’. Research on gratitude toward humans shows that common ways to demonstrate gratitude are directed at the benefactor. These may include a verbal thank you, finding a way to show the benefactor you care about their welfare (e.g. ‘I wanted some way to confirm to her I thought she was the best’. participant from Algoe & Haidt, 2009), connecting/socializing with the benefactor, or directly reciprocating the kind gesture (Chang & Algoe, 2020). Broadly, by either directly acknowledging the kindness of the benefactor or by directly being kind or friendly toward the benefactor, these behaviors can make the benefactor feel understood, validated, and cared for by the grateful person in the moment (Algoe et al., 2016), which sets the stage for a better relationship. In fact, there are very few documented demonstrations of gratitude – relative to demonstrations of other emotions – that are not interpersonal in nature (Algoe & Haidt, 2009; Watkins et al., 2006).

On the other hand, this anthropomorphic view of god coexists and interacts with an abstract and superhuman view of god (Barlev et al., 2019). God is invisible but omnipresent, omnipotent, omniscient, and directly impacts people’s lives (Knight et al., 2004). Of note, all major religions also include a creation story such that god is perceived as an all powerful creator and the ultimate source of everything; this includes everything good in one’s life. Therefore, considering god opens an important area of discovery and theorizing. For example, one cannot ‘repay’ god (what is the value of a life?) or even know that one’s gesture has hit the mark. To be responsive to god in the service of showing care, valuing, and investment in the relationship, grateful people would use cues about what they ‘know’ about god from their belief systems.

One form of behavior that is encouraged in religions is prayer and performing rituals. These behaviors can be construed as opening a direct line of communication with god, which can be used to directly thank or praise god – acknowledge his actions (Schneller & Swenson, 2013. Indeed, in one qualitative study among 21 older American adults, participants reported that it felt important to them to express GTG and that they choose to do so most often in prayer. They also felt inclined to engage in behaviors (e.g. praying for others) to further demonstrate the sincerity of their thanksgiving (Krause et al., 2012).

Another form of demonstration may be to attempt to improve one’s character. For interpersonal gratitude, some studies have demonstrated a small role for grateful people attempting to improve their own character or perceived worthiness (Algoe & Haidt, 2009; Chang & Algoe, 2020). In a religious context, there may be a greater mandate for such behavior. In fact, many have argued that belief in god primarily functions to regulate social interactions and to promote moral behavior (Graham & Haidt, 2010). This could take the form of the person wanting to improve themselves in order to be worthy of the kind actions of god.

Finally, a now classic study showed that grateful people who do not have an opportunity to show their benefactors their gratitude will show it – in the form of greater prosociality – to the next person who comes along (see Study 2, Figure 2, Bartlett & DeSteno, 2006). This has been dubbed a ‘pay it forward’ effect of interpersonal gratitude (e.g. Beeler Duden & Vaish, 2020). Likewise, between one’s inability to directly be kind to god and cultural/scripted mandates about how to honor god, people who feel grateful to god may be more likely to be prosocial to others, or say that they will ‘pay it forward’. Research has already shown that religious people tend to be more prosocial in general (Shariff & Norenzayan, 2007). If god is perceived as an arbitrator for moral behavior, then wanting to please god when one is grateful may very well take the form of paying it forward. In sum, because gratitude’s behavioral manifestations should serve the function within the context (Chang & Algoe, 2020), it is important to explore both similarities and potential differences in demonstrations of GTG vs. to humans.

Whose god?

In the present study, we sampled religious individuals from the three monotheistic major world religions, also called Abrahamic religions, i.e. Islam, Christianity, and Judaism. We also sampled from a fourth type of major world religion, polytheistic, i.e. Hinduism. Each religion places praise and GTG as a central virtue, critical for honoring god (e.g. God, Gd, Allah, Shiva, Vishnu; Emmons & Crumpler, 2000; Schimmel, 2004). From a psychological perspective, they share similar perceptions of who god is and what god can do. Indeed, much of the work we cited above included some samples of participants from these religions. However, we also recognize that the majority of work in the psychology of religion is still based on Christians. It is therefore important to continue the efforts to be inclusive and representative of the world religions when studying gratitude to ‘god’. In sum, although we expect much similarity between the different religions in the frequency, reasons, and demonstrations of GTG, we still find it important to explore similarities as well as nuances among these groups.

The present study

In the present study, we ask participants to complete three tasks that will provide data for a generative foundation for future work, building carefully on recent research on gratitude toward humans. First, we focus participants on a situation – a success in life – and ask them to express gratitude for it, without specifying a target of the potential gratitude. This task provides an opportunity to observe spontaneous mention of god as the target of gratitude. Personal success is a good context to study attributions of gratitude because religious people can easily attribute success to god or to others. One study found such dual attributions using self-report scales (DeBono et al., 2020). Importantly, right before doing this success recall task, participants were randomly assigned to be primed with the concept of their god or not (Shariff et al., 2016). The benefit of the priming conditions is to document the prevalence of spontaneous expressions of GTG (in the Neutral-prime condition) and to study whether it is possible to experimentally increase the occurrence of GTG (in the God-prime condition).

Second, after this task, we had participants explicitly focus on experiences of GTG. They wrote out an experience of GTG so we could code it for the reasons that this experience occurred, and then they generated a descriptive list of how they demonstrate GTG.

Third, we had them explicitly focus on experiences of interpersonal gratitude (GTO). As with GTG, they wrote about such an experience so we could code them for the reasons that this experience occurred and then generated a descriptive list of how they demonstrate GTO.

We pre-registered this study on OSF (first for Christian participants only, then for the remaining religious groups). We hypothesized that the prevalence of attribution of a success to god would be higher in the God-prime condition than in the neutral condition. We planned to explore the results for our other goals (i.e. reasons for gratitude, demonstrations of gratitude, comparisons between religious groups).

Methods

Participants

Our initial sample of N = 1,830 participants (n = 196 Hindus, n = 186 Muslims, n = 193 Jews, and n = 1,255 Christians) consisted of religious people from all over the United States who were recruited through CloudResearch or Qualtrics Panels. The research was advertised as a success reflection study. Participants self-selected to complete the online survey in exchange for a payment of $1.50–$2.00 on CloudResearch, while Qualtrics Panels participants were compensated with incentives that varied by participant (e.g. gift cards, discounts on products). As pre-registered, we aimed to collect a sample of 1,000 Christians and 150 participants for each other religious group. Among Christians, we also aimed to collect data from a racially diverse sample of participants and therefore set to recruit approximately 250 Black/African American participants, 250 Latinx participants, and 500 participants without targeting a specific race/ethnicity (using quotas in Qualtrics and targeting these groups in CloudResearch).

In order to ensure good quality of responses, we used the following preregistered criteria to exclude participants: finishing the survey too quickly (less than 360 seconds), failing to pass the attention check, or having open-ended responses that were identical to those of other participants (suggesting that these participants copied-pasted the same responses, therefore all participants were excluded). Inspection of the data led to the exclusion of 560 participants, resulting in a final sample size of 1,270 (n = 139 Hindus, n = 150 Muslims, n = 145 Jews, n = 836 Christians; 60.7% female, 0.1% Agender; M_age = 39.86, SD_age = 13.93). A slight majority of participants identified as White/Caucasian (54.2%), and the remaining participants as Black/African American (15.8%), South Asian (12.4%), Hispanic and/or Latino/a (6.0%), Biracial/Multiracial (5.6%), East Asian (2.4%), Middle Eastern/North African (2.4%), Other (0.8%; e.g. Nepali, West Indian), American Indian or Native Alaskan (0.4%), or Pacific Islander or Native Hawaiian (0.1%).

Procedure and materials

In an online Qualtrics survey, participants first gave consent and then were randomly assigned to one of two priming conditions: the God-prime condition or the Neutral-prime condition. The God-prime consisted of a task where participants wrote about their relationship with God, while those in the Neutral-prime condition wrote about their favorite colors that were present in their surroundings. Participants were required to write about each prompt for at least two minutes. Next, they completed a success recall task using the following prompt: ‘Please think back to a time when you achieved something that made you feel proud of yourself. Please reflect on, then describe, that event in detail’. Afterwards, they were asked to express in writing gratitude for this success: ‘Now, still thinking about this success, use the space below to express gratitude for it. Please think about how you would like to express gratitude for this success now, not how you expressed it in the past. Please take as much time as needed to write your response’. Participants then completed a series of self-report measures related to whom they attribute their success (Success Attributions; we also measured but do not report on feelings of deservingness, and attribution of religious meaning to the success).

Afterwards, participants were asked to think about how they demonstrate GTG and to aid in this, to first write about personal instances of GTG (‘First, it may help for you to think of recent times, or one strong time when you actually experienced GTG’; see OSM for full prompt). Responses to this question set the stage for the next question on demonstrations and also provided data on the reasons that people mention for what elicited GTG. Then, participants were asked to write up to seven strategies for how they demonstrate their GTG (see Chang & Algoe, 2020): ‘We know gratitude is not only an emotion but also an action. In this part of the study, we are interested in how you show your feelings of GTG. During the next five minutes, we would like for you to name at least 1, and at most 7, strategies that you use to show gratitude when God has been good to you. In your answers, please “name” the strategies and give us a short description of each strategy you pick out’. Note that the word god is used here as a generic term, but was adapted for each religious group. Finally, participants completed the same two tasks, but for gratitude to other people (GTO; see codebook on OSF for full instructions). Information regarding religious beliefs (i.e. importance of religion, see OSM Table S1), religious practices (i.e. frequency of worship and prayer, see OSM Table S1), and demographics was collected last.

This procedure was approved by the first author’s institutional review board. We report all manipulations and data exclusions. In addition to the variables presented here, we also collected data, but do not report it here, on socially desirable responding, plans for Thanksgiving, and names of gods for Hindu participants.

Self-report variables

When examining the open-response data, it became apparent that some participants were not compliant and did not provide thoughtful responses (e.g. writing about unrelated topics). We therefore decided to flag those participants (n = 71) and to not include their self-report data for analyses based on the assumption that they also probably did not pay much attention to the scales. To determine whether a participant was considered as non-compliant, all open-ended responses were evaluated collectively in the Hindu, Muslim, and Jewish samples, and only the open-ended response to the success prompt was evaluated in the Christian sample.

Success attributions (god, self, and other people)

Participants answered nine questions asking how much they attributed their success to god, self, and others (three questions per attribution target), using a Likert scale of 1 (not at all) to 9 (yes, extremely). The Attribution to God scale was modified from Mahoney et al. (2003), whereas Attribution to Self and Others were created for the purposes of the current research. Examples of items include: ‘I was successful thanks to God’, ‘I am responsible for my own success’, and ‘People around me have helped me achieve this success’. Reliability for these scales ranged from good to excellent (Attribution to God: α = 0.98; Attribution to Self: α = 0.75; Attribution to Other People: α = 0.91).

Coded responses

The open-ended responses to the gratitude writing prompts were coded to assess three variables: the targets spontaneously mentioned when expressing gratitude for a success (Targets), the reasons evoked for experiencing GTG and GTO (Reasons), and the ways that people demonstrate their GTG and their GTO (Demonstrations). Coding was completed by independent coders unaware of which priming condition participants were assigned to. For the initial training rounds, teams of two coders coded a subset of between 30–90 responses and discussed in a meeting (facilitated by a third coder) any discrepancies. Discrepancies from the training rounds were resolved through discussion, and those not discussed during meetings were resolved by the third coder. This process was repeated until the two coders reached at least 85% agreement in every category (exceptions noted below), at which point each coder was considered an expert. The remaining responses were then coded by one expert. Full coding manuals with details about each coding scheme can be found on OSF and a shorter version can be found in Table 1. Responses were not coded if they were non-compliant (i.e. did not follow instructions).

Table 1. Summaries of coding categories for gratitude targets, reasons, and demonstrations.

Categories	Descriptions	Examples
Gratitude Targets
God	A god or deity	God, Allah, Vishnu, G-d
Faith	Anything related to faith or religion that is not a god	Congregation elders, Bible study, Jesus
People	Any person (without connection to faith/religion)	Family, friends, co-workers
Self	The self	Me, myself
Experience/Opportunity	Experiences or opportunities given	Going to college, getting a job
Institution	Any institution (not an experience related to an institution)	Workplace, school, country
Other	Anything that does not fit into the above categories	The universe, the COVID-19 pandemic
Gratitude Reasons
Life/Everything	General life circumstances; broad statements expressing gratitude for all the participant has	Being alive, having good things in life, general blessings
Achievement	Life successes and milestones, typically stemming from external validation	Jobs, promotions, graduating, passing exams
Self-Improvement	Improvement or maintenance of positive traits or skills	Becoming a better person, strength, being talented
Health & Safety	Improvement or maintenance of mental or physical health/safety; mitigating a negative health/safety outcome	Family staying healthy during the pandemic, having safe shelter
Relationships	Having particular people in one’s life; maintaining relationships with close others	Birth of a child, having friends in one’s life, working through relationship issues
Self-Transcendence	Experiences that extend beyond the self	Appreciating nature, spiritual experiences, social justice experiences
Gifting/Material Goods	Gifts or general objects that were given/provided by others	Gift cards, birthday presents, being given the opportunity to buy a house
Tangible Support	Doing something for someone, typically that helps them in a tangible way	Having car repair paid for, giving someone a ride, providing food when someone was ill
Health & Safety	Indicates that the act of tangible support resulted in a health/safety outcome	Receiving financial help when in poverty, being cured of a medical condition
Benefactor’s Kindness	Kindness, care, or emotional support from someone; or someone else’s positive traits	God’s unconditional love, someone being there for them during a difficult time, receiving guidance
Health & Safety	Indicates that the act of kindness resulted in a health/safety outcome	Allah improving someone’s depression through support, God guiding someone to bring their child to the hospital
Bonus: Control	Indicates to what degree the reason for gratitude is under the participant’s control (Levels 1–3)	Becoming a better person (Lev. 1), making up with a partner (Lev. 2), recovering from a medical issue (Lev. 3)
Bonus: Others	Indicates whether the recipient(s) of the reason(s) for gratitude is the participant, someone else, or both the participant & someone else	Being given strength by God (Self), family being healthy (Others), participant and their family receiving help cleaning their house (Both)
Gratitude Demonstrations
Reflection	Activities involving reflection or thinking	Reading the Quran, thinking about what was received, meditating
Acknowledgment & Recognition	Explicit expression/acknowledgment, whether verbally or symbolically	Saying ‘Thank you’, prayers of thanksgiving, writing a thank-you note
Celebration	More intense reactions than simple acknowledgment	Singing songs, dancing, praise
Reciprocation^b	Repaying the benefactor; giving or doing something in return (explicit mention of ‘repay’ or ‘in return’)	Giving an offering to god, doing a non-specific kind act for a person to repay their kindness
Tangible Support^b	Doing a kind action for the benefactor that typically helps them in a tangible way	Paying for someone’s dinner, giving an offering in church, giving someone a ride
Pay it Forward	Doing something kind for a person/people other than the benefactor	Volunteering, donating to charity
Moral Self-Improvement	Developing or maintaining moral traits; engaging in moral behaviors	Acting humbly, developing virtues, obedience to god, being kind to others
Skills Self-Improvement	Developing (non-moral) skills or performing a particular activity better	Meeting appearance goals, being successful at work
General Self-Improvement	Engaging in self-improvement, but in ways that can’t be classified into the other two categories	Enjoying life, being content, self-care
Proselytization	Broadcasting to others how great the benefactor is	Telling others about God’s love, teaching kids about religion
Physical Affection	Showing affection to the benefactor through touch	Hugging, cuddling, blowing a kiss to god
Kindness/Care	Providing care to the benefactor that isn’t physical or tangible in nature	Having faith in God, providing emotional support
Fits No Category	Demonstrating gratitude in ways that don’t fit into any of the above categories	Fasting, planting a tree
Connect*	Spending time with the benefactor	Going shopping, watching movies
Gifting*	Demonstrating gratitude by giving the benefactor a material good	Giving a present, giving a gift card
Prayer**	Demonstrating gratitude by praying or talking to god, excluding prayers of thanksgiving	Praying daily, speaking to God
Bonus: Religious Activities**	Indicates whether a given gratitude demonstration involves any kind of religious activity	Praying for someone, puja, mitzvot

Italics indicate secondary categories, which were coded transversally for all responses. Bold text indicates sub-categories, which were coded only when the category above was coded as 1. * = coded for GTO only. ** = coded for GTG only.

^aReciprocation was coded as its own category for GTO but combined with Tangible Support for GTG.

^bTangible Support was coded as its own category for GTO but combined with Reciprocation for GTG.

Targets of gratitude

Frequencies for this coding correspond to the number of times a specific target is mentioned out of all targets. Participants sometimes included more than one gratitude target (e.g. god, their boss) in their expression and each target received a code. Teams for this coding ranged from two to three people. All coding teams had established 85% agreement for each category before becoming experts, with the exception of the People and Other categories, which each had at least 83% agreement for all teams.

Reasons for gratitude

First, one coder evaluated all responses to determine if a clear reason for gratitude was present, which resulted in the exclusion of n = 388 GTG responses, and n = 520 GTO responses. Coders were then trained only on useful responses. Responses were coded for more than one reason for gratitude if they represented distinct categories. The coding scheme was data driven, with the goal of being able to make comparisons between equivalent reasons for GTG and GTO. See Table 1 for descriptions of coding categories. Two primary categories were slightly below 85% agreement (78% minimum) in the final training round (Health & Safety and Fits No Category, for GTG responses only). The two Health & Safety sub-categories had 50% and 81% agreement in the final round for some coding teams. Note that because these sub-categories were coded only for a small subset of responses, agreement across the training rounds was highly variable (often ranging from 50%–100%).

Demonstrations of gratitude

Open-ended responses listing one to seven ways that participants demonstrate GTG and GTO resulted in n = 4,004 and n = 3,679 demonstration descriptions, respectively. Each description was coded using categories that were developed following Chang and Algoe (2020) and expanded upon to account for unique ways that GTG is demonstrated. Some categories differed between GTG and GTO (see Table 1). Teams of three coders performed this coding. Percent agreement was slightly below 85% in the final training round for Reciprocation, Fits No Category, and the secondary transversal Any Religious Activity category (all for GTG only).

Results

To whom do religious people attribute their success? Coded gratitude targets and success attribution ratings

We assessed the various targets of participants’ gratitude expression for a success through coding of their open-ended responses (Gratitude Targets), as well as through self-report attribution variables, with the goal of first exploring potential differences between religious groups (collapsing across priming conditions, see Table 2), and then testing for causal effects of the God-prime on these targets and attributions (collapsing across religious groups, see Table 3). We also explored if there was an interaction between priming conditions and religious groups on each outcome related to god-attribution (i.e. likelihood of the target god, self-reported attribution to god), tests that were not statistically significant (see OSM). See OSM Tables S2– 3 for god-target and god-attribution descriptives by priming conditions fully crossed with religious groups.

Table 2. Gratitude for success: Target frequencies and attribution variable descriptives by religious group.

	Hindus	Muslims	Jews	Christians
Gratitude Expression Coded Targets	n = 97	n = 105	n = 126	n = 643
God	29%a	41%a	12%b	18%b
Faith^a	1%	0%	6%	1%
Experience/Opportunity	22%a	31%ab	36%ab	43%b
Institution	2%	5%	6%	5%
Self	12%	17%	21%	15%
People	54%ab	42%a	60%b	46%a
Other	10%	7%	10%	9%
Self-Report Gratitude Attribution	n = 124	n = 143	n = 142	n = 790
God	6.58b (2.37)	7.92c (1.68)	4.51a (2.80)	7.01b (2.42)
Self	7.12a (1.52)	7.05a (1.50)	7.67b (1.33)	7.41c (1.38)
Other People	7.34b (1.60)	6.57a (2.20)	7.07ab (1.84)	7.11b (1.94)

Multiple targets could be coded for each participant response. Italicized values indicate SDs. Attribution mean scores range from 1–9. Values in bold indicate significant differences at p < .05. Subscripts denote between-group differences from post-hoc tests adjusted using the Benjamini-Hochberg (false discovery rate) method, adjusted p < .05.

^aPost-hoc comparisons are not reported for the Faith target due to extremely low prevalence of this category.

Table 3. Gratitude for success: Target frequencies and attribution variable descriptives by priming condition.

	Neutral-Prime	God-Prime
Gratitude Expression Coded Targets	n = 475	n = 496
God	16%	26%
Faith^a	0%	2%
Experience/Opportunity	42%	36%
Institution	4%	6%
Self	15%	17%
People	51%	46%
Other	8%	10%
Self-Report Gratitude Attribution (1–9)	n = 595	n = 604
God	6.72 (2.62)	6.83 (2.47)
Self	7.46 (1.36)	7.27 (1.46)
People	6.98 (2.01)	7.14 (1.86)

Multiple targets could be coded for each participant response. Italicized values indicate SDs. Values in bold indicate significant differences at p < .05.

^aPost-hoc comparisons are not reported for the Faith target due to extremely low prevalence of this category.

Differences by religious group

Collapsing across priming conditions, we explored potential differences in targets of gratitude between religious groups (see Table 2 for descriptives and statistical tests, including post-hoc comparisons). We first ran a series of chi-square tests to test the association between the prevalence of each gratitude target and religious groups. P-values for post-hoc comparisons were adjusted using the Benjamini-Hochberg (false discovery rate) method.

Religious groups differed overall in the proportion of times they thanked god in written responses (χ² (3, N = 971) = 37.8, p < .001). As documented in details of the post-hoc tests in Table 2, Hindus and Muslims were more likely to mention god as a target of their gratitude than were Jews and Christians. Members of various religious groups also differed in the proportion of times they thanked other people (χ² (3, N = 971) = 11.2, p = .010) and thanked experiences/opportunities (χ²(3, N = 971) = 19.5, p < .001). However, chi-square results were not significant for gratitude towards the self (χ²(3, N = 971) = 3.6, p = .309), towards institutions χ²(3, N = 971) = 2.3, p = .504), or to other targets (χ² (3, N = 971) = 0.9, p = .814). We do not report results from tests for faith-related targets here, given that this category was coded very infrequently (14 times out of 971 responses containing gratitude).

Second, we ran a series of one-way ANOVAs with religious groups predicting each self-reported attribution variable. Religious groups differed in how much they reported attributing their success to god (F(3, 1194) = 56.00, p < .001, η²_p = 0.12), to themself (F(3, 1195) = 6.27, p < .001, η²_p = 0.02), and to other people (F(3, 1195) = 4.03, p = .007, η²_p = 0.01).

Differences by priming condition

Descriptives for gratitude target and attribution variables within each priming condition can be found in Table 3. Note that religious individuals in the Neutral-prime condition spontaneously mentioned god as a target of their gratitude in 16% of the coded responses. Differences between priming conditions on all outcomes were tested using Pearson’s chi-square tests with Yates’ continuity correction. Results indicated that the God-prime condition caused participants to have 1.9× greater odds of mentioning god as a gratitude target compared to those in the Neutral-prime condition (OR = 1.87, 95% CI [1.35, 2.61]), a difference that was statistically significant (χ²(1, N = 971) = 14.7, p < .001). Priming condition did not significantly influence frequency of mentioning the other targets: experiences/opportunities (χ² (1, N = 971) = 3.4, p = .064), institutions (χ² (1, N = 971) = 0.5, p = .469), the self (χ² (1, N = 971) = 0.7, p = .391), other people (χ² (1, N = 971) = 1.7, p = .196), and other targets (χ² (1, N = 971) = 0.7, p = .420).

Second, we ran a series of independent two-tailed t-tests to test whether there was an effect of priming condition on participants’ self-reported attributions to various targets. There was a significant difference in Attribution to Self between the priming conditions (t(1197) = 2.45, p = .015, d = 0.14), such that participants who were primed to think about god reported less responsibility for their own success than those who were given a neutral prime. However, Attribution to God (t(1196) = −0.73, p = .468, d = −0.04) and Attribution to Other People mean scores (t(1197) = −1.37, p = .171, d = −0.08) did not significantly differ between conditions.

Reasons for gratitude

Overall, the most prevalent reasons for GTG were Health & Safety and Life/Everything. Like prior research, the most prevalent reason for GTO was by far Tangible Support (Table 5).

Table 4. Reasons for gratitude to god by religious group.

	Hindus	Muslims	Jews	Christians	Difference Tests
	n = 84	n = 106	n = 97	n = 595
Life/Everything	35%a	42%a	16%b	33%a	χ^2 = 15.4, p = .001
Achievement	23%a	19%ab	8%b	12%b	χ^2 = 12.6, p = .006
Self-Improvement	7%	1%	6%	5%	χ^2 = 5.1, p = .165
Health & Safety^b	31%a	30%a	48%a	36%a	χ^2 = 9.0, p = .030
General^b	77%	84%	79%	73%	χ^2 = 2.1, p = .546
Result of Tangible Support^b	19%	16%	17%	25%	χ^2 = 2.7, p = .444
Result of Benefactor’s Kindness^b	8%	3%	4%	3%	NA
Relationships	11%	13%	15%	15%	χ^2 = 1.2, p = .751
Self-Transcendence	6%	5%	6%	6%	χ^2 = 0.4, p = .947
Gifting/Material Goods	2%	3%	3%	1%	NA
Tangible Support	10%	12%	12%	16%	χ^2 = 4.2, p = .246
Benefactor’s Kindness	6%	10%	12%	14%	χ^2 = 4.8, p = .184
Fits No Category^a	11%a	2%a	3%a	6%a	χ^2 = 8.2, p = .042

All responses could be coded for multiple categories (including sub-categories), therefore percentages will not add up to 100. χ² = chi-square tests, all Ns = 882, all df = 3. Values in bold and not sharing superscripts differ significantly at p < .05 with a Benjamini-Hochberg p-value adjustment.

^aNo post-hoc differences were significant when adjusting p-values for multiple comparisons.

^bCategories and values in italicized text refer to specific sub-categories of the broader Health & Safety category. Values indicate frequency of these sub-categories only out of responses coded for Health & Safety broadly. Note that Result of Tangible Support and Result of Benefactor’s Kindness overlap with Tangible Support and Benefactor’s Kindness. Total n = 320 for all sub-categories.

Table 5. Reasons for gratitude to god versus gratitude to other people.

	GTG	GTO	Difference Tests
	n = 882	n = 750	n = 583^c
Life/Everything	32%	2%	χ^2 = 142.7, p < .001
Achievement	13%	2%	χ^2 = 50.6, p < .001
Self-Improvement	5%	1%	χ^2 = 21.4, p < .001
Health & Safety	36%	12%	χ^2 = 109.1, p < .001
General^a	76%	1%	NA
Result of Tangible Support^a	23%	90%	NA
Result of Benefactor’s Kindness^a	4%	9%	NA
Relationships	14%	5%	χ^2 = 31.4, p < .001
Self-Transcendence	6%	NA	NA^b
Gifting/Material Goods	1%	16%	χ^2 = 68.4, p < .001
Tangible Support	15%	61%	χ^2 = 227.2, p < .001
Benefactor’s Kindness	13%	21%	χ^2 = 14.1, p < .001
Fits No Category	6%	2%	χ^2 = 15.2, p < .001

All responses could be coded for multiple categories (including sub-categories), therefore percentages will not add up to 100. Values in bold indicate significant differences at p < .05. χ² = McNemar test statistic, all df = 1.

^aSee Table 4 for note about subcategories.

^bSelf-Transcendence was not coded for any GTO responses.

^cLower n represents participants who had both GTG and GTO responses that could be coded for reasons and could therefore be compared.

We used the self-generated reasons for gratitude to address novel questions. First, we explored whether reasons for GTG differed by religious group (but see OSM Table S3 for GTO by religious group). Second, we explored whether, collapsing across religious group, reasons differed when considering one’s GTG vs. GTO. We tested the first question using chi-square tests, and p-values for pairwise comparisons between religious groups were adjusted using the Benjamini-Hochberg method (see Table 4). We do not report test results for categories when all frequencies were≤5% and/or multiple expected cell counts for the chi-square were<5. We tested the second question using a series of McNemar’s tests (see Table 5).

Gratitude demonstrations

Overall, religious individuals said they were most likely to demonstrate both GTG and GTO through Acknowledgment & Recognition, while nearly half of all GTG demonstrations involved a religious activity (cf. Any Religious Activity secondary category; see Table 7).

Table 6. Demonstrations of gratitude to god by religious group.

	Hindus	Muslims	Jews	Christians	Difference Tests
	n = 115	n = 130	n = 134	n = 674
Prayer	20%c	24%d	9%a	14%b	H = 51.83, p < .001
Reflection	10%a	11%b	15%bc	8%a	H = 14.6, p = .002
Acknowledgement and Recognition	10%a	19%b	20%bc	23%c	H = 37.94, p < .001
Celebration	4%b	0%a	1%a	7%c	H = 61.53, p < .001
Reciprocation and Tangible Support	21%c	3%a	7%b	6%b	H = 83.76, p < .001
Pay it Forward	17%	18%	22%	16%	H = 5.54, p = .132
General Self-Improvement	3%	2%	3%	3%	H = 1.36, p = .718
Moral Self-Improvement	8%a	12%ab	12%ab	13%b	H = 8.19, p = .043
Skill Self-Improvement	0%a	2%b	2%b	1%a	H = 13.77, p = .003
Proselytization	1%a	1%a	1%a	4%b	H = 29.15, p < .001
Physical Affection	0%	0%	0%	0%	H = 0.56, p = .905
Kindness/Care	2%	1%	1%	1%	H = 1.95, p = .580
Fits No Category	5%	6%	6%	4%	H = 6.13, p = .103
Any Religious Activity (Secondary)	53%b	50%b	33%a	51%b	H = 34.42, p < .001

Values indicate the frequency of each respective gratitude demonstration category out of the total number of demonstrations coded. ns reported here include only participants who had at least one coded demonstration. Values in bold and not sharing superscripts differ significantly at p < .05 according to Dunn’s test with a Benjamini-Hochberg (false discovery rate) p-value adjustment. H = Kruskal-Wallis test statistic, all df = 3.

Table 7. Demonstrations of gratitude to god versus gratitude to other people.

Demonstrations	GTG	GTO	Difference Tests
	n = 1,053	n = 1,044	n = 1,024
Prayer	15%	NA	NA
Gifting	NA	13%	NA
Connect	NA	4%	NA
Tangible Support	NA	13%	NA
Reflection	9%	1%	V = 40,624, p < .001
Acknowledgement/Recognition	21%	29%	V = 75,316, p < .001
Celebration	5%	1%	V = 17,682, p < .001
Reciprocation/Tangible Support	7%	23%	V = 31,167, p < .001
Pay it Forward	17%	7%	V = 105,495, p < .001
General Self-Improvement	3%	1%	V = 6,115, p < .001
Moral Self-Improvement	12%	2%	V = 61,966, p < .001
Skill Self-Improvement	1%	0%	V = 693, p < .001
Proselytization	3%	1%	V = 8,565.5, p < .001
Physical Affection	0%	4%	V = 0, p < .001
Kindness/Care	1%	8%	V = 1,824.5, p < .001
Fits No Category	4%	6%	V = 14,914, p = .023
Any Religious Activity (Secondary)	49%	3%	V = 346,120, p < .001

Values indicate the frequency of each respective gratitude demonstration category out of the total number of demonstrations coded. ns reported here include only participants who had at least one coded demonstration. V = Wilcoxon signed-rank tests, with values in bold indicating significant differences at p < .05.

We used the self-generated demonstrations of gratitude to first explore whether demonstrations of GTG differed by religious groups (but see OSM Table S7 for GTO by religious group), then to explore whether, collapsing across religious group, demonstrations differed when expressing GTG vs. to another person. To test the first question we used Kruskal-Wallis tests, followed by post-hoc comparisons using Dunn’s tests with a Benjamini-Hochberg p-value adjustment (see Table 6). To test the second question we ran a series of Wilcoxon signed-rank tests (see Table 7). We chose to use non-parametric tests because our DVs were percentage data, which are non-normally distributed.

Discussion

Research on gratitude has been exponential in the past two decades and identified many personal and social benefits of expressing this emotion. Yet, such research has missed the opportunity to characterize instances of gratitude targeted to a god despite billions of people believing in a god who is believed to intentionally and positively intervene in people’s lives and the world. In a large study among U.S. Hindus, Muslims, Jews, and Christians, we aimed to document the prevalence of GTG (how much) in a control vs. a god-prime condition, and then to compare the situations eliciting GTG vs. GTO (why), as well as their respective expressions (how).

First, regarding the prevalence of GTG, we established that religious people (i.e. our participants) do spontaneously express GTG in response to a positive outcome, but not as often as we had expected. Specifically, we found that when religious participants are asked to express gratitude following the recall of a success, only 16% spontaneously thanked their god (cf. Neutral-prime condition). More specifically, this number ranged from a high 31% for Muslims followed by 20% for Hindus, 15% for Christians, and a low 5% for Jews (see OSM Table S1). These differences between religious groups may be partly due to differences in how much god is on their mind or overall religiosity (see OSM Table S8 indicating that Jews in our study reported less religiosity than the other groups). Jewish identity is not as much based on religious belief as Protestants (Cohen et al., 2003). Given that some Jews may identify as Jewish because it is an important cultural identity more so than a religious belief, this could explain the lower frequency of GTG in this group (not having god ‘on their mind’ when something good happens). These differences may also be due to different norms in each religious group, which should be further investigated.

When primed with the notion of god, an additional 10% (26% total) of participants thanked their god. In addition, perhaps akin to priming, when directly asked with a self-reported question, these religious participants overwhelmingly agreed that god had something to do with their success (above 6 on a 9-point scale on average) in both priming conditions. Still, collapsing across the priming conditions, the self-reported attribution to god also showed differences between religious groups similar to the coded spontaneous targets. These data suggest that, even in this religious sample, spontaneous GTG is not that prevalent, despite some differences by religious groups. However, when god is primed/on their mind, the success was credited to god at a higher frequency (and less to the self, probably as a result of not wanting to appear prideful).

This last finding is aligned with other studies on religious priming showing that it is possible to increase god’s salience in people’s mind (Shariff et al., 2016). God priming can be a useful technique for future research that aims to maximize people’s reports of GTG. In addition, these results add to the limited extant research on attributions to god following successes. In particular, DeBono et al. (2020) found that when religious participants (predominantly Christian) were randomly assigned to experience success or failure in a game, they were more likely to attribute their successes than their failures to god. Further, in the success condition, participants were similarly likely to attribute their success to god and to the self. Of note, those findings relied on self-reported questionnaires. In contrast, the present research employed a combination of coded spontaneously-generated gratitude targets and self-reported ratings, highlighting a meaningful discrepancy between people’s spontaneous mention of god and attributions provided in self-reported rating scales. Future research should test whether the prevalence of GTG differs depending on context, e.g. is it higher in situations that we have identified as the most common elicitors of GTG (see below)?

Second, regarding the eliciting situations for GTG (reasons), we found that generally, religious people think of a wider array of situational triggers when they think about GTG than when they think about GTO. In particular, events such as being healthy, safe, or simply having a good life were much more frequent elicitors of GTG than GTO. Specifically, we found that Jews were most frequently grateful to god for their health and safety, whereas the other religious groups were similarly especially grateful for their life and everything good in it and to a lesser extent for their achievements. We found that events such as being healthy, in security, alive, or simply having a good life, are more frequently the source of GTG than to other people. Tangible support was, expectedly, much more frequently a reason for GTO than for GTG.

In sum, the GTG situational triggers were in the spirit of what laypeople would otherwise call gratitude for the good things in one’s life. Such type of gratitude has been documented in the broader literature on gratitude (which has tended to focus either on interpersonal gratitude or on appreciation for the good things in life), but scientists never asked if that appreciation was attributed to god as a causal agent. Given the present results, we recommend asking such attribution question as a new practice going forward. Theoretically, this attribution process may refer to the phenomenon that some scientists call the God of the gaps (Lupfer et al., 1996): when religious people don’t have a scientific explanation for an event, they turn to a religious explanation. Here, general good things, perceived as gifts, and that do not have a clear human benefactor, get attributed to god. When other people’s actions cannot explain the benefits, God is invoked as a cause; it’s the God of the good gaps. Of note, we found that god is not only invoked when there are gaps, but also for events that can be attributed to other people’s actions. This explains why we see overall a much wider array of eliciting situations for GTG than for GTO and possibly why religious people are generally more grateful (Emmons & Kneetzel, 2005).

Third, regarding the ways that religious people demonstrate their GTG, we established that they do so in a wide variety of ways, consistent with prior research on GTO (Chang & Algoe, 2020 see also nonverbal expression of GTG in Van Cappellen & Edwards, 2021). We also identified ways that GTG is demonstrated that have never been documented for interpersonal gratitude (e.g. prayer). Specifically, we find that each religious group has their signature demonstration: Hindus demonstrate by reciprocating (e.g. puja and other forms of giving offerings to a god) and praying, Muslims by praying, Jews by paying it forward, and Christians by acknowledging and recognizing (e.g. prayers of Thanksgiving specifically). These differences may signal different norms and suggested ways to please god. Compared to the ways that people demonstrate GTO, specific demonstrations of GTG stand out. Across all demonstration types, GTG is demonstrated through a religious activity in 49% of the cases compared to 3% for GTO (cf. secondary transversal category). More specifically, GTG is more often demonstrated by paying it forward and through moral self-improvement, which had been documented in previous research on GTO but were not as prevalent. In contrast, GTG was much less often demonstrated through reciprocation or tangible support, which are the main ways that GTO gets demonstrated. These results provide clues for behaviors that would be worth measuring in future studies on the personal and social benefits of GTG.

These results call for further theorizing on the functional value of demonstrations of GTG. First, according to the find-remind-bind theory of gratitude (Algoe, 2012), demonstrations of GTG likely serve a key purpose of building a high quality relationship with god. That is, demonstrating gratitude is one way through which religious people can show god that they see and value the contributions god makes to their life and their desire to keep that relationship going. If so, the present results document what religious people find important to fulfill this purpose: to be prosocial (cf. pay it forward) and to be a good person (cf. self-improvement). These are more important that telling others about god for example (proselytization was very low). Not surprisingly, building tight knit communities and maintaining moral order correspond to two main functions of religion (Graham & Haidt, 2010). Other research also showed that religion fosters prosociality (Shariff et al., 2016). GTG may be one mechanism through which ‘religion’ achieves these bigger outcomes.

Second, GTG may be an engine for future religious practice by promoting repeated prayer and other religious behaviors. The Upward Spiral Theory of Sustained Religious Practice (Van Cappellen et al., 2021) suggests that positive emotions, including GTG, experienced during religious practices (a) create nonconscious motives (e.g. positive spontaneous thoughts) for cues associated with the practice and (b) predict long-term maintenance of the practice. The present results suggest that GTG is demonstrated by/promotes repeated prayer and other religious behaviors. These results align with a correlational study finding that more frequent expressions of praise and GTG are associated with more frequent worship participation (Schneller & Swenson, 2013 Future research should more directly investigate the behaviors associated with GTG to illuminate its functional value, similar or different to GTO.

Finally, the present results suggest that the target of an emotion can change the way an emotion is demonstrated. This finding further highlights the importance of taking into account context and target in emotion expression research. We find that religion is one such relevant context given that the vast majority of the world is religious.

The present research also has several limitations. First, despite a generally diverse sample, the samples were still smaller for members of non-Christian traditions than of Christian traditions. Second, given the novelty of the topic, we had to expand existing coding scheme developed for GTO to account for unique responses in the context of GTG. This process was data driven and could look different for other teams working with different samples and data. Much of this research was exploratory and meant to provide a basis for expansion in future research on GTG. Third, we compared GTG to GTO but there are many other relevant comparisons to make such as gratitude for other supernatural agents (e.g. karma) or for targets that maintain moral order like a god (e.g. country or institutions). Finally, our results are limited to what participants self-reported and it remains to be tested whether they actually do demonstrate GTG through the identified behaviors. One area of interest is on the social benefits GTG and on its role in maintaining the system of religion (Eibach et al., 2015).

To conclude, we found that GTG is experienced by people belonging to the major religious traditions (Muslims, Hindus, Jews, and Christians). GTG is triggered by many things, including all good things in life. Such instances of GTG for all good things shed light on a phenomenon described in the broader literature on gratitude but have yet to be connected to attributions to supernatural agents. Finally, we found that GTG expressions may provide one way through which religious beliefs and practices are sustained over time and foster greater prosociality toward others in the world.

Open Scholarship

This article has earned the Center for Open Science badges for Open Data, Open Materials and Preregistered. The data and materials are openly accessible at https://osf.io/fne4k/?view_only=b26f2b57d62e48fdbca3bf08b1cb088b

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data will be made available upon request by contacting the corresponding author.

Additional information

Funding

This study was generously funded by a grant from the John Templeton Foundation, #61513. We thank Amanda Bernal, Bailey Davis, Jesse DeLaRosa, Nour Kanaan, Dav King, Nathalie Hartman, Cai Liu, Callie Murphy, Kerry O’Brien, Jessica Stevens, Neha Vangipurapu, Yvette Yang, Jiaqi Zhang and Bailey Davis for their help with data collection and coding. The data presented are freely accessible through the Open Science Framework: https://osf.io/fne4k/

Notes

1. We use the word ‘god’ throughout to represent the Christian God, Islam’s Allah, Judaism’s Gd, and the deities of the Hindu pantheon. These four religions are the target of the present study.