Exploring the cognitive context of gratitude to God: emotional impact and appraisals of benefits from God

ABSTRACT In two pre-registered studies we investigated the cognitive context of gratitude to God (GTG) and fading positive affect of grateful memories. Although gratitude has been shown to causally impact happiness, few studies have investigated how the nature of the benefactor might affect experienced gratitude. Participants first recalled a positive event, a human-caused benefit, or a God-caused benefit. After recall they reported their current momentary emotions, recalled past emotions from the time of the event and emotion appraisals. Grateful memories showed less fading positive affect over time compared to other positive memories. Appraisals of value and altruism predicted both GTG and gratitude to humans (GTH). Positive spiritual appraisals (e.g. positive interpretations about one’s relationship with God, the sacred giftedness of the event) were uniquely associated with GTG. When it comes to felt gratitude, the nature of the benefactor does matter.

non-specific gratitude; gratitude one feels without referencing any specific benefactor. 2 It is important to highlight that general gratitude likely involves aspects of both religious and non-religious gratitude.In one prospective study, GTG predicted general gratitude over time (P.C. Watkins et al., 2019).Thus, GTG might encourage general gratitude, which has been shown to enhance well-being.
We also contend that GTG should inform the study of gratitude more generally.As shown in the opening paragraph, contrasting benefactors might influence gratitude.Although few studies have investigated this issue, studies suggest that the nature of the benefactor matters to gratitude.For example, Bar-Tal et al. (1977) showed that the nature of the beneficiary's relationship to the benefactor affected the level of gratitude experienced.Somewhat surprisingly, they found that people would experience more gratitude to a stranger than to their mother for an equivalent favor.Furthermore, studies have found that people tend to be more grateful to an individual than to an institution for an equivalent benefit (McLaughlin et al., 2018).In sum, preliminary investigations suggest that the nature of the benefactor matters to gratitude, and therefore comparing human caused to divine caused benefits should be informative to our understanding of gratitude.

Comparing GTG to GTH
How might GTG and GTH be similar, and how might they be different?Just as general gratitude appears to be beneficial to well-being, so too GTG may be beneficial to the well-being of religious/spiritual people, particularly to their spiritual well-being.How might both general gratitude and GTG enhance well-being?The amplification theory of gratitude (P.C. Watkins, 2014;P. C. Watkins, Frederick, & Davis, 2022) proposes that gratitude enhances well-being because gratitude magnifies one's experience of good in one's life.In other words, gratitude supports flourishing by increasing the signal strength of the people and things that are conducive to well-being.If gratitude supports well-being by amplifying the good in one's life, then GTG should increase the perception that God is the giver of good gifts.
From this understanding gratitude enhances the good in several ways.For example, gratitude should amplify the good in memory.Studies have shown that positive memories are more accessible and have a more positive emotional impact for grateful people (P. C. Watkins et al., 2004Watkins et al., , 2005)).Experimental studies show that recalling blessings enhances happiness (Emmons & McCullough, 2003;Seligman et al., 2005), and P. C. Watkins et al. (2015) showed that the impact of recalling blessings on happiness was not just due to remembering positive events -grateful processing of the memories was crucial.
The finding that recalling grateful events enhances positive affect is consistent with the literature on moodcongruent memory.Briefly, mood-congruent memory refers to the finding that one is more likely to recall events that are congruent with their current emotional state (Bower, 1981).Mood-congruent memory has been demonstrated in many studies but seems to be more robust in sadness and depression (e.g.P. C. Watkins et al., 1992Watkins et al., , 2000Watkins et al., , 2005) ) than in anxiety and fear (Mogg & Mathews, 1990).Bower (1981) theorized that moodcongruent memory takes place because emotion nodes exist in memory, and when these nodes are triggered, spreading activation leads to increased accessibility of memories that were encoded during the initiation of a particular emotion in the past.This theory implies that when one is recalling an emotional memory, activation would spread to the relevant emotion node and thus a congruent emotion would be triggered.For example, when asked to reflect on a grateful memory, Bower's theory implies that grateful emotion would be activated (along with other closely related positive emotions).Thus, recalling grateful memories might activate positive affect, which in turn enhances SWB.
Why might grateful memories be fundamental for SWB?Generally, the emotional impact of memories fades over time, called the fading affect effect (Walker et al., 1997(Walker et al., , 2003)).One possibility is that compared to other positive memories, grateful memories maintain more of their positive emotional impact over time.In other words, recalling grateful events might have a more enduring positive emotional impact than other positive memories.Compared to other positive events, the amplification theory of gratitude posits that gratitude amplifies the good in the memory.Recalling an event that activates grateful emotion should increase the signal strength of the positive aspects of the memory, thus encouraging more positive affect -particularly gratitude.Furthermore, gratitude is a cognitively imbued emotion; compared to other positive emotions, gratitude involves a rich array of cognitions.For example, the more one recognizes the goodness of the gift and the goodness of the giver, gratitude is enhanced (P.C. Watkins, 2014).When one recalls a past benefit from someone, this provides an opportunity for the individual to reflect further on this benefit, perhaps enhancing their perception of the goodness of the gift and giver, and thereby enhancing one's current experience of gratitude.This should be true regardless of the benefactor; thus, we predict that grateful memories should show less fading affect than other positive memories for both human and divine benefactors.In short, reflecting on grateful memories might be central to SWB.In the studies described in this paper, we investigated whether grateful memories -both human and divine caused benefits -show less fading positive affect than other positive memories.
As GTH enhances well-being by amplifying the good in relationships (Algoe, 2012), so too GTG might support spiritual well-being by amplifying the good in one's relationship with God.GTG should amplify one's perception of the goodness of God, which should enhance one's sense of closeness to God and thereby increase spiritual well-being.P. C. Watkins, Frederick, and Davis (2022) found that trait GTG predicted increased nearness to God and spiritual well-being.Thus, GTG may be like GTH in that it supports well-being by encouraging healthy relationships.
How might GTG be distinct from GTH?The cognitive context of GTG may be unique.Appraisal theories propose that each emotion is caused by a distinct appraisal or interpretation of a situation (e.g.Giner-Sorolla, 2019;Scherer et al., 2001;Smith et al., 2014).Gratitude results from the appraisal that 'someone has done something important for me'.Understanding the cognitive antecedents of GTG should be a central goal if the science of GTG is to progress.Research has shown that appraisals of value, altruism, and cost are the primary drivers of felt gratitude (Tesser et al., 1968;Wood et al., 2008).Although recent work has questioned the importance of the cost appraisal to gratitude (McLaughlin et al., 2018;P. C. Watkins et al., 2016), we include these appraisals in our investigations.Value and altruism appraisals are likely to contribute to GTG in the same way that they are associated with GTH.However, the altruism appraisal might pose some interesting issues for GTG.More than likely, people expect God to act more altruistically than they expect of their fellow humans.Thus, altruistic acts from people may be more likely to exceed their expectations, causing the altruism appraisal to be more predictive of GTH.Furthermore, most monotheists view God as infinite and all-powerful, so they may see benefits from God as coming from unlimited resources and thus not very costly.
In short, value and altruism appraisals are likely to be similar in their contribution to GTG and GTH.But what appraisals might be unique to GTG? Spiritual appraisals should distinctly characterize GTG.Because God is transcendent, spiritual appraisals such as 'This benefit was an answer to prayer' should predict GTG.More specifically, we propose that construals of sacred giftedness should be unique to GTG.Interpreting a benefit as a sacred gift should be more strongly predictive of GTG than GTH.Thus, spiritual appraisals should be unique to GTG.
Another distinct aspect of the cognitive context of GTG might be related to construal level.In brief, Construal Level Theory (Trope & Liberman, 2010) proposes that things with greater psychological distance create higher level construals.This includes people and situations that are distant from the self in terms of time, certainty, place, and concreteness.Psychologically close objects are those that are more certain, immediate, concrete, and close geographically.When processing something psychologically distant, people reconcile the distance by engaging in high level construals.How do we deal with things and people that are distant from the self?In short, 'We do so by forming abstract mental construals [high level construals] of distal objects' (Trope & Liberman, 2010, p. 440).Because most believers in God conceive of God as transcendent, invisible, abstract, and timeless, Tsang et al. (2021) invoked Construal Level Theory to account for differences between GTG and GTH.GTG may be more associated with higher level construals than GTH.For example, gifts from God may be associated with construals about abstract benefits, benefits that extend into the future, or benefits that extend beyond the self.

Overview of studies
We conducted two studies designed to address two questions: 1) How might GTG be similar and/or distinct from GTH? and 2) Do grateful memories show less fading affect compared to other positive memories?In Study 1 we randomly allocated people to recall one of three memories: a 'happy' memory, a human-caused benefit, or a Godcaused benefit.Participants then reported on their current emotions and recalled their past emotions experienced with the event.They also reported various appraisals associated with the event derived from past gratitude and appraisal research.Study 2 was similar to Study 1, except that the positive comparison recall condition was an event associated with amusement, and in the human-caused benefit recall condition we specified the benefactor as one's mother or another member of their immediate family.Also, in Study 2 we extended the appraisals by including spiritual appraisals and appraisals related to Construal Level Theory.

Overview and design
Participants were randomly assigned to one of three autobiographical recall conditions: happy event, human-caused benefit, or God-caused benefit.After recalling their assigned memory, participants reported on the emotional impact of recalling the event ('present' report) and how the event impacted them at the event's occurrence ('past' judgements).Thus, our primary design was a 3 (recall condition) x 2 (time: past and present) mixed factorial design, with recall condition as a between subjects factor, time was the repeated measure, and extent of emotion was the dependent variable.After judging emotional impact, participants provided appraisals of the event, which also served as dependent variables.

Participants
Undergraduate students (N = 168, M age = 22.75, SD = 4.37) completed the study in exchange for partial course credit.Of these participants, 38 were male and 130 were female.The ethnicity of the participants was predominately white (65.9%), followed by Hispanic/Latino (9.4%), 'Mixed' (4.1%), and African American (1.8%).Other ethnicities comprised less than two percent of the sample and 29 participants chose not to report their ethnicity.Of the sample, 54 completed the 'happy event' recall condition, 60 completed the human-caused benefit, and 54 completed the God-caused benefit recall.In our sample, 22.6% said they did not believe in the existence of God, 57.4% said they did believe in God's existence, and 20.2% said they were unsure.We controlled for belief in God using our dimensional confidence in the existence of God scale (described below) as a covariate.The Institutional Review Board of Eastern Washington University approved this study, and this study conformed to the ethical principles of the American Psychological Association.

Materials
For assessing both current and past emotional impact, we used a modification of Izard's (1977) Discrete Emotions Scale (DES), like that used by Fredrickson et al. (2003).Participants responded to various emotions on a 7-point scale ranging from 'no' to 'very strong' experience.Each emotion is described with several descriptors so as to get a clear response to each discrete emotion (e.g.'grateful, appreciative, or thankful').We assessed the following emotions: contentment, gratitude, joy ('joyful, delighted, or enthusiastic'), gratitude to God, gratitude to humans ('grateful, appreciative, or thankful to other humans' 3 ), indebted ('indebted [feeling like you owe] or obligated to repay'), guilt, amusement, discomfort, love, shame, awe, inspiration, and pride.
Participants also reported on appraisals important to the emotions experienced with the recalled events.Students responded to each appraisal on a 6-point scale ranging from 'Not . . .' to 'Extremely . . .'.Each participant provided appraisals of the value of the recalled event ('How important or valuable was the benefit/blessing to you?'), altruism ('How altruistic [i.e.motivated for your benefit and well-being] were the motivations of the giver of the benefit/blessing?'), cost ('How much did it cost the giver to provide this benefit/ blessing to you?"), exceeded expectations ('How much did this benefit/blessing exceed your expectations?'),deservingness ('How much do you think that you earned or deserved this benefit/blessing?'),owed ('How much do you think the giver owed you this benefit/blessing?'),contribution of other humans ('How much did other humans contribute to this benefit/blessing?'),God's contribution ('How much did God contribute to this benefit/ blessing?'),and self-contribution ('How much did you contribute to this blessing?').Participants also reported on the recency of the recalled event (1=within the last week, 2=within the last month, 3=within the last 6 months, 4=within the last year, 5=within the last 5 years, 6=more than 5 years ago).
Participants responded to the question 'In terms of how you're feeling right now, how confident are you that a loving personal God actually exists?', on a 11-point scale ranging from '0%: I'm completely confident God does not exist' to '100%: I'm completely confident God does exist'.This scale allowed us to control for belief in God using a dimensional scale.Finally, participants completed a demographics form.

Procedure
Because participation in this study was anonymous, students received a written description of the study and were informed that their participation indicated their consent.Questionnaire packets were mostly completed in a group format.A packet of questionnaires was created for each recall condition.All packets were identical, with the exception of the first page which contained their specific recall instructions.Each recall condition was introduced with the following statement: 'For the following, please read the instructions carefully, but if you feel uncomfortable answering any question, or there just doesn't seem to be a reasonable response for you, please feel free to skip that question'.We included this instruction primarily so that atheists assigned to the God-caused benefit condition would be free to skip questions.
Participants were randomly assigned to one of the three benefit recall conditions.In the happy event condition, participants were instructed: 'We would like you to try to remember and vividly imagine an experience that happened to you.Try to remember a happy occasion that brought you pleasure.Please try to imagine this situation as vividly as you can, and write a few sentences about this experience.'In the human-caused benefit recall condition, participants were given identical instructions, with the exception for recall, which read: 'Try to remember a significant benefit or blessing that another person has brought into your life'.Participants in the God-caused benefit recall condition were given identical instructions except for the recall instructions which read: 'Try to remember a significant benefit or blessing that you believe God brought into your life'. 4After providing significant space for participants to write about their remembered benefit, we included a box to endorse if participants did not believe in the existence of a personal God.
The next page of the packet contained participants' report for how their recollection impacted them emotionally, as described in the Materials section.This was introduced with the following instructions: 'As you think about the situation that you wrote about on the previous page, how does it make you feel right now?' After reporting their current emotional state using the modified DES, participants reported on their original reaction to the event with the following instructions: 'Try to think back on the benefit/blessing you wrote about on the first page.How did you feel about that benefit or blessing at the time that it occurred?'After reporting on the same emotions, participants then provided appraisals of the event as described in the Materials section.Participants then completed an experimental state measure of GTG, followed by seven measures not directly relevant to the hypotheses of this study.Finally, participants completed the dimensional scale of belief in God described in the materials section, and demographics.

Fading affect analyses
Factor analysis of the Modified DES revealed a general positive affect factor that included contentment, joy, amusement, awe, inspiration, and pride.For simplicity we combined these items into one scale, henceforth referred to as general PA (α = 82; .86).For our fading affect analyses, we conducted a 3 (recall condition) x 2 (time: past and present) repeated measures analysis for general PA and gratitude.In each analysis, we used confidence in the existence of a loving God and recency of event as covariates.For general PA, a significant interaction for recall condition x time was found, F(2,163) = 5.48, p = .005,partial ղ 2 = .063(means by condition can be found in Table 1). 5Figure 1 illustrates this interaction.'Past' ratings represent participants' judgments as to their initial response to the event in the past, whereas 'Present' ratings are their reports for how the recollection of the event is impacting them currently.Observation of this figure reveals that the interaction was due to happy memories showing the expected fading affect effect, but no such pattern was shown for either of the grateful memories, where positive affect was largely maintained over time.We also found a significant recall condition x time interaction for grateful emotion, F(2,162) = 4.31, p = .015,partial ղ 2 = .051.As seen in Figure 2, this interaction was due to the typical fading affect effect for happy memories, but grateful memories actually showed increasing gratitude over time.For GTG, this interaction did not reach significance (p = .154),but the pattern was like gratitude: happy memories showed fading GTG whereas human-caused benefits showed no change, and God-caused benefits showed increasing GTG over time.There was a main effect for recall condition, F(2,161) = 3.70, p = .027.As expected, pairwise comparisons showed that GTG was higher in the God-caused benefit recall condition than the other two conditions.In sum, grateful memories did not show fading positive affect compared to happy memories.

Effect of recall condition on appraisals
Few significant differences were found on appraisals with the different recall conditions.We conducted a univariate GLM on the various appraisals for recall condition, with ratings of the likelihood of the existence of God as the covariate.No   significant effect of recall condition was found on value, F(2,164) = 0.59, p = .557.Similarly, no significant effect was found for recall condition on cost, F(2,162) = 0.91, p = .404.A nonsignificant trend was observed for recall condition on altruism, F(2,160) = 2.94, p = .055.Post-hoc Bonferroni pairwise contrasts showed that this trend was due to human caused benefits being more altruistically motivated than happy events (p=.055).A nonsignificant trend was also observed for the effect of recall condition on whether the participants felt that they 'earned or deserved' the benefit, F(2,161) = 2.56, p = .080.Posthoc Bonferroni pairwise contrasts suggested that this trend was due people feeling that they earned or deserved the happy event more than the humancaused benefit (p = .132).A significant effect was found for recall condition on participants' ratings of the contribution of God to the benefit, F(2,160) = 5.12, p = .007.Post-hoc Bonferroni pairwise contrasts showed that this was due to participants reporting that God contributed more to the God-caused benefit than the human-caused benefit (p = .007).Somewhat surprisingly, no significant effect of recall condition was found on how much participants felt that they had contributed to each benefit, F(2,163) = 0.76, p = .469.Means and standard errors for appraisals by condition can be found in Table 2.

Correlations of appraisals with emotions
Table 3 shows the relationship between the appraisals and the emotional states under study.For simplicity, instead of including all positive emotion items, we represent these with the general PA factor.As expected, value appraisals significantly contributed to all positive emotions analyzed -gratitude, GTG, and general PA -with the caveat that the relationships were somewhat less strong regarding GTG.A curious pattern emerged with regard to altruistic motivations of the giver.Although this appraisal did not make significant contributions to participants' recollection of their original emotional states ('past' ratings), altruism was significantly related to current positive emotions (p = 017).This suggests that altruism appraisals might be key to the emotional responses after the event has occurred as the person reflects on the benefactor's intention and significance of the benefit.
Although perceived cost to the benefactor is often presented as central to gratitude, this appraisal did not significantly contribute to any of the positive emotional states.Interestingly, the contribution of God to the benefit recalled was predictive of all positive emotions, but as expected, was much more predictive of GTG.
Comparing the correlations of the appraisals to general gratitude and GTG to shows few differences.Thus, value and altruism seem to contribute in a similar way to general gratitude and GTG (value p = .849;altruism p = .080).Indeed, the only clear contrast was with the God contribution appraisal.Not surprisingly, God's contribution appraisal was much more strongly correlated with GTG than with general gratitude (p < .0001).This study showed little in terms of what might be distinct about the cognitive context of GTG.Several issues with this study may have limited the ability to identify the distinct cognitive characteristics of GTG.First, for many individuals, general gratitude involves both GTH and GTG.Second, we included appraisals that  are likely to contribute to both GTG and general gratitude.The distinctiveness of GTG is likely to be found not in appraisals such as value and altruism, but in spiritual appraisals.In Study 2 we attempted to overcome these limitations.

Study 2
In Study 1, we found that grateful memories did not show fading positive affect as did our comparison positive memory.However, one issue with the finding might be that the comparison 'happy' memory was rather non-specific in the recall instructions, thus allowing for participants to recall a qualitatively different kind of memory.In Study 2, we countered by having participants in the positive memory comparison condition recall a more specific positive memory, in this case, an episode of amusement.
Our second question from Study 1 related to the distinct cognitive context of GTG.In our gratitude recall condition, we simply asked participants to recall a benefit that had been provided by another human.Like the 'happy' event recall instructions, this allowed our participants to recall benefits from a wide variety of benefactors, which could have resulted in a more powerful grateful memory.Thus, because God is a specific benefactor, we restricted the human-caused benefit memory benefactor to one's mother (or another immediate family member).By restricting the human benefactor, this creates a more direct comparison of GTG to GTH.In our emotional response measure, we changed the wording on the GTH item to make it clearer that we were interested in the respondent's gratitude to their human benefactor, thus making a more suitable comparison to GTG.
In light of our findings in Study 1, we theorized that spiritual appraisals might be more likely to predict GTG than GTH.Thus, in Study 2 we assessed several spiritual appraisals in addition to appraisals from Study 1.Following Construal Level Theory (Trope & Liberman, 2010;Tsang et al., 2021), we added several high and low level construals.Finally, we also assessed spiritual/ religious variables that should be related to GTG, along with measures of SWB.
Like Study 1, participants were randomly allocated to recall a positive comparison memory (an episode of amusement), a human-caused benefit (provided by one's mother or immediate family member), or a God-caused benefit.Participants then reported on their current emotions, followed by their judgments of their original emotional response to the event.After reporting on appraisals related to their recalled event, they completed measures of SWB and religiosity/spirituality.Following are the preregistered hypotheses for this study: (1) We predicted that spiritual appraisals would be endorsed more in the God-caused recall condition.
(2) We predicted that higher level construals would be endorsed more in the God-caused condition, and lower level construals more in the humancaused benefit recall condition.(3) Value and altruism appraisals would predict both GTG and GTH.(4) Compared with GTH, spiritual appraisals and higher level construals would be more closely associated with GTG than non-spiritual appraisals.(5) The two gratitude recall conditions would show less fading positive affect and fading gratitude than the amusement condition.( 6) State GTG would be more closely associated with experiential than doctrinal concepts of a benevolent God.(7) Momentary emotional state GTG would be more closely associated with experiential than doctrinal nearness to God.

Participants and procedure
In this preregistered study (DOI 10.17605/OSF.IO/QFCEP; https://osf.io/qfcep/),students (N = 166, M age = 21.59,SD = 3.42) completed the study anonymously and received partial course credit for their time.The data from 3 students were excluded because they failed one of the initial attention checks.Of this population, 62% identified as female, 35.5% as male, 1.8% as other, and one participant did not respond.In terms of ethnicity, 51.6% identified as white, 18.1% Hispanic/Latino, 9% Black/African American, 3% Asian, 1.8% Native American, 1.2% as mixed, and 15.7% did not respond.Also, 13.9% of this population expressed no belief in God.
We replicated the procedure of Study 1 with the following exceptions.As in Study 1, participants were randomly allocated to recall one of three episodic memories.In this study, for the positive comparison memory we asked participants to recall an amusement event with the following instructions: 'Try to remember a significant experience where you felt amused'.Also, in the humancaused benefit condition, rather than asking for any benefit, we asked participants to 'Try to remember a significant benefit or blessing that your mother provided for you.If you can't think of a benefit from your mother, please recall a significant event that was provided by another member of your immediate family (father, brother, or sister)'.Following emotion induction research, we changed the writing instructions in each recall condition: 'Please try to imagine this situation as vividly as you can, and describe it as if to someone who had never felt this kind of experience, so that this person would know how it feels to be in this kind of situation'.Writing about their memory was timed for 5 minutes.
Momentary emotion after the recall task was assessed with the DES as in Study 1, but we changed the GTH item to 'In terms of the human other than yourself that contributed most to this benefit, how grateful, thankful, or appreciative do you feel toward that person?'We also added items related to indebtedness (which we defined as 'feeling like you owe') that were targeted to the human benefactor and to God.Participants then completed the GTG-S (P. C. Watkins, Frederick, & Davis, 2022), followed by their recollections as to the emotions that they experienced during the recalled event.
We then assessed 68 appraisals for the recalled event.In addition to adding appraisals in addition to those used in Study 1, we also provided a 'not relevant' box for each appraisal that participants could check.Following advances in appraisal research methodology, this was provided because some appraisals might not be relevant to the recalled event and forcing a participant to respond on such an item might create spurious results.Instructions for appraisals were as in Study 1, but because we added spiritual/sacred appraisals, we included the following definition of sacred: 'Some of the items below refer to the "sacred".The "sacred" could be God, nature, or anything you feel is transcendent -beyond yourself'.The final appraisal list can be found in the supplemental materials.Table 4 shows the derived appraisal factors (described below) and appraisal categories, along with means by recall condition.
Participants then completed several questionnaires (see below), concluding with a demographics form.This study was preregistered on the Open Science Framework website 9 https://osf.io/qfcep/),was approved by the Institutional Review Board of Eastern Washington University, and adhered to the ethical principles of the American Psychological Association.

Measures
Measures are described below and were administered in the order described.The GTG-S (P. C. Watkins, Frederick, & Davis, 2022, α = .97)was administered between the present and past DES.Participants responded to 69 appraisals on a 6-point Likert type scale as described in Study 1.The SHS is a 4-item measure of global feelings of happiness (Lyubomirsky & Lepper, 1999, α = .88).The SWLS is one of the most frequently used and validated measures of the judgment component of SWB and includes 5 items (Diener et al., 1985, α = .79).The SHS and SWLS served as our measures of SWB.To assess one's thoughts and feelings about God we administered the doctrinal vs. experiential version of the Authoritarian/Benevolent God Scales (ABGS; Van Tongeren et al., 2019;Zahl & Gibson, 2012).This questionnaire includes two columns in which participants report first their beliefs about 'what is theologically true about God', followed by their report on the same attribute as to the participants' 'personal experience of what God is like' (Authoritarian doctrinal = 91., experiential α = .92;loving doctrinal α = .95,experiential α = 97.).The doctrinal vs. experiential version of the Revised Nearness to God Scale (NTG-R; Gorsuch & Smith, 1983;Uhder et al., 2010;P. C. Watkins, Frederick, & Davis, 2022) was provided to assess one's relationship with God.We organized this measure similar to the ABGS, with participants providing doctrinal and experiential responses for each of the seven items (doctrinal α = .87,experiential α = .94).After completing the confidence in the existence of God measure (described in Study 1), we administered the Multidimensional Existential Meaning Scale (MEMS; George & Park, 2017).This scale measures three constructs that constitute meaning in life: comprehension (α = .85),purpose (α = .91),and mattering (α = .87,total MEMS α = .92).Finally, to measure religious commitment participants completed the Santa Clara Strength of Religious Faith Questionnaire (SRF; Plante & Boccaccini, 1997, (α = .97),followed by our demographics form.

Fading affect
As in Study 1, we analyzed fading affect with a 3 (recall condition) x 2 (time: past and present emotions) repeated measures GLM for general PA, gratitude, and amusement. 6In these analyses, time was a repeated measure and recall condition was a between-subjects factor.Our general PA variable was as in Study 1, but because our positive event comparison condition was amusement, we did not deem it e (past α = .81,present α = .87).For PA, no main effects for time (p = .105)or condition (p = .166)were found.As predicted, a significant time x condition interaction was found, F (2, 163) = 6.34, p = .002,partial = .072.As seen in Figure 3, this interaction was due to positive affect showing the typical fading affect effect in the amusement recall condition, but essentially no fading positive affect in the two gratitude recall conditions.Means and SEs are found in Table 4.
The GLM analysis for gratitude showed the expected time x recall condition interaction: F(1, 163) = 4.14, p = .018,partial ղ 2 = .048.As seen in Figure 4, this was due to the amusement memory showing the typical fading affect effect for gratitude, but both grateful memories showed flourishing affect: participants reported greater present gratitude than they experienced with the event in the past.The GLM analysis for amusement also showed a significant time x recall condition interaction: F(2, 163) = 4.64, p = .011,partial ղ 2 = .054.As with positive affect more generally, this interaction was due to amusement fading over time with the amusement event, but essentially no fading affect with the gratitude memories.
In sum, these results replicate those of Study 1: compared to other positive memories, gratitude related memories do not seem to fade over time, and in the specific case of gratitude, the emotion appears to grow.

Appraisal analyses by recall condition
Because GTG was central for all remaining statistical tests, following our preregistered analyses, we only used participants who expressed belief in God (n = 138).Because of the number of appraisals we assessed, we conducted data reduction with a Principal Components factor analysis with a Promax rotation.The first factor extracted was related to spiritual appraisals.We factor analyzed the spiritual appraisals and several subfactors emerged.The first factor we titled 'Spiritual Benefit Appraisals' which included 9 items such as 'In this situation, I felt loved by God or the Sacred', 'I felt that this situation enhanced my relationship with God or the Sacred', 'Through this situation, I felt like I was special to God (or the Sacred)'.The appraisal that dealt with God's contribution to the event we deemed more of a construct validity item, and thus this was left out of the calculation of this factor.The second spiritual appraisal factor we termed 'Sacred Gift' and included 7 items such as 'I saw this situation as a sacred gift to me', 'This benefit/blessing had real spiritual or sacred meaning to me', and 'Through this situation, I thought life itself is a sacred gift'.A 'Spiritual Harm Appraisals' factor also emerged which included items such as 'Through this situation, something central to my spirituality was lost'.We did not make predictions by recall condition for this factor, but this scale was used in our correlation analyses described below.
Several non-spiritual appraisal factors also emerged.The first was a Value/Significance factor which included items related to the value of the benefit, as well as the altruistic motives of the benefactor.Other non-spiritual factors also emerged that dealt with themes such as connectedness, deserving or entitled to the benefit, and personal control.Because we did not make predictions regarding these factors, we do not describe them in detail here, but a list of all appraisal items organized by factor is included in the supplementary materials.
To test the effect of recall condition on appraisals, we utilized a univariate GLM with recall condition as the between-subjects factor, confidence in the existence of God as the covariate, and used Bonferroni pair-wise comparisons to explore any main effects.Our first two analyses dealt with Hypothesis 1 which predicted that spiritual appraisals would be more likely in the God-caused recall condition.Indeed, Spiritual Benefit Appraisals were higher in the Godcaused benefit condition, F(2, 135) = 8.03, p < .001,partial ղ 2 = .109.Pairwise comparisons showed that the God-caused benefit condition was significantly higher than the amusement (p < .001)and humancaused benefit (p = .005)conditions.Similarly, we found a significant effect of recall condition on Sacred Gift Appraisals, F(2, 133) = 8.30, p < .001,partial ղ 2 = .111,with Sacred Gift Appraisals significantly higher in the God-caused benefit condition than amusement (p < .001)and human-caused benefit (p = .020).We also conducted this analysis with the Value/Significance factor, and although the effect of recall condition was significant, F(2, 136) = 3.86, p = .023,partial ղ 2 = .054,this was due to the amusement condition being significantly lower than both of the gratitude-relevant conditions (p =.012, p = .023).There was no significant difference between the God-caused and human-caused recall conditions (p = .813).
Hypothesis 2 predicted that higher level construals would be more likely in the God-caused recall condition, and lower level construals would be more likely in the human-caused condition.Although factor analysis did not reveal high or low level construal factors, based on Construal Level Theory, we formed high level and low level construal scales.The high level construal scale consisted of the following appraisals: 'I felt that because of this situation, I would reap benefits in the distant  future', 'I felt that others would benefit from this situation', 'In this circumstance, many people in addition to me received this benefit', 'In this circumstance, I thought about many blessings that I had received in the past', and 'I thought about how the benefits of this situation extend into the future' (α = .79).Low level construals were deemed to be: 'I felt that the situation would benefit me right now', 'I felt that the benefits that I would gain from this situation were very certain', and 'I felt confident that an event like this would happen again in the future' (α = .53).Recall condition did not show a significant effect on high level construals, F(2, 136) = 2.15, p = .121,partial ղ 2 = .031,but pairwise comparisons showed the Godcaused benefit condition to be non-significantly higher than the amusement (p = .097)and the human-caused benefit (p = .063)conditions, providing mixed support for the Construal Level Theory hypothesis.No significant effect of recall condition was found for low level construals, F(2, 135) = 0.98, p = .377,partial ղ 2 = .014.In sum, we did not find clear support for construal levels differentiating the recalled events.

Correlations of GTG with appraisals
Hypotheses 3 and 4 dealt with identifying the distinct cognitive context of GTG.Thus, we conducted bivariate correlations between the relevant appraisals and GTG and GTH.We conducted these analyses across recall conditions for the sake of statistical power and to insure an adequate range of variance for the variables. 7Correlations relevant to our hypotheses are found in Table 5. 8 First, as seen in this table, the value and altruism appraisals predicted GTG and GTH benefactor in a similar fashion, and correlations with the Value appraisal factor bore out this finding.None of these correlations were significantly different (all 1-tailed ps > .16).Supporting Hypothesis 4, the Spiritual Benefit and Sacred Gift appraisal factors were more closely associated with GTG than with GTH.For both past and present emotions, GTG was significantly more closely associated with the Spiritual Benefit Appraisals factor (past: Z = 3.58, p < .001;present: Z = 4.04, p < .001).Similarly, GTG was significantly more closely associated with the Sacred Gift factor than GTH (past: Z = 1.84, p = .03;present: Z = 2.61, p = .004).
Contrary to predictions, high level construals did not predict GTG greater than GTH (p = .369and p = .405).However, although Low level construals showed similar associations with present emotion (p = .810),low level construals correlated more with past GTH than GTG (Z = 1.92, p = .038),as would be expected by Construal Level Theory.Again, mixed support was found for Construal Level Theory predictions.It is possible that the self-report approach for assessing construal level is not ideal, and future research may want to follow the lead of other approaches used by construal level research.
As in Study 1, cost to the benefactor did not correlate with past or present gratitude, and no significant differences between the GTG and GTH correlation coefficients were found.Because cost has been identified as one of the distinguishing appraisals of gratitude, we believe this is a notable result, and we deal with this issue further in our General Discussion.
Somewhat surprising results were found in correlations with the Spiritual Harm Appraisals factor.For reports of present emotion, Spiritual Harm Appraisals were negatively associated with GTH, but no relationship was found with GTG.Indeed, these correlations were significantly different (Z = 1.72, p = .040).Moderation analysis showed that this correlation with GTH was primarily due to the God-caused benefit recall condition (r[45] = α.48,p < .001),and Spiritual Harm Appraisals were positively (but non-significantly) associated with GTH in the amusement recall condition (r[38] = .24,p = .140).Because we did not make this prediction, more research on this finding is needed, but if replicated, it has some interesting implications.In short, when recalling God-caused benefits, Spiritual Harm Appraisals predict less gratitude to human benefactors.
To summarize the findings of appraisal correlations, as predicted, GTH and GTG are similarly associated with value and altruism appraisals.The appraisals that seem to distinguish GTG from GTH are spiritual in nature.Spiritual benefit and sacred gift appraisals are much more predictive of GTG than GTH.

Correlations of GTG with SWB and religiosity/ spirituality variables
Table 6 provides the correlations between GTG, GTH, and SWB and religiosity/spirituality measures.First, note that GTG is positively associated with SWB.Apparently contrasting with past results, GTG was as strongly associated with SWB as GTH.Indeed, GTG was more strongly correlated with the SHS than GTH (Z = 1.82, p = .034).Past research has suggested that general gratitude is much more strongly predictive of SWB than GTG.This contrast highlights a key methodological issue in gratitude research.When researchers assess nonspecific gratitude, they may be assessing several things, including a combination of GTG and GTH.Thus, we believe that the better comparison with GTG is with gratitude felt toward a specific benefactor, which in this case was a human benefactor.
As expected, GTG was more strongly associated with religious/spiritual variables than was GTH.More importantly, as predicted by Hypotheses 6 and 7, GTG was more strongly associated with experiential than doctrinal views of a loving God (Z = 3.04, p = .001),and experiential than doctrinal sense of nearness to God (Z = 5.91, p < .001).This has significant implications for understanding GTG, which we explore in the General Discussion.

Summary of study 2 results
In Study 2 we replicated several results from Study 1. First, the positive emotional impact of grateful memories endures more than other positive memories.Second, value and altruism appraisals predict GTG and GTH in a similar fashion.Third, results raise questions about the importance of cost appraisals to gratitude.Study 2 added to the findings of Study 1 by identifying some distinguishing features of the cognitive context of GTG.In contrast with GTH, Spiritual Benefit Appraisals and appraisals of Sacred Giftedness predict GTG.In this study we also found that GTG was as strongly associated with SWB as GTH.Not unexpectedly, GTG was much more strongly associated with our religiosity/spirituality measures than GTH.GTG was more strongly associated with one's experience of God as loving and near, than with their doctrinal conceptions of these issues.

General discussion
In two studies we explored the unique cognitive context of GTG and investigated how grateful memories fade in emotional impact compared to other positive memories.These studies offer five contributions to the gratitude literature.First, we found that in terms of positive emotional impact, grateful memories endure more than other positive memories.Whereas happy and amusing memories showed fading positive affect, grateful memories did not.Second, results from both studies questioned the importance of the benefactor cost appraisal to gratitude.Thoughts about the cost of a benefit to the benefactor did not correlate with GTG, GTH, or general gratitude.Third, in both studies we found that appraisals of value of the benefit and the altruistic motivations of the benefactor contribute in a similar way to gratitude.Value and altruism appraisals correlated in a like manner to GTG and GTH.Fourth, we found that spiritual benefit appraisals predicted GTG significantly more than GTH.Thus, spiritual appraisals appear to be unique to GTG.Finally, GTG is much more strongly associated with one's experience of God as loving and near, compared to their 'head' doctrinal beliefs about God.We consider each of these contributions in turn.How does gratitude influence SWB?One way that gratitude might enhance happiness is because grateful memories have a positive emotional impact that does not fade over time.In both studies we found that not all positive memories are equal regarding fading positive affect: happy and amusing memories fade over time significantly more than do grateful memories.Indeed, with respect to the specific emotion of gratitude, happy and amusing events show fading gratitude over time, but grateful memories show that gratitude increases with time.
Our findings regarding the cost appraisal also offer a significant contribution to the gratitude literature.In both studies, we found that the cost appraisal did not predict GTG or GTH.In other words, appraising a benefit as more costly to the benefactor did not result in more gratitude.Along with value and altruism, cost has been identified as one of the primary appraisals leading to gratitude, so we view this as a notable finding.Indeed, our results are consistent with recent research with actual significant benefits (P.C. Watkins et al., 2016), and with experimental studies that have manipulated cost in scenarios (McLaughlin et al., 2018).Thus, serious reconsideration of the importance of the cost appraisal to gratitude may be in order.It might be that individuals make a more global appraisal of the 'goodness of the giver' that includes both altruism and cost appraisals (P.C. Watkins, 2014).The fact that cost appraisals correlated with judgments of altruism (r = .214,p = .006),supports this suggestion.Another possibility is that cost appraisals only predict gratitude in unique circumstances.Our cost appraisal findings might be limited to individualistic cultures, but in any case, we await more research on the importance of the cost appraisal to gratitude.
What is similar and what is unique about the two forms of gratitude?GTG and GTH appear to be similar in their relationship with appraisals of value and altruism.The more people valued an event the more GTG and GTH they experienced.The more they viewed a benefit as given with altruistic motives, the more GTG and GTH they experienced.This implies that in many ways GTG and GTH are similar.Even though a divine being is seen as qualitatively different from a human, both forms of gratitude are impacted by psychological valuing, and both are relational emotions.
In Study 2 we included spiritual appraisals to determine if they might be unique to GTG.Indeed, the Spiritual Benefit and Sacred Gift appraisal factors were significantly more predictive of GTG than GTH.When people construe events spiritually, they are more likely to experience GTG.For example, if one interprets an event as indicating that God loves them, they are more likely to experience GTG in response to that event.Because the sacred giftedness appraisals do not explicitly refer to God, we find these relationships informative to theory about the cognitive context of GTG and believe that this construct deserves more investigation.
The question arises as to what prompts a person to spontaneously think about spiritual explanations of an event.Context likely matters.A secular context such as a lab at a state university is unlikely to prompt spiritual explanations, but if one were in a place of worship, spiritual explanations should be more likely.Our findings suggest that encouraging positive spiritual appraisals enhances GTG, and future research might focus on the contexts that encourage spiritual interpretations of events.
One of the limitations of these studies is that people are reporting appraisals retrospectively.Thus, we can only conclude with clarity that positive spiritual appraisals are associated with GTG.Although our findings support the hypothesis that spiritual appraisals cause GTG, clearer support would be provided by experimental designs that manipulate spiritual appraisals about a benefit.One approach that has been successful in experimentally manipulating appraisals is cognitive bias modification for interpretation (Hertel & Mathews, 2011).With this method, interpretations are manipulated by training participants to interpret scenarios in a particular way, and this training has been shown to impact emotions and the interpretation of actual events.Although this approach has yet to be applied to GTG, it has been successfully used to train grateful interpretation biases (P.C. Watkins et al., 2021).Thus, Cognitive Bias Modification for Interpretation might be one way to experimentally manipulate spiritual appraisals to more definitively determine if these cause GTG.
Insightful findings emerged in the correlations of GTG and GTH with SWB.In past research, relationships between GTG and SWB have been notably lower than general gratitude and SWB (P. C. Watkins, Frederick, & Davis, 2022).However, associations of GTG with SWB were comparable to GTH.Indeed, regarding global subjective happiness, GTG was correlated with SWB at a significantly higher rate.We believe that this result has methodological implications for gratitude research.When one asks someone how grateful they are feeling, this untargeted emotion likely involves several factors, including GTG and GTH.Indeed, in Study 2 partial correlations of GTG and GTH with general gratitude indicated that both GTG and GTH contributed independently to general gratitude.Thus, we recommend that when researchers are investigating gratitude in response to events, it would behoove them to include facets of gratitude that target the benefactor(s).
As in past research (P.C. Watkins, Frederick, & Dodson, 2022;P. C. Watkins et al., 2019), GTG was found to be much more highly associated with religiosity/spirituality than GTH.Although this is weak support for the hypothesis that GTG enhances spiritual wellbeing, data from prospective designs have shown that trait GTG predicts increased spiritual/religious wellbeing over time (P.C. Watkins, Frederick, & Dodson, 2022;P. C. Watkins et al., 2019).Experimental designs would offer more definitive evidence for this theory.A novel contribution of our findings is that GTG showed significantly stronger associations with one's subjective experience with God as loving and near than with their doctrinal (i.e.head) beliefs about these attributes.Past research has shown that head beliefs about God vary less than do one's heart knowledge of God (Van Tongeren et al., 2019).Because head beliefs about God are less likely to change, it could be that one's heart beliefs about God are more relevant to their ongoing spiritual well-being than their doctrinal beliefs.This finding adds further support as to the centrality of GTG to one's spiritual experience.
Several additional limitations of these studies should be highlighted.This study asked participants to recall an event along with their emotional experience of that event.Thus, retrospective memory biases could be impacting our results.For example, the mood-congruent memory effect suggests that when recalling a grateful memory, grateful emotion is likely to be activated, and that emotional state could impact participants' recall of their original emotional experience with the event.However, extensive daily diary studies have shown that the fading affect effect is not due to retrospective biases (Walker et al., 1997).Moreover, because we found that fading affect in gratitude memories was significantly less than that of other positive memories, it is difficult to imagine how retrospective memory biases would have impacted gratitude memories more than happy or amusing memories.Nonetheless, it is probably best to interpret the differences in fading affect as participants' perception of their experience until this finding is replicated with other methodologies such as daily diary studies.
Another notable limitation was that that females were oversampled in our studies.Although this is the case with most psychology studies, this presents a number of issues for the generalization of our findings, and future research should recruit more representative samples.Women have been found to be more spiritual/ religious than men and more grateful than men (e.g.McCullough et al., 2002), and these factors could have impacted our results.It is possible therefore, that women recall grateful memories more easily than men.In short, these issues create limitations to the generalization of our results.
In sum, the benefactor matters to one's experience of gratitude, particularly when the benefactor is divine.These studies demonstrated that positive spiritual appraisals are unique to GTG.Moreover, GTG seems to be more important to one's heartfelt experience with God as loving and intimate than with their doctrinal beliefs about these divine attributes.Thus, GTG may be vital to one's relationship with God.Finally, our results showed that over time, participants perceived that their grateful memories held more of their positive affect and gratitude than did memories of happy or amusing occasions, and this might be why grateful recounting is particularly beneficial to SWB.Not all positive memories are equal regarding fading positive affect.Some eventsparticularly grateful memories -seem to have the potential to grow in their ability to produce a positive emotional impact.To draw upon the wisdom of C.S. Lewis; 'A pleasure is only full grown when it is remembered' (Lewis, 1996, p. 73).It might be that remembering grateful events is particularly effective in making the pleasure of a memory 'full grown'.

Notes
1. Admittedly, the fact that orthodox Christian theology holds that Jesus Christ is fully God and fully human makes the contrast between GTG and GTH somewhat problematic.For our purposes, GTH refers to gratitude towards humans other than Jesus Christ.2. As will be seen, we believe that this is an important distinction that bears on the interpretation of gratitude research.3. Upon reflection, we now feel that this was a poor use of words, as in situations with a human benefactor, participants may have been responding to the human benefactor or they may have been responding to humans who were not the primary benefactor.
Because of this problem we did not analyze results with regard to this item.4.Although we acknowledge that God-caused and personcaused benefits are not mutually exclusive, we sought to create explicit comparison recall conditions. 5.In analyses of each of the positive emotions in this factor, contentment, joy, love and pride all showed significant recall condition x time interactions.The pattern was similar to that shown in general positive affect (Figure 1), with the exception that both love and pride showed flourishing affect in the gratitude recall conditions.6.We conducted these analyses with and without confidence in the existence of a loving God as a covariate.
Because the results were essentially equivalent, here we report only the analyses without the covariate.7.For the correlations relevant to our hypotheses, we conducted moderation analyses with PROCESS.These analyses indicated very few qualitative differences in terms of the correlations by recall condition, with the exception of Spiritual Harm Appraisals, which we explain in that context.8.For the critical correlations, we also conducted partial correlations controlling for confidence in the existence of God.These analyses did not change the interpretation of the correlations.

Figure 1 .
Figure 1.Study 1 past and present general positive affect means by recall condition.

Figure 2 .
Figure 2. Study 1 past and present gratitude means by recall condition.

Figure 3 .
Figure 3. Study 2 past and present positive affect means by condition.

Figure 4 .
Figure 4. Study 2 Past and Present Gratitude Means by Recall Condition.

Table 1 .
Study 1 estimated marginal means and standard errors for fading affect.

Table 2 .
Study 1 estimated marginal means and standard errors for appraisals by recall condition.

Table 3 .
Correlations of appraisals with emotions from study 1.

Table 4 .
Study 2 estimated marginal means and standard errors for appraisals by recall condition.
Standard errors are in parentheses.

Table 5 .
Correlations of appraisals with emotions from study 2.

Table 6 .
Correlations of GTG and Gratitude to Human Benefactor with SWB and Religious/spiritual Variables: Study 2.