Abstract
Research demonstrates that people draw on their religions in coping with difficult life circumstances, and that more religious people tend to cope better than less religious people. Lament is a religious prayer practice that constitutes a form of meaning-making coping. Here, we explicate this form of coping and, in a series of three studies with Christian participants, report on the development of a lament scale. In the first study, we develop items and test the items for clarity and for generalizability to diverse Christian groups using expert review and cognitive interviewing with participants representing five Christian groups. In the second study, we report results of exploratory factor analysis using data from primarily White and Protestant MTurk participants (N = 303). In the third study, we report on factor stability and construct validity using data from largely White and African American Protestant and Catholic Prolific participants (N = 346). The studies demonstrated generalizability to diverse Christian groups, a three-factor structure consistent with our theoretical formulation, good construct validity and relationships with well-being. Moderation analyses also indicated that when both Complaint and Praise are either high or low, the severity of the event is inversely related to flourishing. However, when either Complaint or Praise is high, the negative relationship between severity of the event and flourishing is ameliorated. The Lament Scale is a reliable and valid instrument for assessing lament in ways that are consistent with the conceptualization of lament as a meaning-making coping practice.
Recognizing that people often turn to their religions as important sources of meaning when facing difficult life circumstances and that religious variables are often associated with better outcomes (Park et al., 2009), researchers have recently begun investigating religion-specific (i.e., emic) resources for enduring suffering (Hall et al., 2018). The study of emic variables—that is, variables that emerge from within a particular cultural perspective—allows for increased multicultural understanding, more nuanced research results, and, when used in clinical applications, greater ethical sensitivity (DiFonzo & Hall, 2022).
Grounded in Christianity’s Jewish heritage, lament is a form of prayer that constitutes an emic Christian religious practice for coping with suffering, and that takes the form of a meaning-making narrative structure. The intentional practice of lament has been neglected in Christian contexts in recent decades, especially in Western and European American contexts (Brueggemann, 1986). However, there has been a resurgence of interest in reviving this practice in the past few years, as evidenced by a number of major theological works on the topic (Billings, 2015; Harrower & McDonough, 2019), a proliferation of liturgies of lament offered online targeted at specific sources of suffering (e.g., for victims of injustice, Faith for Justice, 2020), and the development of church resources for lamenting social ills such as racism (Muthiah et al., 2021). In this article, we articulate a Christian theological understanding of lament, review related research in psychology, and report on three studies in which we develop a measure of lament and explore its relationship to well-being.
Lament from a Theological Perspective
The practice of lament is a relatively common way of engaging with God often found in the Psalms, where some scholars estimate that up to 40% are psalms of lament (Pemberton, 2012), and quoted extensively in the New Testament. From a theological perspective, “Laments rise to the heavens as a strange combination of complaint, grief, questions, confusion, [admission of guilt,] desire for rescue, and expectation of divine faithfulness” (Kapic, 2017, p. 29). 1 Lament as expressed in scripture and in the historic practice of the Christian Church is distinct from other cultural expressions of grief or complaint primarily because it is personally addressed to God the Father. While in form it overlaps with the Jewish practice of lament (Liska, 2014), in the Christian tradition lament is understood and practiced within the context of the passion and resurrection of Christ (Eklund, 2015).
The following elements are common in the psalms of lament (as illustrated by Psalm 22; Kapic, 2017, p. 37): (a) A cry out to God (e.g., “My God, My God, why have you forsaken me?”); (b) a complaint or voicing of the crisis that drives the poem (e.g., “Why are you so far from saving me, from the words of my groaning . . . All who see me mock me”); (c) A petition in which remedy is yearned for (e.g., “Be not far from me, for trouble is near, and there is none to help”); (d) usually, a claim of confidence in God (e.g., “all the ends of the earth shall remember and turn to the Lord”); and (e) usually, a commitment to praise God (e.g., “I will tell of your name to my brothers; in the midst of the congregation I will praise you”). These last two elements also distinguish Judeo–Christian lament from other cultural forms of lament, which do not end with the desire to claim confidence in and to praise God (Westermann, 1981). Space constraints do not allow for more detailed theological treatment of lament. For more information, see Billings (2015), Pemberton (2012), and Brueggemann (1982, 1995).
Lament from a Psychological Perspective
From a psychological perspective, lament constitutes an emic meaning-making practice for those experiencing distress. This distress can either be one’s own or others’, such as perceived injustice. Lament is intended to shape the sufferer’s experience over time as meaning is formed and to produce hope (Hall, 2016). While psychological research on lament is sparse, some theoretical attempts to explicate lament in psychological terms have been made (Gioielli, 2021; Hall, 2016; Torr, 2019). These articles emphasize the potential for lament to serve as a mechanism for healing from trauma or suffering. Only three empirical studies have directly investigated lament. Two of these (Dickie, 2019, 2020) had small samples and did not use inferential statistics, but noted qualitative and observational findings supporting its effectiveness in trauma populations. Snow et al. (2011) tested a 4-week lament intervention in 200 undergraduate students from Christian colleges. In participants who completed all readings and assignments, the results indicated a significant increase in feelings of closeness with God in the intervention condition but not in the two control groups (active and passive).
In addition to these studies directly focused on lament, two existing psychological constructs have addressed engagement with God in ways that overlap with the complaint element of lament: Hall and Edwards’ (1996, 2002) concept of “realistic acceptance,” and the work by Exline et al. (2012) on “protest.” Hall and Edwards (1996, 2002) drew on object relations theory to create a measure of relationship with God, the Spiritual Assessment Inventory (SAI). One subscale of this measure, entitled Disappointment with God, reflects a number of negative experiences with God. Another, Realistic Acceptance, assesses a person’s capacity to work through difficult experiences with God. High scores indicate continuing engagement with God, while low scores indicate tendencies to blame oneself or give up on the relationship. Like the SAI Realistic Acceptance construct, our own concept of lament involves this capacity to engage with God during difficult times. Lament may also encompass Disappointment with God, but in lament the negative emotions that are brought to God are not necessarily directed toward God.
Exline et al. (2012, 2016) proposed the notion of protest toward God. Exline et al. (2012) found that, after controlling for behaviors indicating distancing from God, protest toward God was associated with seeing oneself as having a close, positive, ongoing, resilient bond with God. In a subsequent study, Exline et al. (2016) studied protest in a short-term longitudinal study of participants in a chronic headache intervention program and found that anger toward God and protest correlated highly with each other. In addition, anger at pretest predicted sustained distress over the treatment period. However, when reported anger was statistically controlled for, protest toward God at pretest predicted decreases in anxiety, depression, and stress at posttest. As a point of contrast, items representing relationship-building with God (e.g., “tried to trust God”) at pretest were not correlated with distress at posttest. These results indicate that engaging with God in protest may be more helpful in alleviating distress than is simply attempting to draw close to God.
In sum, while these areas of research capture the complaint aspect of lament, and are congruent with the concept of lament in suggesting that complaint against God can exist within a warm, engaged experience of relationship with God, no measure yet exists that captures all elements of this emic Christian meaning-making practice.
The Present Studies
From a psychological perspective, based on the theological understanding of lament noted above, the essential elements of lament are: (a) the engagement of God with negative emotions through questioning, complaint, or expressions of longing in the expectation that God will hear and respond; (b) turning to God for help through a request or petition, and (c) a commitment to praise God or expressions of confidence in God that follow closely on the complaint. Using these elements as a framework for construct development, we present three studies documenting the development of the Lament Scale (LS) for use with Christians.
Our work relating lament to well-being requires some clarification. In line with a theological account of lament, we conceptualize lament as a process that can facilitate meaning-making coping. Difficult life experiences can threaten one’s sense of meaning, causing a personal search to cope by making sense of suffering in an attempt to alleviate distress and increase well-being. Successful meaning-making can lead to increased well-being (Park, 2013). However, given the limited use of lament in liturgical contexts noted above, it is unlikely that our study participants have been exposed to the practice of lament in their religious contexts. Given this deficit, along with the cross-sectional nature of the studies described here, we are limited in our ability to study the naturally occurring process of lament. Instead, we will focus primarily on each of the theoretical components of lament.
Study 1: Item Development and Cognitive Interviewing
The purpose of the first study was to develop items to capture the elements of lament and evaluate the initial content validity of the items via expert review and cognitive interviews. Forty-five items were developed to represent the three elements of lament. The items were evaluated by six reviewers with expertise in lament. 2 All held terminal degrees in psychology or theology and had published in related areas. The experts examined the items for face validity, construct fidelity, and clarity of wording. They also provided open-ended feedback about the scale items and suggested additional items, resulting in an initial pool of 63 items.
Using cognitive interviewing, we then reviewed these items for clarity by focusing on the cognitive processes that are engaged in responding to items: comprehension, recall, judgment, and response. An additional purpose was to ascertain whether the range of interpretations associated with the evaluated items varies acceptably between diverse Christian groups, an objective known as semantic equivalence (Daouk-Oyry & McDowal, 2013). To achieve this goal, we drew on methods developed for cross-cultural cognitive interviewing (Willis, 2015).
Method
Varying perspectives rather than representativeness is the goal in cognitive interview sampling. For our purposes, we wanted representation from diverse Christian traditions, and thus recruited two participants from each of the following groups: conservative Protestant, mainline Protestant, Pentecostal, Catholic, and Eastern Orthodox. Our resulting sample size of 10 is within the recommended guidelines (Peterson et al., 2017). After obtaining human subjects ethics approval, personal contacts and snowball sampling were used to obtain participants who were religiously committed, self-reflective, and verbal, and who had faced a life challenge sometime in the recent past; attempts were made to also have ethnic, age, and gender diversity. Interviews, conducted via Zoom, also included items from other scales and lasted approximately 1.5 hours, and participants were compensated with a US$50 gift card. Participants provided informed consent. We used a combination of think-aloud and verbal probe procedures with each item (Peterson et al., 2017), as is normative in cross-cultural cognitive interviewing (Willis, 2015). All interviews were audio-recorded. The resulting sample consisted of four men and six women, ranging in age from 22 to 68 (M = 35.7, SD = 16.89). With respect to ethnicity, six were White/European American (non-Hispanic), three were Hispanic/Latinx, and one was Asian American.
Analysis and discussion
Two researchers reviewed the qualitative data gathered for each item and assessed each for difficulties in comprehension, recall, judgment, and response. They also evaluated any comprehension differences between religious groups. Most items had four or fewer marked difficulties. Two were found to be problematic on one or more of these dimensions by six or more participants and were deleted. No problems emerged with respect to differences in comprehension between religious groups.
Study 2: Item Reduction and Factor Analysis
The purpose of the second study was to reduce the number of items representing the three facets of lament via principal component analysis (PCA), evaluate the factor structure of the reduced set of items with exploratory factor analysis (EFA), and determine internal consistency reliability with Cronbach’s alpha.
Method
After receiving IRB approval (University of Connecticut IRB, X20-0066), participants were recruited from Amazon’s Mechanical Turk (MTurk) and compensated US$2.50. Eligible participants were aged 18 or older, resided in the United States, had at least a 95% MTurk approval rating, and had completed at least 500 successful tasks on MTurk. The project was advertised as an anonymous survey for Christians who are active in their faith and have experienced a serious stressful experience that has caused substantial suffering in the past 6 months. Participants provided informed consent. Items on the LS were rated from 0 (strongly disagree) to 4 (strongly agree). In addition to the items developed for this scale, the questionnaire included items for other measures as part of a larger project.
The MTurk data included attention checks to detect automated responses (i.e., “bots”). The first attention check was performed utilizing participants’ reports about a stressful event. Based on codes suggested by two independent raters, cases were deleted if the responses were not reasonable, such as numbers or “nothing” in text format. Second, we screened duplicate, empty, or any forms of invalid cases. Of the 541 participants, 74 had duplicate or empty responses, 84 were excluded for giving quantitative responses in their reports about the stressful event, and 12 more for giving invalid qualitative answers to this item. Of the remaining valid cases, the number of participants who finished the LS was 303 (56%); only these were used in the analyses.
The average age of participants was 37.82 (SD = 11.16, range = 21–75). Slightly more than half were male (51.8%; n = 157). The majority were White (79.9%; n = 242), 10.6% were Black or African American (n = 32), 5.3% were Asian or Asian American (n = 16), 2% were American Indian or Alaska Native (n = 6), 1.6% Biracial/Multiracial (n = 5), and 0.7% other racial groups (n = 2). As for the participants’ ethnicity, 13.2% were Hispanic or Latinx (n = 40). Most were married or cohabitating (78.2%; n = 237), 17.2% were single (n = 52), 4.3% divorced or separated (n = 13), and 0.3% widowed (n = 1). The participants’ median annual income was US$60,000–US$79,999. The modal participant had a bachelor’s degree (60.4%; n = 183), 17.8% had a master’s degree (n = 54), 11.2% some college (n = 34), 4.6% a high school diploma or GED (4.6%; n = 14), 4.3% an associate degree (4.3%; n = 13), 1.3% a professional degree (i.e., MD, JD; 1.3%; n = 4), and one person held a doctoral degree (0.3%). With respect to religious affiliation, 53.8% reported simply “Christian” (n = 163), followed by Roman Catholic (8.6%; n = 26), Baptist (8.6%; n = 26), Lutheran (4.0%; n = 12), Protestant (2.6%; n = 8), Pentecostal (2.3%; n = 7), or Non-denominational (2.3%; n = 7).
Results and discussion
Item screening
Four rounds of PCA with Oblimin rotation (Delta = 0) were performed to reduce the number of items. The first PCA was conducted using 61 items. After excluding items with loadings less than .50, cross-loadings on multiple components more than .50, and items with duplicate content, eight components were shown and 37 items were removed. The second PCA indicated three components, and one item was deleted; the third round indicated three components and excluded six items; the final round showed three components, and no further items were excluded.
Factor analysis
After item removal processes utilizing PCA, we conducted EFA applied principal axis factoring to detect essential factors explaining lament, using the same three criteria as used above (i.e., loadings less than .50, cross-loadings on multiple components more than .50, and items with duplicate content). Consistent with the PCA, the EFA with the 17 remaining items revealed three factors; no items were dropped (see Table 1). The three factors mapped onto the three elements of lament identified theoretically: Praise, Complaint, and Request. The adequacy of sampling and item correlations were excellent: Kaiser-Meyer-Olkin test (KMO) = 0.89, and χ2 [136] = 2455.83, p < .001. As for the three factors, the reliability of each factor was acceptable (.87, .88, and .78, respectively), as was the lowest item-total correlation (.60, .65, and .36, respectively). The intraclass correlation for single measures was strong (.54, .55, and .42, respectively). In conclusion, the results of the second study suggested a three-factor solution consistent with the theoretical conceptualization of lament, and established the reliability of the three subscales.
Items and Factor Loadings of the Three Factor Solution from Study 2 (EFA) and Study 3 (CFA).
EFA: exploratory factor analysis; CFA: confirmatory factor analysis.
Factor loadings less than .50 was not included in the table. Factor loadings in bold indicate the largest loading for that row.
Study 3: Confirmatory Factor Analysis, Construct Validity, and Relationship to Well-being
The purpose of the third study was to confirm the factor structure and evaluate the construct validity of the scale. We hypothesized that the Complaint subscale would be moderately and positively correlated with the RSS Protest subscale and with the RCOPE Spiritual Discontent subscale; that the Request subscale would be moderately and positively correlated with the SAI Realistic Acceptance subscale, and the RCOPE Pleading for Direct Intercession and Spiritual Connection subscales; and that the Praise subscale would be moderately and positively correlated with RCOPE Spiritual Connection and Active Religious Surrender subscales and intrinsic religiosity, and negatively with Avoidant Attachment. We hypothesized nonsignificant or small correlations with extrinsic religiosity and extraversion.
Because lament is conceptualized as a meaning-making coping practice, we were also interested in its relationship to well-being. As noted earlier, lament consists of a process with three relatively distinct components, so the proposed relationship of the scale to indices of well-being is complex. Consequently, we examined the relationship of the LS to well-being in two ways. First, we looked at the zero-order correlations between the LS subscales and the indicators of well-being. When the subscales are considered individually, we might expect Complaint to be negatively related to well-being if the presence of difficult emotions results in their expression; however, the expression of these emotions in relationship to God might also decrease their intensity. The same might be true for Request. In contrast, we would expect Praise to be positively related to well-being.
Second, we also examined the combined effects of the components of lament. Because, from a theoretical perspective, the combination of the steps in the lament process are all important to the meaning-making process, we tested the potential interactions of the components through moderation analyses. In the interests of decreasing the complexity of the interactions, we used only Complaint and Praise (the first and last steps of the proposed lament process) as moderators. We hypothesized that the interaction of Complaint and Praise would moderate the relations between the severity of the stressor and the four well-being outcomes.
Method
Participants and procedure
After receiving IRB approval (University of Connecticut IRB, X21-0073), participants were recruited from Prolific and were compensated US$6.00. Eligible participants were aged 18 or older and resided in the United States. The project was advertised as an anonymous survey for Christians who are active in their faith and have experienced a serious stressful experience that has caused substantial suffering in the past 6 months. Participants provided informed consent. Our Prolific data contained participant validity checks such that we did not include attention checks to detect automated responses (i.e., “bots”). However, we did perform a “bot” check on reported stressful events. If the response did not make sense as coded by two separate raters, participants that were deemed “bots” were removed (15 of 384 participants; 3.9%), as were duplicate and empty responses (0.5%; n = 2). Of 367 remaining cases, we analyzed data from 346 participants who fully completed the LS (90.1% out of 384). Three weeks after the initial administration, all participants were sent an invitation to participate in a follow-up intended to evaluate test–retest reliability. The survey was closed after responses were received from 100 participants.
Participants reported when their serious stressful experience occurred on a scale of 0 (less than a month ago) to 7 (more than 6 months ago), how severely the serious stressful experience impacted their life on a scale of 0 (not at all) to 4 (a very large amount), and if the experience was chronic or ongoing (yes or no). With respect to amount of time since experiencing the stressor, 23.1% had stressful experiences less than a month ago, 18.2% 1 month ago, and 13.9% 2 months ago. The average severity of the serious stressful experience was 2.56 (SD = 1.07). Slightly less than half of the participants (45.4%) reported their stressful experiences were chronic or ongoing.
Participants’ average age was 34.38 years (SD = 8.86, range = 18–77). The majority of participants were male (73.4%; n = 254). More than half were White (57.8%; n = 200), followed by Black or African American (36.4%; n = 126), Native Hawaiian or Pacific Islander (1.4%; n = 5), Biracial/Multiracial (1.2%; n = 4), Asian or Asian American (0.9%; n = 3), and Middle Eastern (0.3%; n = 1). With respect to ethnicity, 4.6% reported as Hispanic or Latinx (n = 16). The majority of the participants were married or cohabitating (75.7%; n = 262); fewer were single (20.8%; n = 72), divorced or separated (2.6%; n = 9) or widowed (0.3%; n = 1). Median annual household income was in the range from US$80,000 to US$99,999. With respect to education, a substantial minority had a master’s degree (42.5%; n = 147), followed by a bachelor’s degree (34.1%; n = 118), high school diploma or GED (5.5%; n = 19), doctoral degree (4.9%; n = 17), some college education (4.9%; n = 17), associate degree (4.9%; n = 17), or a professional degree (2.6%; n = 9). With respect to participants’ religious affiliation, the largest group was Roman Catholic (41.0%; n = 142); 16.5% self-identified simply as “Christian” (16.8%; n = 58) and smaller numbers were Baptist (7.8%; n = 27), Pentecostal (6.4%; n = 22), and Protestant (5.8%; n = 20).
Measures
Behaviors Toward God Scale-Protest Subscale
The Behaviors Toward God Scale (Exline et al., 2016) consists of 16 items beginning with the prompt, “Have you responded to your suffering in these ways?,” followed by items reflecting four types of behaviors toward God (Protest, Relationship-Building, Disengagement-Exit, Suppression). Response options were not at all, a little, a lot, and a great deal. In this study, we only used the four-item Protest subscale (e.g., “protested to God”). Initial evidence of reliability and validity was found by Exline et al. (2012). Cronbach’s alpha in this sample was .79.
Religious coping
The RCOPE consists of 21 subscales developed to represent various forms of positive and negative religious coping (Pargament et al., 2000). Each subscale consists of five items to which participants respond on a 4-point Likert-type scale ranging from 0 (not at all) to 3 (a great deal). In this study, we used three of the subscales: Spiritual Discontent (e.g., “Wondered whether God had abandoned me”), Pleading for Direct Intercession (e.g., “Pleaded with God to make things turn out okay”), and Spiritual Connection (e.g., “Looked for a stronger connection with God”). Subscale scores range from 0 to 15, with higher scores reflecting greater use of each coping method. Cronbach alphas for the three subscales in the original sample ranged from .81 to .90, and in the present sample from .77 to .87.
Spiritual Assessment Inventory
The SAI (Hall & Edwards, 1996) was designed as a measure of spiritual maturity. The scale has five factors: Awareness, Realistic Acceptance, Disappointment, Grandiosity, and Instability. Only the Realistic Acceptance subscale was used in this study. The SAI consists of 54 items, 7 of which are two-part items; the Realistic Acceptance subscale is calculated using the second item from these 7 two-part items. The items are on a 5-point Likert-type scale that ranges from Not at all true to Very true. The alpha coefficient for Realistic Acceptance was reported as .83, and was also .83 in this study.
Intrinsic/extrinsic religiosity
Orientation toward intrinsic and extrinsic religiosity was assessed using the Intrinsic/Extrinsic-Revised scale (I/E-R; Gorsuch & McPherson, 1989). The 14-item scale uses a 5-point Likert-type scale with anchors of strongly disagree and strongly agree. The scale consists of three subscales, which are the Intrinsic, Personally Extrinsic, and Socially Extrinsic subscales; the two extrinsic subscales were combined in this study. Sample items from the Intrinsic subscale include “I try hard to live all my life according to my religious beliefs” and “My whole approach to life is based on my religion.” Sample items from the Extrinsic subscale include “I pray mainly to gain relief and protection” and “What religion offers me most is comfort in times of trouble and sorrow.” In the original study, the Intrinsic subscale had a reliability of .62 and the Extrinsic subscale had a reliability of .73. Similarly, in this study, the Intrinsic subscale had an alpha of .63, and the Extrinsic subscale had an alpha of .72.
Attachment to God
The Attachment to God Inventory (AGI; Beck & McDonald, 2004) is a 28-item self-report measure of adult attachment tendencies in emotional relationship with God. The AGI consists of two subscales—the Anxiety subscale (measuring anxiety about abandonment; e.g., “I worry a lot about damaging my relationship with God”) and the Avoidance subscale (measuring avoidance of intimacy; e.g., “I am uncomfortable being emotional in my communication with God”). Each subscale contains 14 items, some of which are reverse-scored and all of which involve rating along a 7-point Likert-type scale (ranging from 1 = disagree strongly to 7 = agree strongly). Higher scores on the Avoidance subscale indicate greater difficulty relying on God while higher scores on the Anxiety subscale indicate more anxious preoccupation about the status of the participants’ relationship with God. In this study, the alpha reliability coefficient was .93 for the anxiety subscale and .86 for the avoidance subscale, indicating good to excellent internal consistency.
Extraversion
Extraversion was assessed using the Extraversion subscale of the Big Five Inventory (BFI; John et al., 1991). This subscale consists of eight items in which participants are instructed to rate whether characteristics apply, or do not apply to them, on a 5-point scale ranging from disagree strongly to agree strongly. The characteristics consist of short phrases based on the trait adjectives known to be prototypical markers of the Big Five. In US and Canadian samples, the alpha reliabilities of the BFI scales typically range from .75 to .90 and average above .80; three-month test–retest reliabilities range from .80 to .90, with a mean of .85. Validity evidence includes substantial convergent and divergent relations with other Big Five instruments as well as with peer ratings. Cronbach’s alpha in this sample was .73.
Positive and negative feelings
The Scale of Positive and Negative Experience (SPANE; Diener et al., 2010) was used to assess participants’ positive and negative feelings. The scale includes 12 items: six items for positive feelings and six items for negative feelings. Each item provides feeling-related adjectives (e.g., good, pleasant, bad, sad) and asks participants to report how often they experienced the feelings in the previous four weeks using a 5-point Likert-type scale: 1 (very rarely or never) to 5 (very often or always). The score of each subscale was calculated by adding the items for each subscale. Higher scores mean greater positive and negative feelings. The Cronbach’s alpha in this sample was α = .91 for Positive Feelings and α = .91 for Negative Feelings.
Satisfaction with life
Satisfaction with life was assessed using the Satisfaction with Life Scale (SWLS; Diener et al., 1985). Participants rated the level of agreement with five items describing the current state of their lives. The items include: “In most ways, my life is close to my ideal,” and “The conditions of my life are excellent.” Each item was rated using a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). The total score was calculated by averaging participants’ responses to each item. Higher scores indicate more satisfaction with life. The Cronbach’s alpha in this sample was .92.
Flourishing
Flourishing was measured using The Flourishing Scale (Diener et al., 2010). The scale has eight items describing social-psychological prosperity such as relationships, optimism, and purpose. The items include: “I lead a purposeful and meaningful life” and “My social relationships are supportive and rewarding.” Participants rated the level of agreement with each item using a 7-point Likert-type scale: 1 (strongly disagree) to 7 (strongly agree). The total score was calculated by adding participants’ responses for all eight items. Higher scores mean more psychological resources and strengths. The Cronbach’s alpha in this sample was .98.
Results
Exploratory and confirmatory factor analyses
Study 3 employed an EFA using the same parameters as in Study 2 to examine if the three-factor solution suggested by Study 2 would be consistent in a new sample. We used the same criteria for item retention as in Study 2 (i.e., loadings less than .50, cross-loadings on multiple components more than .50, and items with duplicate content). The results using the 17 items identified in Study 2 did reveal a three-factor solution with loadings greater than .50 for all items. In addition, the results showed good adequacy of sampling and item correlations: KMO = 0.90, χ2 [136] = 3620.92, p < .001. The EFA results explained 67.14% of the variance. The three subscales, Praise, Complaint, and Request, demonstrated good reliability (.93, .88, and .76 respectively), and had acceptable lowest item-total correlations (.72, .61, and .28, respectively) and intraclass correlations for single measures (.69, .55, and .39, respectively).
A confirmatory factor analysis (CFA) was then performed to investigate the fit of the LS items suggested by the EFA results and using the same criteria for item inclusion. We used the lavaan package, version 0.6-9 (Rosseel, 2012) in R version 4.0.2 (R Core Team, 2020). We considered theoretically acceptable fits of three indices: comparative fit index (CFI) greater than or equal to 0.90, root mean square error of approximation (RMSEA) less than or equal to 0.10, and standardized root mean square residual (SRMR) less than 0.10 (Weston & Gore, 2006). The CFA for the three-factor solution indicated that the scale had an acceptable fit with the data: χ2[116] = 409.69, p < .001, CFI = 0.92, RMSEA = 0.09 [0.08, 0.10], SRMR = 0.09. All loadings ranged from 0.20 to 0.89 (see Table 1).
Construct validity
Correlations with theoretically related measures (convergent validity)
As predicted, the Complaint subscale was significantly and moderately correlated with the RSS Protest subscale (r = .48, p < .001) and the RCOPE Spiritual Discontent subscale (r = .55, p < .001). Also as predicted, the Request subscale was moderately and positively correlated with SAI Realistic Acceptance (r = .37, p < .001), the RCOPE Pleading for Direct Intercession (r = .51, p < .001) and Spiritual Connection (r = .44, p < .001) subscales, and the Praise subscale was moderately and positively correlated with RCOPE Spiritual Connection (r = .50, p < .001) and Active Religious Surrender subscales (r = .62, p < .001), and intrinsic religiosity (r = .45, p < .001), and negatively with Avoidant Attachment (r = –.54, p < .001).
Correlations with theoretically unrelated measures (discriminant validity)
As hypothesized, we found nonsignificant or small correlations across all subscales with extrinsic religiosity (r = .11, p = .02, r = .21, p < .001, and r = .30, p < .001, respectively), and extraversion (r = –.09, p = .05, r = .17, p < .001, and r = .30, p < .001, respectively).
Test–retest correlations
Test–retest correlations for all three subscales were good (Praise: n = 88, r = .68, p < .001; Complaint: n = 85, r = .58, p < .001; Request: n = 87, r = .51, p < .001).
Relations with well-being
Zero-order correlations
The hypothesized pattern was clearly evident with positive affect, negative affect, and satisfaction with life, but not with flourishing (see Table 2).
Correlations between All Variables from Study 3.
Subscale of RCOPE.
Subscale of SAI (Spiritual Assessment Inventory).
p < .05, **p < .01, ***p < .001.
Moderation analyses
A series of four moderated moderation analyses were conducted using Hayes’ (2013) PROCESS macro for SPSS Model 3. This tool estimates the regression coefficients and associated statistics for a moderated moderation model as well as the proportion of variance in the criterion variable uniquely attributable to the moderation effects. In each analysis, severity of the stressor was entered as the predictor, Complaint and Praise as the moderators, and the indicators of well-being (i.e., positive affect, negative affect, satisfaction with life, flourishing) as the criterion. To simplify the analyses, the Request component of lament was not included. The moderations were not significant for positive affect, negative affect, or satisfaction with life, but were significant for flourishing (see Table 3). Although the zero-order correlation between severity of the stressor and flourishing was not significant (r = –.08, p > .05), and the correlations of the LS subscales with flourishing were very low (see Table 2), the moderation analysis for flourishing indicated that the relation of severity of the stressor to flourishing is conditional on the interaction of Complaint and Praise, b = –2.30, t(314) = –2.16, p < .05 (see Figure 1). Specifically, the results indicate that when both elements of lament used in the moderation analyses (i.e., Complaint, Praise) are either high or low (i.e., low Complaint and Praise; high Complaint and Praise), the severity of the event is inversely related to flourishing. However, when either Complaint or Praise is high, the negative relationship between severity of the event and flourishing is ameliorated. In fact, in the case of high Complaint, the relation actually becomes slightly positive, with higher severity of the stressor related to higher flourishing.
Summary of Regressions Showing Complaint and Praise Moderating the Relationship between Severity of Stressor and Measures of Well-Being.
SE: standard error.
Interaction between severity of stressor, praise, and complaint in predicting flourishing.
Discussion
In this study, we confirmed the stability of the factor structure for the LS and provided evidence of both convergent and discriminant construct validity. Test–retest reliability was in the moderate range for Complaint and Request, and was stronger for Praise. These results suggest that the tendency to praise may be more stable than the other two, which may be more affected by situational variables.
While the cross-sectional nature of the data limits the ability to draw conclusions about the process of lament over time as well as its potential causal relationship to well-being, some conclusions can be drawn. The correlational results cannot indicate movement through the stages of lament, but they are theoretically consistent with a process in which greater well-being is achieved as lament moves from Complaint, through Request, and to Praise. The non-significance of flourishing in the correlational analysis is explained by its complex relationship to lament as revealed in the double moderation analyses. These effects showed that the relationship between the severity of the stressor and flourishing is dependent on the interaction between Complaint and Praise. Engaging in at least one of these, but not both at the same time, served a protective function, ameliorating the negative relationship between the severity of the stressor and flourishing. While, again, any conclusions about process must be made tentatively given the cross-sectional nature of data, these results are also consistent with a process model of lament. Further longitudinal research is needed to more fully explore the relationships between the components of lament and verify that they function as predicted by the theoretical model.
General Discussion
In this series of three studies, we drew attention to an emic Christian coping practice, lament, and developed it theoretically. We then developed a measure of lament and evaluated whether elements of this practice are related to well-being. The results of these studies support the validity of the scale. The factorial structure from the EFA held in the CFA with an independent sample. The pattern of correlations with constructs that were either conceptually related or conceptually unrelated provided strong support for the scale’s construct validity. In addition, the results of the cognitive interviews support the generalizability of the scale to diverse groups of Christians.
While the study of a process is not possible with cross-sectional data, the results of the third study are consistent with the conceptualization of lament as being composed of three components—complaint, request, and praise—that interact in predicting well-being. We note here, again, that in the context of these studies, the components of lament were studied as they naturally occur, rather than as a part of formal lament practice. Experimental research is necessary to evaluate whether the intentional practice of lament, including moving through the three components we have identified here, results in greater well-being.
Complaint was associated with lower well-being (i.e., positive and negative affect, satisfaction with life) in the zero-order correlations, but in the moderation analyses, complaint served a protective function in relationship to flourishing at lower levels of praise and higher levels of stressor severity. The role of complaint is complex and depends on other elements of lament, as might be predicted by our model. In the zero-order correlations, the negative relation with positive affect and satisfaction with life, and the positive relation with negative affect, are consistent with a model in which greater distress causes people to turn to God with more complaint. However, engaging in complaint to God appears to be better than not engaging. According to the theory of lament, the value of complaint lies in the honest and direct engagement of difficult emotions with God. Bringing these emotions to God may serve many of the interpersonal functions of affect modulation proposed by attachment theory (Granqvist & Kirkpatrick, 2013), which may be further accentuated by the concept of God as powerful and in control. Future research might investigate the influence of attachment style on the potentially beneficial effects of Complaint.
The Request aspect of lament did not emerge as a strong predictor of well-being in our analyses. With small correlations with well-being in the zero-order correlations, it did not seem to play a distinctive role. This is consistent with its place as a midway point in the process of lament. This finding may also be because petitions and requests are typically the most common form of prayer, and may be mechanistically engaged (Spilka & Ladd, 2013), making their effects less distinct. For example, prayers that are said before meals or before sleep are typically petitionary in nature, and may become routine. Research indicates that while use of petitionary prayer is generally associated with positive outcomes (Pössel et al., 2016), it may actually decrease hope in individuals who are spiritually unstable (Paine & Sandage, 2015). Our study did not control for individual differences in religiosity, and this may account for the modest findings for Request. Future longitudinal research is needed to see whether the Request aspect of lament serves a psychologically distinct role in coping or whether this subscale can be dropped from attempts to measure lament without loss of explanatory power.
Praise was consistently related to some indicators of well-being (positive and negative affect, satisfaction with life), although its relationship to flourishing may be more problematic when combined with higher levels of Complaint. It should be noted that our measure of praise specifies that this praise occurs in the context of suffering, which distinguishes it from other measures of worship and praise. Praise, with its focus on God and God’s attributes, may result in experiences of awe, which in its turn may facilitate gaining a larger perspective in which the self and its concerns—including, perhaps, its problems—are diminished (Piff et al., 2015).
In recent decades, the discipline of psychology has increasingly recognized the important role of cultural diversity in human functioning, with multiculturalism sometimes described as the fourth and newest force in the psychology literature (Pedersen, 1999). Religions are deeply intertwined with cultures; recognizing cultural diversity often involves recognizing religious diversity (Cohen, 2009). In the move to adapt interventions to specific cultural contexts, religions can serve as resources for identifying culturally relevant practices. Lament falls into this category. Our results are consistent with the theoretical conceptualization of lament as a meaning-making structure used by Christians when facing suffering and trauma, which, when practiced, can result in greater well-being.
The increased interest in lament in recent years in Christian circles, along with indicators that its practice is related to well-being, make it a promising target for future research. The practice of lament has been neglected in church contexts for decades, if not centuries. Recent attempts have been made to document this through, for example, studies of percentages of lament in hymnals (Pemberton, 2012, p. 39). The LS can be used to document how much practitioners of diverse denominations engage in the different components of lament, as well as documenting differences between various Christian traditions. While it is likely that Praise is prevalent across Christian denominations, it is also likely that there will be much more variability in Complaint. Future research might also document whether certain ways of engaging with religion (e.g., fundamentalism) or beliefs (e.g., prosperity gospel beliefs) can make engaging in lament less likely.
As the practice of lament becomes more prevalent in the general population, the consequences of its practice on psychological, health, and spiritual variables can be traced longitudinally. Lament also holds promise as a resource for religiously accommodative treatment (Worthington et al., 2011). Research is needed to determine whether the relation with well-being indicators is causal in nature, requiring randomized trial studies. As noted previously, laments targeted to specific suffering populations, such as survivors of sexual abuse, are proliferating. These and other church-based lament interventions can similarly be examined for their effectiveness in facilitating meaning-making in specific populations.
This set of studies has several limitations. Although our studies had some demographic diversity, all participants were from the United States and were primarily European American and (in Study 2) Protestant. This problem may have been exacerbated by the use of online data collection platforms, which may be biased on demographic indicators such as age, ethnicity, and gender (McCredie & Morey, 2019). Future research should include samples outside of WEIRD (Western, educated, industrialized, rich and democratic; Henrich et al., 2010) societies and in a greater breadth of cultural contexts. This diversity is particularly important as lament may be practiced more regularly in non-WEIRD contexts. Our studies were also cross-sectional. Consequently, while consistent with the causal theory we have proposed, longitudinal research establishing causal pathways is necessary to verify these cross-sectional findings.
Our intent in developing this measure was to make it broadly applicable to diverse Christian groups. While Christian beliefs formed part of the theological underpinnings of the measure and inform the way lament is practiced in Christian contexts, the resulting items themselves refer only to “God,” rather than referencing specific Christian concepts. Consequently, it is possible that the measure is also relevant to other monotheistic religions, particularly those in the Abrahamic traditions that hold sacred the portions of the Bible known to Christians as the Old Testament. A comparison of the relevant theological underpinnings of lament in those religious traditions and their relevance to the items in the LS is needed before assuming that the measure adequately measures the religious practice of lament as conceptualized in those traditions.
With respect to broad relevance to Christian groups, our cognitive interviews support the validity of the study across the major Christian groups (i.e., Protestant, Catholic, Eastern Orthodox), but Studies 2 and 3 did not have any Eastern Orthodox participants. Further research targeting specific Christian groups is needed to assess the validity of the scale in those populations.
In conclusion, the LS has the potential to contribute to our understanding of a Christian practice that has been used for millennia to cope with suffering and trauma and enhance well-being. The study of culturally relevant coping practices is particularly relevant considering the increase in suffering in the general population due to the ravages of COVID-19, and the need to identify and propagate means of addressing these in both professional and lay contexts. The scientific study of lament is necessary to document its effectiveness, to then implement it more intentionally in assisting people to cope within the contexts of their religious belief systems and communities. Our results suggest that the study of lament is a promising direction for contextually sensitive research and application.
Footnotes
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, MELH, upon reasonable request.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the John Templeton Foundation under Grant 61467 from to M. Elizabeth Lewis Hall and C. L. Park, co-PIs. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the John Templeton Foundation.
