Introduction

Mrdjenovich (2018) explored the practices of religious service attendance and prayer among atheists and agnostic theists through a thematic analysis of commentaries, perspective pieces, and news articles from the popular literature. Mrdjenovich set out to consider two fundamental questions: (1) “Why might non/irreligious individuals engage in religious/spiritual (R/S) practices?”, and (2) “Must individuals be theistic and/or gnostic in order to experience the health benefits of R/S involvement that have been reported in the literature?” The primary objective in terms of the first question was to develop items with adequate face validity that could be included and evaluated psychometrically as part of a questionnaire to assess the functional role of R/S involvement among individuals who are not religious and/or spiritual in the traditional sense. With respect to the second question, Mrdjenovich essentially arrives at the conclusion that the answer is “No”. More specifically, Mrdjenovich submits that “R/S somes” might benefit from R/S practices given underlying processes or influences of health that may be shared by nonbelievers and believers in God. Examples of such processes or influences based on the thematic analysis include ‘‘reflection and reorientation,’’ ‘‘social connectedness,’’ and ‘‘inner discipline.’’

Speed et al. (2020) offered a commentary regarding Mrdjenovich’s (2018) article with particular attention to studies that examine moderators of associations between R/S constructs and health outcomes. I reviewed the Speed et al. commentary, and I identified a number of concerns with their observations. In this paper, I describe those concerns and provide a response.

Qualitative Research Methodology

The first issue pertains to qualitative research methodology. Specifically, Speed et al. (2020) claim that the qualitative piece in Mrdjenovich (2018) was “incidental” and “questionably executed” on the grounds that (1) the sample consisted of 22 articles, “a mere” nine of which were written by atheists or were written from an atheist point of view, and (2) the search terms for the thematic analysis did not include “secular”, “humanist”, or “Sunday Assembly”. In this way, Speed et al. accuse Mrdjenovich of “fishing in a puddle” and “sweeping aside” R/S involvement in nontraditional contexts. Regarding these criticisms, it should be noted that sample size is not an indicator of credibility or rigor when it comes to qualitative analyses (Patton 1999). (In other words, the qualitative paradigm or tradition would advise that one need not cast his/her line in a sea in order to catch fish.) Further, despite the absence of the terms “secular” and “humanist” in the search, the secular humanist perspective is, in fact, represented via illustrative quotations provided in support of the thematic analysis. Mrdjenovich also refers to a theorized model of causal pathways by which secular humanism could influence health (Koenig et al. 2012). Moreover, Mrdjenovich explicitly acknowledges that (1) it is not possible to draw reliable generalizations from the articles included in his thematic analysis, (2) his findings on religious service attendance are delimited (which is not the same as “sweeping aside”) to theistic places of worship, and (3) “Researchers may wish to concentrate on service attendance in nontraditional contexts (e.g., Sunday assemblies [emphasis added] held at atheistic churches) for future analyses” (2018, p. 1503). Indeed, as Speed et al. recommend, ongoing studies are needed to determine the degree to which mechanisms of health that necessarily operate in R/S contexts are applicable in secular contexts. However, Mrdjenovich’s study was not about atheists and agnostic theists engaged in secular contexts; it was about atheists and agnostic theists engaged in traditional R/S contexts.

Religious/Spiritual Constructs

Secondly, Speed et al. (2020) maintain that their position (that there is nothing inherently healthy about R/S constructs) is a “null hypothesis based on the standard scientific principles of parsimony and incremental validity”. Yet, that position appears to have been presented in the literature as an assertive argument at times—not a null hypothesis. For instance, “Religious/Spiritual beliefs and behaviours do not have an inherent benefit, they have a benefit that is contingent on a valuation of those beliefs and behaviours” (Speed and Fowler 2017, p. 987). (Interestingly, such statements imply that R/S beliefs have a benefit, when, according to Speed et al., “... theistic belief has no health implications be it positive or negative”.) More broadly, while the principles of parsimony and incremental validity are generally well taken from a philosophy of science perspective, I would submit that simpler is not always better (Kowalski and Mrdjenovich 2017), and I would mention how difficult it can be to estimate incremental validity in practice (Hunsley and Meyer 2003).

Belief in God

Third, Speed et al. (2020) argue that Mrdjenovich (2018) “recognizes that belief in God has not itself been found to improve health outcomes, but then seemingly concludes that belief content must still play a role in health outcomes regardless”. Allow me to clarify. First of all, the recognition on Mrdjenovich’s part that “belief in God is something so fundamental and paramount to religious traditions” (2018, p. 1505) conveys exactly that; i.e., belief is fundamental and paramount to religious traditions. As Speed et al. imply, such a recognition should not be taken as a conclusion about the role of theistic belief in health. In point of fact, Mrdjenovich is explicit on that issue, “The predominant pattern in the religion-health literature—whereby R/S constructs have a positive rather than negative influence on various aspects of physical and mental health—is not entirely attributable to the domain of belief” (2018, p. 1505). Belief in general, however, demonstrably plays a role in health outcomes (Armitage and Connor 2000; Janz et al. 2002). For example, therapeutic interventions that match patients’ expectations and/or preferences are likely to result in improvements (Kowalski and Mrdjenovich 2013). As Speed et al. might agree, the specific content of belief may be of less consequence (Weber et al. 2012). Accordingly, Mrdjenovich states, “... the religious/spiritual, the non/irreligious, and atheists can all draw on their belief systems for ‘support, explanation, consolation, and inspiration’” (Mrdjenovich 2018, p. 1505, citing Wilkinson and Coleman 2010). I believe this statement is at least somewhat consistent with Galen’s (2018) observation that belief—whether religious or secular—is associated with mental health when the content matches an individual’s worldview.

Religion–Health Relationship

Fourth, Speed et al. (2020) assert various ways in which Mrdjenovich (2018) “misunderstands” the (non)religion–health relationship and the mechanisms involved therein. What Mrdjenovich (a.k.a. the author of the present paper) does understand is this:

  1. 1.

    It is important not to overgeneralize research findings concerning the health benefits of R/S involvement. There is evidence that the salutary effects of R/S constructs are not uniform across groups (Speed 2017; Speed and Fowler 2017).

  2. 2.

    Caution is necessary in the attribution of health effects to specific R/S attitudes and behaviors as opposed to underlying processes or influences such as worldview defense, shared identity, sense of belonging, social engagement, social support, coherence, congruence, and value, which could operate among the religious and the secular alike.

Where Mrdjenovich (2018) may have fallen short during his review of studies that examine associations between R/S constructs and health among the non/irreligious was in his attempt to capture multiple competing perspectives (e.g., the view among some researchers that certain mechanisms confer greater benefits for health when they operate in a R/S context) in a relatively brief discussion, combined with a tendency to equivocate in deference to diverse belief systems, thus leaving out a definitive conclusion in places (to the degree that it was possible to arrive at such a conclusion) or leaving ambiguous his own position on the issues, which may or may not be relevant. For what it is worth, Mrdjenovich’s (my) default position is that there are secular analogs to R/S constructs and common underlying mechanisms of health (e.g., Ai et al. 2004, 2017; Koenig et al. 2012).Footnote 1 This is the perspective from which Mrdjenovich argues that “some mechanisms probably are ‘good for all’”. Importantly, Mrdjenovich hypothesizes that “R/S-somes” might benefit less from mechanisms that operate in a R/S context merely because certain mechanisms may not be part of their R/S “repertoire”. Just as an example, R/S-somes might not utilize social support that is available by way of occasional religious service attendance if they do not feel welcome by other congregants at a traditional (theistic) place of worship. Of course, the same could be said of people who are R/S in the traditional sense and—regardless of an individual’s belief system—the same phenomenon could occur at an atheistic church. This is all to suggest that there may be a common underlying mechanism of social support that could operate (or not) among the religious and the secular alike, both in theistic and secular contexts.

Clearly, the particular subject matter of (non)religion and health is complex and—without careful attention and proper context—misunderstandings can easily arise. Opposing ideas can seem complementary, which is to say that two entirely different positions can be represented by the same statement. For example, if an author were to indicate that findings from health-related studies of religious individuals “do not extend to nonbelievers”, would it suggest that the author is saying that health benefits are not available to nonbelievers (i.e., that whatever R/S mechanisms were involved are superior to secular mechanisms)? Or, would it suggest that data gathered from religious individuals cannot yield information about nonbelievers (e.g., studies that involve large proportions of religious believers are not equipped to assess the efficacy or relative superiority of R/S constructs in general because nonbelievers’ data are being overpowered by the effect of the larger group)? Moreover, complementary ideas can seem different. As an illustration, consider the following statements: “... nonbelief is not unhealthy for nonbelievers” (Mrdjenovich 2018, p. 1505), and “... atheism... is not associated with any substantive health penalty” (Speed and Hwang 2017, p. 7).Footnote 2 Whereas some would read these statements as a tautology, Speed et al. (2020) characterize the former statement as a “specious” argument whereby Mrdjenovich “turns the scientific process on its head”. Perhaps Mrdjenovich does not misunderstand the (non)religion–health relationship as much as Speed et al. misunderstand Mrdjenovich (2018).

Convergence

This brings me to the fifth and final issue, which is that Speed and colleagues (2020) seemingly do not acknowledge instances where the material in Mrdjenovich (2018) converges with what appears to be their own thinking or objectives. For example, just as Speed et al. call attention to the diversity of atheists as a group, Mrdjenovich includes a section on the issue of heterogeneity among the nonreligious. Mrdjenovich also aims for precision and clarity when it comes to defining R/S belief, knowledge, identity/affiliation, and practice categories. Speed et al. go on to say that Mrdjenovich “dismisses the utility of survey data despite it showing promise in addressing why atheists engage in R/S activities” when, in reality, much of Speed’s own work cites substantial conceptual issues and practical challenges related to surveying atheists, not the least of which is that survey respondents who may be counted as “religious” might actually be atheists, which would suggest (to me) that survey data may, in fact, be of limited utility in some cases (Speed and Fowler 2016, p. 305; see also Hwang et al. 2011; Woodberry et al. 2012). Ultimately, Mrdjenovich recognizes that recent studies including those conducted by Speed and colleagues “have the potential to contribute to an understanding of the nuances involved in relationships between R/S constructs and health” (2018, p. 1504). Indeed, researchers are to be commended for disaggregating the R/S–health relationship across groups. When Mrdjenovich indicates that some of the relevant findings may seem obvious, he is actually quoting Speed’s own phrase “degree of obviousness” (Speed 2017, p. 254).

Astonishingly, Speed et al. (2020) maintain that Mrdjenovich’s (2018) description of current atheism–health research as ‘adversarial’ is “unjustified”, when—even in the aforementioned cases where Mrdjenovich agrees with them—Speed et al. seemingly insist on striking a polemical tone. But then, discourse is much the way of science, and perhaps Mrdjenovich fell short by reacting to what he (I) perceived as pejorative language toward the religious/spiritual. One example that comes to mind is Speed’s (2018) likening of religious communities to “chess clubs” in the context of social support. The truth is I cannot say definitively whether others would receive that analogy as being adversarial. I do wonder whether such statements have the potential to lessen Speed’s impact, however, in the sense that Speed and colleagues have called for a more inclusive or pluralistic approach in the (non)R/S–health research field. Ironically enough, one of the themes from Mrdjenovich’s (2018) analysis involved the prospect of ‘‘bridging the worlds of belief and nonbelief.’’ Granted, that is not what the Speed et al. commentary was about (appropriately, their focus appears to have been on science), but I have no doubt or disbelief that inclusivity and pluralism are achievable among authors and readers of this journal and their colleagues in the religion–health research field.

Epilogue: Moving Forward

How do we move forward? Speed et al. (2020) correctly note that a concerted effort is required to investigate religious minorities who are both fewer in numbers and extremely diverse, and their point on that issue is well taken. Although I wonder about the ethical implications of the intervention comparisons Speed et al. recommend whereby religious and nonreligious patients would be assigned to either R/S or secular treatments (see Nielsen 2014), I concur that “future research would best proceed by sampling relatively equal numbers of religious and nonreligious participants and comparing them on levels of engagement in religious and secular activities” (Speed et al.).