This chapter addresses approaches to the measurement of religious concepts for nursing theory and practice based on individuals’ self-reports. Examples of religious concepts encountered in health research include religious affiliation, religious attendance (participation in religious services or activities), religious orientation, private religiousness, religious coping, and religious beliefs, values, and experiences.1 Nurses have used established instruments and developed new ones for the measurement of religious concepts to address questions such as: To what extent do individuals draw on their religion to cope with health challenges? What is the impact of religious beliefs, experiences, and practices on health? What are the implications of individual differences in religious beliefs, experiences, and practices for nursing care? The results of studies that answer these questions vary and must be interpreted in light of the characteristics of the measurement instruments that were used and the populations and purposes for which they were developed. It is therefore important for nurses who draw on these studies to be knowledgeable about the processes and assumptions underlying the measurement validation of religious concepts and the corresponding inferences that may be warranted, which is the focus of this chapter.
A fundamental premise for the measurement of religious concepts is that, although most are inherently latent (i.e., they are not directly observable and hence not directly measurable), their measurements can be inferred from related observations, such as individuals’ reports of their religious beliefs, experiences, and practices.2,3 This is not unique to religious concepts; there are numerous other latent concepts that are not directly observable, such as depression, anxiety, coping, and health. The implication is that the measurement of such concepts requires a theory that relates the concept of interest to the observations from which measurements of the concept are inferred. That is, the measurement of latent concepts requires a theory about the relationship between the conceptual (theoretical) and operational (observable) domains.4 Here, the conceptual domain refers to the latent concepts and the relationships among them that are not directly observable but inferred from related observations. The operational domain refers to observations that, in the context of self-report measurements, consist of individuals’ responses to questions or statements (e.g., items of a measurement instrument or questionnaire) that function as measurement indicators of the latent concept. For example, it has been suggested that a measure of religious well-being (a latent concept) can be inferred from individuals’ responses to questions about their relation with “God.” This is the case in the Spiritual Well-Being Scale that includes 10 questions for the measurement of religious well-being (e.g., “I feel most fulfilled when I am in close communion with God” and “My relation with God contributes to my sense of well-being”) that are rated on a six-point scale ranging from strongly agree to strongly disagree.5 The theory underlying the measurement of a latent concept should provide an explanation of how the measurement indicators represent (or reflect) the concept and the mechanisms by which they produce a measure (i.e., a scaled score) of the degree to which a concept is manifest. The resulting measure is viewed as an “empirical analog” of the concept of interest.6
These considerations are foundational to the validation of self-report measures. Within the field of educational testing, Messick7 offers a comprehensive and influential perspective of measurement validity that he defines as “an overall evaluative judgment of the degree to which empirical evidence and theoretical rationales support the adequacy and appropriateness of interpretations and actions on the basis of test scores or other models of assessment.” In this sense, measurement validation is viewed as a theory-laden activity that involves: (a) the use of empirical approaches (statistical as well as qualitative approaches) to investigate the meaning or interpretation of a measured concept and (b) theoretical and pragmatic considerations about the utility of the measurement in terms of its intended purpose(s) and any related social and ethical consequences.8–11 That is, measurement is not merely a mechanical procedure; it is an inferential process “by which an attempt is made to understand the nature of a variable [or concept].”12 In the context of our discussion, the measurement validation of religious concepts, therefore, refers to the analytical processes (theoretical and empirical) through which we come to understand concepts that pertain to religion, religiosity, religiousness, religious affiliation, and religious beliefs, experiences, and practices. This includes the justification of inferences pertaining to the measurement of these concepts and their relationships with other concepts of relevance to nursing’s theoretical and clinical purposes.
These perspectives of measurement validation have important implications for the measurement of religious concepts. At the operational level, a religious concept is understood through the question(s) and statement(s) (items) from which a measure of the concept is inferred. This necessitates an examination of individuals’ interpretations of the items and whether these are congruent with those of other individuals and the concept that is being measured. The operational level also pertains to the processes by which responses to the items (measurement indicators) are used to produce a measure (scaled score) that is reliable and valid with respect to the intended purpose(s) for which the measure is to be used. At the conceptual level, the measurement of a religious concept (i.e., the interpretation of the measurement scores) requires an in-depth understanding of its relationships with other relevant concepts. This has traditionally been called the “nomological network” 13,14 of a concept and includes knowledge about its antecedents (i.e., concepts that contribute to a change in the religious concept) and consequences (i.e., outcomes or concepts that are affected by the religious concept). An implication of a philosophical nature is that the measurement of a religious concept inevitably reflects a frame of reference by which the relationships between the religious concept and the measurement indicators are understood. This frame of reference includes theoretical, philosophical, theological, religious, or spiritual understandings that inform how the concept is measured (e.g., the wording of questions and statements) and how the measure is to be used. In summary, the measurement of a religious concept requires an in-depth examination of the following: (a) the operational characteristics that define how the concept is measured (i.e., including its measurement indicators and the operations by which a measure of the religious concept is obtained), (b) the conceptual relationships of the religious concept with other concepts of relevance to nursing, and (c) the philosophical underpinnings (frame of reference) by which the religious concept is understood and corresponding measurement instruments are constructed. Examples of considerations pertaining to each of these premises are provided in Table 18.1.
In the following sections, we discuss the above premises with the intent of explicating some of the analytical processes by which nurses have pursued the measurement of religious concepts. Our intention is not to provide a taxonomy or evaluation of particular measurement instruments; extensive reviews of instruments for the measurement of religious concepts have already been provided by other authors.15–18 Nor do we strive to critique particular studies or provide an in-depth account of the analyses and results. Rather, we extracted several examples from published studies to elucidate the types of questions and considerations that guide the interpretation of scores derived from individuals’ responses to questions and statements that have been used for the measurement of religious concepts. We have limited our selection of studies to those that have been conducted by nurses or for nurses and that serve to exemplify the measurement of selected religious concepts in terms of their: (a) operational characteristics, (b) conceptual relationships, and (c) philosophical underpinnings.
TABLE 18.1 Considerations for the Measurement of Religious Concepts |
The operational domain is about understanding how and why individuals respond the way they do to questions and statements used for the measurement of a religious concept. To what extent is the religious concept adequately represented by the questions and statements that are used to measure it? How are responses to multiple questions and statements combined to produce a measure of the religious concept? Do all individuals within the intended target population interpret each of the questions and statements in the same way? The answers to these questions necessitate careful examination of the wording of the items and the statistical models that are applied to examine individuals’ responses to the items. We draw on examples from published studies to illustrate several approaches to answering such questions and highlight particular challenges.
The first example is a study by Ng, Fong, Tsui, Au-Yeung, and Law who examine the validation of a translated Chinese version of the 16-item Daily Spiritual Experience Scale (DES) in a sample of 245 Chinese individuals living in Hong Kong.19 The original English version of the DES was developed by Underwood for use in health studies as a measure of “ordinary experience of the transcendent or sense of the divine” that crosses the boundaries of particular religions or spiritual groups.20 The original instrument was based on in-depth interviews and focus groups with people from various religious backgrounds. Ng et al. set out to examine whether the DES could be translated and meaningfully applied to a different cultural context, namely that of Chinese people living in Hong Kong. The translation process, which involved a team of Chinese mental health practitioners and the use of forward and backward translation techniques, revealed ambiguity regarding the Chinese translation of the term “God” that was used in several of the questions. Specifically, the researchers found that “God” in Chinese could refer to a “humanized God” (as represented in some of the religious rituals of Confucianism, Buddhism, and Daoism) or to a philosophical higher power as reflected in discourses about spiritual transcendence. Ng et al. therefore added an explanatory statement to the measurement instrument to ensure that the term “God” was interpreted in the manner that was consistent with its intended meaning.
Translation challenges, such as the one described by Ng et al.,19 are not surprising considering the central of role of language in the social construction of religious concepts. Similar challenges could arise in situations where questions and statements for the measurement of a religious concept might have different meanings for people with different religious or cultural backgrounds or life experiences. For example, Dunn and Horgas explore gender differences in religious coping styles related to the experience of pain in older adults (N = 200) as one of their study objectives.21 The authors used an adaptation of Pargament’s Religious Problem-Solving Scales (RPSS) that consists of 18 items for the measurement of three religious coping styles (self-directed, deferring, and collaborating) pertaining to an individual’s relationship with God.22 For instance, the item “When it comes to deciding how to manage my pain, God and I work together as partners,” is a measurement indicator of the collaborative religious coping style, which Pargament interprets as reflective of an active partnership between the individual and God that would be particularly representative of some Jewish and Christian traditions.22 Dunn and Horgas observed that the mean score for the collaborative religious coping style, as well as the other religious coping styles, was greater in women than in men. It appears that women use religious coping strategies more frequently than men for the management of their pain. Although there are many possible explanations for this observation, a question of particular relevance to our discussion is whether these findings point toward actual differences in the religious coping styles or whether they are an artifact of the way in which the religious coping styles were measured. The measurement approach that was taken assumes that the measurement indicators represent particular religious coping styles in ways that are equivalent for women and men (i.e., the measurement approach is equivalently applicable to women and men). Although the women had relatively higher average scores for the combined measurement indicators for each coping style, it is not known whether this difference is the same for all of the 18 measurement indicators individually. That is, do the religious coping styles have the same meaning for women and men, and are they manifested (via their measurement indicators) in the same way in both groups?
Similar questions apply when measures of religious concepts are used to compare other groups, such as religious or cultural groups, or when they are applied to a population that is different from the one for which the measurement instrument was developed. For instance, the above example item suggests that the RPSS is specifically representative of monotheistic religious beliefs. The application and use of this instrument for the measurement of religious coping styles in people who have other religious beliefs is likely unwarranted. Like the RPSS, many instruments for the measurement of religious concepts have been developed within particular, often Christian dominant, populations. In such situations, the concept may not be measured in a way that is congruent with other religious, spiritual, or cultural traditions. Researchers who use measures of religious concepts must therefore consider whether the instrument that was used is congruent with the characteristics of the population to which it was applied.
In addition to paying attention to the wording of items used for the measurement of a religious concept, it is important to examine the statistical relationships between the measured concept and its measurement indicators. This is particularly important when multiple items are used as measurement indicators of the same concept. The construction of a scale for the measurement of a religious concept requires that a measurement model of the relationships between the measurement indicators and the measured concept to be established and validated. One of the most common approaches is to use factor analysis techniques to determine whether the measurement indicators are reflective of the same concept.23 The premise of factor analysis is that the measurement indicators and the correlations among them arise from a common source, called a latent factor (i.e., a measure of the concept of interest), plus some degree of error. Specifically, factor analysis addresses the following questions about the relationships between the measurement indicators and the measured concept: (a) Do the measurement indicators reflect a common concept (factor)? and (b) Do the measurement indicators that reflect the same concept do so to the same extent?
An example of a factor analysis is found in a study by Lim and Yi24 who examine the validity of the six-item Spiritual Well-Being subscale of the Quality of Life-Cancer Survivor instrument,25 which they translated into Korean. For the original English translation of the instrument, the responses to the six items could purportedly be combined to produce a measure of spiritual well-being. Lim and Yi examine whether this approach of combining the six items into a single measure was warranted in a sample of breast and gynecological cancer survivors among Koreans living in Korea (N = 110) and in the United States of America (N = 51). The results of their exploratory factor analysis suggested that the six items (measurement indicators) should not be combined in a single measure of spiritual well-being because they reflect two (rather than one) latent factors, which the authors labeled religiosity and spirituality. The religiosity factor was represented by the following three items: “Importance of participation in religious activity,” “Religious or spiritual life change as a result of the cancer diagnosis,” and “Importance of other spiritual activity (meditation).” The other three items for the measurement of spirituality were: “Positive change in life because of illness,” “Sense of purpose or mission for life or a season for being alive,” and “Feeling about uncertainty about the future.” The results also revealed that, although five of the items had relatively strong associations with either the religiosity or the spirituality factor (with factor loadings ranging from 0.80 to 0.92), one of the items for the measurement of spirituality, “Feeling about uncertainty about the future,” had a factor loading of only 0.43. Statistically, this means that this item was a relatively weak measurement indicator of spirituality in this sample. That is, there were sources, other than the latent factor, that were unaccounted for and that influenced the individuals’ responses to this item. These may have included random sources, due to sampling variability, and systematic sources such as cultural or spiritual beliefs that may have influenced individuals’ interpretations and responses to the item.26
Another statistical approach has been demonstrated in a study by Gielen, van den Branden van Iersel, and Broeckaert who use latent class analysis to classify Flemish palliative care nurses based on their responses to a variety of questions about religious or worldview perspectives, affiliation, and various beliefs and practices.27 These researchers found that the sample could be divided into the following five groups (called latent classes): atheists/agnostics, doubters, church-going respondents, religious but not church-going respondents, and devout church-going respondents. What is of interest to the discussion here is that the measurement indicators were used to group respondents in the sample and to characterize those groups in terms of common patterns of responses to the questions. This is different from the previously discussed factor analysis approach, where measurement indicators were used to obtain scaled measures of religiosity and spirituality in the overall sample. That is, the purpose of latent class analysis is to identify and characterize different sampling groups that are not known a priori (it is not known a priori how many groups exist in the sample and to which group an individual belongs). The theoretical premise is that individuals’ responses to the items are reflective of group membership (represented as latent classes). In contrast, the factor analysis approach that was used in the study by Ng et al. seeks to produce a measure (scaled score) of one or more latent factors that is applicable to the overall sample. The theoretical premise of factor analysis is that individuals’ responses to the items are determined by their score on the latent factor (i.e., the extent to which the concept is present). The important point for the discussion here is that both of these approaches can be used to examine various explanations for individuals’ responses to items, which is what measurement is all about.
These statistical approaches illustrate how items could be used to produce a measure of a religious concept (represented as a latent factor) or to identify subgroups (represented as latent classes) within a sample. Whatever approach is taken, a fundamental premise of all self-report measures is that individuals in the sample are consistent in their interpretations of the questions or statements used to measure the concept of interest. Psychometric approaches, such as factor analysis, have been extensively used to examine the measurement validation of religious concepts. However, the possibility of heterogeneity with respect to the individuals’ interpretations of the questions or statements (items) used to measure a religious concept has been much less extensively examined. It is often unclear to what extent people from different backgrounds and with different life experiences and beliefs may variously interpret and respond to questions and statements that are used as measurement indicators. Considering the different populations in which religious concepts are often measured, it is plausible that people may not interpret all of the indicators in the same way (e.g., as may result from the use of the term “God” in measurement instruments such as the DES). Sophisticated statistical techniques, such as differential item functioning and latent variable mixture modeling, should be more extensively used to examine sources of heterogeneity that may help to explain why individuals interpret and respond to questions and statement in different ways.28,29 Qualitative analysis techniques, such as cognitive interviewing, can provide additional valuable insights for understanding potential differences in individuals’ interpretations of the items.30–32 Regardless of the techniques that are used, compelling evidence must be provided to ensure that the religious concept of interest is accurately reflected in the items that are used to measure it, and that individuals are consistent in their interpretations of the items such that their responses can be meaningfully evaluated and compared.
Many studies have examined the associations of religious concepts with various other health-related concepts.33,34 Of interest, here are the processes by which the meaning of a religious concept is understood through its relationships with other relevant concepts, which include antecedents that predict or explain the religious concept and consequences that are affected by the religious concept. Related questions include: To what extent is the religious concept associated with other similar or dissimilar concepts? What are possible explanations for individual differences (variability) in the religious concept? To what extent does the religious concept influence outcomes or other concepts of relevance to nursing? A relatively common approach is to examine measures of religious concepts in terms of their correlations with other measures of the same concept and with measures of different concepts (this corresponds to the examination of convergent and discriminant validity).4 This approach contributes to an understanding of the measured concept in terms of the amount of variance it shares with other measured concepts (or the extent to which the concepts overlap). However, it is important to remember that correlations may be due to the influences of other potentially confounding extraneous variables. Another approach is to examine the extent to which a religious concept uniquely explains (or accounts for) the variance in another concept (often a particular outcome of interest), in the context of other explanatory variables. Conversely, one can examine the extent to which other concepts explain variance in a religious concept (specified as the response variable). This “unique variance approach” is used to adjust for the potentially confounding effects of extraneous variables when examining the association between the religious concept and another concept.35 A limitation of the unique variance approach is that it does not consider the relationships among the explanatory variables. A third approach is to use structural equation modeling, or path modeling, to further examine the associations among the explanatory variables for the purpose of revealing a network of relationships between the religious concept and other concepts of interest. This approach could, for example, be used to reveal hypothesized mechanisms (or paths) by which a religious concept may be associated with an outcome variable through its associations with other (mediating) variables. Whatever the approach taken, the concern of interest to the discussion here is that a religious concept must be understood in terms of its associations with other concepts within a purposefully constructed theoretical framework so as to avoid erroneous conclusions about its meaning and theoretical relevance. We discuss three examples to illustrate these approaches.
The first example is a study by Rohani, Khanjari, Abedi, Oskouie, and Langius-Eklöf who develope and evaluated Persian translations of established English instruments for the measurement of health, sense of coherence, religious coping, and spiritual perspective.36 The authors specifically explored the validation of the translated instruments for the measurement of these concepts by examining the associations among them. Translation was facilitated by the use of recommended procedures, including forward and backward translation, expert reviews, and pilot testing. Religious coping was measured using the translated Brief Religious Coping instrument that consists of 14 items for the measurement of the following two forms of religious coping: (a) “positive religious coping” (seven items; high scores indicate positive religious coping), defined as “an expression of a sense of spirituality, a secure relationship with God, a belief that there is meaning to be found in life, and a sense of spiritual connectedness with others” and (b) “negative religious coping” (seven items; high scores indicate negative religious coping), which purportedly reflects “a less secure relationship with God, a tenuous and ominous view of the world, and a religious struggle in search for significance.” 37 The reliability of the translated instrument was supported by the examination of internal consistency among the items (Crohnbach’s alphas ≥ 0.87 and 0.76 for the positive and negative religious coping scales, respectively) and test–retest reliability estimates (intra-class correlations of 0.80 and 0.74, respectively). The authors used similar procedures to translate the Spiritual Perspective Scale that consists of 10 items addressing a variety of religious and spiritual behaviors and beliefs with higher scores purportedly indicating greater integration of spirituality in one’s life.38
The following discussion focuses on the results pertaining to the correlations between spiritual perspective and the religious coping variables.39 The authors hypothesized that positive religious coping would be positively associated with spiritual perspective, which was supported by the findings. However, the hypothesized negative association between negative religious coping and spiritual perspective was not supported; the correlation was found not to be statistically significant. What are some possible explanations for these findings? One possibility suggested by the authors is that spiritual perspective and negative religious coping may indeed be different and unassociated concepts in this particular sample. This explanation raises questions about the conceptualizations of spiritual perspective and religious coping. It is also possible that the meanings of the measured concepts based on the translated instruments are not the same as those of the English versions. Another possibility is that the results were obtained due to chance or random error. It is always important to remember that inaccurate results from individual studies may have been obtained as a result of sampling variability. Further investigation is needed to determine why the hypothesized association between spiritual perspective and negative religious coping was not observed in this sample. Nonetheless, this study exemplifies that the measurement of religious concepts is a theory-laden activity, where the religious concepts are understood in terms of their associations with other concepts.
Another commonly used approach is to evaluate the extent to which the variance in an outcome of interest is uniquely explained by a religious concept in the context of other potentially relevant explanatory variables. For example, Newlin, Melkus, Tappen, Chyun, and Koenig examine the relevance of religious and spiritual factors in relation to glycemic control in a sample of 109 “Black” women in the United States of America suffering from diabetes (the term “Black” was used to refer to a group of African American women).40 The Spiritual Well-Being Scale was used for the measurement of two factors: religious well-being (10 items) and existential well-being (10 items) (higher scores indicate greater well-being).41 Hemoglobin A1c (HgA1c) was used to assess glycemic control. The authors observed that a relative increase in religious well-being was associated with higher levels of HgA1c, whereas a relative increase in existential well-being was associated with lower levels. What are some possible explanations for these findings? Is it likely that religious and spiritual well-being are predictive of HgA1c? Or can these associations be explained by other variables that were not examined in this study? In correlational analyses, it is important to remember that a relationship between two variables can occur for several reasons. It is possible that the association is due to the impact of confounding variables that were not considered in this study or that the relationships are mediated by other explanatory variables of HgA1c. Another possibility is that the direction of the relationship was misspecified; the positive association between religious well-being and HgA1c may reflect a tendency to rely more intensely on the religious beliefs and practices that are used as indicators for the measurement of religious well-being. The authors used regression modeling to examine several possible explanations, including the possibility that the associations may be mediated by psychosocial factors, such as emotional distress and social support. However, compelling support for these explanations was not achieved. Nonetheless, the study raises important questions regarding the meaning and interpretation of religious and existential well-being scores with respect to a particular clinical outcome. In this case, although a correlation between religious well-being and a clinical outcome was observed, theoretical mechanisms underlying this correlation were not revealed.
There are other studies that have more explicitly examined the mechanisms by which a religious concept may contribute to outcomes of relevance to nursing. For example, Musgrave and McFarlane combine results from several regression models to determine whether the potential effects of intrinsic and extrinsic religiosity on attitudes toward spiritual care in a sample of Israeli oncology nurses was mediated by their spiritual well-being while taking the antecedents of gender, ethnicity, and education into account.42 Here, intrinsic and extrinsic religiosity referred to the motivation underlying religiosity, where the former (internal) pertained to the use of religiosity for the sake of one’s own faith, and the latter (extrinsic) to a “utilitarian approach to religious beliefs.” The researchers used translated versions of the Revised Age University I/E Scale43. and the Spiritual Well-Being Scale34. for the measurement of intrinsic and extrinsic religiosity and spiritual well-being (higher scores indicate greater religiosity or spiritual well-being). They used the Spiritual Care Perspective Survey44. to measure nurses’ attitudes to spiritual care, with higher scores indicating greater regard for spiritual care in nursing. The results provided support for the authors’ propositions that intrinsic and extrinsic religiosity could be viewed as antecedents of nurses’ attitudes to spiritual care and that these relationships were partially mediated by nurses’ reports of their spiritual well-being. Intrinsic and extrinsic religiosity was positively associated with nurses’ reports of their spiritual well-being, which, in turn, was positively associated with their attitudes to spiritual care. However, the negative total effect of extrinsic religiosity on attitudes toward spiritual care suggests that Israeli nurses with higher extrinsic religiosity had relatively poorer attitudes to spiritual care. Although there are many possible explanations for these findings that provide ground for further research, the relevance of this study to the discussion is that the study addresses a hypothesized mechanism that further facilitates the interpretation of religious concepts in terms of their associations with potential antecedents and consequences.
There are many other descriptive studies that have similarly examined the associations between religious concepts and other concepts of relevance to nursing. Some studies contribute valuable insights and most provide ground for new questions about the conceptualization and measurement of religious concepts and their potential value to nursing theory and practice. A general caution applies to all these studies: considering that alternative explanations resulting from measurement bias, confounding, and sampling variability exist for any individual study, some degree of skepticism is called for, at least until an association appears consistently in multiple studies involving samples from different populations.45. The meaningful interpretation of scores pertaining to the measurement of a religious concept necessitates that the measure is examined in different samples and in relation to various other relevant concepts such that consistencies and inconsistencies in the results across several studies are understood.
The philosophical underpinnings refer to the theories, values, and beliefs that comprise the frame of reference underlying the conceptualization and measurement of a religious concept. Related questions include: What are the theoretical and philosophical understandings from which the instrument was constructed? What are the intended and unintended social and ethical consequences of measuring a religious concept in a particular way? These questions are particularly relevant when individuals from different cultural or religious backgrounds or with different life experiences might not be equivalently represented by the questions and statements from which a measure of a religious concept is inferred (e.g., questions using terms such as God, sin, and forgiveness may be relevant for some religious traditions but not others). A related unintended consequence is that particular religious or cultural groups might be essentialized of the use of predefined, and sometimes stereotypical, measurement indicators. Conversely, the existence of purportedly universal measures of religious concepts has been questioned by several authors who suggest that religious concepts should be measured in relation to specific theological or religious traditions.46,47
Hamilton, Crandell, Carter, and Lynn develope the Perceived Support from God Scale with the specific intention of addressing these types of concerns with respect to the shared religious and cultural backgrounds in a sample of Christian African American cancer survivors.48. The authors argue that the spiritual and religious perspectives of Christian African Americans have been uniquely shaped by the shared historical and contemporary contexts of racial oppression. Consequently, their religiosity may not be congruent with many of the purportedly universal measures of religious and spiritual concepts used in health research. To address this concern, Hamilton, Powe, Pollard, Lee, and Felton set out to construct a new measurement instrument that specifically reflects the religious and spiritual experiences and beliefs of Christian African Americans. The resulting instrument consists of 15 items that were based on a qualitative study of 28 Christian African American cancer survivors.49. A factor analysis suggests that these items reflect two factors, one measuring “support from God” (nine items) and the other measuring “God’s purpose for me” (six items). The authors conclude that the instrument provides valid measurements of these religious concepts for Christian African American cancer survivors. However, despite the authors’ laudable efforts to represent accurately the spiritual and religious perspectives of Christian African Americans, there remain potentially unintended consequences that deserve consideration. For example, there is a risk of essentializing the spiritual and religious perspectives of Christian African Americans through the use of potentially stereotypical measurement indicators such as “God allows to suffer” and “Illness made me a better person.” The use of these indicators implies that the Christian African Americans’ perceived support from God would, in part, be defined by their affirmative responses to statements about the extent to which God allows suffering and the extent to which illness made them a better person. Is this necessarily the case for all Christian African American cancer survivors? What does it mean if patients do not respond affirmatively to these indicators? What would nurses do with this information? These types of questions apply to all measures of religious concepts and point to the need for careful consideration of the social and ethical consequences of measuring religious concepts in a particular way.
Examination of the operational characteristics, conceptual relationships, and philosophical underpinnings pertaining to the measurement of religious concepts often reveals challenges and questions about the interpretation of the measurement scores and their relevance to nursing. At times, it is unclear what the measures actually mean and why they are relevant to nursing. Further more, empirical research, both quantitative and qualitative, is needed to understand the “why and how” of individuals’ responses to questions and statements about their religiosity. Are the items and questions used for the measurement of religious concepts interpreted in the same way by all people? We must also understand how the measures of religious concepts relate to other concepts of relevance to nursing. Are the measurement scores relevant with respect to particular nursing outcomes and clinical decisions? Empirical research and philosophical inquiry are needed to evaluate the social and ethical consequences pertaining to the use of religious measures for both theoretical and clinical purposes. What are the implications of using individuals’ selfreports to classify them into predefined religious groups or to determine the extent to which a religious concept is manifest? What are the social and ethical consequences of any inadequately represented religious traditions in the measurement of religion for theoretical and clinical purposes? In clinical practice, are patients who have provided information about their religiosity treated differently from patients who have not provided such information, and does this bring about differences in the outcomes and quality of nursing care? Does knowledge about individuals’ selfreported religiosity contribute to improved decision making, or does it possibly lead to erroneous assumptions and preconceived notions about individuals’ preferences and beliefs? Considering the complexity involved in answering these questions, we recommend a cautionary stance to the measurement of religious concepts to avoid that individuals or groups with particular religious or cultural histories are systematically disadvantaged or advantaged by means of the instruments used for the measurement of religious concepts.
In conclusion, we offer the following considerations that we hope will facilitate the meaningful interpretation and use of measures of religious concepts for nursing’s theoretical and clinical purposes:
A measure of a religious concept must be interpreted with respect to the actual wording of the questions and statements that serve as its measurement indicators. Considering the diversity in conceptual and operational definitions of many religious concepts, there is a significant risk that the results of studies will be misunderstood unless the operational characteristics of instruments for the measurements of religious concepts are explicated and considered in the interpretation of study results.
A measurement model must be specified and examined to determine whether and how various measurement indicators can be combined to produce a measure that is reliable and valid for its intended purposes. This may include factor analyses, analyses of internal consistency among the measurement indicators (Cronbach’s alpha), and other analyses of measurement reliability. These analyses must be performed within the sample that is a representative of the population of interest, and the results must be compared with those of other studies conducted in other samples. There is a significant risk for incorrect inferences based on measurement scores when the statistical properties of the measurement instrument are inadequately examined or poorly understood.
Understanding a religious concept requires that its associations with other relevant concepts are examined and understood. This necessitates examination of potential antecedents and consequences of the religious concept, and the comparison of results of such analyses across multiple studies.
The purposes and contexts to which a measure of a religious concept is applied must be congruent with its philosophical, theoretical, religious, and cultural underpinnings. Nurses who use an instrument for the measurement of a religious concept must be familiar with its frame of reference as the basis for evaluating its appropriateness for particular theoretical and clinical purposes and populations.
The use of an instrument for the measurement of a religious concept must be evaluated in consideration of any intended and potentially unintended social and ethical consequences. Particular attention must be paid to individuals or groups who may be systematically disadvantaged, or advantaged, through the use of a particular measurement instrument.
1. Hall, Daniel, Keith Meador, and Harold Koenig. “Measuring Religiousness in Health Research: Review and Critique.” Journal of Religion and Health 47.2 (2008): 134–63.
2. Berry, Devon. “Methodological Pitfalls in the Study of Religiosity and Spirituality.” Western Journal of Nursing 27.5 (2005): 628–47.
3. Miller, William, and Carl Thoresen. “Spirituality, Religion, and Health. An Emerging Research Field.” American Psychologist 58.1 (2003): 24–35.
4. Viswanathan, Madhu. Measurement Error and Research Design. Thousand Oaks, CA: Sage, 2005.
5. Paloutzian, Raymond, and Craig Ellison. “Loneliness, Spiritual Well-Being and Quality of Life.” Loneliness: A Sourcebook of Current Theory, Research and Therapy. Ed. Lititia Peplau and Daniel Perlman. New York: John Wiley, 1982. 224–37.
6. Stine, William. “Meaningful Inference: The Role of Measurement in Statistics.” Psychological Bulletin 105.1 (1989): 147–55.
7. Messick, Samuel. “Validity of Psychological Assessment: Validation of Inferences from Persons’ Responses and Performances as Scientific Inquiry into Score Meaning.” American Psychologist 50.9 (1995): 741–49.
8. Zumbo, Bruno. “Validity as Contextualized and Pragmatic Explanation, and Its Implications for Validation Practice.” The Concept of Validity: Revisions, New Directions and Applications. Ed. Robert Lissitz. Charlotte, NC: Information Age Publishing, 2009. 65–82.
9. Messick, Samuel. “Test Validity: A Matter of Consequence.” Social Indicators Research 45.1–3 (1998): 35–44.
10. Zumbo, Bruno. “Validity: Foundational Issues and Statistical Methodology.” Handbook of Statistics. Ed. C. Radhakrishna Rao and Sandip Sinharay. Vol. 26. Psychometrics, Amsterdam: Elsevier Science, 2007. 45–79.
11. Hubley, Anita, and Bruno Zumbo. “A Dialectic on Validity: Where We have Beenand Where We are Going.” Journal of General Psychology 123.3 (1996): 207–15.
12. De Ayala, Rafael. The Theory and Practice of Item Response Theory. Methodology in the Social Sciences. New York: Guilford Press, 2009.
13. Cronbach, Lee, and Paul Meehl. “Construct Validity in Psychological Tests.” Psychological Bulletin 52.4 (1955): 281–302.
14. Hubley, op. cit.
15. Hall, Daniel, Keith Meador, and Harold Koenig. “Measuring Religiousness in Health Research: Review and Critique.” Journal of Religion and Health 47.2 (2008): 134–63.
16. Kilpatrick, Shelley, Andrew Weaver, and Michael McCullough, et al. “A Review of Spiritual and Religious Measures in Nursing Research Journals: 1995–1999.” Journal of Religion and Health 44.1 (2005): 55–66.
17. Egbert, Nichole, Jacqueline Mickley, and Harriet Coeling. “A Review and Application of Social Scientific Measures of Religiosity and Spirituality: Assessing a Missing Component in Health Communication Research.” Health Communication 16.1 (2004): 7–27.
18. Hill, Peter, and Ralph Hood. Measures of Religiosity. Birmingham, AL: Religious Education Press, 1999.
19. Ng, Siu-Man, Ted Fong, Elaine Tsui, et al. “Validation of the Chinese Version of Underwood’s Daily Spiritual Experience Scale: Transcending Cultural Boundaries?” International Journal of Behavioral Medicine 16.2 (2009): 91–7.
20. Underwood, Lynn, and Jeanne Teresi. “The Daily Spiritual Experience Scale: Development, Theoretical Description, Reliability, Exploratory Factor Analysis, and Preliminary Construct Validity Using Health Related Data.” Annals of BehavioralMedicine 24.1 (2002): 22–33.
21. Dunn, Karen, and Ann Horgas. “Religious and Nonreligious Coping in Older Adults Experiencing Chronic Pain.” Pain Management Nursing 5.1 (2004): 19–28.
22. Pargament, Kenneth, Joseph Kennell, William Hathaway, et al. “Religion and the Problem Solving Process: Three Styles of Coping.” Journal for the Scientific Studyof Religion 27.1 (1988): 90–104.
23. Gorsuch, Richard. Factor Analysis. 2nd ed. Hillsdale, NJ: L. Erlbaum Associates, 1983.
24. Lim, Jungwon, and Jaehee Yi. “The Effects of Religiosity, Spirituality, and Social Support on Quality of Life: A Comparison between Korean American and Korean Breast and Gynecologic Cancer Survivors.” Oncology Nursing Forum 36.6 (2009): 699–708.
25. Ferrell, Betty, Karen Hassey Dow, and Marcia Grant. “Measurement of the Quality of Life in Cancer Survivors.” Quality of Life Research 4.6 (1995): 523–31.
26. Viswanathan, Madhu. Measurement Error and Research Design. Thousand Oaks, CA: Sage, 2005.
27. Gielen, Joris, Trudie van Iersel, and Bert Broeckaert. “Religion, World View and the Nurse: Results of a Quantitative Survey among Flemish Palliative Care Nurses.” International Journal of Palliative Nursing 15.12 (2009): 590–600.
28. Hancock, Gregory, and Karen Samuelsen. Advances in Latent Variable Mixture Models. The CILVR Series on Latent Variable Methodology. Charlotte, NC: Information Age Pub., 2008.
29. Zumbo, Bruno. “Three Generations of DIF Analyses: Considering Where It Has Been, Where It Is Now, and Where It Is Going.” Language Assessment Quarterly 4 (2007): 223–33.
30. Drennan, Jonathan. “Cognitive Interviewing: Verbal Data in the Design and Pretesting of Questionnaires.” Journal of Advanced Nursing 42 (2003): 57–63.
31. Morell, Linda, and Rachel Jin Bee Tan. “Validating for Use and Interpretation: A Mixed Methods Contribution Illustrated.” Journal of Mixed Methods Research 3.3 (2009): 242–64.
32. Willis, Gordon. Cognitive Interviewing: A Tool for Improving Questionnaire Design. Thousand Oaks, CA: Sage Publications, 2005.
33. Miller, op. cit.
34. Williams, David, and Michelle Sternthal. “Spirituality, Religion and Health: Evidence and Research Directions.” Medical Journal of Australia 186.10 Suppl (2007):S47–50.
35. Miller, op. cit.
36. Rohani, Camelia, Sedigheh Khanjari, Heidar Ali Abedi, et al. “Health Index, Sense of Coherence Scale, Brief Religious Coping Scale and Spiritual Perspective Scale: Psychometric Properties.” Journal of Advanced Nursing 66.12 (2010): 2796–2806.
37. Pargament, Kenneth, Bruce Smith, Harold Koenig, et. al. “Patterns of Positive and Negative Religious Coping with Major Life Stressors.” Journal for the ScientificStudy of Religion 37.4 (1998): 710–724.
38. Reed, Pamela. “Religiousness among Terminally Ill and Healthy Adults.” Research in Nursing and Health 9.1 (1986): 35–41.
39. Rohani, op. cit.
40. Newlin, Kelley, Gail Melkus, Ruth Tappen, et al. “Relationships of Religion and Spirituality to Glycemic Control in Black Women with Type 2 Diabetes.” Nursing Research 57.5 (2008): 331–9.
41. Paloutzian, op. cit.
42. Musgrave, Catherine, and Elizabeth McFarlane. “Israeli Oncology Nurses’ Religiosity, Spiritual Well-Being, and Attitudes toward Spiritual Care: A Path Analysis.” Oncology Nursing Forum 31.2 (2004): 321–7.
43. Gorsuch, Richard, and Susan McPherson. “Intrinsic/Extrinsic Measurement: I/E-Revised and Single-Item Scales.” Journal for the Scientific Study of Religion 28.3 (1989): 348–54.
44. Johnston Taylor, Elizabeth, Martha Highfield, and Madalon Amenta. “Predictors of Oncology and Hospice Nurses’ Spiritual Care Perspectives and Practices.” AppliedNursing Research 12.1 (1999): 30–7.
45. Taveggia, Thomas. “Resolving Research Controversy through Empirical Cumulation: Toward Reliable Sociological Knowledge.” Sociological Methods Research 2.4 (1974): 395–407.
46. Hall, Daniel, Keith Meador, and Harold Koenig. “Measuring Religiousness in Health Research: Review and Critique.” Journal of Religion and Health 47.2 (2008): 134–63.
47. Moberg, David. “Assessing and Measuring Spirituality: Confronting Dilemmas of Universal and Particular Evaluative Criteria.” Journal of Adult Development 9.1 (2002): 47–60.
48. Hamilton, Jill, Jamie Crandell, J. Kameron Carter, et al. “Reliability and Validity of the Perspectives of Support from God Scale.” Nursing Research 59.2 (2010): 102–9.
49. Hamilton, Jill, et al. “Spirituality among African American Cancer Survivors: Having a Personal Relationship with God.” Cancer Nursing 30.4 (2007): 309–16.