Introduction
The role of spirituality and religion in people’s lives has interested researchers over the past 30 years. Studies have focused on the role of spirituality and religion in physical as well as mental health (George, Ellison, & Larson, Reference George, Ellison and Larson2002; Koenig, George, & Titus, Reference Koenig, George and Titus2004; Thoresen & Harris, Reference Thoresen and Harris2002). Although findings are not unambiguous , most studies indicate that various aspects of religiosity and spirituality may enhance well-being (Ellison & Fan, Reference Ellison and Fan2007; Krause, Reference Krause2003); reduce levels of depression and psychological distress (Braam, Van Den Edden, Prince, Beekman, Kivela, Lawlor et al., Reference Braam, Van Den Edden, Prince, Beekman, Kivela and Lawlor2001; Smith, McCullough, & Poll, Reference Smith, McCullough and Poll2003 – meta-analysis); improve cognitive functioning (Hill, Burdette, Angel, & Angel, Reference Hill, Burdette, Angel and Angel2006; Kaufman, Anaki, Binns, & Freedman, Reference Kaufman, Anaki, Binns and Freedman2007); and preserve physical health (George, Reference George2002; Powell, Shahabi, & Thorensen, Reference Powell, Shahabi and Thorensen2003; Seeman, Dubin, & Seeman, Reference Seeman, Dubin and Seeman2003).
In addition, spirituality and religion seem to be more closely linked to psychological well-being among older adults than among younger people (Ellison, Reference Ellison1991). Spirituality and religion emerge as a particularly important resource for people coping with stress in later life (Reyes-Ortiz, Avele, Mulligan, Espino, Berges, & Markides, Reference Reyes-Ortiz, Avele, Mulligan, Espino, Berges and Markides2006). With acute and chronic physical health problems, the loss of close friends and family, approaching death, and other significant life events, spiritual concerns might become increasingly important. Moreover, studies have shown increased spirituality with age in a number of populations (Ahmadi, Reference Ahmadi2001; Thomas, Reference Thomas, Susan, McFadden and Atchley2006; Wink & Dillon, Reference Wink and Dillon2002). According to Dalby (Reference Dalby2006), spirituality corresponds to a search for meaning and purpose at a time of life when earlier sources of meaning and purpose may be diminishing.
One interesting explanation of the link between spirituality and age is the theory of gerotranscendence based on Erikson’s (Reference Erikson1982) model. Tornstam (Reference Tornstam2005) defined gerotranscendence as a shift in meta-perspective, from a materialistic and rational view of the world to a more cosmic and transcendent one, normally accompanied by an increase in life satisfaction. In the empirically based theory, the individual moving towards gerotranscendence may experience a series of gerotranscendental changes or developments. These typically include a redefinition of the self and of relationships with others, and a new understanding of fundamental existential and spiritual questions. Crowther, Parker, Larimore, Achenbaum, Koenig, and Cohen (2002) emphasized that spirituality is an important, but often forgotten, factor in the lives of older people. It is thus important to include aspects of spirituality and religiosity in the psychological care of older people (Ortiz & Langer, Reference Ortiz and Langer2002). For these reasons, we chose to investigate spirituality and religiosity in an elderly population.
Concepts of spirituality and religiosity are often poorly defined in these multidimensional aspects. Hill and Pargament (Reference Hill and Pargament2003) defined religiosity as a reference to external, institutionalized, formal, and doctrinal aspects of religious life, whereas spirituality is a personal and subjective experience including the concept of God and the divine. In general, religiosity refers to group affiliation and practices (Crowther et al., Reference Crowther, Parker, Larimore, Achenbaum and Koenig2002). On the other hand, spirituality is distinguished from material reality and as such refers to the transcendent: something beyond the self (Miller & Thoresen, Reference Miller and Thoresen2003). Spirituality is a search for meaning, involving the questioning of life and the relationship with the sacred or transcendent, which may or may not stem from the development of religious rituals and the formation of a community (Koenig, McCullough, & Larson, Reference Koenig, McCullough and Larson2001; Dalby, Reference Dalby2006; Moreira-Almeida & Koenig, Reference Moreira-Almeida and Koenig2006). Thus, spirituality seems to be a broader term than religiosity. In fact, the two constructs are closely linked; a few people engage in religious activity with no spiritual sense of it, while a small group of people experience spirituality that is not connected to any form of religious belief or activity (Underwood, Reference Underwood2006).
Multiple instruments have been developed to address the construct of spirituality. One such instrument, which seems to be particularly promising, is the Daily Spiritual Experience Scale (DSES). This scale was initiated by the National Institute of Aging and the Fetzer Institute (Fetzer, 1999). The DSES was developed to measure everyday spiritual experiences and is of interest because it is not rooted in any institutionalized or organisational religious activity, nor is it limited to any particular tradition (Ellison & Fan, Reference Ellison and Fan2007; Underwood, Reference Underwood2006; Underwood & Teresi, Reference Underwood and Teresi2002). “This scale is intended to measure a person’s perception of the transcendent (God, the divine) in daily life and his or her perception of his or her interaction with or involvement of the transcendent in life” (Underwood & Teresi, p. 23).
The original version of the DSES was developed from in-depth interviews and focus groups with individuals from many religious backgrounds (agnostics, atheists, Buddhists, Christians, Hindus, Jews, and Muslims). The use of the word God caused some concern; although various words were used to refer to the transcendent or divine, the predominant word used was God. Consequently, the introduction to the scale included advising people who were not comfortable with the word God that they could substitute another word for it that better represented their perception of the divine or holy (Underwood, Reference Underwood2006; Underwood & Teresi, Reference Underwood and Teresi2002). The original scale consisted of 16 items (self-report measures) and demonstrated a high internal consistency (Cronbach’s α = .90–.96 [Ellison & Fan, Reference Ellison and Fan2007; Harris, Sherritt, Holder, Kulig, Shrier, & Knight, Reference Harris, Sherritt, Holder, Kulig, Shrier and Knight2007; Loustalot, Wyatt, Boss, & McDyess, Reference Loustalot, Wyatt, Boss and McDyess2007; Underwood & Teresi]) and good temporal stability (r = .77: Loustalot et al.; r = .85: Underwood & Teresi). In addition, principal component analysis (PCA) supported a one-factor structure (Underwood & Teresi), even though some research divides the DSES items into “God” and “non-God” items (Ellison & Fan; Zemore & Kaskutus, Reference Zemore and Kaskutus2004).
Finally, links were established between the DSES (see http://www.dsescale.org) and health and mental health outcomes. It has been translated into a variety of languages, such as Spanish, Korean, Hebrew, and Vietnamese, and used in Jewish, Muslim, Roman Catholic, and Protestant religious cultures (French, Eisenberg, Vaughan, Purwono, & Suryanti, Reference French, Eisenberg, Vaughan, Purwono and Suryanti2008; Kalkstein & Tower, Reference Kalkstein and Tower2008; Underwood & Teresi).
A six-item scale of the DSES was developed for inclusion in the General Social Survey (GSS, 1998–2004). In the GSS, internal consistency of the short version was .91 (Ellison & Fan, Reference Ellison and Fan2007). In another study, Loustalot et al. (Reference Loustalot, Wyatt, Boss and McDyess2007) evaluated the psychometric properties of this shortened version among African-Americans. Internal consistency and test-retest reliability were good (Cronbach’s α = .81–.86; 2 weeks later, stability r = .79). Although Underwood and Teresi (Reference Underwood and Teresi2002) recommended utilization of a 16-item version of the DSES, recent studies by Ellison and Fan indicate that there is no clear evidence that the long form predicts positive outcomes more effectively than the short form. Cross-version comparisons of the DSES have shown no clear differences in the magnitude or statistical significance of relationships between the DSES and measures of well-being (Ellison & Fan). In addition, concurrent validity between the two versions is similar (r = .85 [Loustalot et al.]). Using the shorter form of the DSES, researchers linked DSES measures with desirable outcomes in studies of inner-city elders, adolescents, and caregiver burnout (Dunn, Chapelski, Stinson, & Massanari, Reference Dunn, Chapelski, Stinson and Massanari2004; Holland & Neimeyer, Reference Holland and Neimeyer2005; Pearce, Little, & Perez, Reference Pearce, Little and Perez2003).
Although the short version of the DSES seems to have adequate psychometric properties, it needs to be validated in different groups. First, because of the need for a short measure, designed especially for use with elderly people, it is important to validate the six-item version. Second, most of the studies and scales in the field of spirituality were developed in the U.S., and there is an urgent need for more trans-culturally validated scales and replication of studies in different countries. Although studies from North America and Europe (Braam et al., Reference Braam, Van Den Edden, Prince, Beekman, Kivela and Lawlor2001) suggest that religiosity and spirituality help to protect against depression and to enhance well-being, there are differences between the two continents with respect to religious traditions. For example, studies have shown that adherence to traditional Christian beliefs is stronger in North America than in Europe (Halman & de Moor, Reference Halman, de Moor, Ester, Halman and de Moor1993. In Europe, particularly in France, a sharp decline has been observed in institutional religion (Braam, Delespaul, Beekman, Deeg, Pérès, Dewey et al., Reference Braam, Delespaul, Beekman, Deeg, Pérès and Dewey2004), which stands in contrast to the vitality of different kinds of religion and a strong belief in God that can be observed in the U.S. (Gallup, Reference Gallup1996; Koenig, Reference Koenig2003; Kosmin & Lachman, Reference Kosmin and Lachman1993). In particular, contextual information about religious phenomena remains problematic in France. Because of the private nature of religious affiliation (only private institutes or research using the quota method can ask for religious affiliation) and the background of mid-20th century history in France, no current public statistical survey records religious affiliation. In 2005, for the first time, a specific question concerning religious affiliation was included in a demographic study (Régnier-Loilier & Prioux, Reference Régnier-Loilier and Prioux2008). Results indicated Catholicism is the predominant religion (80% expressed a link with Roman Catholicism, 5% with Islam, and 2% with Protestantism while 11% stated no link with a religion). Major differences according to age emerged from the study results. The proportion of Catholics was lower among younger people. Of those who had a religion, 88 per cent of the 18 to 24 age group stated that they were Catholic, compared with 95 per cent of the group aged 65 and over. Above all, the proportion of regular attendance at services was different: 47 per cent of the 75 to 79 age group attended at least once a week compared with 17 per cent of the 20 to 24 age group.
The purpose of this study was, therefore, to translate the DSES into French and replicate the psychometric findings with the translated version. Two separate studies were conducted to validate the short version: Study 1 examined the factor structure, internal consistency, and test-retest reliability of the French translation of the short version, while Study 2 provided confirmatory factor analysis, test-retest reliability, and links with mental and physical health.
Study 1: Translation, principal component analysis, internal consistency
Overview
For Study 1, the short version of the DSES was adapted and translated into French. The factor structure of the DSES was examined using PCA. Study 1 also provided data on the internal consistency of the DSES.
Method
Participants
A total of 195 adults, aged 65 and over, were recruited from local senior centers and retirement homes and participated voluntarily in the study. Volunteers were screened for cognitive impairment using the French adaptation of Folstein’s Mini Mental Status Examination (MMSE; Desrosiers & Hébert, Reference Desrosiers, Hébert, Arcand and Hébert1997) and were excluded if they scored less than 20. Mean age of the participants was 77.44 years (SD = 5.12; 65–94) with 120 women (61.53%, M age = 77.81, SD = 5.45) and 75 men (38.47%, M age = 76.84, SD = 4.54). Of those participants, 51.28 per cent were married or had a partner, 34.87 per cent were widowed, and 13.84 per cent lived alone. Regarding religious orientation, the great majority of participants were Catholic (86.73%); the others had no stated religious belief.
Measure
The six-item version of the DSES used in the General Social Survey (GSS) was used. This scale is intended to assess a person’s perception of the transcendent (God, the divine) in daily life. The six items were scored using a Likert scale, in which the response categories were “many times a day, every day, most days, some days, once in a while, and never or almost never”. Lower scores reflected a more frequent DSE.
Translation of the DSES
The DSES was translated into French, on the basis of the methods proposed by Vallerand (Reference Vallerand1989), by two independent professional bilingual translators. These versions were then translated back into English by an independent third translator and compared with the initial English version. A consensus meeting – consisting of one translator, one French researcher, and a professional on spirituality – was organized. These three specialists age-checked and agreed on a version of the DSES that best reflected the linguistic and conceptual matter of the original scale. As Underwood & Teresi (Reference Underwood and Teresi2002) had found, the three found difficulty primarily in the translation process with the use of the word God in some of the items. However, the word God seems to be easily understood, most people being able to “translate” it into their own concept of the divine. As in the original version, the introduction of the DSES advised participants to substitute the word God with another concept that corresponded better to their perception of the divine or holy.
Finally, we pilot-tested the DSES with 10 elderly people. All the participants stated that they had no difficulty understanding the items and expressed their willingness to complete all the items, which took approximately 4 to 7 minutes.
Results
Principal component analysis of the DSES
Bartlett’s chi-square test was used to determine whether the sample size was appropriate for the PCA and whether the data came from a normally distributed sample of the population. This test was statistically significant (χ2 = 879, p < .001).
To establish the factor structure of the DSES, the six items were subjected to PCA with oblimin rotation. This was carried out for all participants together (see Table 1). The principal component factor analysis for the DSES revealed only one factor with an eigenvalue greater than 1. This factor accounts for 67 per cent of the variance, with an eigenvalue of 4.05, and loadings ranging from .55 to .91. All the others factors have an eigenvalue of less than 1.
Table 1: Principal components factor analysis of the DSES (Study 1 and Study 2)
Internal consistency of the DSES
Cronbach’s alpha was .89 indicating a high internal consistency. Removal of one or more weakly correlated items did not significantly affect the alpha value. The mean item correlation for the DSES was .55 (min = .35, max = .86).
Study 2: Confirmatory analysis, internal consistency, test-retest, correlations with health measures
Overview
Study 2 examined the factor structure of the DSES using confirmatory factor analysis (CFA). On the basis of Study 1’s results, this study first examined a one-factor model. Internal consistency and test-retest were evaluated. Measures of life satisfaction, depression, and physical health were evaluated in order to assess the construct validity of the DSES. Correlations with other participants’ characteristics were also examined.
Methods
Participants and Procedure
A total of 338 participants, aged 65 and older, were recruited from local senior centers and advertisements in specialized magazines (aimed at retirement groups). All participants were volunteers. Mean age of the participants was 77.87 years (SD = 5.46; 65–94 years). They included 212 women (62.60%, M age = 78.30, SD = 5.82) and 126 men (37.40%, M age = 76.96, SD = 4.57). Of these individuals, 53 per cent were married or had a partner, 38.33 per cent were widowed, and 8 per cent lived alone. The great majority of participants were Catholic (80%); the others had no stated religious belief.
Two weeks after the initial test, a sub-sample of 40 participants completed the DSES again to test temporal stability. The retest participants did not differ from the full sample on demographics characteristics (sex, marital status, religious orientation), nor did they differ on any variables (life satisfaction, depression, and physical health) examined in this study (all t-test results < 2) except for age. Indeed, the sub-sample of 40 participants was younger than original participants (respectively, 75 years old versus 77.87 years old; p < .001).
Measures
All participants completed the short version of the DSES (developed in Study 1). Life satisfaction was evaluated using the French version of the Satisfaction With Life Scale (SWLS) developed by Diener, Emmons, Larsen, and Griffin (Reference Diener, Emmons, Larsen and Griffin1985) and adapted by Blais, Vallerand, Pelletier, and Brière (Reference Blais, Vallerand, Pelletier and Brière1989). The SWLS is a five-item scale in a seven-point Likert format, with higher scores indicating greater life satisfaction. The French version of the SWLS has good internal consistency (Cronbach’s α =. 81–.82) and stability over time (r = .64 over a 2-month interval [Blais et al.]). Factor analysis revealed a one-factor structure, and construct validity seemed to be acceptable (Blais et al.). The internal consistency of the SWLS in this study was deemed good (Cronbach’s α = .89).
Depression was measured using the geriatric depression scale (GDS; Yesavage, Brink, Rose, Lum, Huang, Adey et al., Reference Yesavage, Brink, Rose, Lum, Huang and Adey1983). French validation of the scale was achieved by Bourque, Blanchard, & Vézina (Reference Bourque, Blanchard and Vézina1990). This is the first depression screening measure developed for, and validated among, older people. The sensitivity and specificity of the GDS have been evaluated in a variety of elderly populations and geriatric in-patients (Shah, Phongsathorn, Bielawska, & Katona, Reference Shah, Phongsathorn, Bielawska and Katona1996) and subjects over age 85 (De Craen, Heeren, & Gussekloo, Reference De Craen, Heeren and Gussekloo2003), with an optimal cut-off point for depression found to be a score of five or above (D’Ath, Katona, & Mullan, Reference D’Ath, Katona and Mullan1994; Weintraub, Oehlberg, Katz, & Stern, Reference Weintraub, Oehlberg, Katz and Stern2006). The French version of the GDS has good internal consistency (Cronbach’s α =.84–.90) and stability over time (r = .83–.70 over a 4-week interval [Bourque et al.]). Factor analysis revealed a one-factor structure, and the GDS has also been shown to correlate with the Beck Depression Scale (QBD; Bourque & Beaudette, Reference Bourque and Beaudette1982). The internal consistency of the GDS in this study was deemed good (Cronbach’s α = .75).
Physical health was assessed by two questions. First, “How is your health in general?” on a Likert-type scale ranging from 1 (poor) to 5 (excellent), higher scores indicating better health. Second, the question used in the Quebec Health Survey: “In general, compared to other people of your age, how good would you say your health is?” on a Likert-type scale ranging from 1 (poor) to 5 (excellent). Researchers have recommended this second question for elderly people to assess their general physical health (Rodin & McAvay, Reference Rodin and McAvay1993).
Results
Confirmatory factor analysis (CFA)
In order to examine the one-factor structure of the DSES (Study1), a confirmatory analysis was conducted using the Statistica. Given the numerous inherent potential problems of the chi-square measure of goodness-of-fit (Tabachnik & Fidell, Reference Tabachnik and Fidell2001), four other indices of the model fit were assessed: (a) the root mean square error of approximation (RMSEA) index (RMSEA values < .05 indicated good fit, and RMSEA < .08 indicated acceptable fit [Hu & Bentler, Reference Hu and Bentler1999]) and its 90 per cent confidence intervals (90% CI); (b) the goodness-of-fit (GFI) index (GFI values > .90 indicated good fit [Tabachnik & Fidell]); (c) the adjusted goodness-of-fit (AGFI) index (AGFI values > .90 indicated good fit [Tabachnik & Fidell]); and (d) the non-normed fit (NNFI) index (NNFI values > .95 indicated good fit [Hu & Bentler]).
To continue exploration of the structure of the DSES, two models were tested. The first was a one-factor model. The second was a two-factor model in reference to recent studies (Ellison & Fan, Reference Ellison and Fan2007; Zemore & Kaskutus, Reference Zemore and Kaskutus2004) that make a distinction between “God Items” (items 1, 4, and 6) and “Non-God Items” (or self-transcendence, non-theist – items 2, 3, and 5) in their exploration of the DSES. Results of the CFA are presented in Table 2. The indices showed the goodness-of-fit of the DSES one-factor model. Analysis showed a GFI of .95, and an NNFI of .96 revealed a good model fit. The RMSEA was slightly above .05 (RMSEA = .08; 90% CI) indicating not a good but an acceptable fit of the model. The AGFI of .88 was slightly under .90. For the two-model factors, only one index was acceptable (GFI = .92). Finally, the one-factor model showed a better fit than the two-factor model.
Table 2: Model indices of Daily Spiritual Experience Scale (DSES)
Likewise, we conducted a principal component factor analysis. As in Study 1, the analysis revealed only a single underlying construct, which accounted for 73 per cent of the variance, with an eigenvalue of 4.38 and loadings ranging from .76 to .91 (see Table 1).
Internal consistency and test-retest reliability
The Cronbach’s alpha value was .92, indicating a high internal consistency for the DSES. Removal of one or more weakly correlated items did not have major consequences on the alpha value. The mean item correlation for the DSES was .65 (min = .53, max = .80).
Test-retest data were collected on a sub-sample of 40 participants (interval time = 2 weeks). Test-retest reliability of the DSES was high (r = .85), indicating good temporal stability.
Correlation with participant characteristics and other health-related measures
To assess the construct validity of the DSES short version, we examined the links between the DSES and measures of life satisfaction, depression, and physical health. In addition, the participants’ characteristics and DSES scores were documented and are presented in Table 3. The DSES scores were significantly correlated with age, with older participants scoring higher. The DSES was significantly associated with life satisfaction and with the two measures of physical health: Participants with higher DSES scores had greater life satisfaction and better health evaluation (in general and in comparison with other elderly people). No significant correlation occurred with sex, marital status, or depression.
Table 3: Descriptive statistics and correlation between the DSES and psychosocial and other health-related measures
* p < .05, ** p < .01.
a Sex: 1 = men/2 = women; Marital status: 1 = single/2: married or living with a partner.
Discussion
The aim of the studies was to validate the DSES short form in an elderly French population. Similar problems were encountered in the use of the word God in the French translation of the scale. In line with Underwood and Teresi (Reference Underwood and Teresi2002), participants were told that they could substitute the word God with another concept representing for them the divine or holy.
Some items are noteworthy. The item “I feel a deep sense of peace and inner harmony” is conceived as a transcendent dimension that may be affected but not determined by events and affect, and this question was designed to elicit something other than positive mood or psychological well-being (Ellison & Fan, Reference Ellison and Fan2007). Even if a sense of wholeness and internal integration is frequently referred to in the spiritual literature, this item can be problematic because many things in addition to religion and spirituality can give a person a deep sense of peace and inner harmony. In addition, the item: “I find strength and comfort in my religion and spirituality” has been described as “social support from God”. This item intends to measure a direct sense of support and comfort from the transcendent, but rather than assess what spirituality is, it could instead represent what spirituality may or may not produce. Because these items are likely to be confounded with both positive and negative aspects of mental health, further research should investigate more thoroughly what, exactly, they measure.
The results of PCA from Study 1 and of CFA from Study 2 indicated that the DSES can be conceptualized on a single general factor. This confirms previous studies (Exploratory Factorial Analysis [EFA]) by Underwood and Teresi (Reference Underwood and Teresi2002) on the long 16-item version of the DSES and by Ellison and Fan (Reference Ellison and Fan2007) on the short form. To our knowledge, our study is the first to have examined the construct validity of the DSES using CFA. The overall goodness-of-fit of the one-factor model is near or above that recommended by conventional standards (Hu & Bentler, Reference Hu and Bentler1999; Tabachnik & Fidell, Reference Tabachnik and Fidell2001). These first results confirm the uni-dimensionality of the short version of the DSES. In particular, items referring explicitly to God did not factor out independently from the others, in contrast to previous results (Ellison & Fan; Zemore & Kaskutis, Reference Zemore and Kaskutus2004). Even if the word God is not always used to refer to the transcendent or divine, the short version of the DSES addresses a possible common ground that transcends religious boundaries. Finally, our results demonstrate strong support for the construct validity of the DSES short form.
The DSES appeared strong psychometrically, evidencing adequate to good internal consistencies and good test-retest reliability. Internal consistency (.89 and .92 for studies 1 and 2 respectively) was higher than that found in the English version (.81–.86 [Loustalot et al., Reference Loustalot, Wyatt, Boss and McDyess2007]). The test-retest Pearson r was high (.85) indicating good consistency of the result over time. This result is also higher than Loustalot’s results for the same time interval. Overall, the results attest to the good reliability of the DSES short form.
With regard to the role of the DSES in mental and physical health, the results are a little less convincing. As expected, higher scores were correlated with good life satisfaction and good self-evaluation of health. Previous studies with the DSES short version have also reported its link with quality of life (Ellison & Fan, Reference Ellison and Fan2007; Underwood & Teresi, Reference Underwood and Teresi2002). “Ordinary” spiritual experiences may enhance the feelings of joy, comfort, love, and spiritual peace in everyday life and so promote good psychological outcomes. In contrast to these findings and contrary to our expectations, the DSES was not correlated with depression. However, most studies indicate that people with high spirituality or religiosity scores are less depressive than others (Braam et al., Reference Braam, Van Den Edden, Prince, Beekman, Kivela and Lawlor2001; Parker, Roff, Klemmack, Koenig, & Baker, Reference Parker, Roff, Klemmack, Koenig and Baker2003; Smith et al., Reference Smith, McCullough and Poll2003 – meta-analysis). Furthermore, Underwood and Teresi established a positive correlation between the DSES and a depression scale (CES-D). Nevertheless, our results are similar to those of another recent study (Ellison & Fan) suggesting a robust association between the DSES (short version) and psychological well-being, while it appears to have little bearing on negative affect. Two points can be highlighted. First, the depression measure depended considerably on the sample and methods used. In our study, we chose the GDS scale because it had been validated among an elderly population. But it is possible that the GDS scale is not the most appropriate tool to assess depression in this population. Depression in old age is often ignored or misdiagnosed (Cole, Bellavance, & Mansour, Reference Cole, Bellavance and Mansour1999; Cole & Yaffe, Reference Cole and Yaffe1996) because older adults frequently present atypical symptoms compared to younger adults (Dreyfus, Reference Dreyfus1998). Further research is needed using recognized indicators of clinically relevant aspects of psychopathology, such as diagnostic measures based on the DSM-IV to confirm this absence of association. Secondly, according to the properties of GDS (D’Ath et al., Reference D’Ath, Katona and Mullan1994; Weintraub et al., Reference Weintraub, Oehlberg, Katz and Stern2006) a cut-off of 5 (or more) is usually acknowledged as indicative of depression in the elderly population. In our sample, the mean score of depression is far lower than 5. Participants are rarely (or not at all) depressed, and this could also explain the absence of links between DSES and depression.
Finally, our results indicated that the oldest people had higher scores on the DSES than the younger participants. This confirms the increase of spirituality with age (Dalby, Reference Dalby2006). As people grow older, they face challenges related to a decline in physical health and changes in social and family networks, and it is during this time that religious and spiritual beliefs may be particularly important in maintaining well-being.
Our study has several limitations. For a better validation of the DSES short form, concurrent, discriminant, and predictive validity need to be tested. Further research on the psychometric properties of this instrument should include the examination of larger sample sizes in different populations. The sample of our studies is heavily Catholic, and it would be interesting to investigate other religious groups. A recent meta-analysis investigated values and religiosity across different countries and regions and religious groups, and it indicated impressive similarities across a variety of religions (Christianity, Islam, and Judaism), denominations (Catholics, Orthodox, and Protestants) and different countries and regions (Europe, U.S., and Middle East) (Saraglou, Delpierre, & Dernelle, Reference Saraglou, Delpierre and Dernelle2004). Differences seem to depend more on the socio-economic development of the countries or regions concerned than on religious affiliation. In the same way, Braam et al. (Reference Braam, Van Den Edden, Prince, Beekman, Kivela and Lawlor2001, Reference Braam, Delespaul, Beekman, Deeg, Pérès and Dewey2004) found no specific pattern between religious denomination and depression , although Kim (Reference Kim2003) found that Protestants were more satisfied with life, followed by Catholics and Buddhists (Korean population). It would also be particularly interesting to explore cross-religious and cross-cultural differences versus trans-religious and trans-cultural constants.
A second limitation is that the combination of the DSES with other religious measures could warrant future work. Studies investigating the association between spirituality and religiosity and depression highlight the impact of different aspects of spirituality and religiosity. More precisely, these studies suggest that the relationships of religiosity with mental health vary with the dimension of religiosity considered (church attendance, religious meetings, prayer, meditation, etc.) (Parker et al., Reference Parker, Roff, Klemmack, Koenig and Baker2003; Smith et al., Reference Smith, McCullough and Poll2003). Finally, because of the implications on health and well-being, it would be extremely helpful to conduct longitudinal studies.
Nonetheless, the results of the studies presented provide preliminary evidence of the psychometric properties of the French translation of the DSES short version, which helps understand the potential benefit of encouraging the spiritual aspects of life, and we hope it will be used in a variety of health-related research.
