Mental Health, Religion & Culture

Vol. 12, No. 1, January 2009, 123

The relationship between spiritual experiences, transpersonal trust,

social support, and sense of coherence and mental distressa
comparison of spiritually practising and non-practising samples
Niko Kohlsab*, Harald Walachbc and Markus Wirtzd
a

Generation Research Program, Human Science Center, Ludwig-Maximilians University

(LMU), Munich, Germany; bBrain, Mind and Healing Program, Samueli Institute, Alexandria,
VA, USA; cThe University of Northampton, School of Social Sciences, Psychology Division,
Northampton, UK; dPadagogische Hochschule Freiburg, Institut fur Psychologie,
Freiburg, Germany
(Received 24 March 2008; final version received 30 March 2008)
We compared the pathways from exceptional experiences (measured with the
Exceptional Experiences Questionnaire, EEQ), transpersonal trust (TPV), social
support (F-SoZu) and sense of coherence (SOC) scales towards mental distress
within a spiritually practising (SP) and a non-practising sample (NSP), using
structural equation modelling. We found a high amount of variance explained for
SOC (SP: R2 0.50; NSP: R2 0.61), a moderate amount for F-SoZU
(SP: R2 0.17; NSP: R2 0.20) and for TPV a very small amount only in the
SP sample (SP: R2 0.04; NSP: R2 0.00). In contrast, for the EEQ, which grasps
positive and negative spiritual, psychopathological, and visionary dream
experiences, a strong relationship was found for the NSP sample (R2 0.53)
but only a moderate relationship for the SP sample (R2 0.28). Further analysis
revealed that the path coefficients from positive, negative spiritual, and
psychopathological experiences to distress were significantly lower in the SP
sample. Thus, as regular spiritual practice seems to alter the pathways to distress
derived from positive and negative spiritual and psychopathological experiences,
unidimensional questionnaires only grasping positive spiritual experiences seem
to be inappropriate for explaining the intrapersonal mechanisms associated with
regular spiritual practice.
[The] unseen region in question is not merely ideal, for it produces effects in this world. When we
commune with it, work is actually done upon our finite personality, for we are turned into new
men, and consequences in the way of conduct follow in the natural world upon our regenerative
charge. But that which produces effects within another reality must be termed a reality itself,
so I feel as if we had no philosophical excuse for calling the unseen or mystical world unreal.
(James, 1904, p. 516)

Introduction
The relationship between spirituality, religion, and health has become a focus of
interest within mainstream health psychology and psychiatry (Culliford, 2002; Miller
*Corresponding author. Email: [email protected]
ISSN 13674676 print/ISSN 14699737 online
2009 Taylor & Francis
DOI: 10.1080/13674670802087385
http://www.informaworld.com

N. Kohls et al.

& Thoresen, 2003). There is meanwhile a bulk of empirical evidence highlighting the
importance of spiritual practice for mental health (Koenig, McCullough, & Larson, 2001;
Weaver & Koenig, 2006), and in a similar line, a robust body of empirical research
demonstrates the positive effects of mindbody practices, which are frequently associated
within or practised as part of a spiritual belief system, for well-being and health
(Walach, Gander, & Kohls, in press).

Uncertainty about the pathways from spirituality to health

There is, however, considerable uncertainty about the mechanisms that drive the
spiritualityhealth connection. It has for example frequently been hypothesized that it is
mainly the social component of religion and spirituality mediating beneficial effects
(Levin, Chatters, Ellison, & Taylor, 1996; Powell, Shahabi, & Thoresen, 2003). Although
this is certainly true for some types of religiosity, the appropriateness of this explanation
for elucidating the intrapersonal effects of introspective techniques that are associated with
spirituality may be debated. To give one example, one of the best studied introspective
techniques is mindfulness-based meditation, which has its roots in an ancient Buddhist
meditation practice. Nevertheless, Mindfulness-Based Stress Reduction (MBSR)
Programmes have been developed in a medical context and are taught in a non-religious
way. Mindfulness can be understood in psychological terms as a mental ability that allows
one to focus on the direct and immediate perception of the present moment with a state of
non-judgemental awareness, while evaluative cognitive feedback is voluntarily suspended
(Hayes & Shenk, 2004). The focus of contemporary MBSR trainings is on integrating
mindfulness into the realities of daily living, and participants are correspondingly
encouraged to practise at home on a regular basis. This approach has in sum proven to be
beneficial for health and well-being (Grossman, Schmidt, Niemann, & Walach, 2004;
Reibel, Greeson, Brainard, & Rosenzweig, 2001). A recent FMRT study for example has
found that individuals taking part in a MBSR Programme seem to be able to alter their
personal psychological model so they can dissociate their self-awareness of the present
from their long-term self image (Farb et al., 2007), which could possibly change the way
they perceive suffering and distress. Thus, there is good reason to assume that the effects
stemming from this or other forms of mediation cannot be merely explained as
consequences of interpersonal or social factors, and to completely elucidate them one
has to take intrapersonal factors into account as well.

Disentangling different layers of spirituality

Spirituality is a complex and multifaceted phenomenon that can correspondingly be
defined in different ways and on many levels. For example, a recent meta-analysis
investigating the relationship of spirituality and quality of life in 62 primary effect sizes
from 51 high-quality studies found that differences in how the concepts of spirituality and
quality of life were operationalized accounted for 27% of the variance in the respective
primary effect size criterion (Sawatzky, Ratner, & Chiu, 2005). This is a clear sign that
there are conceptual divergences in the operationalization of measures that that may
hamper progress in the field.
Although most questionnaires measuring spiritual and religious domains have been
devised on the basis of a Western (Christian) worldview, they are frequently inconsistent

Mental Health, Religion & Culture

from a conceptual standpoint (Edwards, 2003). Often items addressing spiritual and
religious beliefs are lumped together with statements of faith and reports on behaviour,
and both are sometimes mixed with descriptions of spiritual experiences (Hill &
Pargament, 2003; MacDonald, LeClair, Holland, Alter, & Friedman, 2002). Meanwhile,
there is consensus among researchers in the field that further research endeavours
should try to disentangle different layers of concepts (Devon, 2005; Thoresen, 1999;
Thoresen & Harris, 2002), and some authors have thereby explicitly argued for the need
for new instruments assessing spiritual and religious experience instead of faith beliefs or
attitudes (George, Larson, Koenig, & McCullough, 2000). There is indeed good reason to
assume that spiritual practice and experiences arising from regular exercise, rather than
belief sets, attitudes, or behaviour, might be pivotal to revealing the pathways from
spirituality to health: First of all, spiritual experiences as they are reported in the mystical
traditions are arguably at the roots of religion.1 Second, conversion phenomena that
sometimes change the whole trajectory of the life of the respective individuals are based on
extraordinary experiences, interpreted as important divine messages, such as revelations,
apparitions, or miracles. Third, within religious and spiritual systems, meditative or
contemplative practices have been systematically developed in the course of time as venues
for experiencing altered states of consciousness, which are frequently associated with
spiritual or mystical experiences. Correspondingly, spiritual practice may be understood as
any regular activity intended and designed to elicit spiritual experiences, e.g., prayer,
meditation or forms of contemplation.

Own empirical research approach

We believe that an important step towards disentangling the pathways from spirituality to
health is the distinction between direct intrapersonal effects stemming from exceptional
and spiritual experiences as opposed to religious and spiritual beliefs or religious behaviour.
However, we also hold the opinion that the tendency to devise measurement
instruments for grasping spiritual experiences as unidimensional and mainly positive
constructssuch as the Daily Spiritual Experiences Scale (Underwood, 2006; Underwood
& Teresi, 2002)is insufficient for scrutinizing the intrapersonal effects of spiritual
practice. We believe in contrast that it is important to differentiate between positive and
negative spiritual experiences, because it is by no means clear that spiritual experiences are
predominantly positive. This is witnessed by the fact that many individuals reporting
spiritual experiences were actually temporarily suffering distress from their exceptional
experiences (Smucker, 1996).
Hence, we have devised, pilot-tested, cross-validated, and revised a multidimensional
instrument called the Exceptional Experiences Questionnaire (EEQ) for assessing the
frequency and the emotional components of exceptional and spiritual experiences as
opposed to faith and belief statements, which has been described in detail elsewhere
(Kohls, 2004; Kohls, Hack, & Walach, 2008; Kohls & Walach, 2006). In short, the EEQ
captures positive and potentially unfavourable spiritual experiences, psychopathological
experiences, as well as visionary dream experiences (see Measures section for sample
items). We could show that individuals with regular spiritual practice report both more
positive and negative spiritual experiences. We could additionally show that a class of
items describing psychopathological experiences of mainly delusionary character as well as
visionary dream experiences can be psychometrically separated from positive and negative
spiritual experiences.

N. Kohls et al.

For analysing the pathways from the four factors of the EEQ towards psychological
distress, linear regression analysis was previously utilized (Kohls & Walach, 2007),
comparing two sociodemographically balanced nonclinical subsamples of spiritually
practising and non-practising individuals. Although spiritually practising individuals
reported significantly more positive and negative spiritual experiences, they accounted
for only 7% of psychological distress (as measured with the Brief-Symptom
Inventory (BSI)) in the spiritually practising sample, but for 36% of distress in
individuals with lack of spiritual practise. A comparison of the respective regression
weights between spiritually practising and non-practising individuals revealed only a
significant difference for the pathway from experiences of ego loss to psychological
distress: Experiences of ego loss had no effect on psychological distress in the group of
individuals with regular spiritual, contemplative, or meditative practice, while they
exhibited significant impact on distress in individuals with lack of spiritual practice.
In contrast, no significant contribution was found for positive spiritual experiences in
both groups.
Based on these findings, we have suggested that spiritual practice could be considered
to be a specific coping strategy for the distress caused by experiences of ego loss. Thus, our
analysis suggested that instead of interpreting spiritual practice as a health resource, lack
of spiritual practice should rather be regarded as a distinct risk factor. We have replicated
this finding in a sample of chronically ill patients (Kohls, Walach, & Lewith, submitted).
In sum, our research findings are obviously in contrast with existing research, which
shows that mainly positive spiritual experiences have a protective effect upon health
(George et al., 2000). Our outcome rather suggests that a key mechanism of regular
spiritual practice seems to be that a distressing impact of negative spiritual experiences can
be annihilated or at least gradually suspended.

Open questions and structural equation modelling

Empirical research shows that positive spiritual experiences have only a low or moderate
impact on well-being (Underwood & Teresi, 2002). Thus, the question arises, whether our
finding is a new insight that was made possible through the multidimensional design of the
EEQ, which takes both positive and negative spiritual experiences into account.
Alternatively, this result may alsoat least partlybe a methodological artefact, because
linear regression analysis, which we have used (Kohls & Walach, 2007), does not account
for intercorrelations between predictor variables. However, the four subfactors of the EEQ
are naturally intercorrelated (Kohls, 2004; Kohls & Walach, 2006). We therefore decided
to reanalyse our full nonclinical data set by means of structural equation modelling
(SEM). This analysis technique explicitly allows the definition and estimation of complex
model structures by combining the approaches of factor and regression analysis with path
analysis and by allowing interrcorrelations between predictor variables. Furthermore, if
multiple indicator variables are present, theoretically derived constructs can be modelled
by means of so-called latent or structural variables that possess desirable psychometric
properties (Bollen, 1989). One of the advantages of latent variables is that the
measurement error can be explicitly separated from the true variance. Thus, latent
variables, in contrast to observed manifest indicators, do not suffer from systematic
restrictions in measurement quality, given that certain criteria are met (Hair, Anderson,
Tatham, & Black, 2004).

Mental Health, Religion & Culture

Scope of the paper

In this paper:
(1) in order to obtain indications for the intrapersonal mechanisms associated with
regular spiritual practice, we analyse the predictive value of exceptional experiences
on mental distress for individuals with a regular spiritual practice and individuals
without such a practice, applying SEM;
(2) we investigate the amount of variance in mental distress explained by the EEQ with
the respective amount explained by well-established constructs such as social
support, sense of coherence, and transpersonal trust;
(3) we compare the pathways from exceptional experiences, social support, sense of
coherence, and transpersonal trust to mental distress.

Method
Participants
We have tested both a sample of individuals with a regular spiritual practice (spiritually
practising [SP]; N 350; 71% women) and individuals without such a practice
(non-practising [NSP] N 299; 69% women). All samples are convenience samples,
and more detailed demographics can be derived from Table 1.
In short, samples were comparable with regard to their mainly Christian denomination
(Catholic: SP: 30%; NSP: 32%; Protestant: SP: 31%; NSP: 42%; no denomination:
SP: 33%; NSP: 22%) and their high degree of education (university-entrance diploma
SP: 78%; NSP: 81%). Individuals in the SP sample (mean age: SP: 44.9 years, SD 12.3)
were older than those in the NSP sample (mean age NSP: 34.1 years, SD 13.1) by an
average of almost 11 years. The difference in age is mainly due to the fact that in the NSP
sample, many students were included, naturally also affecting the family status of this
cohort (single: SP: 40%; NSP: 60%; married: SP: 43%; NSP: 29%; divorced: SP: 15%;
NSP: 10%). In order to obtain two sociodemographically matched subsamples for
analysing intersample differences, in previous analyses post-hoc controlling by means of
propensity score matching was used (Kohls & Walach, 2007, 2008). However, in order to
account for a sufficiently high sample size necessary for utilizing SEM, the full data set will
be used in this analysis. This is warranted as a sensitivity analysis with our propensityscore matched sample and regression analysis has shown that the results between the full
and the matched sample are comparable.

Procedures
The study was conducted at the University Hospital in Freiburg. Participants from
Germany and Switzerland were recruited in public campaigns, university lectures and
courses as well as in meetings, congregations and conferences, and also by word of mouth
over a period of 2 years. For recruiting spiritually practising individuals, we additionally
addressed individuals from spiritually interested groups like religious communities of
Christian background, courses of Zen or Viapassana meditation and the German Spiritual
Emergence Networks.
Participants were presented with a set of paper and pencil questionnaires twice within a
6-month interval. Return envelopes were paid in advance and addressed to ensure

N. Kohls et al.

Table 1. Sociodemographic data for subsamples.

Spiritually practising
N
Sex
Women
Men
Mean age
Family status
Single
Married
Divorced
Widowed
Own children
Life situation
Living alone
Living in parental home
Living with a partner
Flat share
Denomination
Catholic
Protestant
Free Churches
Moslem
Jewish
Hindu
Buddhist
No denomination
Other
Education
None
Still in school
Secondary school
Secondary modern school
University-entrance diploma
Qualification
No formal qualification
Still qualifying
Apprenticeship
Vocational college
University diploma

350

Spiritually non-practising
299

247 (71%)
103 (29%)
44.9
(SD 12.3)

206 (69%)
92 (31%)
34.1
(SD 13.1)

141
149
52
7
170

(40%)
(43%)
(15%)
(2%)
(49%)

178
87
29
4
106

(60%)
(29%)
(10%)
(1%)
(36%)

134
5
179
32

(38%)
(1%)
(51%)
(9%)

81
28
124
63

(27%)
(10%)
(42%)
(21%)

103
110
8
0
0
1
4
116
3

(30%)
(31%)
(2%)
(0%)
(0%)
(0%)
(1%)
(33%)
(1%)

96
124
5
1
1
0
1
66
6

(32%)
(42%)
(1%)
(0%)
(0%)
(0%)
(0%)
(22%)
(2%)

1
1
16
52
278

(0%)
(0%)
(5%)
(15%)
(78%)

0
7
13
37
240

(0%)
(2%)
(4%)
(13%)
(81%)

5
36
34
47
224

(1%)
(10%)
(10)%
(14%)
(65%)

5
130
30
20
98

(2%)
(46%)
(11%)
(7%)
(35%)

Note: Figures in this table are rounded up to nearest whole if they are 40.5 and rounded down if
50.5; therefore, the cumulated percentage may differ slightly from 100%. Missing data are not
included. Parts of this table are taken from Kohls, N., & Walach, H. (2006). Exceptional experiences
and spiritual practice a new measurement approach. Spirituality and Health International, 7(3),
125150. John Wiley & Sons Limited. Reproduced with permission.

confidentiality. For the first survey, a total of 2000 questionnaires were disseminated and
N 7052 replied, leading to a response rate of approximately 35%. After 6 months, the
same questionnaire battery was distributed to those 642 participants who had given
written consent for the follow-up study, and N 451 replied, leading to a response rate
of approximately 70% for the second survey. Additionally, interviews for determining
inter-method validity were conducted with 35 selected individuals that are reported in
detail elsewhere (Kohls, 2004; Kohls et al., 2008).

Mental Health, Religion & Culture

Spiritual practice was operationalized as regular practice of any one spiritual discipline
such as meditation, prayer, contemplation, thai chi, or chi gong, or several kinds of yoga
techniques. Subjects were assigned to the spiritually practising sample if they had
answered the questions Do you practise meditative or spiritual techniques on a regular
basis? in the positive. All individuals gave informed consent prior to the participation in
the study.

Measures
Exceptional Experiences Questionnaire (EEQ)
This is a 25-item instrument developed by us for the measuring exceptional and spiritual
experiences (Kohls, 2004; Kohls, Hack & Walach, 2008; Kohls & Walach, 2006).
Every item of the EEQ describes a potential exceptional experience, and respondents are
requested to consider both frequency and individual evaluation of these experiences as
additional information. A principal-component factor analysis, which was based on the
prevalence data, extracted four factors that explain 49% of the variance. The first factor
contains positive spiritual experiences (item example: I am illumined by divine light and
divine strength), the second factor describes experiences of ego loss and deconstruction
(item example: My world-view is falling apart), the third factor includes psychopathological experiences (item example: I clearly hear voices, which scold me and make
fun of me, without any physical causation), and the fourth factor is pertaining to
visionary dreams (item example: I dream so vividly that my dreams reverberate while I
am awake). The instrument shows adequate discriminant validity with Sense of
Coherence, Social Support, as well as Mental Distress and in some aspects convergent
validity with Transpersonal Trust. The analysis of first-order correlations between
our questionnaire and other scales confirms the hypothesis that spiritual and psychopathological experiences represent different classes of experiences, and that they are
separated by our questionnaire. The 25-item short form of the instrument shows good
psychometric properties (range for Cronbachs alpha: r 0.670.89, range for testretest
reliability after 6 months r 0.660.87).
Transpersonal Trust Scale (TPV)
This is an 11-item scale with good psychometric properties measuring one
dimension of trust in the processes of life, in some larger purpose of life or some higher
being like God (Belschner, 2000, 2001). The scale has been gauged in a representative
sample of the German population (Albani et al., 2003) and has been successfully
used to predict improvement of therapy in a psychotherapeutic inpatient setting
(Belschner, 2003). Two examples for representative items are I feel connected
with a higher reality/with a higher being/with God. Even in hard times I can trust
this reality. and Sometimes in my life I have the impression that I am led by a
higher insight.
Sense of Coherence (SOC 13)
This is a concept originally introduced by Antonovsky to describe whether a person finds
their environment and life circumstances understandable, manageable, and predictable
(Antonovsky, 1993; Langius, Bjorvell, & Antonovsky, 1992). Within health psychology,
SOC is regarded as an important source of resistance against and resilience towards

N. Kohls et al.

various stressors and frustrations in life. Sense of coherence has frequently been associated
with spirituality or has been used as a measure for gauging spirituality (Delgado, 2007).
For the sake of parsimony, we used the newly constructed, validated, and gauged German
short-form version with 13 items (Schumacher, Gunzelmann, & Brahler, 2000;
Schumacher, Wilz, Gunzelmann, & Brahler, 2000). Two examples for representative
items are Do you have the feeling that you are being treated unfairly? and Until now
your life has had: Scale: 1 no clear goals or purpose and 7 very clear goals
and purpose.
Social Support (F-SoZu)
Social Support is one of the most important constructs predicting health outcomes
and quality of life in a variety of diseases (Barker & Pistrang, 2002; Hogan,
Linden, & Najarian, 2002; Uchino, Cacioppo, & Kiecolt-Glaser, 1996). We measured it
using the 14-item short form of one of the most widely used German scales
(Fydrich, Sommer, & Brahler, 2002). Two sample items are There are people that
stand by me both in good times and bad times and There are people who accept me
without limitation.
Brief Symptom Inventory (BSI)
The 53-item short form of the Symptom Checklist (SCL 90) is one of the most widely used
screening instruments to briefly assess psychological disturbances on nine subscales
(Derogatis & Melisaratos, 1983). It uses a frequency rating of common symptoms of
disturbances to assess whether psychiatrically relevant symptoms of distress are present.
We used the newly developed abbreviated German version (BSI) which gives one Global
Severity Index (GSI) of distress (Franke, 1995; Klaiberg, 2002).

Analytic plan
In this paper, data collected at the first point of measurement are analysed, and with
regard to the EEQ only prevalence data have been used.3 For describing differences
between the spiritually practising and non-practising sample, we used SPSS 11.0 for
calculating t-tests for independent samples (p 5 0.01). For descriptive purposes, Cohens d
as a measure of effects size was also computed.
The SEM models were estimated with AMOS 4.01 software using the maximum
likelihood minimization (Bollen, 1989). In order to allow for model estimation despite the
presence of missing data within the data matrix, means and intercepts were modelled as
elements of the covariance matrix. We assessed the model fit using the ratio of chi-square
value to degrees of freedom (CMINI/df), the Comparative Fit Indices (CFI), the
TuckerLewis Index (TLI), and the root mean square error of approximation (RMSEA)
(Kline, 2005). For the SEM analysis, no ad hoc changes, e.g., correlation of error terms to
improve model fit, have been made.
Applying SEM, we determined four models to investigate the straightforward
pathways from exceptional experiences, social support, sense of coherence, and
transpersonal trust as exogenous variables to mental distress as measured by the BSI as
endogenous variable. Each model was tested both for the spiritually practising SP and
NSP. Thereby, each of the applied constructs has been operationalized as a latent variable
as follows.

Mental Health, Religion & Culture

Exogenous variables
EEQ (multivariate model)
SEM analysis was based on the prevalence data, and we used all items of the 25-item
version for estimating the latent factor construct. Hence, factors 13 are estimated by
seven items, whereas factor 4 is estimated by four items. Intercorrelations between all four
latent factor constructs were also permitted in the SEM model in order to accommodate
correlations between the four factors found in the PCA analysis (Kohls, 2004; Kohls
& Walach, 2006).
TPV, SOC, F-SoZu (univariate models)
As all constructs were operationalized as one-dimensional concepts by the short scales,
we used all respective items of the scale for estimating the respective latent construct
variable, i.e., 11 items for transpersonal trust, 13 for sense of coherence, and 14 for
social support.

Endogenous variable
BSI: We calculated the means for the nine subscales of the BSI and used them as a basis
for estimating the Global Severity Index (GSI) as a latent variable.

Testing for moderating effects

In order to test for moderating effects of regular spiritual practice, relevant pathways
from the four factors of the EEQ towards psychological distress were tested by a
multigroup-comparison approach. With this approach, we tested whether group-invariant
parameter constraints on the unstandardized b-regression weights result in a significant
decrease in model-fit, thus indicating group-specific parameter values (Arbuckle
& Wothke, 2003; Bollen, 1989; Homburg & Giering, 2001; Kline, 2005). In the Results
section, both the standardized beta regression weight and the unstandardized
b-coefficients are reported. To allow for the interpretation of absolute effect sizes and
for the sake of graphicness, in Figures 1 and 2 only the standardized beta coefficients have
been reported.

Results
Mean differences in sample characteristics
Table 2 depicts the mean characteristics for the SP and the NSP sample for the EEQ-25,
TPV-11, SOC-13, FSoZu-14, and the BSI-53 as well as results from an independent
samples t-test and effect sizes using Cohens d.
With regard to the EEQ, the SP sample reports both more spiritual
experiences (factor 1) and more experiences of ego loss/deconstruction (factor 2) as well
as visionary dream events (factor 4) than the NSP group. There is also a significant
interdifference for transpersonal trust. In sum, both the EEQ and the TPV
can discriminate between spiritual practising and non-practising individuals, whereas
SOC-13, F-SoZu, and BSI-53 cannot make this distinction (Kohls, 2004; Kohls &
Walach, 2006).

Figure 1. SEM analysis for SP sample modelling pathways from exceptional experiences to mental distress. Note: Rectangles indicate observed indicator
variables for the EEQ and computed subscales for the BSI. Ovals indicate unobserved latent variables. Numbers printed in bold next to the single-headed
arrows indicate standardized regression weights (for unstandardized b-regression weights, see Table 3), and numbers next to the double-headed arrows
indicate intercorrelations between the factors of the EEQ. Numbers at the upper right corner next to rectangles indicate squared multiple correlation
coefficients.

10
N. Kohls et al.

Figure 2. SEM analysis for NSP sample modelling pathways from exceptional experiences to mental distress. Note: see Figure 1.

Mental Health, Religion & Culture

11

0.35
0.66
0.70

0.30
1.44
4.23

0.52

4.90
0.38

0.78

0.60

0.66

1.22

4.38

0.85

SD

1.75

0.53

4.79

4.43

3.03

1.26

0.24

0.81

0.83

0.44

0.85

0.60

1.01

0.74

0.30

0.61

0.72

SD

Spiritually non-practising
(N 299)

0.70

0.09

0.33

50.01

50.01

50.05

50.01

50.01

p (independent samples t-test)

0.02

0.13

0.08

1.19

0.24

0.17

0.62

1.08

Cohens d

Notes: Parts of this table are taken from Kohls, N., & Walach, H. (2006). Exceptional experiences and spiritual practicea new measurement approach.
Spirituality and Health International, 7(3), 125150. John Wiley & Sons Limited. Reproduced with permission.

BSI: Mental Distress

SOC: Sense of Coherence

F-SoZu: Social Support

(low)
(high)
(low)
(high)
(low)
(high)
(low)
(high)
(low)
(high)
(low)
(high)
(low)
(high)
(low)
(high)

EEQ: Prevalence
Positive Spiritual Experiences
EEQ: Prevalence
Loss of Ego/Deconstruction
EEQ: Prevalence
Psychopathology
EEQ: Prevalence
Visionary Dreams
TPV: Transpersonal Trust

0
4
0
4
0
4
0
4
1
5
1
5
1
7
0
4

Range

Characteristic

Spiritually practising
(N 350)

Table 2. Mean characteristics for the spiritually practising and spiritually non-practising sample.

12
N. Kohls et al.

Mental Health, Religion & Culture

13

SEM analysis for modelling pathways from exceptional experiences on mental distress
In a first step, the pathways from the four factors of the EEQ to mental distress (GSI) were
investigated. Figures 1 and 2 depict the SEM model for the SP and the NSP sample
analysing pathways from the four factors of the EEQ on the GSI.
The overall fit for the model was satisfactory for both samples. Although chi-square
values indicate significant differences between the observed and model implied covariance
matrix (2 [1034 df] 2103.058; p 5 0.001), according to the measures of approximate fit
(CMIN/df 2.034; CFI 0.953, TLI 0.946, RMSEA 0.040) a sufficiently close
approximation of empirical associations is achieved by the model (Hair et al., 2004;
Kline, 2005).
While the four factors of the EEQ explained 53% of variance in psychological distress
as measured by the GSI for the NSP sample, the respective amount explained for the SP
sample was only 28%.

SEM analysis for modelling pathways from transpersonal trust, social support, and sense
of coherence on mental distress
Similar to the SEM models depicted in Figures 1 and 2, we also calculated SEM models
analysing unidimensional pathways from transpersonal trust, social support, and sense of
coherence on the GSI for each sample separately. The results as well as the fit indices are
depicted in Table 3 for both the SP and NSP sample.
The overall index fits for all three models, which were not trimmed for a better fit, were
moderate for both samples but can still be regarded as satisfactory. The sense of coherence
is able to explain the highest amount of variance in the GSI (SP: R2 0.50;
NSP: R2 0.61) followed by social support (SP: R2 0.17; NSP: R2 0.20). In contrast,
there are no crucial pathways from transpersonal trust to the GSI (SP: R2 0.04;
NSP: R2 0.00).

Testing moderating effects of regular spiritual practice vs. lack on practice on structural
path coefficients
A more detailed analysis of the SEM model focusing on the structural path coefficients
reveals that for both samples a negative structural path coefficient points from the positive
spiritual experiences factor towards the GSI (SP: 1 0.32; b1 0.11; p 5 0.001;
NSP: 1 0.49; b1 0. 31; p 5 0.001) as opposed to the remaining three factors. The two
positive structural path coefficients concerning ego loss and psychopathology are also
smaller within the SP sample (deconstruction/ego loss for SP: 2 0.28; b2 0.11;
p 0.006; NSP: 2 0.35; b2 0.21; p 0.056; psychopathology: SP: 3 0.37, b3 1.21;
p 0.002; NSP: 3 0.66; b3 2.19; p 0.023). Only the path coefficient pertaining to
visionary dream experiences is, although generally speaking weak, larger within the SP
sample (SP: 4 0.06; b4 0.03; p 0.394; NSP: 4 0.02; b4 0.01; p 0.78).
Moderating effects of regular spiritual practice on the pathways from exceptional
experiences on psychological distress were to be expected, because spiritual practice can be
understood as a means for facilitating spiritual experiences. A visual ad hoc comparison of
the two SEM models also suggests that the influential pathways made up by positive
spiritual experiences, experiences of ego loss/deconstruction and psychopathological
experiences are distinctly buffered within the SP sample. As potential moderating effects
may help in understanding the intrapersonal mechanisms associated with regular spiritual

Chi-square [482 df] 1549.485

p 5 0.001, CMIN/df 3.215;
CFI 0.965, TLI 0.963,
RMSEA 0.060

Chi-square [458 df] 1396.971

p 5 0.001, CMIN/df 3.042;
CFI 0.978, TLI 0.973,
RMSEA 0.056

Chi-square [338 df] 1002.106,

p 5 0.001, CMIN/df 2.965;
CFI 0.972, TLI 0.965,
RMSEA 0.055

Chi-square [1034 df] 2103.058,

p 5 0.001,
CMIN/df 2.034; CFI 0.953,
TLI 0.946, RMSEA 0.040

Model fit indices (general model)

R 0.50
 0.71; b 0.724; p 0.001

R2 0.17
 0.41; b 0.415; p 5 0.001

R2 0.04
 0.21; b 0.079; p 5 0.001

1 0.32 (b1 0.110; p 5 0.001)

2 0.28 (b2 0.108; p 0.006)
3 0.37 (b3 1.21; p 0.002)
4 0.06 (b4 0.028; p 0.394)

R 0.61
 0.78; b 0.903; p 0.001

R2 0.20
 0.45, b 0.250; p 5 0.001

R2 0.00
 0.02; b 0.005; p 0.717

1 0.49 (b1 0.314; p 5 0.001)

2 0.35 (b2 0.213; p 0.056)
3 0.66 (b3 2.19; p 0.023)
4 0.02 (b4 0.013; p 0.780)

R2 0.53

R2 0.28

**Chi-square difference significant at the 1% level; *chi-square difference significant at the 5% level.

Equality Constraint

Sense of Coherence
(SOC)

Equality Constraint

Social Support
(F-SoZu)

Equality Constraint

Transpersonal Trust
(TPV)

Equality Constraints for

Positive Spiritual Experiences (1)
Deconstruction/Ego Loss (2)
Psychopathology (3)
Visionary Dreams (4)
All factors (14)
Factors 13
Factors 2 and 3

Exceptional Experiences
(EEQ)

SEM for

Spiritually non-practising (N 299)

Variance explained (R2)/
(standardized unstandardized
path coefficient to GSI)

Spiritually practising (N 350)

Variance explained (R2)/
(standardized/unstandardized
path coefficient to GSI)

6.295 [
df 1]; p 0.012*

3.896 [
df 1]; p 0.048*

10.738 [
df 1]; p 0.001**

6.973 [
df 1]; p 0.008**

0.545 [
df 1]; p 0.460
1.058 [
df 1]; p 0.304
0.077 [
df 1]; p 0.078
12,342 [
df 4]; p 0.015*
12,287 [
df 3]; p 0.006**
7,241 [
df 2]; p 0.027*

Chi-square difference if
path coefficient(s)
imposed as equal

Table 3. Summary for SEM models analysing pathways to mental distress (GSI) for transpersonal trust, social support and sense of coherence.

14
N. Kohls et al.

Mental Health, Religion & Culture

15

practice, in a next step, differences between the two samples in the structural path
coefficients pointing towards GSI were tested for significance for every factor of the
exceptional experiences scale as well as the transpersonal trust, social support, and sense of
coherence scale.4 The results can be found in the far right column in Table 3. With regard
to the EEQ, upon constraining the four path coefficients for all four factors to be equal, we
found a significant difference at the p 0.05 level (
2 12,342 [
df 4]; p 0.015).
When neglecting factor 4 as a result of the small regression weight (critical ratio of the
maximum likelihood estimation for both samples 5  1.96) and only constraining factors
13 as equal, the respective chi-square difference test is significant at p 0.01
(
2 12.287 [
df 3]; p 0.006). However, when constraining only one single factor,
the chi square difference test is significant only for the positive spiritual experiences factor
(
2 6.973 [
df 1]; p 0.008), while it becomes non-significant for the experiences of
ego loss (
2 0.545 [
df 1]; p 0.460) as well as the psychopathology factor
(
2 1.058 [
df 1]; p 0.304). This is probably due to the high intercorrelation
between these two factors (SP: r23 0.67; NSP: r23 0.81), because if equality constraints
are imposed on the structural path coefficients for the two respective factors, the chisquare difference test becomes significant (
2 7,241 [
df 2]; p 0.027).
Additionally, when analysing the three unidimensional constructs, we found significant
differences only for transpersonal trust at the p 0.01 level (
2 10.738 [
df 1];
p 0.001). However, this finding is negligible from a practical point of view, as the
respective path coefficient is not significant for the NSP sample. Moreover, the structural
path coefficient from social support (
2 3.896 [
df 1]; p 0.048) as well as sense of
coherence (
2 6.295 [
df 1]; p 0.012) showed only a significant difference at the
p 0.05 level.

Discussion
In this paper, we have presented an alternative way of analysing the pathways from
exceptional experiences, transpersonal trust, social support, and sense of coherence on
psychological distress harnessing the advantages of SEM. Additionally, the amount of
variance explained in distress was compared with the well-established constructs social
support, sense of coherence, and transpersonal trust.
We have previously published a conventional analysis based on a linear regression
analysis, where we have compared two sociodemographically balanced subsamples of
spiritually practising and non-practising individuals (Kohls & Walach, 2007). In sum, the
results from the SEM analysis presented in this paper corroborate the previous findings
achieved by linear regression analysis. However, the SEM also revealed interesting new
details that are missed by the classical approach:
(1) First, the overall amount of variance in psychological distress explained by the four
factors of the EEQ was higher in the SEM analysis: Whereas the conventional
linear regression analysis was able to explain 7% or 36% of the variance in distress
in the samples with and without spiritual practice, the respective amount of
variance explained by the SEM model was 28% and 53%. The larger amount of
variance explained in the SEM model is most likely due to the fact that latent
variables were used in the model, which allow one to explicitly take the
measurement error into account. While the total amount of variance explained
differs considerably between the SEM and the linear regression analysis, the
difference in the amount of variance explained between the two samples is

16

N. Kohls et al.
comparable: it is 29% in the linear regression analysis and 25% in the SEM
analysis. This is a clear sign that the intersample difference in the amount of
variance explained in distress cannot be attributed to a methodological artefact; it
shows rather that regular spiritual practice moderates the pathways from spiritual
experiences to health.
(2) Second, the structural path coefficients of the SEM analysis differed from those
found in the linear regression analysis. It is necessary to take a closer look: When
the SEM analysis was independently computed with equality constraints
sequentially imposed4 on every factor of the EEQ, a significant intersample
difference was only found for the positive spiritual experiences factor.
Interestingly, the negative regression weight pointing from positive spiritual
experiences towards distress was higher in the spiritually non-practising sample,
thereby indicating that individuals without spiritual practice benefit more from
positive spiritual experiences. In contrast, the linear regression analysis suggested a
significant difference in the regression weights for negative spiritual experiences.
Here, the beta weight of experiences of ego loss predicting psychological distress
was not significant in the subsample of spiritually practising individuals, whereas it
was highly significant in the sample with lack of spiritual practice. However, when
equality constraints were imposed on both the negative spiritual experiences
factor and the psychopathology factor, a significant difference in the model fit was
also found in the SEM analysis. This is probably a consequence of high
intercorrelations between the EEQ ego loss and psychopathology factor of r 0.81
in the NSP and r 0.67 in the SP sample (Figures 1 and 2), which is explicitly
accounted for in our SEM model. This very likely allows the algorithm to
distribute and compensate variance across both pathways, if inequality
constraints4 are imposed on only one pathway. Thus, this finding is very likely a
consequence of accounting for intercorrelations between the predictor variables,
which werein contrast to the linear regression analysisexplicitly modelled in
our SEM analysis.
(3) It is additionally important to recall that the linear regression analysis was based
on a subset of sociodemographically matched subsamples (Kohls & Walach, 2007),
whereas the SEM analysis was based on the full sample. However, the overall
findings seem to be comparable, although the results of the SEM analysis are more
sophisticated, because they explicitly allow testing for differences in pathways.
Specifically, although regular spiritual practice seems to increase both the
frequency of positive and negative spiritual experiences, through SEM analysis,
it has become clear that spiritual practice apparently buffers the impact of a
distinct subset of exceptional experiences comprising positive and negative spiritual
as well as psychopathological experiences.

Taken together, the findings found in the SEM analysis give a much clearer and more
differentiated picture about the intrapersonal effects of spiritual practice than the
conventional regression analysis. We nevertheless acknowledge the fact that the two
samples used in the present analysis could not be matched for sociodemographic variables,
because the matching procedure that we have employed for the regression analysis was
only able to provide 100 (Kohls & Walach, 2007) or 120 (Kohls & Walach, 2008)
well-matched cases and discards the rest, which is not a number high enough for running
an SEM analysis. Thus, it would be desirable for future projects to collect enough data so
that post-hoc controlling procedures can be combined with SEM. This would necessitate a

Mental Health, Religion & Culture

17

sample larger by a factor of 3, i.e., around 18002000 cases. Nevertheless, we believe that
the SEM analysis has provided more useful insights than the conventional regression
analysis: First of all, the comparison of the path coefficients between the two samples
suggests that both the stress-annihilating impact of positive spiritual experiences and the
stress-augmenting impact of negative spiritual and psychopathological experiences are
reduced by regular spiritual practice. Thus, the SEM analysis seems to reveal a paradoxical
finding at first glance: Both the stress-annihilating impact of positive spiritual experiences
and the stress-inducing effect of experiences of ego loss and psychopathological
experiences are buffered by regular spiritual practice. Correspondingly, one could be
inclined to assume that the double-barrelled effects of regular spiritual practice on
psychological distress are in sum self-annihilating, as both positive and negative impact is
diminished by regular spiritual practice. However, individuals with a lack of spiritual
practice seem to suffer much more distress from negative spiritual and psychopathological
experiences than individuals with regular spiritual practice. This could be a hint that some
experienceslike losing oneself, losing ones coherent picture of the world, etc.that
are indicative of cognitive deconstruction, when hitting the individual unprepared, can be
detrimental, while they can be viewed and reframed in a more positive manner when met
within a spiritual context. From a psychological perspective, while altering the self concept
in a less ego-centred way, many spiritual techniques seem to buffer the impact of negative
spiritual experiences on mental distress by a gradually suspending negative impact of
deconstruction and even psychopathological experiences. The flip side of this process is
that the stress-reducing impact of positive spiritual experiences is also partially diminished.
However, in sum, the reduced distress annihilating impact of positive spiritual experiences
by no means outweighs the resilience against distress as it is derived from experiences
of ego loss.
Thus, in order to get the full picture of the intrapersonal mechanism of spiritual
practice, it is necessary to scrutinize the impact on distress of positive and negative spiritual
experiences. By taking only positive spiritual experiences into account, one might actually
find misleading if not contradictory results: if a key psychological function of regular
spiritual practice seems to be the ability to integrate exceptional experiences and thereby
particularly deconstructive experiences into the self model more easily, how can the finding
that the stress annihilating impact of positive spiritual experiences is lower in samples with
spiritual practice be properly explained? This finding alone seems at first glance not only
illogical but also completely counterintuitive at second thought. In order to make sense out
of it, one needs also to scrutinize the pathways from experiences of ego loss and distress in
a complementary way in order to fully grasp the mechanism apparently associated with
regular spiritual practice that annihilates stress: Regular spiritual practice buffers the
impact of positive and negative spiritual experiences on health.
It is noteworthy that one-dimensional scales that grasp only positive spiritual
experiencessuch as the Daily-Spiritual-Experiences Scalehave only found low
correlations with health-related parameters (Underwood, 2006; Underwood & Teresi,
2002). However, based on our findings, the assumption that only positive spiritual
experiences have beneficial effects on health would appear to be incorrect. These
suppositions may have occurred because many of the instruments used to measure
spiritual experience are unidimensional and so only appear to record positive experiences
and their impact on health. Thus, one should be wary of promoting only positive spiritual
experience as a route to well-being. Instead, for gauging the impact of spiritual experiences
on health, it seems to be helpful to assess both positive and negative spiritual experiences
as well as their impact upon distress.

18

N. Kohls et al.

The distress annihilating impact of regular spiritual practice seems thereby to be an

important mechanism for maintaining health and well-being that should not be neglected.
Comparing the amount of variance explained in distress by our multivariate model with
the three external unidimensional constructs, a relative assessment of the impact of
exceptional experiences and the moderating role of regular spiritual practice is possible:
First of all, it becomes clear that only sense of coherence was able to explain more variance
(61%) in mental distress for the spiritually non-practising sample. Recall that sense of
coherence has been conceptually criticized for inversely assessing anxiety and depression
(Geyer, 1997). Hence, it might simply be a redundant operationalization of distress.
Additionally, for transpersonal trust and social support, although statistically significant,
no comparable differences were found in the amount of variance explained between the
two samples. In direct comparison with the respective model for social support, our
multivariate model predicting the impact of exceptional and spiritual experiences on
mental distress is able to explain about two and a half times as much variance as social
support for the spiritually non-practising sample. In contrast, within the spiritually
practising sample, the amount of variance in mental distress explained by exceptional and
spiritual experiences is comparable to social support. Interestingly, the reliable and valid
questionnaire measuring transpersonal trust does not explain any variance in psychological distress, although it has the highest effects concerning differences in mean for the two
samples with regard to both effect size and difference in structural path coefficients.
Both sense of coherence (Larsson & Kallenberg, 1996; Richardson & Ratner, 2005)
and social support (Barker & Pistrang, 2002; Hogan et al., 2002; Uchino et al., 1996) are
among the best-studied constructs of health psychology and are widely accepted for their
importance, mostly for the management of chronic diseases. However, although social
support and sense of coherence are able to differentiate between a spiritually practising
and a non-practising sample, the difference in the amount of variance explained in mental
distress is much higher for the exceptional experiences scale. Hence, fostering spiritual
practice could be even more promising as a preventive or therapeutic intervention, than,
for instance, improving social support. Clearly, this possibility ought to be studied
more widely.
The effect of spiritual practice, annihilating distress, which we discovered, might be due
to psychological habituation or moderated by the psychophysiological effects of the
experiences themselves. It does not seem to be due to a particular faith, belief, cognitive
framework, or world view, although Transpersonal Trust, a construct measuring such a
framework, did in fact show differential effects in the pathways to mental distress between
our two subsamples. However, it did not explain a large amount of variance in mental
distress. Hence, our results cannot be used to argue for or against any belief system,
cognitive system, or creed. They rather open a new venue of research: the impact, effect,
and nature of positive and negative spiritual experiences and the moderating role of
spiritual practice.
Our findings imply three possibly interrelated, interpretations: First, persons with
spiritual practice seem to be able to make more sense of their spiritual experiences.
Thus, experiences are more easily integrated and accepted, which possibly buffers their
effect on mental health in both directions (positive and negative). Second, within the
spiritually practising sample, the influence on distress of negative exceptional experiences
(experiences of ego loss and psychopathological symptoms) is smaller. If experiences of
ego loss are less correlated with psychological distress, then reduced psychological distress
will possibly reduce the presence and impact of psychopathological experiences and hence
reduce the impact of the full model on psychological distress as a consequence. Thus, even

Mental Health, Religion & Culture

19

the lower explanatory power of the model for the spiritual sample is indirect evidence for
the importance of spiritual experiences for psychological health, and spiritual practice
seems to be a key factor.
Third, it seems to be experiences that are supportive, not attitudes or beliefs.
That assumption was already proposed by William James (1904). It is consistent
with recent findings that analysed the impact of spirituality in cancer patients, which
used a newly constructed scale, the FACIT-Spiritual Well-being Scale, and
differentiated Meaning/Peace from Faith, the latter being a set of cognitive attitudes.
While Meaning/Peace was predictive of quality of life, even after adjusting for other aspects
of quality of life, and emotional aspects of the disease, the influence of Faith was smaller
(Brady, Peterman, Fitchett, Mo, & Cella, 1999; McClain, Rosenfeld, & Breitbart, 2003).
Our study, although based on a sufficiently large sample to render most estimations of
the models stable, has several limitations which should not be overlooked: First, we did
not recruit a representative sample of the population and we did not utilize sample
matching in order to sustain a sufficiently high sample size. The next step would be to
validate the EEQ and replicate the results in a large representative population sample.
In the same vein, it would be necessary to study clinical populations. Second, as we had to
rely on the full sample set, we were naturally not able to control for parameters. Hence,
although our data suggest a causal impact of spiritual experiences on mental health
moderated by regular spiritual practice, for the time being our interpretation should be
regarded as a hypothesis that needs to be replicated. To investigate the direct effects of
spiritual practice on the frequency and evaluation of exceptional and spiritual experiences
as well as their impact on psychological distress, controlled trials with baseline matching
are necessary that introduce spiritual practice as a treatment variable.
We believe that the strengths of this studythe large sample size, the good
psychometric properties of the instruments used, and the modelling of intercorrelationsoutweigh weaknesses and allow us to draw valid, albeit tentative
conclusions: Spiritual practice is an important protective factor for psychological health.
Together with other aspects measured by our instrument, it explains 53% of the variance
in psychological distress for spiritually non-practising individuals but only 28% for
spiritually practising persons. Spiritual practice seems to be a buffer for positive but also
disturbing spiritual experiences. It is time to study the ramifications of this finding more
broadly, especially in clinical populations, chronically ill subjects, and the general
population at large. If our findings are replicated, we have made a start in establishing
a hitherto overlooked risk factor for well-being: lack of spiritual practice.

Acknowledgements
This work was part of NKs Ph.D. thesis at the University of Freiburg, Institute of Psychology, and
was supervised by HW. It was supported by a scholarship awarded to NK by the Institute of
Frontier Areas of Psychology and Psychohygiene (Institut fur Grenzgebiete der Psychologie und
PsychohygieneIGPP), Freiburg, Germany. We are especially grateful to its late director, Johannes
Mischo, for his support. We dedicate this work to his memory. We thank Eberhard Bauer und
Dieter Vaitl for their support of this study, and Cosima Friedl and Gudrun Kress for helping with
the revision of the instrument. HW and NK are sponsored by the Samueli Institute, Alexandria,
USA. NK collected the data, conducted the psychometric analysis and the statistical analysis of the
Structural Equation Models, and participated in the interpretation of the results, as well as in writing
and revising the manuscript. HW developed the general idea of the study and supervised it.
He suggested details for the final analysis, participated in interpreting the results, and wrote parts of
the manuscript. MW provided support for the SEM analysis.

20

N. Kohls et al.

Notes
1. Although it is widely accepted that spiritual experiences are completely dependent on social and
religious context (Katz, 1992), these arguments do not seem to be in line with the
phenomenology of at least some spiritual experiences (Hufford, 2005). We will assume in the
following that at least some spiritual experiences can be seen as prior to and foundational of
formal religions, and not necessarily following from religious doctrine.
2. We also collected data from a small clinical sample that is not reported in this paper, because it
cannot be regarded as representative for clinical populations, mainly because it consisted of
spiritually practising or at least spiritually interested individuals with mental disorders
(see Kohls, 2004 and Kohls & Walach, 2006 for details).
3. We restrained from analysing the evaluation data of the EEQ, because N for evaluation data
varies for each item depending on the item difficulty based on prevalence. This is because
we have asked our participants to exclusively assess the evaluative component of an
exceptional experience if they had personally encountered this experience (Kohls, 2004; Kohls
& Walach, 2006).
4. We tested every path coefficient for significant intersample differences by defining a nested
model where the respective path coefficient(s) is (are) restricted by equality constraints between
the two subsamples. Correspondingly, the only difference between the nested and the general
model is that the respective path coefficients are imposed to be equal for both samples in the
nested model, whereas they are allowed to differ in the general model. Thus, technically
speaking, the general model has one (or n) degree(s) of freedom more than the nested model,
because one parameter has to be estimated instead of two (or n). Due to that fact, the chi-square
value will always be higher for the nested model. The question is whether the difference in the
chi-square value between the general and nested model is statistically significant. The model fits
of the two models can be easily tested for significance by an overall chi-square difference test
which is compared with a chi-square distribution with one (or n)degree(s) of freedom. For
example, with regard to one degree of freedom difference the critical value for the difference in
chi-square at the p 0.05 level is 3.84 and at the p 0.001 level is 6.63. However, only if the
imposed constraints lead to a significant decrease in data fit as indicated by a significant
difference of the chi-square test (Homburg & Giering, 2001), can the corresponding subgroup
specific model components be considered important. A significant difference in the chi-square
values can then be interpreted as an indicator for intersample differences in the respective
structural path coefficient(s), where equality constraints have been imposed. We compared the
nested model with the general model for each construct using both p 01 and p 0.05 levels.

References
Albani, C., Bailer, H., Blaser, G., Geyer, M., Brahler, E., & Grulke, N. (2003). Psychometrische
Ueberprufung der Skala Transpersonales Vertrauen (TPV) in einer reprasentativen
Bevolkerungsstichprobe [A psychometric investigation of the transpersonal trust scale in a
sample representative of the German population]. Transpersonale Psychologie und
Psychotherapie, 9(1), 8698.
Antonovsky, A. (1993). The structure and properties of the Sense of Coherence Scale. Social Sciences
& Medicine, 36, 725733.
Arbuckle, JL, & Wothke, W. (2003). AMOS 5.0 users guide. Chicago: Small-Waters Corporation.
Barker, C., & Pistrang, N. (2002). Psychotherapy and social support: Integrating reserach on
psychological helping. Clinical Psychology Review, 22, 361379.
Belschner, W. (2000). Integrale Gesundheit: Zur Integration einer Transpersonalen Psychologie in
die Akademische Psychologie [Holistic health: Working towards an integration of transpersonal
psychology into academic psychology]. In W. Belschner, & P. Gottwald (Eds.), Gesundheit und
Spiritualitat (pp. 71118). Oldenburg: BIS.
Belschner, W. (2001). Tun und Lassen: Ein komplementares Konzept der Lebenskunst [Acting and
resting: A complementary concept for balancing life]. Transpersonale Psychologie und
Psychotherapie, 7(2), 85102.

Mental Health, Religion & Culture

21

Belschner, W. (2003). Ergebnisse empirischer Forschung zur Transpersonalen Psychotherapie.

[Transpersonal psychotherapy some empirical results]. In J. Galuska (Ed.), Den Horizont
erweiternDie transpersonale Dimension in der Psychotherapie. [Broadening ones horizons the
transpersonal dimension within psychotherapy] (pp. 93135). Berlin: Verlag Ulrich Leutner.
Bollen, K.A. (1989). Structural equations with latent variables. New York: Wiley.
Brady, M., Peterman, A., Fitchett, G., Mo, M., & Cella, D. (1999). A case for including spirituality
in quality of life measurement in oncology. Psychooncology, 8, 417428.
Culliford, L. (2002). Spiritual care and psychiatric treatment: An introduction. Advances in
Psychiatric Treatment, 8(4), 249258.
Delgado, C. (2007). Clinical scholarship sense of coherence, spirituality, stress and quality of life in
chronic illness. Journal of Nursing Scholarship, 39(3), 229234.
Derogatis, L., & Melisaratos, N. (1983). The Brief Symptom Inventory: An introductory report.
Psychological Medicine, 13(3), 595605.
Devon, B. (2005). Methodological pitfalls in the study of religiosity and spirituality. Western Journal
of Nursing Research, 27(5), 628.
Edwards, A.C. (2003). Potential sources of bias in questionnaires in researching transpersonal
psychology. Transpersonal Psychology Review, 7(1), 315.
Farb, N.A.S., Segal, Z.V., Mayberg, H., Bean, J., McKeon, D., Fatima, Z., et al. (2007). Attending
to the present: Mindfulness meditation reveals distinct neural modes of self-reference. Social
Cognitive and Affective Neuroscience Advance Access, 2(4), 313322.
Franke, G. (1995). Die Symptom-Checklist von Derogatis (SCL-90-R) [Deragatis Symptom
Checklist]. Weinheim: Beltz.
Fydrich, T., Sommer, G., & Brahler, E. (2002). F-SoZUFragebogen zur sozialen Unterstutzung
(F-SoZU). [F-SoZu: A scale for measuring social support]. In E. Brahler, J. Schumacher, &
B. Strau (Eds.), Diagnostische Verfahren in der Psychotherapie [Diagnostic instruments for
psychotherapy] (pp. 150153). Gottingen: Hogrefe.
George, L.K., Larson, D.B., Koenig, H.G., & McCullough, M.E. (2000). Spirituality and
health: What we know, what we need to know. Journal of Social and Clinical Psychology, 19,
102116.
Geyer, S. (1997). Some conceptual considerations on the sense of coherence. Social Science and
Medicine, 44(12), 17711779.
Grossman, P., Schmidt, S., Niemann, L., & Walach, H. (2004). Mindfulness based
stress reduction and health: A meta-analysis. Journal of Psychosomatic Research, 37,
3543.
Hair, J.F., Anderson, R.E., Tatham, R.L., & Black, W.C. (2004). Multivariate data analysis.
Upper Saddle River, NJ: Prentice-Hall.
Hayes, S.C., & Shenk, C. (2004). Operationalizing mindfulness without unnecessary attachments.
Clinical Psychology: Science and Practice, 11, 249254.
Hill, P.C., & Pargament, K.I. (2003). Advances in the conceptualization and measurement of religion
and spirituality: Implications for physical and mental health research. American Psychologist, 58,
6474.
Hogan, B.E., Linden, W., & Najarian, B. (2002). Social support interventions: Do they work?.
Clinical Psychology Review, 22, 381440.
Homburg, C., & Giering, A. (2001). Personal characteristics as moderators of the relationship
between customer satisfaction and loyaltyan empirical analysis. Psychology and Marketing, 18,
4366.
Hufford, D.J. (2005). Sleep paralysis as spiritual experience. Transcultural Psychiatry, 42, 1145.
James, W. (1904). The varieties of religious experiencesa study in human nature. London:
Longmans: Green & Co.
Katz, S.T. (1992). Mysticism and Language. New York: Oxford University Press.
Klaiberg, A. (2002). Brief Symptom Inventory von L.R. Derogatis. In E. Brahler, H. Holling,
D. Leutner, & F. Peterman (Eds.), Brickenkamp Handbuch psychologischer und padagogischer
Tests (pp. 822825). Gottingen: Hogrefe.

22

N. Kohls et al.

Kline, R.B. (2005). Structural equation modelling. New York: Guilford Press.
Koenig, H., McCullough, M., & Larson, D. (Eds.). (2001). Handbook of religion and health.
New York: Oxford University Press.
Kohls, N. (2004). Aussergewohnliche ErfahrungenBlinder Fleck der Psychologie? Eine
Auseinandersetzung mit aussergewohnlichen Erfahrungen und ihrem Zusammenhang mit geistiger
Gesundheit [Exceptional experiences: The blind spot of psychology? Investigating the relationship between exceptional experiences and health]. Munster: Lit-Verlag.
Kohls, N., Hack, A., & Walach, H. (2008). Measure the unmeasurable by ticking boxes and actually
opening pandoras box? Mixed methods research as a useful tool for thinking out of
the box while investigating exceptional human experiences. The Archive for the Psychology of
Religion, 30, 155187.
Kohls, N., & Walach, H. (2006). Exceptional experiences and spiritual practice a new measurement
approach. Spirituality and Health International, 7(3), 125150.
Kohls, N., & Walach, H. (2007). Psychological distress, experiences of ego loss and
spirituality: Exploring the effects of spiritual practice. Social Behavior and Personality, 35(10),
13011316.
Kohls, N., & Walach, H. (2008). Validating four standard scales in spiritually practicing and nonpracticing samples using propensity score matching European Journal of Psychological
Assessment, 24(3), 165173.
Kohls, N., Walach, H., & Lewith, G. (submitted). Mindfulness buffers the distress of experiences of
ego loss in a sample of chronically ill patients.
Langius, A., Bjorvell, H., & Antonovsky, A. (1992). The sense of coherence concept and its relation
to personality traits in Swedish samples. Scandinavian Journal of Caring Sciences, 6, 165171.
Larsson, G., & Kallenberg, K. (1996). Sense of coherence, socioeconomic conditions and
healthInterrelationships in a nation-wide Swedish sample. European Journal of Public
Health, 6(3), 175180.
Levin, J.S., Chatters, L.M., Ellison, C.G., & Taylor, R.J. (1996). Religious involvement, health
outcomes, and public health practice. Current Issues in Public Health, 2, 220225.
MacDonald, D.A., LeClair, L., Holland, C.J., Alter, A., & Friedman, H.L. (2002). A survey of
measures of transpersonal constructs. In H.L. Friedman, & D.A. MacDonald (Eds.), Approaches
to transpersonal measurement and assessment (pp. 1579). San Francisco: Transpersonal
Institute.
McClain, C., Rosenfeld, B., & Breitbart, W. (2003). Effect of spiritual well-being on end-of-life
despair in terminally-ill cancer patients. Lancet, 361, 16031607.
Miller, W.R., & Thoresen, C.E. (2003). Spirituality, religion, and health: An emerging field.
American Psychologist, 58, 2435.
Powell, L.H., Shahabi, L., & Thoresen, C.E. (2003). Religion and spirituality: Linkages to physical
health. American Psychologist, 58, 3652.
Reibel, D.K., Greeson, J.M., Brainard, G.C., & Rosenzweig, S. (2001). Mindfulness-based stress
reduction and health-related quality of life in a heterogeneous patient population. General
Hospital Psychiatry, 23, 183192.
Richardson, C., & Ratner, P. (2005). Sense of coherence as a moderator of the effects of stressful life
events on health. Journal of Epidemiology and Community Health, 59, 979984.
Sawatzky, R., Ratner, P., & Chiu, L. (2005). A meta-analysis of the relationship between spirituality
and quality of life. Social Indicators Research, 72(2), 153188.
Schumacher, J., Gunzelmann, T., & Brahler, E. (2000). Deutsche Normierung der Sense of
Coherence Scale von Antonovsky [Antonovskys Sense of Coherence Scale A German
validation study]. Diagnostica, 46, 208213.
Schumacher, J., Wilz, G., Gunzelmann, T., & Brahler, E. (2000). Die Sense of Coherence Scale von
AntonovskyTeststatistische Uberprufung in einer reprasentativen Bevolkerungsstichprobe
und Konstruktion einer Kurzskala [Antanovskys Sense of Coherence Scale: Psychometric
investigation in a sample representative of the German population and construction of a short
form]. Psychotherapie, Psychosomatik, Medizinische Psychologie, 50, 472482.

Mental Health, Religion & Culture

23

Smucker, C. (1996). A phenomenological description of the experience of spiritual distress. Nursing

Diagnosis, 7(2), 8191.
Thoresen, C. (1999). Spirituality and health: Is there a relationship? Journal of Health Psychology,
4(3), 291300.
Thoresen, C., & Harris, A. (2002). Spirituality and health: Whats the evidence and whats needed?
Annals of Behavioral Medicine, 24(1), 313.
Uchino, B.N., Cacioppo, J.T., & Kiecolt-Glaser, J.K. (1996). The relationship between social
support and physiological processes: A review with emphasis on underlying mechanisms and
implications for health. Psychological Bulletin, 119, 488531.
Underwood, L.G. (2006). Ordinary spiritual experience: Qualitative research, interpretive guidelines,
and population distribution for the daily spiritual experience scale. Archive for the Psychology of
Religion, 28, 181218.
Underwood, L.G., & Teresi, J. (2002). The Daily Spiritual Experience Scale: Development,
theoretical description, reliability, exploratory factor analysis, and preliminary construct validity
using health-related data. Annals of Behavioral Medicine, 24(1), 2233.
Walach, H., Gander, M., & Kohls, N. (in press). Mindbody-practices in integrative medicine.
In J. Giordano (Ed.), Complementary and integrative approaches to chronic pain. Boston: Jones
& Bartlett.
Weaver, A.J., & Koenig, H.G. (2006). Religion, spirituality, and their relevance to medicine: An
update. American Family Physician, 73(8), 13361337.