Validation of the Interreligious Forms of the Centrality of Religiosity Scale in Taiwan: Perspectives from Selected University Students

: This study validated three versions of the Centrality of Religiosity Scale (CRSi-7, -14, and -20), which propose an adequate assessment tool for the diversity of religious belief systems co-existing in Taiwan’s society. The sample ( N = 331) was drawn from the selected undergraduate university students of the country. Descriptive statistics and Cronbach’s alpha values were calculated for the ﬁve subscales (ideology, intellect, experience, private and public practice) of the Centrality of Religiosity Scale. The factor structure of the interreligious Centrality of Religiosity Scale was tested using conﬁrmatory factor analysis. The current study utilized the CRSi-14 model 3 as the basis for later analysis. All items have loaded signiﬁcantly in the different subscales with internal consistency within the acceptable values. Findings show that the selected Taiwanese youth are religious.


Introduction
The island nation of Taiwan's constitution inscribes the right to freedom of religion, which promotes the flourishing of a diversity of religious beliefs and practices. Taiwan is considered a religious and spiritual sanctuary due to highly diversified religious beliefs and practices, such as Buddhism, Taoism, Yiguandao, Christianity, Mormonism, Islam, Eastern Orthodoxy, as well as various folk religions and native sects (Tourism Bureau 2020). The majority of the 23 million people in Taiwan practice a combination of Buddhism and Taoism, colored with a Confucian worldview, which often is collectively termed as Chinese folk religion.

Brief History of Religions in Taiwan
Clart and Jones (2003) edited a book on religion in modern Taiwan, tracing the tradition and innovation in a changing society. Prior to the 17th century, Taiwan was inhabited by the aboriginal groups of the Austronesian family, who traditionally practiced animism. Protestantism was introduced to the Taiwan aborigines in 1624 by the Dutch colonists (Chiu 2008), and Catholicism was introduced later on by the Spanish colonists (Mateo and Eugenio 2002).
The Han Chinese had begun to arrive in Taiwan in the 1660s and formed the Taiwanese Chinese ethnic group, with the simultaneous exchanges taking place between the indigenous religion of the aborigines and the Chinese traditional religions (Clart and Jones

Taiwan's Multifarious Religious Ecology
In many respects, Taiwan is a sort of proving-ground in which various religions freely adapt to the contextualized realities (Vermander 1999, p. 129). The Taiwanese government recognizes about 26 religions, including Buddhism, Taoism, Yiguandao, folk religions, Protestantism, Catholicism, Mormonism, Islam, and the Church of Scientology. The American Institute in American Institute in Taiwan (2020) on religious freedom in Taiwan reported that based on a comprehensive study conducted in 2005 by the Religious Affairs Section of the Ministry of the Interior, about 81% of the population adhere to some sort of religion, and about 19% of the population is unaffiliated or non-religious. The estimates suggested that 35% of the Taiwanese population considers itself Buddhist, followed by 33% percent Taoist and folk religion, 3.5% Yiguandao, 2.6% Protestants, and 1.3% Catholics. Although the Ministry of the Interior has not tracked population data on religious groups since the 2005 study, it stated that this estimate remains mostly unchanged.
Pew Research projections for 2020 on Global Religious Futures (Pew Research 2020) indicated that the majority of the people in Taiwan are affiliated with folk religions (43.8%), followed by Buddhism (21.2%), and other religions (15.5%). About 13.7% of Taiwanese are not affiliated with any religion. The non-affiliated population includes atheists, agnostics, and those who do not identify with any particular religion. The Christian affiliation (5.8%) includes Catholics and all other Christian denominations. A 2019 report of the Ministry of Ministry of Interior (2020) on the general condition of the religions in Taiwan indicated Taoism and folk religion as the prominent religions, followed by Buddhism, Yiguandao, various Christian denominations, and Catholicism.

Dialogical Religious Ecosystem and Religious Consumerism in Taiwan
In its various forms, Buddhism is the most followed religion in Taiwan and is placed within a complex religious milieu that includes Taoism, folk religions, Protestantism, Catholicism, and various new religious movements. Buddhism engages with this diverse religious ecosystem, showcasing its contextualized practices within a broader Taiwanese and global religious landscape. Several interfaith dialogues and mutual influence activities, such as the case of Tzu Chi organization, were established by the Buddhist nun Cheng Yen after she met with some Catholic nuns (Pacey 2020).
In Taiwan, the cultural influence of Christianity goes beyond charitable work; Christian concepts and values appeal to the general public (Vermander 1999, p. 129). Major religions and emerging small-scale religious movements with their emphasis on meeting spiritual needs and providing psychological support have been attracting simple people, especially millennials (Vermander 1999, p. 133).
In addition to mutual dialogue and influence among various religions, there appears to be a form of religious consumerism (Vermander 1999, p. 131), religious economy (Lu et al. 2008) or religious capital (Hu and Leamaster 2013, p. 258) in Taiwan, to the extent that creeds and beliefs, and religious and spiritual practices are adapted and consumed competitively (Vermander 1999, p. 131). Lu and colleagues (2008, p. 140) suggested that the religious economy model assumes a religious market in which religious organizations and institutions compete to retain the faithful or even attract the new adherents by providing religious products and services.
A 2019 report of the Ministry of Ministry of Interior (2020) said that there are 12,279 temples and 2896 churches in Taiwan. Vermander (1999, p. 131) said that everything religious seems to induce Taiwanese people to assert the position of a community, preserve "one's face," secure moral leadership, engage in religious practices, and proselytize activities, organizing religious ceremonies, and investing in temple construction and ornamentation. Hence, at the economic level, such a consumeristic practice affects religious activities and increases the vitality of religious markets. At the personal level, the religious competition seems to account for high levels of religiosity (Lu et al. 2008, p. 140).
Given this long history of religions in Taiwan, the scope for the existence of diverse religious demographics, constitutional right to freedom of religion, the continuous flourishing of religious beliefs and practices, records of successful incidents of interreligious dialogues and mutual influence, and the impact of religious ecology on the public system and in the personal life of adherents, it is worth validating various versions of the Centrality of Religiosity Scale in a Taiwanese religious context.

Centrality of Religiosity Scale
The versions of the Centrality of Religiosity Scale (CRS) of the present validation are the interreligious forms with seven items (CRSi-7), 14 items , and 20 items (CRSi-20) focusing on a group of Taiwanese university students with a mean age of 20.59 years old. The scale is interreligious in scope, which is appropriate considering the diversity of religious belief systems co-existing in Taiwan society. Huber and Huber (2012) describe the CRS as an instrument that measures the "centrality, importance or salience of religious meanings in personality" (Huber and Huber 2012, p. 711). The religious life is represented in five core dimensions: ideology, intellect, public practice, private practice, and religious experience (Huber and Huber 2012). "From a psychological perspective, the five core dimensions can be seen as channels or modes in which personal religious constructs are shaped and activated. The activation of religious constructs in personality can be regarded as a valid measure of the degree of religiosity of an individual" (p. 710). The CRS rests on two assumptions: "(1) The measurement of the general intensity of the five core dimensions allows a representative estimation of the frequency and intensity of the activation of the personal religious construct system and (2) The probability of a central position of the religious construct-system in personality increases with the overall intensity and frequency of its activation" (Huber and Huber 2012, p. 715).
The CRS also has three validated versions for Abrahamic religions: five items (CRS-5), 10 items (CRS-10), and 15 items (CRS-15). However, for other religions, specific modifications of the CRS were developed to reflect their openness to polytheistic concepts and practices. Hence the term "God or something divine" in Items #2, #5, #10, and #15 of the CRS were extended to "God, deities, or something divine." "This enhances the scope of personal religious constructs that can be measured in the ideological and experiential dimensions" (Huber and Huber 2012, p. 719). These changes are reflected in CRSi-7, CRSi-14, and CRSi-20, which bring the total number of CRS versions to six.

The Centrality of Religiosity Scale Versions in Different Contexts
The Centrality of Religiosity Scale has been validated and applied in numerous studies. However, it is still interesting to examine current validations of CRS versions drawn from different religious and sociocultural contexts. In this section, a particular focus on the Asian context will be discussed. In Hong Kong, researchers Lee and Kuang (2020) validated the CRS Chinese version among primary and secondary school teachers. They found that "the single-factor solution of five items (CRS-5) had better fit indices than the seven-item version (CRSi-7), which, in turn, was better than CRS-15 with a five-factor solution" (p. 1). In the Philippines, del Castillo et al. (2020) validated the CRSi-7, -14, and -20 among Filipino youth and found out that they show a good global fit. However, the researchers defer to the CRSi-20 model with correlated factors since it is a simpler model. In Pakistan, Abbasi et al. (2019) translated and cross-language validated the Centrality of Religiosity Scale from English (source language) to Urdu (target language). They discovered that the "Urdu translation of CRS has satisfactory alpha reliability (α = 0.77) and the three subscales [renamed] "exclusive," "inclusive," and "collective" religious beliefs significantly correlated with each other." In addition, the confirmatory factor analysis of their research showed that the final model of CRS in the Urdu language with 11 items and three dimensions is the best fit for Pakistani culture to assess the level of religiosity. In India, Dua et al. (2020) presented the adaptations required in CRS-20 for the Indian population. The researchers found out that CRS's use in the Indian context requires some adaptations to suit the various religious practices. Furthermore, in Japan, Kambara et al. (2020) examined the psycholinguistic feature of religious words associated with the CRS dimensions. Their findings suggest that psycholinguistic features of religious words contribute to the detection of religiosity.

Measuring the Centrality of Religiosity Among Taiwanese Youth
The Centrality of Religiosity Scale has been applied in more than 100 studies in religious studies, sociology of religion, and psychology of religion in different countries with more than 100,000 participants. Moreover, the CRS items have been translated into more than 20 languages (Huber and Huber 2012). The scale has also been applied in various studies in Asia. However, looking at the list of countries because the CRS was utilized, the researchers found out that the interreligious forms of the Centrality of Religiosity Scale (CRSi-7, -14, -20) has not been validated in the Taiwanese milieu and utilized to illuminate the salience of religion in the personality of the Taiwanese youth. By validating the CSR, the researchers can ascertain that the measure does not carry delimiting objectives or contexts that are very specific to its design and development. Additionally, the CRSi-20 will allow the researchers to answer the following questions: (1) What CRSi version is valid in the Taiwanese context? (2) What is the position of the religious construct-system among selected Taiwanese university students? and (3) How does the centrality of religion influence the selected Taiwanese university students' subjective experience and behavior?

Participants
Participants were 331 undergraduate university students from Taiwan (193 or 58.3% females and 138 or 41.7% males). The mean age of the participants was 20.59 years old, with a standard deviation (SD) of 1.48 years. Most of the participants practiced Eastern religious beliefs (Taoism, Buddhism, and folk religions) with n = 158 (or 47.7%), Western religious beliefs (Roman Catholic and Protestant) with n = 33 (or 10.0%), and atheists with n = 140 (or 42.3%).

Context and Methodology of Instrument Validation in this Study
Höllinger and Eder (2016) noted that the notions of religiosity between the Western and the Eastern Asian societies are quite different (p. 6). Furthermore, within a predominantly non-Christian country such as Taiwan (Shiah et al. 2016), the measurement for the religiosity of individuals should be in relation to a wider general concept or meaning, rather than restricted to a more narrow specific definition (Ju et al. 2018). Hence, the CRSi-20 (Huber and Huber 2012) is used in this study.
The English version of the CRSi-20 (Huber and Huber 2012), with permission from Professor Stefan Huber, was translated into Mandarin Chinese (See Appendix A) and then translated back into English further to validate the instrument (Behling and Law 2000). In addition, some items were revised to encompass a more general concept such as "God or something divine (or supernatural)" (Ju et al. 2018). Furthermore, the concept of praying was also expanded to include incense holding (and lighting); wherein within the Eastern Asian societies, it can be equivalent to communicating with or praying to the deities (Arthur 2018;Yao and Zhao 2010).
Convenience sampling was used to collect the data for the study. Participants were volunteer students enrolled within a comprehensive university in the Northern Taipei area. A total of 400 surveys were distributed with a total valid return of 331, which were encoded and analyzed. As mentioned, Taiwan is a predominantly non-Christian country (Shiah et al. 2016). However, its religious beliefs can be said to be quite diverse (Tourism Bureau 2020). A more detailed breakdown of the participants' religious belief is as follows (in descending number of respondents): Taoism n = 78 or 23.6%, Buddhism n = 40 or 12.1%, folk religion n = 28 or 8.5%, Protestant n = 24 or 7.3%, Roman Catholic n = 9 or 2.7%, atheist n = 140 or 42.3 %, and others n = 12 or 3.6% (might include Hinduism, Islam, Judaism, Mormonism, the Unification Church, Yiguandao, and many others).

Instrument
The centrality of religiosity (CR) of Taiwan university students was measured using the Interreligious Centrality of Religiosity Scale (CRSi) versions CRSi-7 (7 items), CRSi-14 (14 items), and CRSi-20 (20 items) (Huber and Huber 2012). Within the CRSi framework, five subscales are used to define the concepts of religiosity, namely: intellect, ideology, public practice, private practice, and religious experience. A five-point Likert (1932) type scale was used to collect the perceived importance of items #1 to #8 (not at all, not very much, moderately, quite a bit, and very much so). This is followed by three items with eight-point Likert (1932) type scale frequency of practice (#9 to #11); several times a day, once a day, more than once a week, once a week, one to three times a month, a few times a year, less often, and never). Lastly, a five-point Likert (1932) type scale was used to collect the perceived experience of items #12 to #20 (never, rarely, occasionally, often, and very often). The CRSi-7, CRSi-14, and CRSi-20 are unique, whereby two of the core dimensions (private practice and religious experience) are assessed with the use of two different religious patterns; interactive and participative (Huber and Huber 2012, p. 716). Table 1 shows the sample items and the number of items per subscale of the CRSi-7, CRSi-14, and CRSi-20.

Data Analysis
To validate the CRSi-7, CRSi-14, and CRSi-20, confirmatory factor analysis or CFA was accomplished. Within CFA, the number of factors and the relationship between the factors and measured variables is already known (Ullman 2006). Hence, the objective of the analysis was to examine the hypothesized structure and test the competing theoretical models about the structure. To accomplish this, structural equation modeling or SEM was used. The SEM is quite informative, wherein it extends the possibility of relationships among the latent variables (which are the inherent common component within the items) and at the same time includes two important components, namely: a measurement model (which is the CFA) and a structural model (Schreiber et al. 2006).
Descriptive statistics were computed with the use of the Statistical Package for Social Science (SPSS) software version 20. Analyses such as the means and standard deviations (SD) of the five subscales of the interreligious Centrality of Religiosity Scale were computed. In addition, the measure of skewness and kurtosis were also reported to further define the distributions of the subscale scores. Lastly, Cronbach (1951) alpha reliabilities were also computed for each of the subscales to establish their internal consistencies.
CFA using SEM was achieved with the use of the Analysis of Moment Structures (AMOS) software version 20 (Arbuckle 2011). The preliminary analysis for the normality of the data was achieved by examining the skewness and kurtosis of the data distribution (Kline 2005). The results of the univariate normality test indicated that the values are within the accepted ranges of ±2.0 for skewness and ±7.0 for kurtosis (Cunningham 2008). However, many have proposed that it is not sufficient to validate for the univariate normality, but it should instead include the assumption of multivariate normality (Byrne 2010;West et al. 1995). Hence, Mardia's (1970) coefficient was computed, resulting in values above the accepted norm, therefore violating the assumption of multivariate normality (Byrne 2010). To fix this, the bootstrapped method was incorporated within the subsequent CFA analyses (Zhu 1997;Yung and Bentler 1996;West et al. 1995). For the CFA, the following fit indices and criteria were used to establish model fit: a non-significant chi-square, root-mean-square-error-of approximation (RMSEA; <0.08) with a 90% confidence limit (Browne and Cudeck 1993;Schreiber et al. 2006), standardized root mean square residual (SRMR) < 0.08, and goodness of fit index (GFI), comparative fit index (CFI), and Tucker-Lewis index (TLI) > 0.90 (Hair et al. 2005). In most practices, numerous goodness-of-fit indicators were used together to assess the model fit (Ching et al. 2014). In addition, to further improve the models, modifications were accomplished with the residuals and were covaried within the model (Byrne 2010). Furthermore, during the model evaluation process, modification indices larger than χ 2 = 3.84 on one degree of freedom were considered a point of iteration within the model modification until the models fit the set-up global fit criteria (Whittaker 2012).
The core model for the CRSi-7 is built around the centrality of religiosity, which is defined as a latent variable with five reflective indicators, each denoting one core subscale of CRSi (Huber and Huber 2012). The succeeding models CRSi-14 and CRSi-20 are built around the five core subscales, which are correlated and receive two (CRSi-14) or three (CRSi-20) reflective indicators (for more details, see Table 1). CRSi-14 and CRSi-20 models with a second-level factor with five reflective indicators of the core dimensions were tested as well.

Validation of the CRSi-7, CRSi-14, and CRSi-20
To validate the CRSi-7, CRSi-14, and CRSi-20, the collected survey data were first encoded and analyzed accordingly. Table 2 shows the means, SDs, skewness, kurtosis, and Cronbach (1951) alpha reliabilities for the CRSi-7, CRSi-14, and CRSi-20 and their corresponding subscales. Mean value of the CRSi-20 shows that the centrality of religiosity of the selected Taiwanese youth is around 2.72 with a SD of 0.58 (CRSi-14 mean = 2.70 and SD = 0.56). The finding shows that the selected Taiwanese youths can be considered as religious, according to Huber and Huber (2012, p. 722), wherein a mean value of 1.0 to 2.0 is non-religious, 2.1 to 3.9 is religious, and 4 to 5 is highly religious. In addition, skewness was computed at 0.43, indicating that the distribution is slightly inclined towards the right side of the graph. Meanwhile, kurtosis was computed at 1.11, indicating a leptokurtic distribution. Cronbach (1951) alpha reliability of CRSi-20 was computed at 0.93, while the rest of the subscales had values ranging from 0.71 to 0.90, indicating acceptable to good internal consistency (Cohen et al. 2007). Similar patterns were also found for the CRSi-14 and CRSi-7.
Correlations among the CRSi-20 and CRSi-14 subscales were also computed. Tables 3  and 4 show the various inter-correlations for the CRSi-20 and CRSi-14 subscales. Findings revealed that all of the subscales were positively correlated with each other. Note that no inter-correlations above the 0.85 threshold were found within the subscales.   A total of seven models were tested in the validation process. Model 1 was the single-factor CRSi-7, Model 2 was the single-factor CRSi-14, Model 3 was the five factors (subscales) CRSi-14, Model 4 was the five factors CRSi-14 with one higher-order factor, Model 5 was the single-factor CRSi-20, Model 6 was the five factors CRSi-20, and Model 7 was the five factors CRSi-20 with one higher-order factor. Table 5 shows the summary of the fit indices for all of the seven models. As noted, due to the large sample size, chi-square values are easier to reach statistical significance (Bollen and Stine 1992). Among the models, only CRSi-14 model 3 showed an adequate fit with RMSEA = 0.08, SRMR = 0.05, GFI = 0.92, CFI = 0.93, and TLI = 0.93. In practice, numerous goodness-of-fit indicators were used to assess a model. However, different indices appear to work well with various sample sizes, types of data, and ranges of acceptable scores to decide whether a good fit exists (Hu and Bentler 1999;MacCallum et al. 1996). In most cases, if the majority of the indices indicate a good or acceptable fit, then there is probably a good fit (Schreiber et al. 2006). In sum, the current study utilized the CRSi-14 model 3 as the basis for later analysis (see also Figures 1-7 for the structural diagram of the models with their corresponding standardized loadings and estimates).   --------Note: Model 1 = single-factor CRSi-7, model 2 = single-factor CRSi-14, model 3 = five factors CRSi-14, model 4 = five factors with one higher-order factor CRSi-14, model 5 = single-factor CRSi-20, model 6 = five factors CRSi-20, and model 7 = five factors with one higher-order factor CRSi-20. X 2 = Chi-square, RMSEA = root-mean-square-error-of approximation, SRMR = standardized root mean square residual, GFI = goodness of fit index, CFI = comparative fit index, and TLI = Tucker-Lewis index. Table 6 shows the various standardized (β) and unstandardized (B) coefficients of the models, together with their corresponding z values and measurement errors (SE) (Fornell and Larcker 1981). To further improve the model, some residuals were covaried within the model (Byrne 2010). Residuals for item #9 and item #8 of the subscale private practice were covaried ( = 0.179, z = 5.713, p < 0.001). Similarly, residuals for item #5 and item #8 of the subscale private practice were also covaried ( = 0.192, z = 6.119, p < 0.001). For the subscale religious experience, residuals for item #18 and item #14 ( = 0.074, z = 2.915, p < 0.001), and item #18 and item #19 ( = 0.167, z = 5.271, p < 0.001) were also respectively covaried. In addition, all of the items have loaded significantly on different subscales with internal consistency within the acceptable values; Cronbach's (1951) alpha reliability ranging from 0.62 to 0.88 (see Table 6). Overall Cronbach's (1951) alpha of the CRSi-14 was computed at 0.90.          Table 6 shows the various standardized (β) and unstandardized (B) coefficients of the models, together with their corresponding z values and measurement errors (SE) (Fornell and Larcker 1981). To further improve the model, some residuals were covaried within the model (Byrne 2010). Residuals for item #9 and item #8 of the subscale private practice were covaried (δ x 9 x 8 = 0.179, z = 5.713, p < 0.001). Similarly, residuals for item #5 and item #8 of the subscale private practice were also covaried (δ x 5 x 8 = 0.192, z = 6.119, p < 0.001). For the subscale religious experience, residuals for item #18 and item #14 (δ x 18 x 14 = 0.074, z = 2.915, p < 0.001), and item #18 and item #19 (δ x 18 x 19 = 0.167, z = 5.271, p < 0.001) were also respectively covaried. In addition, all of the items have loaded significantly on different subscales with internal consistency within the acceptable values; Cronbach (1951) alpha reliability ranging from 0.62 to 0.88 (see Table 6). Overall Cronbach (1951) alpha of the CRSi-14 was computed at 0.90. Note: β = standardized estimates, B = unstandardized estimates, SE = standard error. z = obtained z value, and p = significant value.

Religiosity Profile of Taiwanese Youth
The religiosity profile of the participants based on the results of CRSi-14 revealed that the majority of the participants were religious (n = 293 or 88.50%), some were non-religious (n = 31 or 9.40%), and a minority were highly religious (n = 7 or 2.10%). Interestingly, crosstabulation shows that even though participants consider themselves as to being an atheist, the majority of them are probably practicing folk religion (see Table 7). Folk religion is quite varied, and there is not much consensus of what constitutes Taiwan folk religion (Yang and Hu 2012). Nonetheless, Taiwan folk religion is very much diffused with its cultural heritage and traditions (Wong 2011). Diffused in the sense that it is already mixed with the people's everyday life (Li 1998). Hence, participants rarely realized that they are probably practicing folk religion.  Table 8 shows the means and SDs of the various items and subscales using the CRSi-14 categorization. Findings showed that the subscale ideology scored the highest as compared to the other subscales. This suggests that participants acknowledge and believe in the presence of a higher being. In contrast, the subscale religious experience scored the lowest subscale mean of 2.25 (SD = 0.83), while items #8 and #11 scored the lowest with a mean of 2.02 (SD = 0.80 and 0.67, respectively). Seemingly, participants are not keen on meditation. Within the translation of the word meditation, the Mandarin Chinese terminologies used were jìng sī (静思) and jìng zuò (靜坐); these terms might coincide with some more profound sense of religiosity or faith. As for taking part in religious services, it was hypothesized that participants might consider such behaviors more relative to cultural heritage and traditions. These issues would warrant further investigation, which is currently beyond the objective of the present study. To further understand the religiosity profile of Taiwanese youth, independent samples T-test were also conducted for gender differences and ANOVA for the comparison of the three religious affiliations; Eastern religious beliefs: Taoism, Buddhism, and folk religions with n = 158 or 47.7%, Western religious beliefs: Roman Catholic and Protestant with n = 33 or 10.0%, and atheists with n = 140 or 42.3%. Independent samples T-test showed no significant gender differences with all of the CRSi-14 subscales, signifying that the religiosity of Taiwanese youth is relatively homogeneous between biological genders. Table 9 shows the results of the ANOVA analyses with significant differences between the three religious affiliations for all of the CRSi-14 subscales. Furthermore, Cohen (1988) f was computed to have values ranging from 0.30 to 0.57, denoting medium to large effect sizes. As a rule of thumb Cohen's f values lower than 0.25 are categorized as small, between 0.25 to 0.40 as a medium, and values higher than 0.40 as large effect sizes (Cohen 1988). Post-hoc analyses were also provided to clarify further the significant relationships between the three religious affiliations within each of the CRSi-14 subscales. Findings suggest that individuals who established either an Eastern or a Western religious belief still scored significantly higher within all of the CRSi-14 subscales. Note: N = 331, G1 (Eastern religious beliefs n = 158), G2 (Western religious beliefs n = 33), and G3 (Atheist n = 140). ** p < 0.01 and *** p < 0.001.

Conclusions
This study validated the interreligious forms of the Centrality of Religiosity Scale (CRSi-7, -14, and -20) by using samples from selected university students in Taiwan. The results showed that the CRSi-14 is a valid instrument to measure the centrality of religiosity among selected Taiwanese youth. According to the CRS Index, the selected Taiwanese university students are "religious," which means that while religion is present in their life, it does not play an important role in their decisions. Interestingly, this study also revealed that "religion" is still present in the personal construct-system of the respondents who reported themselves to be atheists. One possible reason for this is folk religion which is very much diffused in the cultural heritage and traditions. As such, individuals rarely realize that they are already practicing religion.
This paper also showed that the selected Taiwanese youth are not keen on meditation despite its relevance to the realm of eastern religious traditions. In further researches on the religiosity of Taiwanese youth, the Mandarin Chinese terminologies for meditation jìng sī (静思) and jìng zuò (靜坐) can be investigated if they coincide with some more profound sense of religiosity or faith. It is also worth considering Yang and Hu's (2012) conceptualization of folk religions which can influence an individual's religiosity profile and religious or non-religious affiliation. This issue warrants further investigation and can be the topic of future research.
Funding: This research received no external funding.

Institutional Review Board Statement:
The study was conducted in accordance with the guidelines of the Declaration of Helsinki. Ethical review and approval were waived for this study, since data were collected within an accepted educational setting and the information obtained is encoded in such a manner that the identity of the human subjects cannot readily be ascertained, directly or through identifiers linked to the subjects.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgments:
The authors are deeply grateful to Stefan Huber for allowing us to use the Centrality of Religious Scale in this study.

Conflicts of Interest:
The authors declare no conflict of interest.