Expanding the CYRM-R to include domains of spirituality and religiosity for use with children and youth in Indonesia

Abstract This study aimed to develop a tool to assess resilience in children and youth in Indonesia holistically. We used the Indonesian adaptation of the Child and Youth Resilience Measure-Revised (CYRM-R). We extended it to include domains of spirituality and religiosity, which are important protective factors for young people in Indonesia. Following focus group discussions with Indonesian children and youth, 26 items were developed and tested with students aged 10–18 years (N = 5,504, 63.5% female, 98% Muslim). Exploratory factor analysis produced a two-factor 12-item model. This was then validated together with the CYRM-R using confirmatory factor analysis, which involved invariance tests with sex, age, religion, and school groups. Further analyses confirmed the convergent and incremental validity of the extended measure (correlating and predicting measures of well-being, respectively), and tests of internal consistency also found that each subscale of the extended measure had a high level of reliability. This expanded version of the CYRM-R, which now includes subscales assessing spirituality and religiosity, works well for measuring important protective factors implicated in the resilience of children and youth in Indonesia. We named this expanded measure the “CYRM-R Plus Spirituality and Religiosity”.

This study presents the development of an instrument in Bahasa Indonesia intended to assess resilience in children and youth in Indonesia.Results showed that the Child and Youth Resilience Measure Revised (CYRM-R) Plus Spirituality and Religiosity demonstrates good psychometric properties and may be considered a robust and contextually sensitive measure of resilience for use in Indonesia.Researchers and practitioners may use this instrument to develop a comprehensive account of the resilience of children in Indonesia.

Introduction
Investigations into the effect and associations of spirituality and religiosity are becoming more numerous in developmental studies of children and adolescents (Hardy et al., 2019;Schnitker et al., 2021), particularly in adversity contexts (Vinueza, 2017;Wilson & Somhlaba, 2018).However, there are concerns about the coherence of this growing body of knowledge.One essential criticism is the failure to consistently and clearly define the two constructs (Damiano et al., 2019;Vitorino et al., 2018).Some regard religiosity and spirituality as indistinguishable and use the terms interchangeably (Hill & Pargament, 2003;King et al., 2013;Strelhow & Henz, 2017), while others argue that the constructs have important differences (Koenig et al., 2012;Koh, 2018;Walach & Kohls, 2019).It is said that spirituality concerns a belief in a power apart from one's existence and implies a connection with a universal force transcending everyday sense-bound reality (Connor et al., 2003), as well as engagement with the holy, divine, or that which is beyond the material world (Miller & Thoresen, 2003).In contrast, spirituality has been defined as a comprehensive belief system focusing on intangible elements that impart vitality and meaning to life's events (Maugans, 1996), providing direction to people to help them understand their lives given their ultimate meaning and value (Muldoon & King, 1995).While spirituality may or may not have a connection with specific religious beliefs, religiosity may be thought of as the behavioural expression of spirituality through different activities and practices which are intertwined with a particular religious denomination (Campesino & Schwartz, 2006;Koh, 2018).Religiosity has more specific parameters than spirituality and often involves practising religious rituals, such as reading holy books, praying, visiting holy places, or participating in a pilgrimage.

Indonesian contexts of religiosity and spirituality
Although research into religiosity and spirituality is developing, most research has been conducted in Western countries and tends to focus on Christian populations (Abu-Raiya, 2017;Kosher & Ben-Arieh, 2017).Abu-Raiya (2017) has gone as far as to suggest that this collective knowledge can not be considered a psychology of religion but rather the psychology of Christianity.It is still being determined whether the results of these studies in Christian and Western countries apply more generally to other populations (Kosher & Ben-Arieh, 2017).Several measures based on Western cultures and Christianity may overlook salient aspects of Islam and non-Western cultures (Kosher & Ben-Arieh, 2017;Tiliouine et al., 2009) and may not apply to Islam (Tiliouine et al., 2009).
To our knowledge, limited research has been conducted in Indonesian contexts regarding resilience, spirituality, and religiosity (Borualogo, 2021).Although Indonesia is not defined as an Islamic country, approximately 90% of its citizens are Muslims, and it holds the world's largest Muslim population (British Council, 2021;Indonesia Investments, 2021).Six religions are officially recognised in Indonesia: Islam (87.2%),Protestantism (6.9%), Roman Catholicism (2.9%), Hinduism (1.7%), Buddhism (0.7%), and Confucianism (0.05%) (Portal Informasi Indonesia, 2021).Religion plays an essential role in many aspects of Indonesian society.It is a required subject in both public and private Indonesian schools, where children are exposed to practises at an early age.Regarding family life, Indonesian parents attempt to instil religious values in their children through childrearing techniques (Abubakar et al., 2015).In Islamic families, parents teach their children to read the Quran, pray, and practice fasting on Ramadhan.Children attend Sunday schools for the smaller number of Christians in Indonesia to learn more about Christian values and worship God.Since religion is a significant part of the lives of Indonesian people, major religious celebrations of all religions in Indonesia are celebrated as national holidays.
A national ideology that guides the way of life for Indonesians is known as "Pancasila".Pancasila translates to "The Five Principles", which are (1) belief in the One and Only God, (2) justice and humanity, (3) the unity of Indonesia, (4) democracy guided by the inner wisdom of deliberations amongst representatives, and (5) the realisation of social justice for all the people of Indonesia.The importance of religion in Indonesian peoples' lives is reflected in the first principle of Pancasila.This is also shown in the Year 2000 World Values Survey, which stated that 100% of the Indonesian sample between 18 and 24 indicated that religion is significant in their daily lives (Hosen, 2001).

Spirituality, religiosity, and resilience
Several studies on religiosity and spirituality in children have suggested that the constructs play essential roles in positive development.Kosher and Ben-Arieh (2017) stated that religion is influential in forming children's subjective well-being, where religion is seen as an aspect of culture.In such contexts, children who are religious and place a high value on their faith have higher subjective well-being scores when they live in a country where others do as well (Kosher & Ben-Arieh, 2017).Religious affiliation has also been found to significantly affect life satisfaction, perceptions of support, hope, emotional well-being, and distress in adolescents (Wilson & Somhlaba, 2018).Mountain (2006) stated that children perceive prayer as helping them clarify and articulate their feelings and can be used to aid during significant moments in life.Another study suggested that religious adolescents avoid involvement in delinquent behaviour because their religiosity helps to strengthen their moral beliefs (Massarwi et al., 2019).
Many studies into religiosity and spirituality share ways in which the constructs can also play a role in facilitating well-being despite experiencing significant adversities (Constantine et al., 2000;Koh, 2018;Massarwi et al., 2019;Taufik et al., 2022).Such an ability to overcome, adapt to, and recover from adversity, is known as resilience (Masten, 2018).Resilience relies not only on individuals' abilities but also on the interaction between individuals and their environment.A modern and multisystemic approach to resilience includes protective factors that span psycho-socialecological domains, such as peer support, a sense of belonging, and safe living conditions (see, Ungar, 2021).
The mechanism that underlies resilience processes involving religiosity and spirituality has been examined in Indonesia.Participants who reported that they experienced adversities in life indicated that their adaptation brings them closer to God (Borualogo, 2018(Borualogo, , 2021;;Taufik et al., 2022).This can involve exploring the belief that God allows them to experience problems because God believes they can cope with them (Borualogo, 2021).Through engaging in prayer and worship, individuals reported that they also remained connected to the spiritual power which transcends the self, which helped them believe that they could overcome all adversities.In Indonesia, it is also prevalent for communities to engage in mass prayers when children and youth are about to experience stressful events in life, such as national final graduation tests.These mass prayers are known as istighosah, which means a request for help from God when the community is facing significant challenges (Berita Kaltim, 2019;Latief, 2010;Rozak, 2019).

AIM of the study
Emerging evidence indicates that religiosity and spirituality may be important domains for younger individuals, helping children and youth manage and overcome their adversities (see, Benavides, 2017).However, concise tools to measure the extent to which individuals possess strengths in spirituality and religiosity still need to be improved.A review of resilience measures indicates that there currently needs to be tools available for resilience in children and youth that involve spirituality or religiosity.
Rather than develop an entirely new measure of resilience to include domains of spirituality and religiosity, we decided that it would be more appropriate to build on an existing measure of resilience which covers other areas of resilience and has already established good psychometric properties.The Child and Youth Resilience Measure Revised (CYRM-R; Jefferies et al., 2019) was considered most appropriate for this purpose, as it has already been translated and tested with children in Indonesia (I.S. Borualogo & Jefferies, 2019).Therefore, this study aimed to develop a measure of resilience suitable for use with children in Indonesia, which includes spirituality and religiosity and to do so by extending the CYRM-R to include these domains.

Design
To extend the measure of resilience, we followed the methods developed by Boateng et al. (2018), who recommend a series of processes focusing on item and scale development and scale evaluation.These processes are described below but were preceded by a careful review of the literature to confirm that there were no existing measures of religiosity and spirituality in children and youth as domains of resilience that would be suitable for use in Indonesian contexts.

Participants
To test the expanded CYRM-R measure, the research team recruited a convenience sample of 6,002 elementary, middle and high school students from 29 provinces in Indonesia.A link to a Google Form containing the questionnaire was sent to parents from the schools via WhatsApp.The link navigated to the Google Form, which included an introduction to the study and requested consent for their children to participate.Once consent was obtained, the form instructed parents to pass the questionnaire to their children to complete.The children were informed that they were not required to answer the questions if they did not want to and that their data would be treated confidentially.
Since the survey was administered online, as a data-quality step, we removed 8% of the data of participants who gave the same response to every item of the measures, as it would be improbable that participants endorsing the same response for every item would be responding accurately or truthfully (see Leiner, 2019).The final sample comprised 5,504 individuals (63.5% female) aged 10-18 (M = 14.33,SD = 1.64).Most of the sample was also Muslim (98%), close to recent census demographic figures (British Council, 2021).Table 1 presents characteristic of the participants.
The data supporting this study's findings are available from the corresponding author upon request.This study gained ethical approval from Nusantara Scientific Psychology Consortium (Konsorsium Psikologi Ilmiah Nusantara, K-PIN).

Measures
This study extends the CYRM-R to include spirituality and religiosity (Jefferies et al., 2019).The CYRM-R is a 17-item self-report measure of social-ecological resilience, consisting of two subscales: personal (individual) and caregiver resilience.Personal resilience includes items measuring resilience in intrapersonal and interpersonal relationships.Caregiver resilience relates to characteristics associated with the critical relationship shared with a primary caregiver (Resilience Research Centre, 2018).Individuals indicate how much they believe the item represents them on a 5-point Likert scale, where 1 = "Not at all" and 5 = "A lot".Higher summed scores on the CYRM-R indicate higher levels of resilience.The CYRM-R is widely used in international studies of child and youth resilience (see Resilience Research Centre, 2018).It is a Rasch-validated revision of the original 28item CYRM, developed using data from 14 communities from 11 countries (Liebenberg et al., 2011) and found to have good psychometric properties (see cyrm.resilienceresearch.org/properties).The CYRM-R has been translated into the Indonesian language and demonstrated a strong reliability level (α = .902;I.S. Borualogo & Jefferies, 2019).
The survey also consisted of two measures of well-being.These were the Children's Worlds Subjective Well-Being Scale (CW-SWBS) and the Children's Worlds Psychological Well-Being Scale (CW-PSWBS).Well-being measures are commonly used in resilience studies as outcome status indicators (e.g., Chmitorz et al., 2018;Panter-Brick & Leckman, 2013).Both measures have been developed to assess well-being, specifically in children (Casas & González-Carrasco, 2021), where the former is a measure of subjective well-being (e.g., "I have a good life", "I enjoy my life"), while the latter focuses specifically on psychological well-being (e.g., "I like being the way I am", "I feel positive about my future").Both have been translated and tested in Indonesian contexts (see, Borualogo & Casas, 2019, 2022).The measures use an 11-point Likert scale ranging from 0 to 10, where 0 = "Do not agree" and 10 = "Totally agree".Scores on the measures are summed before converting to a percentage.Both the CW-SWBS and CW-PSWBS have been found to have good psychometric properties, validated using confirmatory factor analyses (Borualogo & Casas, 2019, 2022).They have convergent validity with similar measures (Moreira et al., 2022), negative correlations with bullying (Borualogo & Casas, 2021) and positive correlations with school sense of belonging (Kühner et al., 2021).

Item development
To generate items that could extend the CYRM-R, we drew on data from focus group discussions (FGDs) conducted as a part of a related study by the lead author (see, Borualogo, 2018).These involved 33 individuals aged 15-23 who had experienced different adversities in their life (e.g., bullying, human trafficking, divorced parents, child labour, child sexual abuse) and who shared what they believed helped them to recover from stressful life challenges.Although this sample included individuals slightly older than the age group of the sample the measure was intended for, the original CYRM was designed for individuals aged 10-23 (Liebenberg et al., 2011), and we considered that the older individuals might be better able to reflect on experiences of adversity.Indeed, many mentioned how spiritual resources helped them overcome adversities.Having a spiritual connection to a higher power (i.e., God) and practising prayer or worship helped them through their challenges.Based on these discussions, 26 items related to religiosity and spirituality were developed.These items were then reviewed by four Indonesian academic experts who specialised in psychology, resilience, and spirituality.The experts received a definition of religiosity (Campesino & Schwartz, 2006;Koh, 2018) and spirituality (Maugans, 1996;Miller & Thoresen, 2003;Muldoon & King, 1995) and were asked whether they thought the items aligned with either definition (or both).They were also asked to judge whether any of the items were too similar and, if any were, which they felt should be removed and which should be retained.Last, the experts were requested to include any additional items if they felt an important quality related to spirituality or religiosity was missing.Several rounds of expert input were conducted, sharing feedback until a consensus was reached.A total of 26 items were eventually retained, half of which addressed religiosity and half spirituality (see Supplementary Information).To test the legibility of the items, we discussed them with 14 school children aged 10-18.All of the children understood the meaning of the items.

Scale development
To check the dimensionality of the novel items, we conducted an exploratory factor analysis (EFA) on the 26 items using data from the online survey.Prior to this, the dataset was randomly halved (n = 2,752) for the EFA so that the other half could be used for a subsequent Confirmatory Factor Analysis (CFA; n = 2,752).The data was checked and appeared to be suitable for the analysis (Kaiser-Meyer-Olkin test for sampling adequacy = .96;Bartlett's test of sphericity<.001;Tabachnick & Fidell, 2019).The EFA itself used a minimum residual extraction and oblimin rotation, and given the relatively large number of items, Comrey and Lee's (1992) threshold of ≥.63 was used as a cutoff to retain items that were considered "very good" indicators of a factor.Multiple factor models were requested and re-run with low-loading items removed.We contrasted models by reviewing their fit statistics (see below), scree plots, eigenvalues, and whether a particular factor composition appeared sensible (Henson & Roberts, 2006, p. 399).

Scale evaluation
The selected model was validated through CFA using the other half of the dataset.The CFA confirms the factor structure found in the EFA, indicating construct validity (Devins et al., 1988).For both EFA and CFA, a suitable model involves fit statistics that include a root mean square error of approximation (RMSEA) <.08 (Kenny, 2015), a root mean square residual/standardised root mean square residual (RMSR/SRMR) <.08 (Hu & Bentler, 1999), a non-significant (>.05) chi-square test of model fit (Barrett, 2007), and Comparative Fit (CFI; CFA only) and Tucker-Lewis indices >.90 (Hox et al., 2010;Hu & Bentler, 1999).Chi-square statistics are also generated by the analysis but are known to be sensitive to large samples, and there are recommendations against using them as fit criteria (Schumacker & Lomax, 2004).
As part of the CFA, measurement invariance was tested across sex, age groups (10-14, 15-18 years), religious affiliation, and type of school.Configural/form, factorial/metric, and scalar/ intercept invariance have been identified as particular tests of measurement invariance (Lee, 2018;Milfont & Fischer, 2010).These tests produce additional models and can be judged using the same initial CFA fit criteria.ΔCFI is also provided, and Cheung and Rensvold (2001) recommend that changes in CFI should be ≤-.01 to indicate invariance across groups.
We also reviewed the internal consistency of the expanded measure by checking Cronbach's alpha and McDonald's Omega, ensuring values were ≥.7 (Hair et al., 2009;Hinton et al., 2004).Convergent validity was checked by correlating the expanded measure with the CW-SWBS and CW-PSWBS.Incremental validity was checked through hierarchical regression.The 17-item CYRM-R was entered as a predictor of CW-SWBS scores before a second model was entered, including both CYRM-R and the additional items, to check whether the inclusion of the expansion explained a significantly higher proportion of variation in CW-SWBS scores.
Finally, we explored scores on the measure by group (sex, age, school type, religious affiliation) to see if any significantly differed.These involved ANOVA and t-tests to check for differences.All analyses were performed in R (v4.0.0;R Studio Team, 2020) using R studio (v1.2.5042; R Studio Team, 2020).The analysis script can be obtained on request by contacting the lead author.

Results
Missing data analysis indicated that fewer than 1% of data was missing, so no imputation was required.

Exploratory factor analysis
Of the models produced by the EFA, a two-factor model appeared most appropriate, based on the fit statistics and the composition of the factors.After respecifying the model without low-loading items (two further iterations), a two-factor model was produced that consisted of 12 items (RMSEA = .07[90% CI: .067-.073],RMSR = .02,TLI = .96;see, Table 2).

Confirmatory factor analysis and invariance testing
The two-factor model produced by the EFA was then tested with the second half of the dataset using CFA to validate the structure.The overall model, consisting of the two new factors and the original two of the CYRM-R, was found to have a good fit (CFI = .91,TLI = .90,RMSEA = .065[90% CI: .063-.066],SRMR = .054;Figure 1; also see Supplementary Information).The invariance tests similarly determined a good fit for the model across the groups (see, Table 3).
To explore the incremental validity of the expanded measure, we ran a hierarchical linear regression using well-being (CW-SWBS) scores as the outcome and CYRM-R scores (the original 17-item scale) as a predictor (Table 5).This model, which contained the original scale, explained 35% of the variance in well-being scores (R 2 adj = .35;F[1,5502] = 2943, p < .001).A second model containing the 12 new items as an additional predictor explained 36% of the variance (R 2 adj = .36;F [2,5501] = 1542, p < .001),and comparison of these models indicated that the additional items were a significant inclusion (ΔR 2 adj = .01;F[1,5501] = 91.83,p < .001),despite only increasing the predictive power by 1%.

Discussion
Several studies foreground the association between spirituality and religiosity and resilience processes (Cahyani & Akmal, 2017;Oktavia & Muhopilah, 2021;Taufik et al., 2022;Uyun & Rumiani, 2012).However, this study is the first to develop a measure of resilience that includes an assessment of spirituality and religiosity for use with children and youth in Indonesian contexts.
We named this expanded measure the "CYRM-R Plus Spirituality and Religiosity" (the full measure can be found in the Supplementary Information).It was found to work well with both male and female children, older and younger children, those attending religious and non-religious schools, and those who identified as Christian and Muslim.With promising psychometric properties, this expanded measure enables practitioners and researchers to measure young people's resilience in Indonesia comprehensively.
Several measures of resilience have already been used in studies involving Indonesian populations.The most popular are the CYRM-R (see, Agustine & Borualogo, 2021;Höltge et al., 2021;Idriyani, 2021;Rakhmadianti et al., 2021;Wahab et al., 2021) and Connor-Davidson Resilience Scale (see, Kaloeti et al., 2019;Karin et al., 2020;Ramadianto et al., 2022;Setiawati et al., 2021).However, resilience measures' content and specific focus vary significantly and require scrutiny to ensure they are appropriate to a study context (for a parallel with anxiety measures, see, Wall & Lee, 2022).In particular, the abovementioned measures lack a specific assessment of spirituality and religiosity and therefore neglect these critical protective factors, which are particularly important in Indonesia.
Although the measure worked well for both sexes, males demonstrated lower spirituality and religiosity subscale scores.These results align with Baker and Whitehead (2016), Collett and Lizardo (2009), Robinson et al. (2019), andSchnabel (2017), who found that females tend to be more religious than males in different Western countries.However, these studies typically involve adults who identify with Christianity.While we found that religious affiliation was not associated with a difference in spirituality scores, being Muslim was associated with higher religiosity scores, a trend also supported by Aditya et al. (2021).The differences we observed may indicate different trajectories in the spiritual and religious life of children and youth in Indonesia.Further research may explore differences in measures on these domains between sexes of different religious affiliations.
The findings of this study also indicated that younger children had higher religiosity scores than their older peers.Dyer and colleagues (2020) similarly found that religiosity declines across adolescence, arguing that as children go through adolescence, they may experience greater religious doubt.A decline in religious behaviours during this period has already been observed (Hunsberger et al., 2002).However, despite the differences in scores on the measure between the subgroups of our sample being statistically significant, further research is required to determine whether these differences are meaningful and the extent to which they genuinely impact the resilience of children in Indonesia (for instance, the extent to which the nearly 2.5-point difference on the 36-point spirituality subscale that we observed between males and females is impactful).For now, users of the expanded measure should be mindful that these differences exist, foregrounding the potential importance of analysing samples concerning these demographic features.

Conclusion
The CYRM-R Plus Spirituality and Religiosity is an expanded measure of resilience that can be with children aged 10 to 18 years old.It demonstrates good psychometric properties, and its inclusion of spiritual and religious components suggests it may be of interest to researchers and  practitioners interested in discovering a more comprehensive account of the resilience of children in Indonesia.

Limitations and future research
This study was conducted in Indonesia, and although the extended measure might be appropriate for use in other countries and cultural contexts, it is essential to test this.It would also be interesting to examine whether the additional domains of the measure work well as extensions of the adult version of the CYRM (the Adult Resilience Measure; ARM), given the continued importance of spirituality and religiosity in adulthood in Indonesia.
We also note that although the expanded measure explained a significantly greater proportion of variance in well-being scores than the initial CYRM-R, it was only a 1% improvement, which is small.Upon further scrutiny, some of the more distinct spirituality or religiosity items omitted from the measure's final version could be reintroduced to examine any improvements to predictive capacity.However, this study concerned a preliminary validation of the measure.A more thorough examination could involve a longitudinal design where an individual's level of functioning (such as well-being) is assessed before and after an experience of adversity to determine the extent to which individuals with higher scores on the expanded measure return to their previous level of functioning (thereby indicating resilience).In the current study, a fundamental limitation was not knowing functioning levels prior to any disclosed adversity, which would involve good quality resilience studies (Ungar, 2019).
A further important point is that the sample involved in the development and preliminary validation of the extended measure was predominantly Muslim.While the CFA indicated measurement invariance for the Christian subgroup, the measure may have involved different elements if solely developed with non-Muslims (including those with no religion).Future research may look a little more closely at face and content validity with Indonesians who do not identify as Muslim to confirm the measure's relevance with such other groups.You are free to: Share -copy and redistribute the material in any medium or format.Adapt -remix, transform, and build upon the material for any purpose, even commercially.The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms: Attribution -You must give appropriate credit, provide a link to the license, and indicate if changes were made.You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.

No additional restrictions
You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
and dialog with, expert editors and editorial boards • Retention of full copyright of your article • Guaranteed legacy preservation of your article • Discounts and waivers for authors in developing regions Submit your manuscript to a Cogent OA journal at www.CogentOA.com

Table 4 . Descriptive and correlational statistics for the subscales of the measure Min score Max score
† All correlations were p < .001;Spearman's correlations were used due to negative skew.