Resilience, shaped by both internal and external factors, has been explored through various theoretical frameworks. Potential determinants of resilience include genetic, epigenetic, developmental, demographic, cultural, economic, and social variables. It heavily relies on personal characteristics, external support systems, and coping strategies, and can be strengthened at multiple levels, including individual, family, community, and cultural contexts. Evidence suggests that resilience is linked to the ability to flexibly employ a range of coping strategies tailored to specific challenges and to use corrective feedback to refine those strategies (Southwick et al., 2014).

Coping strategies involve cognitive and behavioral efforts to manage demands that exceed personal resources (Lazarus and Folkman, 1984; Endler and Parker, 1990; Smith et al., 2016). Some strategies reduce stress and lead to positive outcomes, while others increase stress and result in negative outcomes (Parker and Endler, 1992; Endler and Parker, 1994; Smith et al., 2016). The effectiveness of a coping strategy can also depend on personal resilience (Lazarus and Folkman, 1984; Zeidner and Saklofske, 1996; Folkman and Moskowitz, 2000; Smith et al., 2016; Zengyan and Liu, 2024). Higher personal resilience is associated with greater use of task-oriented coping strategies, which in turn are linked to more adaptive outcomes. Conversely, higher resilience is linked to less reliance on nonconstructive emotion-oriented strategies, which are associated with poorer psychological outcomes (Smith et al., 2016).

In a recent scoping review, Ballard et al. (2024) identified six commonly used self-report resilience scales for young adults - Connor-Davidson Resilience Scale (CD-RISC; Connor and Davidson, 2003), Resilience Scale (RS-25; Wagnild and Young, 1993), Child and Youth Resilience Measure (CYRM; Ungar and Liebenberg, 2011), Resiliency Scales for Children and Adolescents (RSCA; Prince-Embury, 2008), Resilience Scale for Adolescents (READ; Hjemdal et al., 2006) and the Brief Resilience Scale (BRS; Smith et al., 2008). Among these, CD-RISC and RS are most commonly used; however, they were initially designed for adult resilience research and later adapted for youth. In contrast, CYRM, RSCA, and READ were all developed specifically for children and adolescents (Ballard et al., 2024).

In our study, we focus on CYRM-12 and BRS due to their absence in the Slovenian context. The availability of additional widely utilized questionnaires would enable more straightforward comparisons of our findings with those of international studies, facilitating effective cross-cultural comparisons across diverse populations, aiding in understanding how resilience manifests itself in relation to other constructs in various contexts. For instance, the CYRM-12 has seldom been validated in studies conducted outside of English-speaking contexts (Mu and Hu, 2016). Furthermore, Ballard et al. (2024) reported that BRS has rarely been reviewed or applied in studies on adolescent resilience. Additionally, while CYRM-12 and BRS scales stem from different theoretical rationales for the construct of resilience (as described in the following sections), the relation between them (and to other resilience measures) has been rarely explored in empirical studies.

The CYRM was developed with the aim of capturing and understanding resilience across different cultures and contexts, based on a socio-ecological interpretation of resilience. During its development phase, a mixed-methods approach was employed, involving adolescents and adults from 11 countries (Ungar and Liebenberg, 2011). The original self-report scale (CYRM-28) measures three major categories of resilience: individual resources (personal skills, peer support, social skills), relationships with caregivers (physical and psychological caregiving) and contextual resources (spirituality, culture, education; Liebenberg et al., 2012). The shorter, 12-item measure can be used as a screening tool in clinical and non-clinical settings and is also suitable for inclusion in omnibus surveys. It was validated on a sample of adolescents exposed to higher risk as well as on a sample of school children aged 10–18 years. Data for certain analyses were collected from a sample of young people up to 22 years old. Liebenberg et al. (2013) report sufficient validity of the measure.

The BRS was developed to assess resilience as the ability to bounce back from stress. The reliability and validity of the scale was tested on four different samples, with two groups comprising students and two comprising adults afflicted with chronic conditions. The results suggested that BRS is a reliable and valid measure (Smith et al., 2008).

CYRM-12 was developed by Liebenberg et al. (2013), who proposed a model in which the 12 questionnaire items load onto a single latent factor. After allowing four pairs of items to covary, the one-factor structure exhibited satisfactory fit indices and reliability (for more information, see Liebenberg et al., 2013). Since then, the scale has been validated in several languages, for example Chinese (Mu and Hu, 2016) and Persian (Mohammadinia et al., 2024), with both studies supporting one-factor structure. In Arabic (Panter-Brick et al., 2021), the online version of CYRM-12 showed good validity and reliability. However, the factor structure was not tested. Supporting its multi-purpose aims, the questionnaire has also been validated in a disadvantaged community sample (Russell et al., 2021) and among Syrian refugee youth (Mohammadinia et al., 2024). Based on the one-factor structure of the original short-version questionnaire (Liebenberg et al., 2013) and further support for this structure in other languages, we set our first hypothesis below.

The authors of BRS also proposed a one-factor structure of the questionnaire, which was tested and confirmed in their initial study (Smith et al., 2008) and other studies using English-speaking samples (e.g., McKay et al., 2019). So far, the measure has been translated and validated in many languages. Brazilian-Portuguese (da Silva-Sauer et al., 2020) and German (Chmitorz et al., 2018) validations showed a good fit of the unidimensional structure after accounting for another factor in the model only with negatively worded items. Similarly, Turkish (Haktanir et al., 2016), Korean (Choi et al., 2019), and French (Jacobs and Horsch, 2019) validations also showed a good fit of the unidimensional structure. On the other hand, validations on Spanish (Rodríguez-Rey et al., 2016), Greek (Kyriazos et al., 2018), Dutch (Soer et al., 2019), Chinese (Fung, 2020), Polish (Konaszewski et al., 2020), and Arabic (Baattaiah et al., 2023) samples tested a structure with two factors represented by positively and negatively worded items and showed a good fit. Still, since the initial study proposed a unidimensional structure, which has successfully been replicated in several studies, we set the second hypothesis.

Smith et al. (2008) found a moderate positive correlation between the BRS scale and the CD-RISC scale. A moderate positive association was also found between CD-RISC and CYRM-12. Similarly, a moderate positive correlation between BRS and CYRM-12 was reported in Soliman. We expect the following associations between the resilience scales included in the study.

As mentioned in the introduction, resilience is closely related to coping strategies skills (Southwick et al., 2014). Min et al. (2013) reported positive reappraisal and refocus on planning to be the two most important predictors of resilience among all tested coping strategies. Positive correlation with planning ranged from small to large, and positive correlation with positive reappraisal was in moderate range (Smith et al., 2008; Min et al., 2013; Lee et al., 2019). Following from this, we set hypotheses below predicting convergent validity of resilience scales.

Following from defining resilience as the ability to adapt positively in the face of adversity or stress, previous research has found a negative association between resilience and mental health (Yıldırım et al., 2020). Resilience has been tested in the original article by Smith et al. (2008) on four different samples. Negative associations with both depression and anxiety ranged from moderate to large. Therefore, we set the two hypotheses below.

Similar to adaptive coping strategies, maladaptive coping strategies, such as self-blame and catastrophizing, also showed associations with resilience. Negative associations between resilience and self-blame ranged small to moderate (Lee et al., 2019; Min et al., 2013), while associations with catastrophizing were moderate negative (Min et al., 2013). Based on the associations found in the literature, we propose the following set of hypotheses linked to discriminant validity.

When people assign positive meaning to negative experiences, which is indicative of high resilience, this act can have a positive effect on their perception of their position in life (Arslan and Yıldırım, 2021; Tempski et al., 2015). Previous studies have reported a significant positive association of resilience with all four subdimensions of quality of life; moderate association for physical and social, and large association for psychological and environmental quality of life (Keener et al., 2021), which is why we propose the same in our last hypothesis below.

A total of 389 participants started filling out the survey. Some had to be excluded from the final sample due to dropping out before finishing the survey (n = 21; 5.4%), not meeting age-related inclusion criteria (n = 11; 2.8%) or failing two or more attention checks (n = 27; 6.9%). The final sample, hence, consisted of 330 participants who attentively filled out all relevant questionnaires. Of those, 209 (63.3%) were female, 116 (35.2%) male, while 6 (1.7%) identified as ‘other’ or preferred not to answer the question about gender. Participants were, on average, 21.2 years old (SD = 1.85, min = 18, max = 24), and most of them had completed some form of secondary education (n = 229; 69.3%), followed by those with completed tertiary education (n = 75; 22.7%) and those with primary education (n = 26; 7.9%). The socioeconomic status of participants was, for the largest share of them, a lot lower than average (n = 145; 43.9%), followed by those with lower than average (n = 90; 27.3%), average (n = 76; 23.0%), higher (n = 15; 4.5%) and a lot higher (n = 4; 1.2%) than the average Slovenian net salary at the time of the study.

Coping strategies were assessed via selected dimensions of Cognitive Emotional Regulation Questionnaire (CERQ; Garnefski et al., 2001), i.e., two dimensions assessing positive coping strategies, namely refocus on planning (2 items, e.g., “I think of what I can do best”) and positive reappraisal (2 items, e.g., “I think I can learn something from the situation”), and two assessing negative coping strategies, namely self-blame (2 items, e.g., “I think about the mistakes I have made in this matter”) and catastrophizing (2 items, e.g., “I continually think how horrible the situation has been”). Items are answered on a 5-point scale (1 – “(Almost) never”, 5 – “(Almost) always”). The scale exhibited satisfactory internal consistency in the original validation study for all dimensions used in this study (α = 0.72–0.81).

Anxiety was assessed with the Generalized anxiety disorder questionnaire (GAD7; Spitzer et al., 2006; Jelenko Roth and Dernovšek, 2011). It consists of 7 items asking about the frequency of generalized anxiety disorder symptoms that participants were bothered by in the last 2 weeks (e.g., “Worrying too much about different things”). Items are answered on a 4-point scale (0 – “Not at all”, 3 – “Nearly every day”). The scale exhibited excellent internal consistency in the original validation study (α = 0.92).

Depression was measured with the Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001; Cesar et al., 2021) that consists of 9 items asking about the frequency of symptoms of depression that participants were bothered by in the last 2 weeks (e.g., “Feeling tired or having little energy”). Items are answered on a 4-point scale (0 – “Not at all”, 3 – “Nearly every day”). The scale exhibited good internal consistency in the original validation study (α = 0.86–0.89).

Quality of life (QoL) was measured with the World Health Organization Quality-of-Life Scale (WHOQOL-BREF; WHOQOL Group, 1998). It consists of 26 items. 24 items form domain scores that reflect individuals’ perceptions of their psychological QoL (7 items, e.g., “How much do you enjoy life?”), physical QoL (6 items, e.g., “Do you have enough energy for everyday life?”), social relationships (3 items, e.g., “How satisfied are you with your personal relationships?”), and environment (8 items, e.g., “How healthy is your physical environment?”). The remaining two items refer to the overall perception of quality of life and health, respectively, and were omitted from the analyses of this study. Items are answered on a 5-point scale with labels varying between different subsets of items (e.g., 1 – “Very poor”, 5 – “Very good”; 1 – “Not at all”, 5 – “An extreme amount”). The scale exhibited acceptable to good internal consistency in the validation study (physical: α = 0.80–0.84; psychological: α = 0.75–0.77; social relationships: α = 0.66–0.69; environment α = 0.80).

In addition to the questionnaires described above, participants were asked to provide information on their socio-demographic characteristics and respond to three attention check items spread throughout the survey, all in the form of directed questions (e.g., “This is a control question. Mark “Strongly disagree” and continue”; Maniaci and Rogge, 2014).

The authors translated all questionnaires that were not already available in Slovenian (i.e., CYRM-12, BRS, and WHOQOL-BREF) using the translation-back translation method. Specifically, two researchers independently translated the items into Slovenian, and these translations were merged by a third researcher. The merged Slovenian items were translated back to the original language by a fourth researcher who was not familiar with the original items. The original and back-translated items were compared, with all discrepancies resolved by discussion. The translators had a psychology or psychiatry background with additional expertise in the social sciences methodology and had (close-to) native proficiency in both languages.

The study was conducted in Slovenia. The participants were recruited via the Valicon online panel¹ that is focused on marketing and other types of research and data collection, primarily in Slovenia. The data was collected in March and April 2024. Participants had to be between 18 and 24 years old to be eligible for participation (M = 21.2, SD = 1.85). First, they were informed about the aims and characteristics of the study and on the voluntary and anonymous nature of their participation. After consenting to participate, they proceeded to an online survey on the 1 ka platform.² The survey took 16.5 min to complete on average. Participants received small monetary compensation for their participation as agreed between them and the online panel. Ethical approval was not sought for this study, as per institutional and national guidelines.

We first cleaned the database by excluding participants who did not meet our inclusion criteria, dropped out during the study, or failed more than one attention check. As the remaining participants had no missing values, no data needed to be imputed. We also checked for common method bias by calculating Harman’s single factor score that loads all items used in the study onto a single factor. The results showed that the single factor explains 26.90% of total variance, which is below the recommended threshold of 50.00% (Harman, 1960). As such, there is no evidence of common method bias in our study.

In the next step, we calculated the relevant descriptive statistics, tested the assumptions, and conducted confirmatory factor analyses (CFA) in the Mplus 8.0 software. Maximum Likelihood (ML) was used as the default estimator. However, since some items of the CYRM-12 exhibited larger deviations from the normal distribution, ML was replaced with the unweighted least squares (ULS) estimator in confirmatory factor analyses of this questionnaire (Li, 2016). The model fit was assessed by examining the Comparative Fit Index (recommended CFI ≥ 0.90), Tucker-Lewis Index (recommended TLI ≥ 0.90), Root Mean Square Error of Approximation (recommended RMSEA ≤0.080), and Standardized Root Mean Residual (recommended SRMR ≤0.080; Kline, 2005). We additionally investigated the factor loadings and standard errors (SE). Moreover, we investigated the modification indices (MI) to identify potential areas of misfit.

The remaining analyses were conducted in IBM SPSS Statistics 28. We first assessed the internal consistency of the central measures and other included questionnaires with α coefficients. Next, we calculated the factor scores, descriptive statistics, and bivariate associations between the constructs (using Pearson’s r). Lastly, we conducted multiple regression analyses to investigate the predictive value of the central questionnaires in predicting different quality of life aspects (controlling for other predictors of these outcomes).

We tested the convergent, discriminant, and incremental validity by calculating the bivariate Pearson’s correlations between the CYRM-12 and BRS scores and other relevant constructs and performing hierarchical regression analyses.

The implications of the present study highlight robust psychometric properties of the Slovenian versions of the CYRM-12 and BRS resilience measures among youth. Our findings support the unidimensional structure of both scales, along with their convergent, discriminant, and incremental validity, suggesting that these measures can effectively assess resilience in this demographic. Notably, the CYRM-12 demonstrated a stronger predictive value for quality of life outcomes compared to the BRS, indicating its potential as a more comprehensive measure for evaluating resilience in relation to these outcomes. The positive associations with adaptive coping strategies and negative correlations with depression and anxiety further underscore the importance of resilience in mental health among Slovenian youth. These results pave the way for future research on resilience, enabling cross-cultural comparisons and enhancing the resources available to practitioners for evidence-based interventions aimed at fostering resilience in young individuals.

The present study has some limitations. First, as the collected data are solely based on self-report, the study is not immune to the well-known shortcomings of this approach, including social desirability, which can create false relationships or obscure relationships between variables (van de Mortel, 2008) and recall biases, which can lead to under- or overestimation of positive or negative past emotional experiences (Colombo et al., 2020). Second, as the study was conducted online, we could not control the external factors that may disrupt participants, such as noise, interruptions, or other distractions. Third, we recruited participants via an online panel, which could affect different aspects of data quality, such as attention, comprehension, honesty, and reliability (Peer et al., 2021). However, this problem was mitigated by employing attention checks in the form of directed questions (Maniaci and Rogge, 2014). Fourth, we employed a convenience sample of young individuals willing to participate in the study in exchange for a smaller reward. While the resulting sample is relatively diverse, it is only partially representative as, for example, demonstrated by a larger share of women (63.3%) than men (35.2%). This may hinder the generalizability of our findings. Lastly, even though our findings are consistent with those seen in other validations of the primary questionnaires, we did not specifically assess the measurement invariance of the translated and original versions.

Future research may replicate and extend our study by recruiting a representative sample of young individuals or investigating the psychometric characteristics of the two central questionnaires in general population samples. Furthermore, future studies may focus on empirically investigating the measurement equivalence of the Slovenian versions and the original versions of both scales. Beyond psychometric studies, future research can build on our study and use the translated measures to investigate the protective role of resilience in coping with different stressors, explore cross-cultural differences in youth resilience, and quantify the effectiveness of resilience-building interventions. It is worth noting that some of these ideas will be explored in the EU funded project “Supporting mental health in young people: Integrated Methodology for clinical decisions and evidence-based interventions (SMILE),” which will deploy digital tools capable of enhancing the resilience of young individuals in seven countries including Slovenia.