Cross-Cultural Differences Between Spain and Morocco in Depressive Symptomatology During Adolescence

1. Introduction

Adolescence is characterised by the occurrence of multiple individual and new and more complex social dynamics (Viejo y Ortega-Ruiz, 2015). Adaptation to these changes and the performance of new tasks is an important source of opportunities, but also of difficulties, which, if not properly managed, can increase the risk of suffering from maladjustments at this stage. In this context, depression emerges as a particularly relevant and concerning disorder (Garaigordobil et al., 2023) due to its high prevalence and profound, long-lasting impact on health and well-being, especially considering that, according to the World Health Organization (2025), depressive disorders affect approximately 1.3% of adolescents aged 10–14 years and 3.4% of those aged 15–19 years worldwide.

Although there has been considerable interest in recent decades in the study of cross-cultural variations in the epidemiology of depression, we still know little about the nature of depression and variations in its frequency and manifestations in different cultures (Birtel y Mitchell, 2023; Xiang et al., 2024). Several studies have found that industrialised countries have more cognitive-affective symptomatology while developing countries have more somatic-vegetative symptomatology. More specifically, less somatic symptomatology has been found in the North American or European population than in the Latin American (Guerrero, 2014; Koenig et al., 2021), Arab (Dardas et al., 2016; Sulaiman et al., 2003) and Eastern (Chang et al., 2017; Koenig et al., 2021; Stewart et al., 2004) population. Although there are also studies that contradict these results both in the East (Kim et al., 2021) and in Africa (De la Torre-Luque et al., 2020).

In addition to this lack of consistency in findings, there is a major gap in the scientific literature examining the phenomenon of cross-cultural differences in depressive symptomatology: most studies have been based on clinical samples of adults, so there is very little information on cross-cultural patterns in the normal adolescent population.

Depression during adolescence has unique characteristics that differentiate it from depression in other age groups. During this developmental stage, signs of depression can be difficult to detect because adolescents may express symptoms differently from adults. They may be irritable, experience changes in appetite or sleep, lose interest in activities they previously enjoyed, or present recurrent physical complaints with no apparent medical cause. Although these symptoms may also occur in adults with depression, in adolescents they are often mistaken for typical developmental or behavioural changes. As a result, these signs are frequently misunderstood or minimised (Garaigordobil et al., 2023), delaying diagnosis and appropriate intervention. This is particularly concerning given that the duration of untreated depressive illness is an important predictor of recurrence in adulthood and may increase the risk of suicide (Patton et al., 2014), making early detection essential (Pelkonen et al., 2008).

Therefore, the general objective of this study was to carry out a cross-cultural comparison of depressive symptomatology among adolescents born and residing in Spain and adolescents born and residing in Morocco. The Moroccan population was selected because Moroccans represent the largest immigrant population in Spain and one of the largest in Europe. Given the limited availability of assessment instruments with demonstrated validity for this population (El-Ammari et al., 2023), the first specific objective was to linguistically and culturally adapt the Beck Depression Inventory-II (BDI-II), one of the most widely used depression measures internationally, to the Moroccan context (French and Arabic versions) and to analyse its psychometric properties in a sample of Moroccan adolescents living in Morocco. The second objective was to compare adolescents from Spain and Morocco, two geographically close but culturally distinct contexts, through structural equation modelling and prevalence analyses. Ultimately, this study aims to contribute to more appropriate psychological care by promoting accurate assessment free from language barriers and sensitive to the cultural differences between countries of origin and destination.

2. Methods

2.1. Research Design

This was a descriptive, correlational, and cross-sectional study approved by the Ethics Committee of the University of the Basque Country (UPV/EHU; reference number: M10_2018_185).

2.2. Sample and Sampling

The study was cross-cultural and multicentre in nature. A non-probabilistic convenience sampling strategy was used and conducted in two stages, first in Spain and subsequently in Morocco. Schools were selected to ensure heterogeneity in terms of school type (public, publicly subsidised, and private) and geographical context (provincial capitals and smaller towns). Educational centres were contacted based on their accessibility and willingness to participate in the study.

All provisions of the General Personal Data Protection Regulation (EU) 2016/679 were complied with, i.e. the personal data obtained by completing the battery of instruments were processed with the consent of each participant, school management and family members, and for the sole purpose of scientific promotion and dissemination.

A total of 1,141 students participated, 514 residents in Spain, from 4 different educational centres, and 627 in Morocco, from 5 centres, aged between 14 and 19 years (M = 16.09; SD = 1.45) (see Table 1). Most participants (93.2%) came from urban environments, and the majority were classified as having a medium-high socioeconomic status based on the sociodemographic information collected during the study.

2.3. Instruments

A booklet was generated in which 3 questionnaires were included:

- Ad hoc questionnaire assessing sociodemographic and school-related characteristics (e.g., sex, age, place of residence, type of school, and participants’ and parents’ place of birth).

- BDI-II. Beck Depression Inventory (Beck, Steer, y Brown, 1996) in its Spanish version (Sanz et al., 2003). It is a self-report questionnaire that consists of 21 items in which the severity of each symptom during the previous two weeks is assessed on a descriptive scale of four options (0-3); the higher the score, the greater the severity. For its correction, all the answers that have been marked are added together. For interpretation, scores between 0 and 13 indicate minimal depressive symptomatology, between 14 and 19 mild symptomatology, between 20 and 28 moderate symptomatology and above 29, severe symptomatology. It is completed in 5- 10 min. It has shown good psychometric properties in Spanish, Chilean, Mexican and American adolescents.

- Depression subscale of the Behavioural Assessment System for Children and Adolescents (BASC-3S) (Reynolds y Kamphaus, 1992) in its Spanish adaptation (González et al., 2004). It consists of 185 true or false statements (0-1) grouped

into 14 scales and 3 global dimensions: personal adjustment, clinical maladjustment and school maladjustment. It takes 30-45 minutes to complete. It has shown adequate psychometric properties in Spanish, Basque and Persian samples.

2.4. Procedure

Following the guidelines of the International Test Commission (ITC, 2017), the intellectual property rights of the questionnaires have been checked and the process of linguistic, conceptual and cultural adaptation has been carried out entirely in Morocco in 3 steps:

1) Linguistic adaptation of the instruments: A bilingual Moroccan translator with experience in education produced the first version of the instruments in French and Moroccan Arabic. A second translator validated them following a back-translation procedure. Similarities and discrepancies are assessed taking into account Hambleton and Zenisky's (2011) checklist for the quality control of the translation-adaptation of the items.

2) Adaptation to the cultural context: This was done by expert judgement in a multidisciplinary committee consisting of a methodologist, three teachers, a head teacher and a mother. The relevance of the content of the tests and their local comprehensibility were assessed.

3) Analysis of metric properties: This involved the layout and paper-based administration of the instruments, contact with schools and approval of participation, group data collection in classrooms during school hours and always in the presence of the main researcher, as well as quality control, data entry, and data analysis.

2.5. Data Analysis

Initially, an analysis was carried out to assess the presence and type of missing and outliers.

The metric properties of the instruments were then evaluated in the samples of interest. In order to obtain a first approach to the behaviour of each of the items, descriptive statistics were calculated (% ground, % ceiling, mean and its 95% confidence interval, standard deviation, skewness, kurtosis and homogeneity index). Secondly, the relationships between the test items and their agreement with the theoretical model used in their construction were analysed by means of confirmatory factor analysis using the weighted least squares robust mean and variance adjusted (WLSMV) estimation method, which is suitable for ordinal variables with few categories, reliable with small samples and robust to non-normality. The assessment of the fit of the models to the data was based on the value of the chi-square/df ratio (2 /df), together with information provided by the incremental goodness-of-fit index (CFI), the root mean squared error of approximation (RMSEA) and its standardisation (SRMS). Models with values equal to or less than 5 in the ratio2/df, equal to or greater than .90 in CFI and equal to or less than .08 in RMSEA and SRMS (Tabachnick y Fidell, 2005) were considered acceptable. To obtain evidence of external convergence validity, the association with scores on the BASC-3S Depression scale was analysed using Spearman's correlation coefficient (r). Ordinal alpha (αo) and McDonald's omega (ω) reliability coefficients were also calculated for each dimension, thus providing indicators of the internal consistency of the scores in the two language versions. The minimum value considered good is .70 for both indicators.

Finally, the description of the questionnaire scores based on the dimensionality found and the analysis of possible cross-cultural differences were done in two steps. In a first step, a stepwise invariance analysis of the instruments between the groups (by gender, age and country) was carried out taking into account their ordinal nature (Tse et al., 2023). As an acceptance criterion for metric, scalar and strict invariance models we used the variation in fit indices, i.e. the change in CFI (CFI1 - CFI2 < .01), in RMSEA (RMSEA2 - RMSEA1 < .015), and in SRMR (SRMR2 - SRMR1 < .030) between nested models (Cheung y Rensvold, 2002). In a second step, differences in latent mean scores derived from the above models were calculated as a function of country and other socio-demographic characteristics. Effect sizes were assessed using Cohen's d parameter, splitting the latent mean difference by the pooled standard deviation across countries (Hong et al., 2003), according to the procedure described by Hancock (2001). Values lower than 0.20 were considered small, while values higher than 0.8 were considered large differences, resulting in Cohen's values in the interval [0.2, 0.8] judged as moderate (Cohen, 1988). The Spanish sample was set as the reference group, so its latent mean values were set to zero.

The analyses were carried out in the R environment (R Development Core Team, 2022).

3. Results

3.1. Adaptation Process

The linguistic adaptation to the Moroccan context was carried out through translation, back-translation, and expert review procedures. During the cultural adaptation stage, a multidisciplinary committee composed of researchers, educational professionals, and a parents’ representative evaluated the appropriateness of the questionnaire items for administration in the Moroccan school context. Following this review, the items referring to suicidal thoughts and loss of sexual interest were removed from the Moroccan French and Arabic versions due to concerns about their cultural and contextual acceptability in school settings. This decision was not based on assumptions regarding religious affiliation, but rather on recommendations aimed at facilitating the implementation of the assessment in the participating educational centres and ensuring adolescents’ and families’ acceptance of the instrument. The adapted bilingual version used in this study can be found in Appendix 1.

3.2. Psychometric Properties of the Beck Depression Inventory-II (BDI-II) in Spain and Morocco

Despite the presence of missing values, these did not exceed 10% of the cases in any item and did not follow a defined pattern. It was therefore decided not to use imputation procedures to replace them.

Table 2 shows the formal description of each of the items that make up the final questionnaire in Spain and Morocco. In both countries the averages are well below the theoretical mean of 1.5 points in all items and the skewness and kurtosis statistics, together with the presence of a floor effect, report a distribution of scores with a tendency to accumulate in the low scores (asymptomatic pole). All items showed homogeneity indices equal to or above the accepted minimum.

When the internal structure of the questionnaire was tested, both in Spain and Morocco, the model that showed the best results was the one with two highly correlated factors (as opposed to the unidimensional and the two-dimensional with a second order factor) with values of the ratio2 /df below 2, in CFI above 0.95, in RMSEA below .05 and in SRMR at .08. The associations between Cognitive-affective and Somatic- vegetative symptomatology were positive and high. See Figure 1 for more details.

Associations between scores on the Cognitive-Affective and Somatic-Vegetative symptomatology dimensions of the BDI-II and the BASC-3S depression dimension were positive, moderate and statistically significant in Spain (r_CA = .73, p < .001; r_SV = .51; p < .001) and Morocco (r_CA = .64, p < .001; r_SV = .52; p < .001).

In terms of reliability, internal consistency was found to be very high for both Cognitive-Affective and Somatic-Vegetative symptoms in both countries, exceeding .70 for all coefficients.

3.3. Cross-Country Differences in Cognitive-Affective and Somatic-Vegetative Symptom Dimensions

Regarding subgroup differences, the analysis started with the preliminary study of stepwise invariance, with scalar invariance being the minimum requirement to be able to compare means (latent and observed).

In the case of Spain, the fit indices obtained (Table 3) allowed us to accept the equivalence of the measurement model between sexes. Adding restrictions to the regression coefficients, the values listed in the table and the differences between ꭓ2 (Δꭓ2= 31.05, p = .014), CFI (ΔCFI = -.009), RMSEA (ΔRMSEA = .008) and SRMR (ΔSRMR= .024) led us to accept the metric invariance model, which allowed us to assess the equivalence between the values of the intercepts. The values obtained allowed us to accept the scalar invariance (Δꭓ2 = 35.64, p <.001, ΔCFI = -.012, ΔRMSEA = .003, ΔSRMR = .002) and test for strict invariance by adding restrictions on the residuals, which was also satisfied (Δꭓ2 = 21.9, p = .010, ΔCFI = -.015, ΔRMSEA = -.001, ΔSRMR= .017). Something similar occurred with the analyses as a function of age range, which allowed us to conclude the fulfilment of the configural invariance, metric (Δꭓ2 =-20.94, p = .369; ΔCFI = -.001, ΔRMSEA = .001, ΔSRMR = .016), scalar (Δꭓ2 = 20.04, p= .115, ΔCFI = -.006, ΔRMSEA = .000, ΔSRMR = .001) and strict (Δꭓ2 = 16.65, p =.454, ΔCFI = -.005, ΔRMSEA = .000, ΔSRMR = .007).

For Morocco, the fit indices (Table 3) allowed us to accept the strict equivalence of the model between sexes (metric: Δꭓ2 = -0.54, p = .277, ΔCFI = .016, ΔRMSEA =.001, ΔSRMR = .007; scalar: Δꭓ2 = 30.30, p < .001, ΔCFI = .001, ΔRMSEA = -.002, ΔSRMR = .001 and strict: Δꭓ2 = -93.88, p < .005, ΔCFI = - .009, ΔRMSEA = .002, ΔSRMR = .006). Strict invariance as a function of age range was also found (metric: Δꭓ2 = 36.05, p = .320, ΔCFI = -.009, ΔRMSEA = .008, ΔSRMR = .024; scalar: Δꭓ2 = 35.64, p = .019; ΔCFI = -.012, ΔRMSEA = .003, ΔSRMR = .002 and strict: Δꭓ2 = 35.52, p = .296, ΔCFI = -.015, ΔRMSEA = -.001, ΔSRMR = .017).

Cross-country invariance analyses yielded fit indices that made it possible to accept strong or scalar invariance (metric: Δꭓ2 = 4.85, p = .396, ΔCFI = -.009, ΔRMSEA = .007, ΔSRMR = .012; scalar: Δꭓ2 = 199.82, p < .001, ΔCFI = -.010, ΔRMSEA = .015, ΔSRMR = .012) but not strict (Δꭓ2 = 610.26, p < .001, ΔCFI = -.114, ΔRMSEA = -.032, ΔSRMR = .065).

These findings indicated that comparisons by gender, age group and between countries were feasible.

In Spain, there were no statistically significant sex differences in Cognitive- affective symptomatology (∆M = 0.07; ∆SEM = 0.03; Z = 1.96; p = .050), but there were significant differences in Somatic-vegetative symptomatology, which was higher among women (∆M = 0.15; ∆SEM = 0.05; Z = 3.11; p = .002), as can be seen in Table 4. There were no significant differences according to age range in Cognitive-affective symptomatology (∆M = 0.03; ∆SEM = 0.04; Z = 0.79; p = .430) or Somatic-vegetative symptomatology (∆M = 0.03; ∆SEM = 0.05; Z = 0.68; p = .498) either when attending to the sex*age range interaction. Somatic-vegetative symptomatology was significantly higher than Cognitive-affective symptomatology (∆M = 0.17; ∆SEM = 0.05; Z = 3.65; p< .001).

However, in Morocco, Cognitive-affective symptomatology was significantly higher among girls (∆M = 0.24; ∆SEM = 0.10; Z = 2.47; p = .013) but there were no significant sex differences in Somatic-vegetative symptomatology (∆M = 0.15; ∆SEM = 0.09; Z = 1.58; p = .114). There were also no differences due to age in Cognitive- affective (∆M = 0.16; ∆SEM = 0.09; Z = 1.66; p = .096) or Somatic-vegetative (∆M = 0.10; ∆SEM = 0.09; Z = 1.06; p = .276) symptomatology. There were no differences as a function of the sex*age range interaction or between Cognitive-affective and Somatic- vegetative symptomatology (∆M = 0.04; ∆SEM = 0.02; Z = 0.41; p = .168).

The cross-country comparison yielded large differences between Spain and Morocco in Cognitive-affective (∆M = 0.59; ∆SEM = 0.05; Z = 12.19; p < .001) and Somatic-vegetative (∆M = 0.45; ∆SEM = 0.05; Z = 9.05; p < .001) symptomatology, with Moroccan adolescents showing greater distress. This translated into a higher prevalence of adolescents at the highest levels of depressive symptomatology in Morocco (2(3) = 157.97; p < .001; VCramer = .42). n Spain, the majority of participants (89.6%) reported minimal symptomatology, with only a small proportion (7.0%) exhibiting slight symptoms, and an even smaller percentage showing moderate (1.6%) or serious (1.8%) symptoms. In contrast, Morocco showed a different distribution, with a significantly lower percentage (54.2%) of participants reporting minimal symptoms. The proportions of participants experiencing slight (16.4%), moderate (21.3%), and serious (8.1%) symptoms were considerably higher in Morocco compared to Spain.

4. Discussion

Childhood and adolescent depression is not only an individual burden, but also a significant challenge for society as a whole. The negative effects of depression during these early stages of life may have long-term repercussions on academic performance, interpersonal relationships, physical health, and socio-emotional development in both affected individuals and their families (Clayborne et al., 2019).

The first step for an adequate intervention is to ensure a correct assessment using culturally and linguistically adapted instruments with good metric properties. The results showed a behaviour of the final 19-item inventory (after eliminating the items related to suicidal ideation and decreased sexual desire) similar to that found in the original version, showing an optimal fit to the two-dimensional model and high internal consistency in the Affective-cognitive and Somatic-vegetative symptomatology dementias, in the Spanish and Moroccan samples. These findings are in line with those found by other authors (Contreras-Valdez et al., 2015; Melipillán et al., 2008) and support the use of the questionnaire with metric guarantees; contributing to alleviate the scarcity of validated tools in the Moroccan adolescent population (El-Ammari et al., 2023) and to increase the cross-cultural projection of the BDI-II, favouring the standardisation of the assessment of depression in adolescence. Furthermore, the confirmation of factorial invariance allows us to assume the similarity in the attribution of meanings to symptomatology between groups and to carry out the comparisons necessary to achieve the second objective.

In both Spain and Morocco, more depressive symptomatology was found among women compared to men, with no significant differences according to age group. This is in line with different studies that have shown a preponderance of depression among women compared to their male counterparts (Lewis et al., 2015; Wartberg et al., 2018). In non-Western countries, such as Morocco, the differences may be exacerbated by the greater social pressure, cultural constraints and religious forces faced by young women compared to their male counterparts (Rohlinger et al., 2020).

In the comparison between countries, more cognitive-affective and somatic- vegetative symptomatology was found in Morocco than in Spain. This resulted in a higher percentage of adolescents with moderate (21.3% vs. 1.6%) and severe (8.1% vs. 1.8%) depressive symptomatology in Morocco than in Spain. These results are consistent with those reported by a study on Moroccan adolescents in which the prevalence of major depressive disorder was estimated to be 26.5% (Kadri et al., 2010). Studies in Spain, on the other hand, estimated a prevalence of between 2.3% (Jaureguizar et al., 2015) and 19.5% (Aláez-Fernandez et al., 2000) for adolescent depression.

These differences between countries could be influenced by a variety of factors, including socioeconomic, cultural, and access to health care (Gautam et al., 2019). In general, Spain has a higher level of economic development than Morocco, which could imply that the adolescent population has access to more resources and opportunities, which may help reduce the risk of depression. In addition, Spain may have greater public awareness of depression and other mental disorders, as well as a more developed health care system with more extensive educational programs aimed at prevention and treatment.

In line with these findings, it is worth noting that variation in the domain of somatization is considered one of the most consistent findings in cross-cultural studies of depression (Dardas et al., 2016). In particular, people from non-Western countries such as Eastern and Arab cultures report a greater emphasis on the somatic component while those from Western cultures emphasise the emotional or psychological aspects (Chang et al., 2017; Chang y Jetten, 2015; Dardas et al., 2016; Ryder et al., 2008; Sulaiman et al., 2003). In addition, the stigma of having a mental disorder exists in all Arab countries and is associated with significant negative effects on affected individuals and their families (Abdullah y Brown, 2011; Dardas et al., 2016) which may explain why people in this context are inclined to describe mental health problems as physical symptoms.

Therefore, although Spain and Morocco may share similar challenges in terms of adolescent mental health, the differences found alert us to the need to take into account the idiosyncrasies of each culture in order to develop more specific assessment and intervention tools with the aim of increasing their effectiveness.

Limitations

The main limitations are the possible lack of representativeness of the sample derived from the incidental sampling, which compromises the generalizability of the results, and the fact that the fit of the factor models was not investigated in adolescents from the clinical population diagnosed with any of the depressive disorders included in the DSM-5.5.

5. Conclusions

In response to the need identified by El-Ammari et al. (2023) to provide culturally adapted assessment tools for the Moroccan population, the present study offers a valid, reliable, and culturally adapted version of the Beck Depression Inventory-II (BDI-II) for assessing depressive symptoms among Moroccan adolescents. Additionally, the use of this tool in Spain without language barriers for the Moroccan immigrant population will favour more accurate assessments and equal access to psychological care, which will reduce health inequalities and help in their school and social adaptation.

Although depression is a widespread clinical mental health disorder, it is essential to consider cultural and social context, as understanding these differences can help design more effective preventive interventions and improve psychological assessment and care.