Protective factors have been established as moderators in the association among adverse experiences and future outcomes as suicidality in adulthood, performing child-to-parent violence or exhibiting trauma-related responses, therefore establishing the need to explore protective factors and their characteristics. The aim of this study is to identify profiles among protective factors in children and adolescents at risk, and to relate these profiles to several sociodemographic variables such as age, gender, country of origin (native as opposed to immigrant) and the type of family structure (being a single parent family or a bi-parent family). Data was collected from professionals involved in the Children Protective Services (CPS). Sample was composed by 635 children and adolescents involved in the CPS. Protective factors were assessed by the Adolescents and Children Risk of Abuse and Maltreatment Protective Factors Scale (ACRAM-PFS). A Latent Profile Analysis (LPA) was estimated. Six profiles were retained. Membership to these profiles was associated to gender, age, family structure and country of origin. Country of origin showed significant association to several profiles. Information provided in this study is novel and can help to improve quality of interventions from an ecological perspective.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Introduction
It is well established in the scientific literature that children and adolescents who are exposed to adverse childhood experiences such as child maltreatment or neglect are at risk of developing a number of negative outcomes as trauma-related distress, academic and social problems, substances addiction, poor long term health, risky sexual behavior, suicidality and generational perpetuation of maltreatment (Afifi & Macmillan, 2011; Kisely et al., 2018; Madigan et al., 2019; McKay et al. (2021); Norman et al., 2012; Racine et al., 2020).
Risk assessment and intervention have been traditionally based on deficit models, in the way that only risk factors were assessed and interventions were only focused on diminishing them, while protective factors were conceptualized as the absence of risk (Navarro-Pérez & Pastor, 2017; Ross & Vandivere, 2009). However, with the emergence of other theoretical models, such as positive psychology in the late 90 s, theories of positive growth related to children as prosocial behaviors and emotional intelligence became popular (Catalano & Hawkins, 1996; Cyrulnik, 2011). As a result, this rise of positive psychology led to an increase in the utilization of interventions grounded in its principles. This shift was prompted by the recognition that the previous approach, which solely focused on risk factors and overlooked protective factors, restricted the scope of diagnosis and limited the potential interventions available (Cress et al., 2016). Nowadays, both prevention and research fields have stressed the need to explore protective factors and their characteristics, as their moderating effect has been well established in the association between adverse experiences and children’s outcomes (Afifi & Macmillan, 2011; Panisch, et al., 2020; Racine et al., 2020; Ridings et al., 2017).
Protective factors are a complex construct since they exist at multiple levels of the child’s ecology. It is possible to distinguish among personal, family, or community protective factors (Cicchetti & Toth, 2016). A parallelism of this same conceptualization can be drawn to Bronfenbrenner’s (1979) and Belsky’s (1993) ecological model of maltreatment, in which causal factors are located in three levels which are nested among each other: the individual, the proximate and the distant system. Thus, the individual system would correspond to the personal resources understood as protective factors of the child or adolescent; the proximate system would correspond to the family environment, relations, and resources; and finally, the distant system would refer to the environment and the resources available in the environment of the child or adolescent.
Protective factors are defined as the characteristics of a child, a family, or the environment that diminishes the probability of suffering from adverse experiences such as child maltreatment, abuse, or neglect (Cress et al., 2016; Sprague-Jones et al., 2019) and improve individual’s response to adverse experiences that would typically lead to a negative outcome (Rutter, 1985). It has been shown that protective factors reduce the risk of suicidality in adulthood (Janiri et al., 2020), and the probability of performing child-to-parent violence (Beckmann et al., 2021). Additionally, another recent study performed by Racine et al., (2020) explored the moderating role of protective factors in the association between childhood adversity and child trauma-related distress, concluding that protective factors successfully buffered the manifestation of trauma-related distress after experiencing cumulative adversity as maltreatment or household dysfunction.
It has been established in research that children and adolescents at risk are a highly heterogeneous collective as they face different adversities and have different resources available to face them (Fitton et al., 2020; Guyon-Harris et al., 2021). For instance, the study performed by Guyon-Harris et al., (2021) explores the association between experiences of childhood maltreatment during pregnancy and the risk for disrupted parenting behavior before the birth of a child. The study focuses on the potential impact of different types or combinations of childhood maltreatment experiences on later parenting behaviors. Their results support this same statement as they obtain four different profiles of child maltreatment: low exposure, high exposure, high sexual maltreatment, and high physical and emotional maltreatment. Additionally, Fitton et al., (2020) conducted a comprehensive review and meta-analysis to examine the association between childhood maltreatment and different violent outcomes, their results also show that different types of maltreatment do not lead to the same violent outcomes. Thus, it is possible to distinguish different sub-populations which can be divided according to their gender, family environment, age, or country of origin. In this line, differences among genders have been found in protective factors as shown by Godbout et al., (2019) and Hartman et al. (2009), who have found that girls tend to score higher on personal resources than boys. However, there is no information on gender differences regarding other types of protective factors as far as we are concerned.
Additionally, children and adolescents at risk can also be subdivided based on the different family compositions. Although we have not found studies regarding family structure and protective factors, studies assessing risk have shown that a single-parent family might be at higher risk of child or adolescent maltreatment or neglect than a bi-parent family (Oliver et al., 2006), other more recent studies, however, point out that the risk is higher when the family structure is non-nuclear and not necessarily a one-parent family (Assink et al., 2019). A non-nuclear family means that the family does not have a regular structure, in the way that it could one a lone parent, but also could be that the child or adolescent is taken care of by any other proximate family member or that children or other partners have been included in the family structure. Similarly, age can be another key variable to distinguish this type of population, especially when it comes to risk and protective factors, as has been shown by Sun and Stewart (2007) who have found that younger children score higher than their older peers on variables that can be considered personal protective factors as empathy and communication.
Lastly, in the case of foreign children, and specifically in the case of unaccompanied asylum-seeking children, who comprise a specific type of sample considering that the family unit might be experiencing changes due to the migration process, we can point out that there are several protective factors on all three levels which can protect this high-risk group from developing negative outcomes, as stated by a recent systematic review performed by Höhne et al., (2020). In their study, they identified a stable environment and the type of accommodation act (both corresponding to the previously mentioned distant system) as protective factors. Secondly, social support and family contact, related to the proximate system also consistently appear as protective factors in this type of children. Lastly, personal characteristics as their gender or their cultural competences were highlighted as relevant against developing negative outcomes.
Apart from the relative lack of studies on protective factors, there is also a methodological issue relevant to this study. Regardless of the theoretical model in which child maltreatment, abuse, and neglect literature is positioned (measuring only risks or combining risks and protective factors), the analytical strategy in virtually all social sciences has been the variable-centered approach (Howard & Hoffman, 2018; Roesch et al., 2010). This approach focuses on the descriptions of the relationships among variables based on the assumption that the population of reference is a uniform group (Laursen & Hoff, 2006). However, there is enough argumentation on the heterogeneity of children and adolescents at risk (Fitton et al., 2020; Guyon-Harris et al., 2021; Swartout & Swartout, 2012) to suggest that there might be more informative approaches to be used to explore this population. The person-centered approach is an alternative analytical strategy. This approach emerges as a very useful technique for identifying groups of individuals among a population based on their response patterns (Wang & Wang, 2012). It might be highly informative in the context of children and adolescents at risk as it explores the multidimensionality of the constructs at the same time as the heterogeneity of response patterns of the population (Milne et al., 2021).
As far as we are concerned, there are only a small number of studies employing this approach in the protective factors from the risk of maltreatment, abuse, or neglect context. We found some studies on the profiles of the resilience construct. Green et al. (2021) examined the influence of individual, family, and contextual factors on the development of resilience profiles in children assisted by child protection services. This research concludes that living in a higher socio-economic status neighborhood, being of non-Indigenous origin, having a low likelihood of being maltreated, reporting higher levels of perceived support, and having parents with no criminal offenses were associated with stress resistance profile. In contrast, these indicators were not associated with the emerging resilience profile. Cases that were included in this profile presented a non-normative development. For its part, the research carried out by Oshri et al. (2017) found four specified classes concerning social skills: unresponsive-maladaptive, breakdown, emergent resilience, and stress-resistant. The first typology presented low social skills on the first wave with a slightly negative slope. The second one presented a high intercept and a large negative slope. The emergent resilience typology showed a low intercept and a large positive slope. Lastly, the stress-resistance class showed a high intercept and a slightly positive slope. This study concluded that differences among typologies were due to the access to resources. In short, resilience is a dynamic construct that is built and modified by the disposition of protective resources. Concerning behavioral adjustment, Proctor et al. (2010) found different trajectories for both the internalizing and externalizing behavior variables. In the first case, three trajectories were found: stable adjustment, mixed/decreasing adjustment, and increasing adjustment. In the case of externalizing behavior, four trajectories were found: stable adjustment, mixed adjustment, increasing adjustment, and stable maladjustment. In fact, behavior adjustment is understood from an ecological perspective as a protective factor of the individual against maltreatment.
Copeland-Linder et al. (2010) examined the relationship among community violence exposure, protective factors, and mental health in a sample of African-American adolescents. Protective factors were composed of self-worth, parental monitoring, and parental involvement. In their results, they obtained a three-profile solution composed of a vulnerable profile, a moderate risk/medium protection profile, and a moderate risk/high protection profile. These three classes had the capacity to successfully predict differences in depressive symptoms but not in aggressive behavior, in both genders, leaving therefore the future suggestion to utilize more specific measures on the variables and to expand this study to other community samples (Copeland-Linder et al., 2010). Another study performed by Brody et al., (2013) examined the cumulative socioeconomic status (SES) risk, allostatic load, and psychological adjustment conceptualized as a protective factor in a sample of 443 African-American youths. They obtained five distinguishable profiles which can be combined into two more general types of profiles named focal profiles by the authors: a physical health vulnerability profile characterized by high SES risk, high allostatic load, and low psychological adjustment problems; and a resilient profile characterized by high SES risk, low allostatic load, and low adjustment problems. However, these two studies have in common that they do not focus on children or adolescents considered at risk of maltreatment, abuse, or neglect.
Thus, a closer study of our reference population was performed by Anthony and Robbins (2013), who tried to extract different classes of early adolescents living in public housing neighborhoods based on their patterns of resilient development. They obtained three different classes: a “substance use and delinquency” class, a “limited support” class, and a “family risk but many resources” class. According to their results, we can perceive that all three classes score low on variables which can be considered protective factors, such as self-esteem, coping, academic efficacy, school commitment, and neighborhood cohesion, however, the clearest differences are in parental supervision and discipline, support, substance abuse, and delinquency.
In sum, we can conclude that more research on protective factors with a person-centered approach could add new light to the child maltreatment literature. Being able to identify and explain the different patterns and profiles in protective factors among children and adolescents at risk can shape more effective and tailored interventions that would prevent them from long-term suffering and social, personal, and economic negative consequences (Tufford et al., 2021; van der Put et al., 2017). Therefore, the aim of this study is to identify profiles among protective factors in children and adolescents at risk, and consequently relate these profiles to sociodemographic variables.
Method
Procedure
The data comes from the first wave of Development and Validation of the Adolescent and Children in Risk of Abuse and Maltreatment Scale, a longitudinal study aimed at assessing children and adolescents’ risk of maltreatment and protective factors. Data were collected from professionals involved in the Children Protective Services (CPS) in the Valencian Community (Spain). In order to ensure a comprehensive range of viewpoints, we included professionals from different positions, including psychologists (24.74%), educators (21.95%), social workers (19.98%), and other technical roles (33.33%). Professionals had medium years of experience in the field of 9.43 years (SD = 7.71). This diverse composition of participants allowed for a more comprehensive understanding of the subject matter. Professionals were asked to complete an electronic version of the Adolescents and Children Risk of Abuse and Maltreatment Instrument (ACRAM). They were required to complete the assessment for either the cases they were currently monitoring or for recently closed cases in which they had previously intervened (within the range of one month). In order to ensure accurate completion, professionals were provided with specialized online group training and precise guidelines on how to utilize, interpret, and distinguish unique aspects of the cases being evaluated. Additionally, each professional was instructed to evaluate a minimum of four cases, preferably with varying profiles. All participating professionals were given a two-week timeframe, starting from the day of training, to complete the online tool, as well as respond to several demographic questions. This study was approved and supported by the Valencian Government and met APA ethical standards. Professionals signed informed consent and were clearly informed about the anonymity of their answers.
Sample
The sample was composed of the assessment of 635 children and adolescents involved in the CPS. From this sample, 19 cases were discarded due to missing data, and therefore final sample had 616 children and adolescents. From these, 42.5% identified as females, 56.8% as males, and 0.6% as other. Mean age was 12.14 years (SD = 5.22). Additionally, 77.3% of the sample had Spanish nationality and the remaining 22.7% were coming from other countries than Spain. Additionally, Table 1 provides more detailed information on the sample characteristics.
Table 1
Descriptive variables
N
%
Nationality
Spanish
490
77.2
Non-Spanish
145
22.8
Gender of the child
Female
266
42.2
Male
365
57.8
Type of family structure
Single Parent Family
238
37.5
Bi-parent family
232
36.5
Others
151
23.7
M
SD
Age of the child (years)
12.16
5.24
ACRAM-PFS
F1. Children and adolescents’ Resources
1.79
0.65
F2. Family Resources
1.83
0.61
F3. Community Resources
2.09
0.64
N = n° of participants, M= Mean, SD= Standard Deviation,
ACRAM-PFS Adolescents and Children Risk of Abuse and Maltreatment Protective Factors Scale, M mean, SD standard deviation
Instruments
Protective factors were assessed by the Adolescents and Children Risk of Abuse and Maltreatment Protective Factors Scale (ACRAM-PFS; García-Mollá et al., 2023), a section of the modular instrument ACRAM (Navarro-Pérez et al., 2023). ACRAM is a comprehensive instrument composed of 97 risk or protection indicators divided into three general sections: (1) Risk factors associated to parental/caregiver behavior (ACRAM-PS); (2) Risk factors associated to the community (ACRAM-CFS); and 3) Protective factors (ACRAM-PFS). Additionally, there is another scale that deals with risks particular to foreign unaccompanied minors. Specifically, ACRAM-PSF is composed of three subsections: (1) children and adolescents’ resources (“The child or adolescent is able to channel their own emotions properly”), (2) family resources (“The family has social support and community networks”), and (3) community resources (“The child or adolescent is integrated into school and it is a positive factor for her/his socialization”). ACRAM-PSF is composed of 18 items. Response format has three options: “there is clear evidence it does not occur” (0), “there are signs it might occur, but it cannot be confirmed” (1), and “there is clear evidence it does occur” (2). For psychometric assessment of the scale “Information was not gathered” is treated as a missing value, because there was no option to get information on the indicator. In this sample, this scale showed an adequate internal consistency measured by McDonald’s omega of ω = 0.917 for the first factor, ω = 0.914 for the second, and ω = 0.798 for the third factor. Considering the aim of this study, only the ACRAM-PFS with its three factors was utilized.
Additional sociodemographic indicators were employed to validate the resulting protective factors profiles. Sociodemographic indicators included age, gender, country of origin (native as opposed to immigrant), and the type of family structure (being a single-parent family or a bi-parent family).
Statistical analyses
Descriptive statistics on focus variables were calculated to explore the data. Then, a latent mixture model was estimated, specifically, Latent Profile Analysis (LPA) with the three types of protective factors included as indicators for the latent profiles. This LPA was assessed according to the criteria stipulated by Lukociené et al., (2010) but also taking into account the interpretability of the profiles. The following test statistics were considered: Lo-Mendell-Rubin Likelihood Ratio (LMR LR) test, Bootstrapped Likelihood Ratio Test (BLRT), Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and entropy. LMR LR Test and BLRT both compare a certain model with k profiles to a model with k-1 profiles. A non-significant p-value would indicate no significant improvement of the k-profiles model compared to the k-1-profiles, more parsimonious, model (Wang & Wang, 2012). Lower values of AIC and BIC indicate improvement in relative model fit. Among all these indices and tests, the BIC and the BLRT have been shown to perform better (Nylund et al., 2007). Finally, entropy is a measure of classification quality, with values ranging from 0 to 1, and higher values indicating better classification. A value of 0.80 or higher is considered adequate (Clark, 2010). The LPAs were estimated using Robust Maximum Likelihood estimation (MLR).
Once a decision about the number of profiles to retain was taken, latent profile membership was extracted and related to the aforementioned sociodemographic variables. In order to do so, inferential tests as χ2 tests and Analysis of Variance (ANOVA), and their corresponding effect sizes, were calculated. Descriptive statistics, χ2 tests, and ANOVAs were computed using SPSS 26, while LPAs were estimated in MPlus 8.7 (Muthén & Muthén, 1998‐2021).
Results
Latent profiles of protective factors
We estimated LPAs from 1 profile (all people into a single group, maximum homogeneity, or baseline model) up to 8 profiles. Table 2 shows test statistics and fit indices for the eight LPAs. Values of entropy began to be acceptable in the models with 6, 7, and 8 profiles. BIC and AIC indices were lower in each consecutive model. Nevertheless, the 7 and 8 profile solutions were not statistically better than the 6-profile solution (in terms of the LMR Test), and had some very small size profiles difficult to separate from others. On the other hand, the 4-profiles solution had worse fit indices but was close in fit to the 6-profiles solution. However, a careful examination of the showed that three of the four profiles were only quantitatively different as they represent low protection, medium protection, and high protection. There was only one profile (with only 14.5% of the children) that showed a qualitatively different tendency: high family and community protective factors and low personal resources. If a clustering technique discerns only people in terms of how ranked they are in several variables, then they do not make much difference to use the quantitative differences between children in a variable-centered approach. On the other hand, the 6-profiles solution as we will see in the results showed two profiles as low and high, but the other four profiles had important qualitative differences, and this was a main point in order to make the final decision to retain six profiles. Therefore, we retained the 6-profiles model because it was better in terms of interpretability and it presented acceptable fit indexes according to the previously established criteria.
To interpret these protective profiles, we used estimated sample means across the three types of protective factors. Profile types can be graphically seen in Fig. 1. Based on the interpretation of these estimated means, profiles were labeled the following way: Low in all three protective factors (Profile 1); Medium in personal and community resources but low in family resources (Profile 2); Medium in personal and community resources but high in family resources (Profile 3); High in personal and community resources but low in family resources (Profile 4); High in family and community resources but low in personal resources (Profile 5); and finally, High in all three protective factors (Profile 6). The majority of the sample was comprised of profile 1 (27.3%) and profile 2 (22.7%). Profile 3 comprised 11.7% of participants, profile 4 contained 9.9% of participants, while profile five accounted for 17.4% of participants and finally, profile 6 was composed of the remaining 11%.
Fig. 1
Estimated means of types of protective factors based on the profile
×
Relations between the latent protective profiles and socio-demographic information
Membership to protective factors profiles was associated to several demographic variables. Results from the χ2 tests between protective factors profiles and gender (χ2 = 16.89, df=10, p = 0.077, V= 0.196), and type of family (χ2 = 15.51, df=15, p = 0.415, V= 0.092) were not significant and presented small effect sizes. However, there was one statistically significant association between protective factors profiles and country of origin (χ2 = 23.74, df=5, p < 0.01, V= 0.196) which also had a small to moderate effect size.
Percentages in each category and standardized corrected residuals of the significant association between protective profiles and country of origin are shown in Table 3. It can be appreciated that children and adolescents who come from foreign countries have a higher probability of belonging to profiles 4 (High in personal resources and community resources but low in family resources) and 6 (High in all three protective factors) while children and adolescents who belong to profile 1 (low protection in all three factors) have a higher probability to have Spanish nationality Table 4.
Table 2
Fit of models from 1 to 8 profiles
#C
AIC
BIC
Entropy
LMR Test
p
BLRT
p
1
3515.11
3541.65
–
–
–
–
–
2
3261.70
3305.93
0.691
251.62
<0.001
−1751.56
<0.001
3
3164.55
3226.47
0.828
101.21
<0.001
−1620.85
<0.001
4
3099.40
3179.01
0.794
70.41
<0.001
−1568.27
<0.001
5
3072.96
3170.27
0.794
33.14
>0.05
−1531.70
<0.001
6
3029.26
3144.27
0.809
49.76
<0.001
−1514.48
<0.001
7
3017.01
3149.71
0.806
19.49
>0.05
−1488.63
<0.001
8
2998.09
3148.49
0.816
25.91
>0.05
−1478.51
<0.001
AIC Akaike Information Criterion, BIC Bayesian Information Criterion, LMR Test Lo-Mendell-Rubin Likelihood Ratio test, BLRT Bootstrapped Likelihood Ratio Test
Table 3
Percentages and standardized corrected residuals for the association between protective factors profiles and country of origin
Profile
Country of origin
Foreign country
Spanish Nationality
1
Low protection in everything
%
14.9%
85.1%
Corrected residual
−2.8
2.8
2
Medium in PR and CR but low in FR
%
22.1%
77.9%
Corrected residual
−0.2
0.2
3
Medium in PR and CR but high in FR
%
25.0%
75.0%
Corrected residual
0.5
−0.5
4
High in PR and CR but low in FR
%
36.1%
63.9%
Corrected residual
2.6
−2.6
5
High in FR and CR but low in PR
%
16.8%
83.2%
Corrected residual
−1.6
1.6
6
High in PR, FR and CR
%
38.2%
61.8%
Corrected residual
3.2
−3.2
Total
140
476
%
22.7%
77.3%
PR personal resources, FR family resources, CR community resources
Table 4
Descriptive statistics and percentages of the resulting profiles
Profiles
N
%
F1. Personal resources
F2. Family resources
F3. Community resources
M
SD
M
SD
M
SD
1
Low protection in everything
168
27.3
1.14
0.18
1.24
0.27
1.60
0.52
2
Medium in PR and CR but low in FR
140
22.7
1.87
0.20
1.57
0.34
2.01
0.55
3
Medium in PR and CR but high in FR
72
11.7
2.15
0.18
2.44
0.30
2.37
0.48
4
High in PR and CR but low in FR
61
9.9
2.68
0.22
1.46
0.32
2.46
0.51
5
High in FR and CR but low in PR
107
17.4
1.28
0.23
2.39
0.34
2.18
0.64
6
High in PR, FR and CR
68
11
2.87
0.16
2.57
0.30
2.64
0.42
N = n° of participants
M mean, SD standard deviation
An Analysis of Variance (ANOVA) comparing the six profiles in age was performed. The ANOVA did not show any statistically significant differences in age across protective profiles and effect size was small: F (5, 610) = 1.729, p =0.126, η2 = 0.014.
Finally, the relationships between the socio-demographics and the profiles were also analyzed at the multivariate level. We performed a multinomial logistic regression with the profiles as the dependent variable, and gender, type of family, country of origin, and age. Overall, the four predictors had a Nagelkerke’s R-square value of 0.154. Country of origin produced several statistically significant differences among the profiles. Specifically, children from a foreign country were more likely to be in the high protection profile, compared to nationals who were more likely to be in the low protection profile (odd ratio= 0.343, p= 0.002), medium protection but with low family protectors (odd ratio= 0.419, p= 0.008), and the low personal protection profile (odd ratio= 0.289, p< 0.001). Additionally, the multinomial regression also found that as people age the likelihood of being high in protection increases (odd ratio= 1.12, p< 0.114), and that males were more likely to be low in personal protection than females (odd ratio= 2.87, p< 0.001).
Discussion and conclusion
The current study aimed to analyze latent profiles of protective factors among children and adolescents involved in child welfare services in the Valencian Community (Spain). In carrying out this study, we follow the line of previous studies such as those carried out by Copeland-Linder et al. (2010) and Anthony and Robbins (2013). As mentioned above, the most similar study to date is a latent class analysis of resilient development among early adolescents living in public housing performed by Anthony and Robbins (2013) although not comparable due to the different nature of the analysis and the variables utilized. Therefore, this study is the first to independently analyze the latent profiles of protective factors against child maltreatment by using an assessment instrument that specifically measures protective factors related to the individual, family, or caregiver, and environmental resources.
In our results, we retained six protective profiles. It is expected to obtain the three logical profiles indicating low, medium, and high ratings in these variables. However, we obtained other conceptually meaningful profiles, as the children and adolescents who are both medium in personal and community resources but differ in their family resources (profiles 2 and 3); the participants who are high in personal and community resources but low on their family resources (profile 4); or the ones who are high on their family and community resources but are low in their personal resources (profile 5). These findings confirm what is already known in scientific literature and professional settings: that not all children and adolescents at risk are the same and they do not face adversities with the same resources (Fitton et al., 2020; Guyon-Harris et al., 2021). In fact, as there are empirically distinguishable profiles, this information should be utilized to tailor more effective intervention programs in order to boost the protective factors that are the lowest in each case.
Consequently, we explored the relationship of the different protective factors profiles with the country of origin of the children or adolescents, their gender, their type of family composition, and lastly, their age. We attempted to understand if some socio-demographic characteristics make one or other profile more likely. However, we had non-significant results for gender, type of family, and age. These results indicate that profile’ differences are not likely depending on family type, age, or gender. This is not in accordance with literature on risks of child maltreatment where gender differences were found, which may lead to believe they might have different resources (protectors) to face maltreatment (Godbout et al., 2019). Specifically in protective factors, Hartman et al. (2009) also reported gender differences in these factors. These differences may be due to the fact that in the studies mentioned above, protective factors refer to personal resources or characteristics or the individual’s ability to find support from people close to them that have an impact on the reduction of mental health or behavioral problems. In contrast, this study also takes into account the resources that the family and the environment can provide to the individual to avoid risk situations or reduce their impact. Additionally, there is also evidence of age differences for child maltreatment as shown by Sun and Stewart (2007), who found that the youngest individuals scored higher in communication, empathy and help-seeking, school support, prosocial peers, meaningful participation in school activities and autonomy experiences than their oldest partners. Nevertheless, there is evidence about the uniformity among age groups of a protective factor as resilience (Meng, et al., 2018) which in fact is in accordance with our results.
However, there was a significant difference in the proportion of participants who were born in Spain and the ones who came from a foreign country among the different protective factors. Specifically, there were significantly more children and adolescents from Spain who scored low in all three protective factors than participants from other countries. Albeit, the proportion of foreign children and adolescents was higher in the profile of high personal and community resources but low family resources, and high in all three factors. This finding may have several explanations depending on whether they are children and adolescents who migrated with their families or whether they are unaccompanied asylum-seeking youths. On the one hand, the fact that families who have just arrived to a new country usually have no social and community support network to rely on and in most of cases need to invest efforts into achieving economic stability, therefore resulting in neglectful behaviors towards their children, who have been labeled as the “home alone generation” or “euro-orphans” (Levai et al., 2018). However, this result is also evidence for the healthy immigrant paradox, as there are more immigrant children and adolescents who are high in all three protective factors compared to native participants (Millett, 2016). On the other hand, in the case of unaccompanied asylum-seeking children and adolescents, they did not decide to migrate on their own will but are forced by their families because they are in a situation of poverty, mistreated by their family, or live in a war situation in their country of origin as stated by The United Nations High Commissioner for Refugees (2020). Given the cultural nature of the conception of childhood and adolescence, in most cases, the referent adults of these children force them to migrate to Western countries independently and with the perspective of allowing them to become self-sufficient, which is why they have lower scores on family-related protection resources (Navarro-Pérez et al., 2021). In this line, the sense of self-determination and the acquisition of autonomy (Deci & Ryan, 2000) does not compete with the need to feel secure after having participated in this type of migration process. Therefore, it is important to highlight the resilience of this population (Jafari et al., 2022; Pieloch et al., 2016) and their means to find social support from their peers (Keles & Oppedal, 2022). Additionally, these differences among the country of origin of the subjects corroborate the results reported by Yu et al. (2020), since the protective factors respond to cultural variables.
Although the main strength of this research is its novelty and possible applications in real-life professional settings, it also has limitations: Firstly, the fact that this study has been performed in a sample composed of institutionalized children and adolescents of child protective services selected with non-probabilistic methods since the information was completed by the professionals in the care of children. Secondly, the definition of the protective factors used in the scale, as the ACRAM scale was developed based on the judgment of child protective services workers in the development and content validation of the ACRAM (Navarro-Pérez et al., 2023). This fact implies certain difficulties for the generalization of results given the specific legislation and conceptualization of protective factors in this territory. Thirdly, as the ACRAM-PFS is a hetero-administered instrument, there is always a certain margin of subjectivity, especially when it comes to cases which are not currently worked with but have to be recalled from the close past.
The results of this study can be applied in several ways: Firstly, this information can be utilized in elaborating tailored intervention plans for children and adolescents as being aware of the strengths of the individual, the family, or the community can be a starting point to intervene with. Secondly, these findings can also be useful in the elaboration of prevention plans, as being knowledgeable of the different profiles can help strengthen the protective factors that are not so high and effectively prevent adverse experiences from occurring. Thirdly, the knowledge gained from this study can be incorporated into training programs for professionals working in child welfare services. Professionals can be equipped with a deeper understanding of protective profiles and their implications for intervention. This can enhance their assessment skills, intervention planning, and overall effectiveness in supporting at-risk children and adolescents. Lastly, the results of this study are novel as it is performed with a representative sample, utilizing a comprehensive protective factors measurement, with a person-based approach, and with a specific focus on the practical implications of the knowledge. By highlighting the actionable implications of the findings, the study bridges the gap between research and practice, providing valuable insights for professionals working in the field of child welfare.
Finally, we strongly suggest that future research should explore how latent profiles evolve over time in relation to individual, family, and contextual factors, in line with research by Green et al. (2021). For its part, studies focusing on individual factors, such as those performed by Oshri et al. (2017) and Proctor et al. (2010), form the basis for future lines of research on individual differences in protective factors against maltreatment. More research in different countries and samples is needed to be able to draw more reliable conclusions. It is concluded that the study of these trajectories is necessary for the generation of intervention plans based on information from children and adolescents (Proctor et al., 2010).
In conclusion, while additional research is still needed, results from this investigation shed light on the way protective factors against child maltreatment personally relate to individuals. This information is novel and relevant for the development of future interventions tailored to the individual characteristics of the cases as both weaker and stronger points of children. Adolescents and their families can be targeted and therefore effectively addressed from an ecological perspective, as suggested by Begle et al. (2010), who support the use of comprehensive methods and tools for the detection and analysis of child maltreatment.
This work was supported by the University of Valencia (OTR2021-07233CATED). Additionally, the researcher Sylvia Georgieva is receiving a Ph.D. grant from the Valencian Government (ACIF/2021/175). Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.
Compliance with ethical standards
Conflict of interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Met BSL Psychologie Totaal blijf je als professional steeds op de hoogte van de nieuwste ontwikkelingen binnen jouw vak. Met het online abonnement heb je toegang tot een groot aantal boeken, protocollen, vaktijdschriften en e-learnings op het gebied van psychologie en psychiatrie. Zo kun je op je gemak en wanneer het jou het beste uitkomt verdiepen in jouw vakgebied.
