Initially, 619 responses were recorded in LimeSurvey, a secure online platform. Of these, 328 were incomplete and were therefore discarded, leaving 291 complete responses for review. Strict inclusion and exclusion criteria were applied to obtain a homogenous sample aligned with the objectives of this study. Inclusion criteria required participants to be biological mothers or fathers of Spanish nationality, with children aged 4 to 19 years, and without diagnosed developmental disorders or mental health conditions. Exclusion criteria included non-Spanish respondents (n = 42), adoptive or foster caregivers and professionals working in foster care (n = 23), and parents who reported that their children had neurodevelopmental disorders (n = 15), mental health problems (n = 10), or high abilities (n = 2). Exclusion reasons were not mutually exclusive; when more than one exclusion criterion applied, the case was counted once in the total number of exclusions (n = 91) but was included under each relevant reason. Specifically, cases excluded for neurodevelopmental problems included ADD/ADHD (n = 9), autism spectrum disorder (n = 5), and intellectual disability (n = 2). Further exclusions were made for children with anxiety or comorbid conditions (n = 9), dyslexia (n = 1), selective mutism (n = 1), and conduct disorder (n = 1).

Therefore, the final sample comprised 200 Spanish parents of children and adolescents aged 4–19 (mean [M] = 11.61; standard deviation [SD] = 4.23) from the general population. Among the children and adolescents, 53% were female (n = 106). Among the children and adolescents assessed, 12.5% (n = 25) were in early childhood education (ages 4–5), 32.0% (n = 64) were in primary school, 34.5% (n = 69) were in compulsory secondary education (ESO, ages 12–16), and 21.0% (n = 42) were enrolled in upper secondary education (Baccalaureate or vocational training).

Parents’ ages ranged from 25 to 62 years (M = 45.66; SD = 6.60). Regarding responding parents, 84% were mothers/females (n = 168) and 16% were fathers/males (n = 32). Regarding educational level, 54% (n = 108) held a university degree, 26.5% (n = 53) had completed upper secondary education (Baccalaureate or vocational training), 17% (n = 34) had finished compulsory secondary education, and 2.5% (n = 5) had not completed mandatory schooling.

The sample size was determined by calculating the number of cases required for each dimension to analyze the multivariate model of the associations among the items comprising each subscale. A 95% confidence level was achieved, and De Moivre’s theorem was satisfied, which requires an estimated frequency greater than 10 to provide stable estimates (Hazewinkel, 2001). In addition, specific guidelines for confirmatory factor analysis (CFA) were followed: for models with five or fewer constructs, a minimum of 100 cases is recommended (Kline, 2016), and at least 10 participants per estimated parameter or per item are advisable (Boomsma, 1982; MacCallum et al., 1999). Given that the RPQ comprises 10 items, a sample of 200 participants meets and exceeds these conventional criteria, supporting the adequacy of our sample size for CFA.

Ad hoc questionnaires were developed to collect sociodemographic and clinical information from both parents and children or adolescents. These questionnaires captured key data, including the caregiver’s country of residence, nationality, age, gender (male, female, or other), relationship to the child or adolescents (e.g., mother, father, or legal guardian), and highest level of education completed. Parents provided information on the children’s and adolescent’s age, gender, nationality, current educational level, mental or neurodevelopmental health conditions (such as anxiety, depression, ADHD, intellectual disability, or autism spectrum disorder), physical health conditions (e.g., asthma, migraines, or allergies), medications taken for physical or psychological health, and recognized disability status. For children or adolescents reported to have been placed in foster care, information was collected on the duration of time spent in the social protection system and the reasons for the protective measures (e.g., abuse, neglect, abandonment or parental relinquishment).

This study used the English version of the RPQ, a tool for screening symptoms related to attachment disorders, completed by parents and teachers (Minnis et al., 2002). The instrument is available in both an 18-item version and a 10-item version. Given the strong psychometric properties demonstrated in previous studies (Minnis et al., 2007; Minnis et al., 2009; Schröder et al., 2019; Vervoort et al., 2013), the higher response rate associated with reduced fatigue, (Kling et al., 2019) the chance to cover large populations (Urzúa and Navarrete, 2013), and the opportunity to include several instruments in a single survey (Rammstedt and Beierlein, 2014), the 10-item version was used in the present study.

The RPQ is based on the diagnostic criteria for attachment disorders outlined in the DSM-IV (APA, 1994) and has been used in both general and clinical populations, as reported by Minnis et al. (2007) and Vervoort et al. (2013), respectively.

The RPQ has a two-factor structure, as supported by findings from several studies (Minnis et al., 2002; Minnis et al., 2007; Minnis et al., 2009; Schröder et al., 2019). The disinhibited subscale comprises four items (1, 2, 3, and 6) (e.g., “Gets too physically close to strangers”), while the inhibited scale comprises six items (4, 5, 7, 8, 9, and 10) (e.g., “There is a false quality to the affection they gives”). Each item is rated on a four-point scale: “Not at all like”, “A bit like”, “Like”, and “Exactly like”, scored 0, 1, 2, and 3, respectively. The total score is obtained by summing the scores of items 1–10.

The RPQ demonstrates strong psychometric properties. It has good internal consistency, with Cronbach’s alpha values ranging from 0.75 to 0.85 for both the inhibited and the disinhibited subscales. Its two-factor structure has been validated through factor analysis, proving its reliability in distinguishing these behaviors. Measurement invariance tests confirmed its consistency across multiple informants, such as parents and teachers. The RPQ correlates well with other established RAD assessment tools and shows adequate sensitivity and specificity for identifying attachment disorders (Vervoort et al., 2013; Schröder et al., 2019). Different adaptations of the questionnaire have adequate psychometric properties, such as the Persian version (Heshmati et al., 2018) with a Cronbach’s alpha of 0.79 and a construct validity of .50, or the German version with a Cronbach’s alpha of 0.82 (Schröder et al., 2019).

Regarding translation, the Spanish version of the RPQ was developed through a rigorous cross-cultural adaptation process, following established international guidelines (Beaton et al., 2000; World Health Organization, 2016). Two bilingual psychologists independently translated the original English version into Spanish. The translations were synthesized into a single version through consensus, balancing linguistic accuracy and cultural relevance. This version was then back-translated into English by two native English-speaking professionals who were unfamiliar with the original instrument. The back-translations were compared with the original version to identify and resolve any semantic or conceptual inconsistencies. A panel of experts reviewed the final adapted version to ensure equivalence and validity before its application.

Behavioral and emotional problems were evaluated using the SDQ (Goodman, 2001), in its Spanish version adapted by Ortuño-Sierra et al. (2015). The SDQ is a screening questionnaire commonly used for assessing various emotional and behavioral issues related to mental health in children and adolescents. In the present study, the parent-completed version was used. The SDQ consists of 25 items divided into five subscales: Emotional Symptoms (e.g., “Many worries”), Conduct Problems (e.g., “Often has temper tantrums or hot tempers”), Hyperactivity (e.g., “Easily distracted, concentration wanders”), Peer Problems (e.g., “Rather solitary, tends to play alone”), and Prosocial Behavior (e.g., “Considerate of other people’s feelings”). Each item is rated on a 3-point Likert scale, with each subscale scoring between 0 and 10 points. ‘Somewhat True’ is consistently scored as 1, while ‘Not True’ and ‘Certainly True’ vary with each item.

The total difficulties score is calculated by summing the scores from all subscales except for the prosocial scale, resulting in a total score ranging from 0 to 40. If any of the four items comprising a subscale is missing, the Total Difficulties score is also recorded as missing. The externalizing score, which ranges from 0 to 20, is calculated by summing the conduct problems and hyperactivity scales. Similarly, the internalizing score, also ranging from 0 to 20, is calculated by adding the emotional symptoms and peer problems subscales. The classification of scores in the SDQ is based on cutoff points that categorize results into different groups (‘close to average’, ‘slightly raised’, ‘high’, or ‘very high’ groups), providing an understanding of how the child’s behavior compares with the general population in Spain (Barriuso-Lapresa et al., 2014).

The SDQ has demonstrated solid psychometric properties (Goodman, 2001; Stone et al., 2010), including good internal consistency, with Cronbach’s alpha generally above 0.70 across most subscales, although the Peer Problems and Conduct Problems subscales occasionally show lower values. Its five-factor structure—emotional symptoms, conduct problems, hyperactivity, peer problems, and prosocial behavior—has been widely replicated. Additionally, the SDQ has demonstrated robust validity, supporting its reliability for assessing emotional and behavioral difficulties in children and adolescents in Spain (Ortuño-Sierra et al., 2016; García et al., 2000).

This study employed a cross-sectional design, using anonymous self-report questionnaires completed by parents of children and adolescents. Participants were recruited through internet platforms and social networks using a non-probabilistic snowball sampling method. All respondents were informed about the study’s purpose, the voluntary nature of their participation, the anonymity of their responses, and the confidentiality of the data collected prior to participation. Participants were also assured that no personally identifying information—either theirs or the child’s—would be recorded and that they could withdraw from the study at any time without any consequences.

Participants provided informed consent before accessing the questionnaire. Data collection was conducted online via LimeSurvey, which was accessible from any internet-enabled device (e.g., smartphone, tablet, or computer). The surveys were distributed through the authors’ Instagram and LinkedIn accounts. The survey was available between December 19, 2023, and December 1, 2024, and took approximately 10 minutes to complete. The study was approved by the Ethics Committee of the University of Valencia (reference code: 2023-MAG-2735572). All data were anonymized and securely stored on a computer belonging to the institution. Response patterns were examined (e.g., uniform responses across all items, inconsistencies between reverse-coded items and regular items), and no cases of careless or invalid responding were detected. These procedures ensured that the final sample was robust and aligned with the study’s aim of validating the RPQ in typical Spanish family contexts.

Before full data collection, a pilot survey was administered to a small group of Spanish-speaking parents (n $\approx$ 15) to assess the clarity, cultural appropriateness, and usability of the Spanish version of the RPQ. Feedback from this pilot phase prompted minor wording adjustments that improved item comprehension. These revisions were implemented prior to launching the final version of the instrument.

The RPQ was selected based on its strong psychometric performance in prior studies, including its use in clinical populations with suspected RAD. It has demonstrated robust internal consistency, factorial validity, and clinical utility in identifying RAD and DSED in children and adolescents (Talmón-Knuser et al., 2024; Vervoort et al., 2013).

Data analysis was conducted using several tools. SPSS software (Version 29; IBM Corporation, Armonk, NY, USA) was used for descriptive analyses and calculation of internal consistency, while EQS (version 6.3; Multivariate Software, Inc., Encino, CA, USA) was used for CFA and average variance extracted (AVE). Microsoft Excel was used to calculate composite reliability. CFA was validated with the Satorra-Bentler index and maximum likelihood estimation. At the same time, goodness-of-fit was evaluated using the comparative fit index (CFI), incremental fit index (IFI), and root mean square error of approximation (RMSEA). Values below 0.08 were considered acceptable. Sample size adequacy for the CFA was determined a priori using common guidelines ($\ge$10 participants per item). With 10 items and N = 200, the sample satisfied this criterion (Boomsma, 1982; MacCallum et al., 1999).

The RPQ consisted of 10 items distributed across two dimensions. Item analysis indicated that both dimensions demonstrated adequate reliability (α for the disinhibited subscale = 0.89; α for the inhibited subscale = 0.82). Moreover, the results indicated that removing any item would not improve the reliability indices (Table 1).

After analyzing the psychometric properties of the items, the internal validity of the RPQ was evaluated using CFA. Since the RPQ does not yield a total score and consists of independent subscales, the suitability of the data was assessed separately for each subscale through the Kaiser-Meyer-Olkin (KMO) test and Bartlett’s test of sphericity. Results indicated that the data were suitable for factorial analysis. For the Disinhibited subscale, the KMO index was 0.800, and Bartlett’s test was statistically significant ($\chi$² = 511.615, df = 6, p 0.001). Accordingly, for the Inhibited subscale, the KMO index was 0.822, and Bartlett’s test was also significant ($\chi$² = 409.219, df = 15, p 0.001). These values confirm the adequacy of the data for factorial validation.

An EFA was not performed since the factorial structure has been previously theoretically established. Instead, CFA was performed to test whether the proposed structure adequately fit the data.

Although this study makes an important contribution to the Spanish-speaking context by adapting and examining the psychometric properties of the RPQ, several limitations warrant consideration. First, although online data collection allowed access to a larger population in a shorter time frame and facilitated participant responses at a time and place of their choosing (Chuey et al., 2021), the nonprobabilistic nature of the sample and its limited size restrict the generalizability of the findings to the broader population. For the same reason, the percentile reference table included in the manuscript should be interpreted with caution. Our intention was not to provide definitive normative data, but rather to offer preliminary reference values that may serve as a guide for researchers and clinicians until larger normative studies become available. Although the child sample was evenly split by sex, the informants were predominantly mothers. This participation pattern is common in parent-report studies (Valero-Moreno et al., 2021) and may limit the generalizability of the findings to fathers and other male caregivers.

Additionally, the age range of the sample did not include children under 4 years of age, even though symptoms related to RAD typically emerge between 9 months and 5 years of age (APA, 2022). Future studies may address this limitation by including younger participants and exploring age-specific differences in symptom presentation. Although the sample size (N = 200) met conventional adequacy criteria for CFA (approximately 10 participants per item), it was insufficient to test measurement invariance across subgroups with adequate power. This remains a goal for future research.

Second, this study was conducted in a nonclinical community sample and was not intended for diagnostic decision-making. Although the RPQ demonstrated adequate psychometric properties, further research using clinical samples is essential to assess its diagnostic utility. Specifically, establishing clinically meaningful cutoff points and conducting ROC curve analyses would require the comparison with gold-standard diagnostic tools, whereas the present study was limited to a normative sample. Therefore, this study does not aim to establish diagnostic thresholds but rather to provide preliminary evidence of the RPQ’s reliability and validity in the Spanish context. Future research should also examine the RPQ in populations with neurodevelopmental disorders, as attachment difficulties often co-occur with conditions such as ADHD or autism spectrum disorder (Talmón-Knuser et al., 2023). This approach could provide valuable information about the RPQ’s sensitivity and applicability in more complex clinical presentations. Additionally, future longitudinal studies would be important to assess test–retest reliability and further strengthen the psychometric evidence of the instrument.

Finally, although atypical response patterns were checked and no invalid cases were identified, future research should incorporate formal infrequency measures, such as the Oviedo Infrequency Scale, to further strengthen data quality control. Furthermore, only limited sociodemographic information (e.g., parents’ gender, age, and educational level) was collected, restricting a more detailed characterization of the sample. Future research should also examine the convergent validity of the RPQ using multiple informants (e.g., teachers, clinicians) or complementary observational methods to better understand the expression and development of attachment-related symptoms across different environments. Additionally, validating the instrument in other cultural contexts would be valuable for examining potential cross-cultural variability in symptom patterns and caregiver perceptions.