The Italian version of the Reflective Functioning Questionnaire: Validity within a sample of adolescents and associations with psychological problems and alexithymia

Abstract Objectives This study aims to test the psychometric proprieties of the Reflective Functioning Questionnaire (RFQ) applied to younger (13–16 years) and older (17–20 years) Italian adolescents examining (1) the factorial structure of RFQ; (2) its invariance across age and sex; (3) correlations between RFQ subscale scores, as well as the associations of the RFQ with (4) psychological problems and alexithymia dimensions. Methods A cross‐sectional study was conducted with 593 adolescents between the ages of 13 and 20 years old recruited from the community within Italy. These participants completed the RFQ, Symptom Checklist‑90, and Toronto Alexithymia Scale. Results The two‐factor structure of the RFQ was confirmed. However, higher internal consistency of RFQ was obtained by removing two items that seemed problematic within this sample. Using a six‐item version of RFQ, the two‐factor structure was invariant across adolescent age and sex. Significant correlations among RFQ subscale scores, and between RFQ subscales with both psychological problems and alexithymia dimensions were found. Conclusions Preliminary results reveal a short version of RFQ (six‐item) is a suitable measure to assess mentalizing in adolescents in the Italian context.

This version of the RFQ has also been used with Polish adolescents (Gambin et al., 2021), to test the link between mentalizing difficulties, schizotypal personality features, and thought problems in a sample of community adolescents (Salaminios et al., 2020), as well to investigate the longitudinal associations between RF and empathy on potential changes in externalizing behaviors over time . Among Italian samples, the RFQ has not yet been used with adolescents but only validated for adults (Morandotti et al., 2018). This represents a gap in the literature.

| RF, adolescence, and psychological adjustment
Adolescence is a sensitive period characterized by rapid changes in biological, psychological, and motivational systems and interpersonal relationships. There is a social reorienting wherein the opinions of peers become more important than those of family members. The psychological adjustment of adolescents is influenced by the interaction between the development of socio-cognitive capacities and challenges from the environment (Blakemore & Mills, 2014), and mentalizing is one such socio-cognitive capacity that has gained much attention in the last two decades (Fonagy et al., 2002). Mentalizing develops across adolescence before stabilizing in the early 20s and reflects changes in neurocognitive strategy and/or neuroanatomy (Blakemore & Mills, 2014). This is supported by some studies that find lower mean overall RF scores in adolescents as compared with adult samples (Borelli et al., 2015, as well as studies demonstrating that adolescents' RF is related to their sex (Protic et al., 2020), or otherwise is related to age (Cropp et al., 2019).
Studies on adolescents' RF also suggest that the increase of psychopathology in this transitional period (Kessler et al., 2005) may be linked to mentalization deficits (e.g., Fonagy et al., 2002). Specifically, mentalizing difficulties in adolescence are associated with borderline personality pathology (Bo & Kongerslev, 2017;Duval et al., 2018;Sharp et al., 2013), as well as externalizing Sharp & Venta, 2012), and internalizing disorders (Bizzi et al., 2019;Chow et al., 2017). Psychopathic traits of adolescents are correlated with aggressive behavior/externalizing symptoms in the case of low or moderate RF (Cropp et al., 2019;Taubner & Curth, 2013). In addition, several studies (e.g., Badoud et al., 2015;Gambin et al., 2021;Morosan et al., 2020;Salaminios et al., 2020) using the RFQ version for adolescents show that higher scores on the Uncertainty subscale (i.e., uncertainty about mental states of self and others) and lower scores on the Certainty subscale (i.e., certainty about mental states of self and others) are associated with higher difficulties in emotion regulation, higher levels of alexithymia, borderline personality traits, schizotypal features, as well as internalizing and externalizing symptoms in adolescence.
Although these previous studies have highlighted a link between mentalizing deficits and clinical problems and have also examined the distinctions between mentalizing in adolescence as compared with adulthood, it remains to be seen whether mentalizing varies by according to age and sex within adolescence and whether difficulties in mentalizing may constitute a vulnerability factor also in healthy adolescents. If the RFQ is related to age, sex, clinical variables, and alexithymia within a community sample in Italy, it would establish that is an important construct to study within this population. Similarly, the age and sex differences would help to understand how RF is distributed within this culture and help to know potential confounds that should be measured in future studies. Therefore, this study aims to fill these gaps by testing the psychometric properties of RFQ in early and older adolescents.

| Aims and hypotheses
Since the RFQ was developed in the English language (Fonagy et al., 2016), and has since been translated into French and administered to adolescents in France (Badoud et al., 2015) and in some European countries (Gambin et al., 2021;Morosan et al., 2020;Salaminios et al., 2020), our first goal was to administer the RFQ within Italy, and to test preliminary hypotheses regarding its psychometric properties among two groups of adolescents, younger (13-16 years) and older (17-20 years) adolescents. Specifically, we expected to find evidence for a two-factor structure of the RFQ, referring respectively to the degree of subjective confidence (i.e., Certainty, RFQc) or doubt (i.e., Uncertainty, RFQu) that actions are mentally driven. Furthermore, we predicted that this two-factor structure would be invariant across (i) younger and older adolescents, (ii) male and female adolescents. Subsequently, following prior validity work on the French version of RFQ (Badoud et al., 2015), we sought to examine correlations between RFQ subscale scores with clinical variables (i.e., psychological problems), as well as psychological capacities (i.e., alexithymia dimensions) that have been previously linked with RF both theoretically and empirically. We hypothesized that there would be negative associations between the degree of certainty concerning mental states (RFQc), psychological problems, and alexithymia dimensions, as well as positive associations between these variables and the degree of uncertainty concerning mental states (RFQu).

| Participants
We recruited N = 593 Italian adolescents (334 [56%] males) between the ages of 13 and 20 (M = 16.36, SD = 1.57) from the community. All participants were of national origin and had heterosexual parents. Inclusion criteria included being an adolescent within our target age range (between 13 and 20 years) and speaking fluent Italian. The final sample was divided by age into a group of 378 younger adolescents (209 males, aged 13-16 years old; M AGE = 15.35, SD AGE = 0.87) and a group of 215 older adolescents (125 males, aged 17-20 years old; M AGE = 18.12, SD AGE = 0.73). Younger and older adolescents differed significantly in terms of participant age (t(592) = −39.53, p = .000), but they did not differ significantly in terms of participant sex (male: 55% in younger vs. 58% in older adolescents; χ 2 = 0.45, p = .501) or family composition (biparental family: 87% in younger vs. 90% in older adolescents; monoparental family: 13% in younger vs. 10% in older adolescents; χ 2 = 1.15, p = .283).

| Measures
The Reflective Functioning Questionnaire (RFQ; Fonagy et al., 2016) is a questionnaire used to evaluate mentalizing abilities by measuring the degree of certainty and uncertainty with which individuals utilize mental state information to understand their own and others' behavior. The Certainty about Mental States (RFQc) subscale consists of six items focusing on the extent to which individuals disagree with statements such as "I don't always know why I do what I do." All items are scored by participants on a 7-point Likert-type scale, ranging from "completely disagree" to "completely agree." Items are subsequently rescored to capture more extreme levels of certainty, so that very low agreements on this scale reflect hypermentalizing, while some agreement reflects adaptive levels of certainty about mental states. To this end, these items are recorded to 3, 2, 1, 0, 0, 0, 0. The Uncertainty about mental states (RFQu) subscale, which in the extreme captures hypomentalizing, also consists of six items scored on the same 7-point Likert-type scale. Responses to items such as "Sometimes I do things without really knowing why," are recoded to 0, 0, 0, 0, 1, 2, 3, again to ensure that high scores reflected a stance characterized by an almost complete lack of knowledge about mental states, while lower scores reflect an acknowledgment of the opaqueness of one's mental states and that of others, typical of genuine mentalizing. The items are described in Table 1. For this study, we used the Italian version of the test provided by Fonagy et al. (2016) and available at https://www.ucl.ac.uk/ psychoanalysis/research/reflective-functioning-questionnaire-rfq.
The Symptom Checklist-90-Revised (SCL-90-R; Derogatis & Savitz, 1999;Prunas et al., 2012) is a self-report questionnaire designed to assess psychological problems and psychopathological symptoms in individuals 13 years and older. This original measure, as well as the validated Italian version (Prunas et  The Toronto Alexithymia Scale (TAS-20; Bagby et al., 1994;Bressi et al., 1996) is a 5-point Likert-type self-report 20-item questionnaire to assess the alexithymia (i.e., a personality trait characterized by the subclinical inability to identify and describe emotions experienced by one's self or others) in the total level and the three aforementioned factors, as difficulties in identifying feelings (DIF), difficulties in describing feelings (DDF), and lack of focus on internal emotional experiences (EOT). In this study, the internal consistency has been reported as good for all subscales (Cronbach's alpha values between .81 and .89).

| Procedure
The procedure was approved by the local University Research Ethical Committee and complied with the ethical standards of the international scientific community. Data were collected within 2020. The adolescents were recruited from two high schools (a high school and a technical institute) in northern Italy (from the first to the fifth grade); every student in each grade was contacted at school by a medical resident belonging to the research team for a total of 605 participants, of which only 12 did not participate because they were absent at the time of the data collection. Therefore, the final sample (N = 593) represented 98% of those eligible.
All the parents and participants were informed of the research goals. In line with ethical requirements, it was emphasized that participants' cooperation was voluntary and that their answers were confidential and used only for the study. Written informed consent was received from the participants (and the parents of adolescents if they were younger than 18 years old). The assessment was administered to the adolescents at school in a data collection session lasting around 45 min.

| Statistical analyses
Statistical analyses were carried out with Statistical Package for the Social Sciences (SPSS, Version 24.0; IBM Corp.) and JASP (JASP Team, 2020). Demographic variables were assessed using descriptive statistics (i.e., χ 2 test and Student's t tests). According to our aims, analyses were assessed following three steps. First, we examined the | 507 supposed two-factorial structure of RFQ using confirmatory factor analysis (CFA). Second, we used multigroup CFA to investigate the factorial invariance of RFQ in both groups (male vs. female adolescents; younger vs. older adolescents) testing for configural, metric, scalar, and strict measurement invariance. We used fit indices to test model fit using cut-off values generally indicating a good fit (Hu & Bentler, 1999): χ 2 /df ratio (<3 acceptable), comparative fit index (CFI over 0.90 acceptable), Tucker-Lewis fit index (TLI over 0.90 acceptable), root mean square error of approximation (RMSEA between 0.05 and 0.08 acceptable), and standardized root mean squared residual (SRMR < 0.08 acceptable). A comparison between different factor solutions has been made with a χ 2 difference test and a drop in CFI greater than 0.005 (Chen, 2007). Third, Pearson correlations were used to test the association between RFQ, demographic features, clinical variables (psychological problems), and alexithymia dimensions.
As the scale's reliability was not supported, we deleted items 2 and 7, testing a six-item RFQ model, which did not  Table 4).  Table 5). About younger and older adolescents, we tested configural invariance (Model 3), metric invariance (Model 4), scalar invariance (Model 5), and strict invariance (Model 6) obtaining a good model fit in every model (see Table 5 for the different models' fit values), and thereby our findings confirm factorial invariance across both younger and older adolescents' groups. About male and female adolescents, we tested configural invariance (Model 7), metric invariance (Model 8), scalar invariance (Model 9), and strict invariance (Model 10), obtaining a good model fit in every model (see Table 5 for the different models' fit values), and thereby our findings confirm factorial invariance across both male and female adolescents' groups. In total, our findings confirm the two-factorial solution of six-item RFQ and indicate factorial invariance across both younger and older adolescents' groups and male and female adolescents' groups.

| Correlations with demographic variables, psychological problems, and alexithymia dimensions
Means and standard deviation of SCL-90 (to assess psychological problems) and TAS-20 (to investigate alexithymia dimension) are reported in Table 6.
Using the six-item version of RFQ, neither RFQ subscale was related to demographic features (age, sex, family composition). Correlations with psychological problems and alexithymia dimensions obtained from self-report measures are shown in Table 7

| DISCUSSION
The present study examined the psychometric proprieties of the Italian versions of the RFQ with a community sample of adolescents. We first predicted we would find evidence for a two-factor structure, referring respectively to the degree of subjective confidence (i.e., certainty, RFQc) or doubt (i.e., uncertainty, RFQu). Consistent with our expectations and with prior research (Badoud et al., 2015;Fonagy et al., 2016), results revealed a two-factor structure (RFQc and RFQu subscales) of the RFQ. Therefore, our study revealed support for these different aspects  of mentalizing in Italian adolescents. Further, we predicted that the two-factor structure of RFQ in adolescence was invariant across younger (13-16 years) and older adolescents (17-20 years), as well as across males and females.
The results confirmed the expectations. However, higher internal consistency of RFQ was obtained by removing two items (item 2 and item 7) that seemed problematic within this population. In particular, item 2 was formulated with a negative sentence, and for cultural reasons, this item might be challenging to understand and might not assess what it is supposed to assess. In addition, the use of the adverb "always" in both items could confuse the answers and the inclusion of the word "always" could be more complicated also in the sentence with the negation where "not always" could correspond to "sometimes." Therefore, from this perspective, a six-item RFQ model significantly improved the internal structure of RFQ as well as its clarity in understanding. Despite the multiple developmental changes that are inherent to adolescence and the differences in terms of sociodemographic data (i.e., sex) found in previous studies of reflective ability (Cropp et al., 2019;Protic et al., 2020), no differences were found comparing younger and older adolescents nor males and females on the RFQ. This finding, which contradicts the theoretical framework of mentalization may suggest that the RFQ is not sufficiently sensitive to sex differences or may be due to the characteristics of our sample and should be verified in future studies. Overall, our findings show that the six-item version of RFQ is a suitable measure to assess mentalizing capacity in adolescents on a large scale (i.e., easily applied with large samples).
Our last hypothesis was that the RFQ subscale scores were related to clinical variables (i.e., psychological problems) and psychological capacity (i.e., alexithymia). Results revealed negative associations relatively modest in size (Cohen, 1988) between the degree of certainty concerning mental states (RFQc) with psychological problems and alexithymia dimensions, as well as a reverse pattern of associations between the factors listed above and the degree of uncertainty about mental states (RFQu). These results are consistent with the mentalizing framework (Fonagy et al., 2016) suggesting that the RFQ_u may be a good marker of typical features associated with clinical problems, while the RFQ_c may be a marker of psychological adjustment for adolescents. In this sense, the poorer an adolescent's mentalizing, the more problems concerning affect dysregulation and emotional problems can be expected (Fonagy et al., 2016).  Badoud et al. (2015) findings, we found associations between all subscales of the alexithymia questionnaire (also the externally oriented thoughts subscale) and each of the two RFQ subscales, suggesting that as discovered by Gambin et al. (2021), despite using the Difficulties in Emotion Regulation Scale rather than the TAS-20, mentalizing abilities play a very important role in emotion regulation by providing awareness and understanding of own and others mental states, as well as ability to correctly identify and manage emotional states. Similarly, but in the opposite direction, adolescents' poor mentalizing seems to reflect general emotional distress, including a difficulty in recognizing and distinguishing between feelings and bodily sensations, in describing feelings to others, and restricted imagination, marked by the paucity of fantasies, dreams, and daydreaming. However, these findings need to be interpreted with caution as the clinical problems and alexithymia do not reach the clinical cut-off.
We believe it is useful to contextualize the contribution of our study in terms of its strengths and limitations. A strength of this study is the large number of participants, as well as the focus on different age ranges within the broader adolescent phase. However, some methodological limitations must be mentioned. First, our analyses and conclusions are exclusively based on self-report questionnaires. As a result, studies are now needed to investigate mentalizing in adolescence using multi-method approaches. A further weakness is the lack of an appropriate criterion in comparing our results; therefore, our conclusion that the RFQ measures RF can be only provisional (Badoud et al., 2018). Besides, the results on the six-item version of RFQ need to be replicated within clinical samples, testing if this version is a valid method to assess mentalizing also in clinical adolescents. In this regard, future studies may explore the pattern highlighted in this study between mentalization, alexithymia, and psychological problems to illustrate the absence or presence or greater strength of these relationships in clinical populations. Moreover, this study does not include scales measuring externalizing constructs and other personality disorder traits (i.e., borderline or schizotypal features) that previous literature has shown to be linked to mentalizing difficulties. Lastly, RFQ's overlapping of the questions which make up the two scales might be an issue, but nonetheless, our findings suggest that RFQ subscales are associated with psychological problems and alexithymia dimensions in opposite directions, suggesting a negative covariance in accordance with prior studies (Badoud et al., 2015;Fonagy et al., 2016;Morandotti et al., 2018). These limitations help to point the way toward the next frontier in this line of research.
In sum, our results provide preliminary evidence that in the Italian context a short version of RFQ (six-item RFQ) holds solid psychometric properties, and it can be used to assess mentalizing in adolescence. Besides, this study provides basic information about the assessment RF tool in healthy adolescents supporting that the six-item RFQ version can evaluate mentalizing independently to age and sex within adolescence and that is able to individuate mentalizing difficulties which may contribute (or limit) the vulnerability to psychological problems and alexithymia through exploring of specific RFQ subscales. In other words, this instrument can be considered as a screening tool able to measure individuals' reflective abilities in adolescence, detecting possible difficulties or skills in healthy samples. In this way, this measure represents a very useful tool in terms of providing information on reflective abilities in adolescence and has crucial utility in the assessment process. Mentalizing abilities could be an important target for interventions aimed at improving emotion regulation strategies and psychological well-being in youth (Gambin et al., 2021). We suggest a systematic use of the RFQ (opportunely adapted to the different contexts) to increase our knowledge about mentalizing in adolescence.

CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.

ETHICS STATEMENT
All procedures were in accordance with the ethical standards of the institutional research committee, the international standards, and the 1964 Helsinki Declaration and its later amendments. Informed consent was obtained from all individual participants included in the study (and the parents of adolescents if below 18 years old).

PEER REVIEW
The peer review history for this article is available at https://publons.com/publon/10.1002/jclp.23218

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author.