1 Introduction

The study of teachers’ social-emotional competence has gained particular importance over the last few decades (Lozano-Peña et al., 2021; Yin et al., 2019), and has been approached as a key factor in the conceptual and experimental organisation of teaching (Pérez-Bonet & García-Domingo, 2024), with emphasis being placed on the importance of teachers’ well-being vis-à-vis forging and maintaining a positive atmosphere in the classroom and positive student–teacher relationships (Jennings & Greenberg, 2009). Models have been created to explain the control and regulation of the affective and social variables that enhance teacher-student relationships (Aldrup et al., 2024), to gauge how a better climate of co-existence may be achieved (Gimbert et al., 2023), to provide support and the ability to respond to individual needs that will enable teachers to understand and regulate both their students’ as well as their own emotions (Jennings & Greenberg, 2009), and as predictors of teaching quality (Aldrup et al., 2022; Llorent et al., 2020). These models have even been deemed to reduce the likelihood of stress and teacher burnout (Jennings, 2015) and have also been seen to boost students’ academic performance (O’Hare et al., 2020). Social-emotional skills are gradually shaped during initial teacher training, albeit not always to their full potential (Corcoran & O'Flaherty, 2022). Although aware of their importance, some teachers admit to not having sufficiently well-developed social-emotional skills (Mrnjaus & Vignjevic, 2021), even though they have received specific training in such skills (Martinsone & Damberga, 2017). This leads to differences in teachers’ social-emotional development that can go beyond mere individual variability. One example is teachers working in the early stages of education exhibiting higher levels of emotional competence (Yin et al., 2019), whereas teachers at higher levels of education evidence greater perceived support from supervisors in addition to displaying greater job satisfaction (Mérida-López et al., 2022). Likewise, female teachers are seen to exhibit greater emotional regulation than their male counterparts (Papoutsi et al., 2022).

1.1 Teachers’ social-emotional competence

Social-emotional competence refers to a person’s knowledge, skills, and motivation that are required to master social and emotional situations (Elias et al., 1997). It is a wide-ranging and multifactorial construct of skills, strategies and knowledge which–when applied by teachers within the context of education–determines how instructional processes are applied (Duckworth & Putnam, 2022). Despite the general consensus concerning the importance of this skill, there is less agreement with regard to its conceptualisation and structure (Lozano-Peña et al., 2021). A number of differing models have been put forward. Some of these consider social-emotional competence to be a component of emotional intelligence, such as Bar-On’s emotional intelligence model (Bar-On, 1997) or Mayer and Salovey’s emotional intelligence model (Mayer & Salovey, 1997), whereas others see it is a construct in itself. The latter include Gross’ model of the emotional regulation process (Gross, 1998), which focuses on emotional regulation. Finally, other proposals focus on acquiring and enhancing social-emotional competence in the classroom. Specifically, the social-emotional learning model (SEL, CASEL, 2020) stresses the importance of educators in students’ social and emotional development (Zins & Elias 2007), while the prosocial classroom model (Jennings & Greenberg, 2009), which underscores the teacher-student interactions that establish the emotional and organisational climate in the classroom, and which embraces social and emotional aspects that are deemed pivotal to forging a positive atmosphere in the classroom.

Social-emotional competence amongst teachers involves both intra as well as interpersonal aspects (Aspelin & Jonson, 2019) such as emotional regulation and relationship management, both of which are required in order to adequately cope with the emotional and social demands involved in the teaching profession. Indeed, teachers have a very specific way of expressing emotions (Sutton, 2004), and the appropriate regulation thereof impacts the emotional education provided to students (Jennings & Greenberg, 2009; Kang, 2022). As a result, the way in which teachers relate interpersonally with their students predicts to a great extent the emotions of the students themselves (Mainhard et al., 2018). The most socially and emotionally competent teachers are those who are best able to manage their classrooms, given that they perceive more acutely the emotional states of their students and skilfully apply emotional expressions and verbal support for their students rather than punishment (Jennings, 2015). This leads to beneficial effects such as improved student–teacher relationships, more positive and realistic expectations on the part of the teacher, and enhanced academic performance (Frenzel et al., 2021). In contrast, when teachers experience social-emotional difficulties, there may be negative consequences, both indirectly by generating a worse classroom atmosphere (Sutton & Wheatley, 2003), and directly on students themselves by leading to poor academic performance and worse problems of behaviour (Jennings & Greenberg, 2009), less academic satisfaction and low self-concept due to poorer student–teacher relationships (Frenzel et al., 2021), demotivation and higher drop-out rates (Sutton & Wheatley, 2003). The situation may even trigger increased student stress and anxiety (Kurdi & Archambault, 2018). Whatever the outcome, the positive or negative effects of teachers’ social-emotional competence depend on the context in which the interaction occurs (Frenzel et al., 2021).

As a result–and following Aldrup et al., (2020) in order to measure teachers’ social-emotional competence, it is important to focus on emotional regulation and on handling relationships, given that–although important–an awareness of both one’s own as well as others’ emotions is not by itself sufficient in terms of achieving the right social-emotional performance.

1.2 Teachers’ emotional regulation

Analysis of teachers’ emotional regulation has been closely linked to coping with stress (Aldrup et al., 2024), which relates to teacher well-being (Mérida-López et al., 2022). Emotional regulation refers to “the processes by which individuals influence which emotions they have, when they have them, and how they experience and express these emotions” (Gross, 1998, p. 275). As a result, this refers to teachers’ ability to control their own emotions (Waber et al., 2021), to distinguish between them and to use this information to shape their own thoughts and behaviours (Mayer & Salovey, 1997). This involves internal regulation strategies –which affect how the situation is evaluated– and external regulation strategies –which affect the situation– as well as strategies which impact the emotional response (Aldrup et al., 2024). Internal regulation strategies include cognitive re-evaluation processes (Dryman & Heimberg, 2018), modifying how the teacher interprets the situation in order to minimise the emotional impact (Gross, 1998). Other strategies involve diverting attention away from the situation through the use of mechanisms such as avoidance, and rumination (Taxer & Gross, 2018). With the exception of situations that are uncontrollable, these internal regulation strategies are not considered to be effective (apart from cognitive re-evaluation) since they enable specific situations to be dealt with, yet fail to provide a long-term solution (Aldrup et al., 2024).

External regulation strategies concern how to cope with the characteristics of the situation. For example, modifying the situation (Gross, 1998), which involves making changes to the environment or to the external circumstances in order to influence the emotions experienced, and which entails (a) identifying the situation, (b) assessing the viability of being able to change it, (c) devising strategies, (d) implementing the strategy, and (e) gauging its effectiveness. This strategy is considered to be effective in the long term, provided that the environment can be controlled, in contrast to others such as situation selection, or confrontation avoidance, since these seek to avoid having to face the situation (Aldrup et al., 2024), which is not always possible in an educational context.

Finally, there are strategies that focus on emotional response, such as the search for support or finding an outlet through colleagues (Sutton, 2007) or suppressing emotional expression (John & Gross, 2004) when, for instance a teacher is worried about a personal issue but does not wish to express it in order to maintain a good learning environment in the classroom (Sutton, 2004); or the selective expression of emotions (Sutton, 2007), in which the teacher regulates their emotional expression depending on the context and the intentionality (e.g. the start of the lesson when students’ interest needs to be aroused vs. a situation involving bad behaviour that needs to be dealt with). Nevertheless, these strategies are by no means easy to apply since, although the teacher tries to hide their real emotions, students may perceive very subtle signals that lead them to think that the emotion being expressed is not the real one (Sutton & Wheatley, 2003).

1.3 Teachers’ relationship management

Relationship management is “the ability to respond to other people’s needs while asserting one’s own goals” (Aldrup et al., 2020, p. 2), and includes interpersonal or social type variables, such as conversation initiation, negative assertion, emotional support, disclosure and, of course, conflict solving (Aspelin & Jonsson, 2019). This influence can occur through a number of different ways. Frenzel et al., (2021) find that the intrapersonal effects of teachers’ emotions on their students occur through instruction strategies (Duckworth & Putnam, 2022), given that emotions affect teaching performance, whereas interpersonal effects may occur directly (the emotion experienced by the teacher inducing an effect on the student, e.g. students taking an interest in a certain subject matter when they see their teacher motivated or interested in the topic); through the quality of the relationship (Mainhard et al., 2018), or by the type of non-verbal social messages conveyed by the teacher, and which will depend on teachers’ expectations (Poulou, 2017), and which impact students’ causal attributions (Waber et al., 2021). Frenzel et al., (2021) contend that students’ emotions also influence their teachers’ emotions. This is a mutually reinforcing process, since students’ resulting behaviour helps teachers in their efforts to establish a prosocial atmosphere in the classroom. In sum, teaching performance is closely linked to teacher-student relationships, particularly in terms of how teachers perceive their students’ emotions (Poulou, 2017).

1.4 Evaluating social-emotional competence

Efforts have been made to design instruments to measure teachers’ social-emotional competence, although many of these have been devised and validated for the population as a whole, principally for adults and teenagers (Cooper & Petrides, 2010; Gross & John, 2003), with very few having been specifically adapted to measure the social-emotional factors of teachers –except for the context of university education (Llorent et al., 2020). Many of these instruments have been designed following the previously described theoretical explanatory models of social-emotional competence, such as (Lozano-Peña et al., 2021): (a) the Emotional Quotient Inventory (EQ-i 2.0, Bar-On, 2011), based on Bar-On’s emotional intelligence model; (b) The Trait Meta-Mood Scale (TMMS, Salovey et al., 1995), and the Mayer Salovey Caruso Emotional Intelligence Test (MSCEIT, Mayer et al., 2002), following Salovey’s emotional intelligence model (1997); (c) the Emotional Regulation Questionnaire (ERQ, Gross & John, 2003), corresponding to Gross’ model of the emotional regulation process; or (d) the Social Emotional Competence Questionnaire (SEC-Q, Zhou & Fe, 2012), based on the Social, and Emotional Learning (CASEL, 2020). There is no specific questionnaire based on the prosocial classroom model (Jennings & Greenberg, 2009), and it is a model that is used more as a guide to design programmes to boost social-emotional competence.

It is also worth remembering the limited range of formats currently available for evaluation tools or the tests for measuring social-emotional competence, as well as the scant methodological corpus (Keefer, 2015). Self-report tools have traditionally played a leading role in measuring social-emotional skills. As a result, early tools for assessing emotional intelligence were self-reports (Sánchez-Camacho & Grané-Oro, 2022) wherein subjects expressed –on a Likert scale– the extent to which they agreed or disagreed with a proposition concerning their mastery of certain social-emotional skills and abilities. Self-reports are easy to apply but provide a subjective appraisal of social-emotional behaviour, which entails significant bias in the results obtained, such as situational variability bias, social desirability bias, or how questions are interpreted (Keefer, 2015). Examples of this kind of assessment include the Trait Meta-Mood Scale (TMMS, Salovey et al., 1995), the Emotional Quotient Inventory (EQ-i, Bar-On, 1997), and its updated version (EQ-i 2.0, Bar-On, 2011), the Schutte Self-Report Inventory (SSRI, Schutte et al., 1998), the Emotional Competence Inventory (ECI, Boyatzis & Burckle, 1999), the Emotional Regulation Questionnaire (ERQ, Gross & John, 2003), and the Trait Emotional Intelligence Questionnaire (TEIQue, Petrides & Furnham, 2003), and its abbreviated version (TEIQue–SF, Cooper & Petrides, 2010).

Tests gradually began to incorporate measures of skill or execution that involved people solving certain problems or conflictive situations of an emotional nature (Lozano-Peña et al., 2021), in which the challenge is to resolve problem situations by evaluating the appropriateness of the responses given. This has led to the creation of evaluation tests such as the Multifactor Emotional Intelligence Scale (MEIS, Mayer et al., 1999), as well as its abbreviated version, the Mayer Salovey Caruso Emotional Intelligence Test (MSCEIT, Mayer et al., 2002). These tests were developed at the same time as the so-called external or 360º tests, such as the Emotional Quotient-360 (EQ-360, Bar-On & Handley, 1997), which focused on measuring behaviour that enabled complexity to be quantified, since they were evaluated by external observers.

Finally, mention should also be made of so-called situational judgement tests (SJTs, Shaw, 2021) in which individuals must evaluate a given social situation, from which their reaction is then assessed (Webster et al., 2020), and which merge intrapersonal evaluation and interpersonal type elements. The format followed by these tests is to present situations similar to those faced by subjects in the context being evaluated (normally, work contexts) together with a series of possible actions/responses that subjects might adopt. Subjects must evaluate the effectiveness of each of these actions when faced with the situation. SJTs overcome many of the limitations inherent to self-reports, since they reach beyond the self-perception of a skill formulated generically, and force subjects to choose a specific solution to a situation which –albeit hypothetical– is one they are likely to have to deal with in reality. SJTs have been shown to be less prone to dishonesty than self-report measures (Kasten et al., 2018), and they also provide a more accurate evaluation by linking specific behaviours to given situations, added to which they display less bias than self-report scales for subgroups such as gender and ethnic groups (Weekley & Jones, 1999). Finally, they may be applied for training purposes by evaluating the consequences of each response option in the context to which the situations refer, although this use is less common (Webster et al., 2020). Nevertheless, designing them tends to be more laborious, particularly when it comes to determining the scoring criteria for the responses (Zhang & Wang, 2021), since scoring is not necessarily simply a matter of adding up the responses. This is because some responses may be scored depending on answers given to other questions, since some reactions may be more effective than others, or may not be totally ineffective, but neutral. Moreover, as with many other types of test, there are doubts as to the long-term predictive usefulness. Nevertheless, the results of SJTs do tend to be positively related to professional performance (Klassen et al., 2020). As a result, they do provide a suitable alternative that is able to overcome the limitations of self-reports, also bearing in mind the importance of offering measuring tools specifically designed for teachers: hence the design and adaptation to the Spanish context of the Test of Regulation in and Understanding of Social Situations in Teaching –TRUST (Aldrup et al., 2020)–which is based on duality in intra and interpersonal factors, conceptualised through emotional regulation and relationship management.

1.5 Test of regulation in and understanding of social situations in teaching (TRUST, Aldrup et al., 2020)

This evaluation test measures teachers’ knowledge of the strategies involved in two distinct components within the construct of social-emotional competence (Aldrup et al., 2020): emotional regulation (teachers’ ability to change their experiences and emotional expressions when faced with emotionally challenging teacher-student interactions), and relationship management (teachers’ ability to build positive relationships between teachers and students and to maintain them when faced with difficulties), specifically in situations in which there is interaction with students, with these skills being considered the principal facets that define teachers’ social-emotional competence (Mayer & Salovey, 1997). Each factor is made up of different scenarios that correspond to a situation experienced by a teacher and which involves conflict solving. When designing it, the authors identified key situations in teachers’ emotions and in the teacher-student relation. This yielded four thematic areas, in line with the teacher emotion model of Frenzel et al. (2016): student motivation, student social-emotional behaviour, student performance, and the above-mentioned teacher-student relation.

For each scenario, four items are presented that represent possible reactions when dealing with the conflict in question. Teachers must score each item in terms of its degree of effectiveness. After analysing its validation, the test was finally made up of 15 scenarios, with seven (28 items) belonging to the subtest of emotional regulation, and eight (32 items) to the subtest of relationship management.

1.6 The current study

Following Frenzel et al. (2021), one of the current challenges involves overcoming the excessive trust placed in the results of self-reports, with there being a need to delve deeper into alternative procedures to self-reports for evaluating social-emotional competence for those working in basic education (Kang, 2022), which are the levels at which emotional management and the development of social skills are key variables vis-à-vis achieving high levels of quality in student learning (Gill, 2021).

This work thus seeks to adapt into Spanish and to evaluate the reliability, factorial validity and convergent validity of the Test of Regulation in and Understanding of Social Situations in Teaching (TRUST) amongst teachers of primary education and secondary education in Spain. This test is designed specifically for teaching, and is based on the SJTs method, such that it provides an alternative to tools based on self-perception, and thereby overcomes the latter’s limitations. The hypotheses are: (a) the reliability and validity results will be similar to those obtained with the original instrument used in Germany, (b) the test will prove valid for both primary education and secondary education teachers, as well as for men and women alike; and (c) the test will correlate moderately with other measures of social and emotional competence obtained using self-reports, and which are aimed at the general public.

2 Method

2.1 Participants

In order to establish how many participants should be included in the study to obtain reliable and meaningful results, the sample size was first determined. To do this, a confidence level α = 0.05 was adopted, together with a margin of error of 5%, which are the levels usually taken in educational research (Cohen et al., 2018). Taking into account that the population size was 25,284 teachers (data from the Regional Ministry of Education at the Regional Government of Castilla y León), the required sample size was 379.

The study involved 503 teachers from schools who teach in primary or secondary education in Castilla y León, Spain (Table 1), which entails a final margin of error of 4.3% in the sample size. Teachers in primary education (n = 198) were aged between 23 and 68 (M = 43.2, SD = 9.94), and had between one and 45 years’ teaching experience (M = 16.7, SD = 10.32). Those in secondary education (n = 305) were aged between 24 and 64 (M = 46.3, SD = 8.92) and had between three months and 42 years’ teaching experience (M = 16.8, SD = 10.09).

Table 1 Sample distribution according to educational stage
Full size table

2.2 Measures

Test of Regulation in and Understanding of Social Situations in Teaching (TRUST, Aldrup et al., 2020). This is a tool used to measure social-emotional competence based on situational judgement. It has a two-factor oblique structure (emotional regulation and relationship management), and displays a good fit in the study carried out by the authors of the test, χ2(89) = 135.10, p < .001, CFI = 0.92, TLI = 0.91, RMSEA = 0.05, SRMR = 0.05.

It offers 15 scenarios (seven from emotional regulation and eight from relationship management). From each of the scenarios, four items are presented that represent possible reactions when dealing with the conflict in question. These reactions may be effective, ineffective, or ambiguous, and the latter may be either effective or ineffective. Teachers must score each item in terms of its degree of effectiveness on a five-point scale ranging from 1 (very ineffective) to 5 (very effective).

Test correction is based on organising reactions into three large groups (see an example in Fig. 1): very effective, very ineffective, and ambiguous. Very effective strategies are those which clearly prove to be beneficial (subjects are awarded one point if they correctly rate a very effective strategy as 5 = very effective, and half a point if they rate it as 4 = slightly effective), whilst very ineffective strategies are clearly damaging to emotional and optimal social functioning (subjects are awarded one point if they correctly rate a strategy as very ineffective; 1 = very ineffective, and half a point if they rate it as 2 = slightly ineffective). Ambiguous strategies involve strategies that are effective to a limited degree, or which are even ineffective but not damaging. In these strategies, intermediate responses (2 = slightly ineffective, 3 = neutral, and 4 = slightly effective) may be deemed partially correct, making a comparison between the evaluation given to the item representing an ambiguous response with that given to the other items in each scenario. For this, the authors established a more complex codification procedure in which points were awarded if respondents correctly distinguished ambiguous strategies from those which were very effective or very ineffective. Specifically, ambiguous responses, which tend towards being effective, are scored depending on: (a) how close their score is to that given by the subject to the effective reactions contained in each scenario, and (b) the distance from the score given to the ineffective reactions contained in each scenario. If the ambiguous response tends towards being ineffective, the same procedure is applied, but in reverse.

Fig. 1
figure 1

Example of a scenario of the sub-test of emotional regulation that includes the four possible reactions, and the assessment of their effectiveness (Aldrup et al., 2020, p. 5): Reaction (a). Corresponds to an effective response. If the respondent chooses the option very effective, they are given 1 point. If they choose the option slightly effective they are given ½ point. The remaining options score 0 points. The same applies to reaction (c); Reaction (d). This corresponds to an ineffective response. Respondents are given 1 point if they choose very ineffective, ½ point if they choose slightly ineffective. The rest of the options score 0 points; Reaction (b) is an ambiguous response, albeit tending towards effective. This kind of reaction is scored by making a comparison between peers, with the options chosen in the remaining reactions. In this case: Comparison between reaction (b) and (a). One point for scoring correctly (b), at least one unit worse than (a). The same occurs when comparing reactions (b) and (c). Comparison between (b) and (d). One point for scoring (b) at least two units worse than (d). The total score in this example would be six points. In fact, it is five points because the score from the comparison between (b) and (d) was excluded from the final version of the test due to the low correlation between the item and the total

Full size image

In order to test convergent validity, the following instruments were also applied. Although not specifically designed for teachers, they are frequently used for this purpose in Spain, and measure interpersonal as well as intrapersonal variables. Intrapersonal variables are expected to correlate with the dimension of emotional regulation, whereas interpersonal variables are expected to do so with relationship management.

Brief Form of the Interpersonal Competence Questionnaire (ICQ-15, Coroiu et al., 2015, adapted to Spanish by Salavera & Usán, 2018). This is the abbreviated form of the Interpersonal Competence Questionnaire (Buhrmester et al., 1988). The ICQ-15 is a self-report containing 15 items in which each person describes an everyday interpersonal situation (e.g. “finding and suggesting things to do with new people and which prove interesting and appealing"), asking subjects to rate their own self-perceived competence to manage each situation on a 5-point Likert scale, from 1 (I’m poor at this) to 5 (I'm very good at this), evaluating interpersonal competence in five dimensions: (a) the ability to initiate relationships (Initiation, e.g. “Finding and suggesting things to do with new people whom you find interesting and attractive”), (b) the ability to assert displeasure with others (Negative assertion, e.g. “Confronting your close companion when he or she has broken a promise”), (c) the ability to provide emotional support and advice (Emotional support, e.g. “Helping a close companion get to the heart of a problem s/he is experiencing”), (d) the ability to disclose personal information (Disclosure, e.g. “Confiding in a new friend/date and letting him or her see your softer‚ more sensitive side”), and (e) the ability to manage interpersonal conflict (Conflict management, e.g. “Being able to admit that you might be wrong when a disagreement with a close companion begins to build into a serious fight”). All the dimensions display acceptable internal consistency, with alpha coefficients in the range [0.77, 0.87]. Salavera and Usán (2018) adapted and validated this scale in the Spanish population and found that it demonstrated good factorial validity and an index of internal consistency of α = 0.85. The reliability coefficients obtained with the sample in this study are acceptable and are within the range [0.71, 0.84].

Trait Emotional Intelligence Questionnaire –Short Form (TEIQue-SF, Petrides, 2009, adapted to Spanish by Laborde et al., 2016). This is made up of a total of 30 items on a 7-point Likert scale, from 1 (completely Disagree) to 7 (completely Agree), evaluating emotional intelligence. It consists of four factors: (a) well-being (e.g., “I feel that I have a number of good qualities”), (b) self-control (e.g., “I usually find it difficult to regulate my emotions”), (c) emotionality (e.g., “Expressing my emotions with words is not a problem for me”), and (d) sociability (e.g., “I'm usually able to influence the way other people feel”). In addition, these provide a total emotional intelligence score. It displays high reliability, with alpha values in the range [0.70, 0.87]. This questionnaire was translated into and adapted to Spanish by Laborde et al., (2016), and exhibits acceptable factorial validity. With our sample study, we also obtained acceptable reliability coefficients in the range [0.74, 0.83].

Emotional Regulation Questionnaire (ERQ, Gross & John, 2003, adapted to Spanish by Cabello et al., 2013). This is a self-report made up of 10 items on a seven-point Likert scale, from 1 (totally disagree) to 7 (totally agree). It has a two-factor structure, comprising a cognitive part (reappraisal) that assesses the process of re-evaluation or reinterpretation of a situation (e.g. “I control my emotions by changing the way I think about the situation I’m in”), and the actual emotional response in terms of how emotions are handled or controlled through their non-expression (suppression, e.g. “I control my emotions by not expressing them”). It displays high reliability (α = 0.79 for reappraisal and α = 0.73 for suppression). The Spanish adaptation of the questionnaire designed by Cabello et al. (2013) was used, with this also showing suitable psychometric properties. The reliability obtained with this study sample was also high (α = 0.81 for reappraisal and α = 0.80 for suppression).

2.3 Procedure

This study was approved by the Committee on Ethics in Non-Clinical Research Involving Human Subjects and their Data or Samples (CEISH, University of Valladolid, Spain), delegate in CEIm Research Ethics Committee (Drug Research Ethical Committee, East Valladolid Health Area, University Clinical Hospital of Valladolid, SACYL, Spain), code PI-21-2286. First, we looked at whether the envisaged scenarios were applicable to primary and secondary education. In order to adapt the TRUST to Spanish, the International Test Commission for test adaptation guide was followed (Hernández et al., 2020): (a) the authors’ permission was sought for the adaptation and psychometric study of the test, and we were provided with the test in English as well as the correction procedure, (b) a group of primary education and secondary education teachers studied the situations referred to in the scenarios and evaluated the possibility of them being applied to all teachers working in compulsory education (primary education and secondary education); (c) three specialists in psychology and education were asked to make an independent translation, formulating the scenarios that could be applied to primary education and to secondary education (see the example in Fig. 1), after which coincidences and discrepancies were examined, taking into account the original wording, (d) two native speakers of English –specialists in translating psychology and scientific education texts– were asked to make a translation of the Spanish version in order to ensure that the sense of the text had been maintained vis-à-vis the original wording; (e) application of the Spanish version to 12 teachers of primary and secondary education who, in addition to completing the test, also evaluated its linguistic, psychological and cultural suitability. As a result of the responses and the group discussion, (e) the test was applied to a random sample of teachers by sending an e-mail to a random sample of schools in the region of Castilla y León (Spain), and which reached at least 1,264 teachers who might have responded. Of these, 503 teachers did finally complete the questionnaires.

2.4 Data analysis

We conducted confirmatory factor analysis (CFA) in order to pinpoint the internal structure of the TRUST, and robust maximum likelihood (MLR) was applied. We also calculated Bentler’s simplicity index. Various multivariate goodness of fit indices were estimated in this analysis, such as estimating the maximum robust likelihood, CFI, TLI, RMSEA, and Standardized Root Mean Square Residual (SRMR). We also calculated internal consistency with McDonald’s Omega coefficient, since this overcomes some of the limitations of the Alpha coefficient, such as the number of items or the number of response alternatives. Nevertheless, the Alpha coefficient was also calculated in order to compare it to those obtained by the authors of the test when applied in Germany. The model was also compared with a single-factor model in order to determine which offered the best fit. For this, we used the AIC and BIC Criterion (in which lower values indicate a better model fit). To determine whether the model is also valid for men and women, we studied configural, metric, scalar, and factor mean invariance through multigroup analysis, with the chi square difference test. We adopted the criterion of Cheung and Rensvold (2002), calculated as the difference between the CFI values, and considering that invariance can be accepted if this difference is less than or equal to 0.01 in favour of the less restrictive model. We also analysed factor invariance between the sample of teachers from primary education and secondary education. We used MPLUS software, v.8.8.

In order to estimate convergent validity, we calculated the Pearson correlation coefficients between the factors of TRUST and the factors of ICQ-15, TEIQue-SF and ERQ. The cut-off points adopted were: (a) very weak correlation (r <  = .19), (b) weak correlation (r = .20 to .39), (c) moderate correlation (r = .40 to .59), (d) strong correlation (r = .60 to .79), and (e) very strong correlation (r > .79). For this purpose, we used the IBM SPSS Statistics statistical package, v.29 (2022).

3 Results

3.1 Confirmatory factorial analysis

This analysis aims to test the fit of the envisaged scenarios to the theoretical factor put forward in the original design of the test. The assumption of multivariate normality was violated (Mardia’s coefficient = 6.05), such that robust tests were applied. Moreover, skewness or kurtosis values were within normal parameters, since none of the items evidenced values of over two or seven, respectively (Table 2). Factor analysis proves pertinent since the assumptions of the Kaiser–Meyer–Olkin index (0.90) and Bartlett sphericity test, χ2(105) = 1783.5, p =  < .001 are met. Bentler’s simplicity index is high, S = 0.98. Significant correlation was found between the two subtests of the test (r = .58), such that it is expected to fit an oblique factor model.

Table 2 Means, Standard deviations, Skewness and Kurtosis between teachers from primary education versus secondary education in emotional regulation and relationship management
Full size table

The indices exhibit an acceptable fit, χ2(89) = 198.49, p < .001, CFI = 0.932, TLI = 0.919, RMSEA = 0.049, 90% CI [0.040, 0.059], SRMR = 0.042 (Fig. 2). However, when analysing the goodness of fit of a single-factor model, the indicators are not acceptable, since the TLI value is below 0.90, and the CFI, AIC and BIC values are lower than those of the two-factor model (Table 3).

Fig. 2
figure 2

Factorial Structure of the Spanish Version of TRUST

Full size image
Table 3 Fit indices of the possible models
Full size table

Composite reliability is high in both factors (CRER = 0.83, CRRM = 0.80), and the reliability indices measured with the alpha and omega coefficients are adequate –both when considering the scenarios (with the global scores obtained in each of the 15 scenarios) and the items (with the scores obtained in each of the reactions in each scenario)– and are also higher than those obtained in the study carried out in Germany (Table 4).

Table 4 Reliability coefficients of German version and Spanish version
Full size table

Nevertheless, improvement analyses indicate three scenarios that present cross-loading problems; specifically, two items in the emotional regulation subtest (E1 and E2) and another item in the relationship management subtest (R1). In any case, the factor loading is much higher in the scenario to which they in theory belong.

Table 5 Model summary for multi-group test of measurement invariance
Full size table

3.2 Invariance analysis

In order to test whether the resulting model is equally valid for both men and women as well as between primary education and secondary education teachers, we subsequently analysed factorial invariance (Table 5), conducting multigroup analysis. As regards gender, the fit indices of the configural model were also acceptable. When restricting the factor loadings of the items of this model (metric invariance), we obtained acceptable data. The difference between the CFI values of the models was acceptable (ΔCFI = 0.004), and the chi-square test between the configural and metric model proved to be non-significant, χ2(10) = 16.26, p = .092, showing that metric invariance was fulfilled. The following nested model adds to the former models the restriction of the intercepts in order to determine possible scalar invariance. We also obtain acceptable data. The difference between the CFI values of the models was acceptable (ΔCFI = 0.006), and the chi-square test between the configural and scalar model was non-significant, χ2(20) = 27.96, p = .110. Moreover, scalar invariance holds, such that the model is valid for comparing social-emotional competence between females and males.

The same procedure was then applied, although in this case comparing teachers who teach primary education to those who teach secondary education. The configural model presents acceptable goodness of fit indices. Metric invariance is met, given that the difference between the CFI values of the models does not exceed the criterion value (ΔCFI = 0.002), and the chi-square test between the configural and metric model was non-significant, χ2(10) = 13.40, p = .202. The same occurs with scalar invariance: (ΔCFI = 0.009), with the chi-square test proving to be non-significant, χ2(20) = 32.11, p = .052.

3.3 Convergent validity

In order to estimate convergent validity, the factors of the Test of Regulation in and Understanding of Social Situations in Teaching were correlated with those from the Interpersonal Competence Questionnaire (ICQ-15), Trait Emotional Intelligence Questionnaire –Short Form (TEIQue-SF), and Emotion Regulation Questionnaire (ERQ).

In general, significant correlations were found with many of the dimensions of the questionnaires applied, although these correlations are only weak (Table 6). These correlations are, in general, higher with relationship management than with emotional regulation, when it is to be expected that intrapersonal variables correlate more with emotional regulation than with relationship management. Specifically, significant correlation indices were obtained for the two factors of TRUST with all the dimensions of interpersonal competence (ICQ-15). Particularly worthy of note is the correlation of the emotional support dimension with relationship management (r = .39), and with emotional regulation (r = .38), indicating the importance of the two variables when we wish to help others emotionally, since this entails possessing social skills, such as empathy, but also controlling and managing our own emotions. Something similar occurred with disclosure, which displayed a similar correlation to emotional regulation (r = .22), and with relationship management (r = .21). In contrast, initiating relationships correlates more with relationship management (r = .22) than with emotional regulation (r = .14). It should be remembered that approaching another person and initiating social contact involves the effective use of social skills. The correlations between the two dimensions of TRUST with negative assertion and conflict management are very weak, although higher than emotional regulation (r = .12 and r = .17, respectively) than with relationship management (r = .10 and r = .13, respectively).

Table 6 Pearson correlations between the factors of the Spanish version of TRUST and ICQ-15, TEIQue-SF, and ERQ
Full size table

As in the previous model, the same occurs with emotional intelligence. The total score for the TEIQue-SF questionnaire correlates significantly with the two factors of the TRUST. Surprisingly, and in contrast to what might be conjectured, the correlation is higher with the relationship management factor (r = .27) than with the emotional regulation factor (r = .23). In any case, the correlation is very weak in both instances. The same occurs with the factors of the TEIQue-SF. Specifically, the emotionality factor displays the highest correlations, both with emotional regulation (r = .24) and with relationship management (r = .29).

Finally, the correlation between the Emotional Regulation Questionnaire (ERQ) and the TRUST was examined. Again, in contrast to what might be hypothesised, lower significant negative correlations are obtained between emotional suppression with emotional regulation (r = –.18) than are obtained with relationship management (r = –.29). However, the reappraisal factor does not correlate with either of the two components of the TRUST.

4 Discussion

Recognising, expressing and understanding one’s own emotions as well as those of others is essential for teachers (Jennings & Greenberg, 2009; Mrnjaus & Vignjevic, 2021). As a result, measuring social-emotional competence has gained particular importance, both in the field of research and in professional practice (Lozano-Peña et al., 2021), and is linked to evaluating and fostering teaching skills (Corcoran & Flaherty, 2022; Pérez-Bonet & García-Domingo, 2024). Yet there are very few valid and reliable instruments which can assess this skill amongst teachers and which are not based on self-reports. This study therefore sought to adapt and validate the TRUST to Spanish –both for primary and secondary school teachers– measuring the components of social-emotional competence from the perspective of situational judgement rather than the usual self-reports (Kang, 2022). The test used is a tool that contains a range of different items which are representative of scenarios similar to those teachers may encounter in their everyday professional practice and which are grouped into two subtests: emotional regulation and relationship management.

The goodness of fit data obtained in this test are acceptable, and are similar to those acquired during research in the context of Germany (Aldrup et al., 2020), and indeed are even slightly better. Likewise, a significant correlation was found between the two subtests of the test, which reflects the two-fold nature of social-emotional competence, encompassing processes inherent to emotional regulation and handling of social relations (Petrides & Furnham, 2003), with the two constructs being related, but independent (Rodríguez-Pérez et al., 2021). High reliability coefficients were also obtained, and the test also confirms the two-factor structure of teachers’ emotional behaviour, which is similar to that found in other groups and ages (Rodríguez-Pérez et al., 2021).

Nevertheless, a more detailed analysis indicates that three items also exhibit this load, albeit to a lesser degree, in the other subtest; specifically, in the subtest of emotional regulation, the scenario concerning teacher disappointment due to lack of cooperation, and student behaviour. Respondents probably attached greater value to solutions geared towards enhancing student behaviour and motivation rather than to those concerned with improving their own feelings when faced with the adverse situation (Aldrup et al., 2022). Indeed, these two scenarios have response options which, although emotional, involve seeking help from others. Much the same might be true of the scenario in which teachers feel irritated because of poor student behaviour while they are explaining. In contrast, in the relationship management subtest –the scenario in which the teacher receives negative criticism of their lessons, particularly as regards the unfair treatment some students believe they receive from teachers– certain respondents consider that this refers more to emotional regulation than to how personal relations are handled (Mainhard et al., 2018).

The test presents suitable convergent validity, with significant although not very high correlations being obtained with other instruments that measure variables related to social-emotional competence, but not specifically associated to teaching, and that are based on self-reporting. It should be remembered that correlations between the SJTs and the self-report tests tend to be moderate or low (Rodríguez-Pérez et al., 2021) because the answer format is aimed at different facets. Self-reports often ask about frequency or success, while the SJT measures the ability to assess the long-term effectiveness of different behavioural alternatives. Nevertheless, even though intrapersonal variables are expected to correlate more with emotional regulation, and interpersonal variables more with relationship management, overall we find higher correlations in all of the variables with relationship management when compared to emotional regulation. Worthy of particular note was the correlation of the emotional support dimension of interpersonal competence with the two components of the TRUST. It should be remembered that the items of the ICQ-15 questionnaire that correspond to this dimension include both emotional and interpersonal aspects, which are very closely related to prosocial behaviour (Coroiu et al., 2015). Something similar occurs with the disclosure dimension; revealing our feelings to others involves regulating one’s own emotions, yet mediated through our trust in others (Waber et al., 2021).

In addition, both the overall score for emotional intelligence and its dimensions correlate similarly with the two components of the TRUST. Worthy of note, however, is the fact that higher correlations are obtained with the relationship management factor than with the emotional regulation factor, although the differences between the correlation coefficients are small. Specifically, both emotional regulation and relationship management are particularly linked to emotionality (Cooper & Petrides, 2010), offering others emotional support (Mainhard et al., 2018), the ability to initiate social relations with others, emotional expression (Boyatzis et al., 2000), or trusting in and opening up emotionally to others (Schelhorn et al., 2023). However, they are negatively linked to emotion suppression. This result might be because the suppression dimension refers to lack of expressiveness and to concealing emotions from others (Waber et al., 2021), such that it has a high social load (Gross & John, 2003). However, the reappraisal dimension does not correlate with the TRUST. In addition to the questions not being specifically targeted at teaching, this might also be explained by the fact that the reappraisal dimension focuses on one single strategy –cognitive restructuring (Dryman & Heimberg, 2018)– whereas the TRUST envisages other strategies, some of which may not be fully optimal but which are, nevertheless, acceptable. Overall, the model evaluated presents factorial invariance when comparing data from primary education teachers to data for secondary education teachers, and also between male and female teachers (Papoutsi et al., 2022).

In sum, the TRUST is a good alternative to measurement questionnaires for social-emotional competence that are based on self-reports, since the limitations linked to the latter are overcome, such as: (a) giving non-valid responses, whether stemming from a lack of sincerity or due to social desirability (Kasten et al., 2018), in the sense that respondents may feel inclined to answer in a more socially acceptable manner; (b) less precise responses as a result of items being posited generically (Weekley & Jones, 1999), (c) less motivating, since they fail to place the subject in a situation which, although hypothetical, is closer to reality (Shaw, 2021), and (d) having less long-term predictive power in terms of professional performance (Klassen et al., 2020).

5 Limitations

The first limitation is that teachers come exclusively from the region of Castilla y León. Although it is the largest region in Spain, the research fails to take account of other Spanish regions. Moreover, the sample also contains more secondary than primary education teachers. Secondly, when adapting the instrument, the situations posited are closer to the realities found in secondary education, which means that teachers working in the early years of primary education would need to make a greater effort when responding to the scenarios put forward. In sum, although the scenarios are common, the formulation could be adapted to more specific primary education situations. Finally, the test evidences certain limitations inherent in SJTs, such as the complexity of the scoring system (Zhang et al., 2021), the transcultural validity of the test (Klassen et al., 2020), and the “testing effect”, when the test is known beforehand. Moreover, it should be remembered that there may be situations envisaged in the scenarios that could prove to be more determinant, more influential and/or more common than others usually encountered in a classroom environment. Efforts should be made to gain further insights into this by attaching greater weight to those situations which are more important to teachers. Likewise, the test considers possible responses given by teachers to specific situations, but does not consider others that might be the result of group consensus, such as rules established by schools or others agreed on by the teaching staff, and which may even be determined by the nature and quality of teachers’ social relations with their colleagues. Finally, it should be borne in mind that the test requires a great deal of effort on the part of teachers when compared to self-reports, such that teacher motivation and interest must be taken into account before it is applied, particularly if it is not put to teachers as a voluntary activity.

6 Conclusion

The present study contributes to the adaptation and validation to Spanish of a tool used to evaluate teachers’ social-emotional competence beyond what is provided by self-reports. This tool may prove useful both for training preservice teachers –given that initial training does not seem to improve the social-emotional skills linked to educational practice (Corcoran & O'Flaherty, 2022)–and for teachers already working –given that teachers’ emotional skills need to be enhanced by promoting their social competence (Waber et al., 2021). It should be remembered that it is difficult for teachers to enhance their students’ social-emotional competence unless the teachers themselves are properly competent therein (Mainhard et al., 2018). This test may serve as an initial evaluation in teachers’ ongoing training in social-emotional variables in terms of encouraging them to subsequently reflect on reactions to conflictive situations that are common in the classroom. It may even serve as a guide to create activities based on case studies to be included in intervention programmes geared towards improving education through the management of emotions, interpersonal relations, and coping strategies, focused on responses to situations which, although specific, are commonplace. It also poses scenarios that are common to schools in different countries and cultures (Aldrup et al., 2020) as has been seen in the results when applied to the German and Spanish contexts. It may also help in teacher recruitment processes, which is precisely one of the most common uses of situational judgement tests. It may also aid other research focused on deepening current understanding of the link between social-emotional competence in teaching and other educational psychology variables, and even with regard to convergent validity when developing other measurement instruments. As a result, it is an instrument which may prove useful in Spanish-speaking countries and communities, such as Latin-America, or certain communities in the USA.