Emotional intelligence, spelling performance and intelligence quotient differences based on the executive function profile of schoolchildren

The aim of this study was to determine the correlation of intelligence quotient with spelling performance, executive functions and emotional intelligence and to analyse the profiles in executive functions presented by the participants and the differences at the level of intelligence quotient, spelling performance and emotional intelligence. The sample consisted of 125 Spanish primary school students (58 girls and 67 boys) in a rural environment, with a mean age of 10.92 (±.68) years. Intelligence quotient and spelling performance were measured using the BADyG test; executive functions were assessed using the NIH EXAMINER battery and emotional intelligence was assessed using the Bar‐On questionnaire. The results show a high correlation of intelligence quotient with spelling performance, executive functions and two emotional intelligence dimensions (adaptability and interpersonal). Two clusters were observed: a high executive functions profile (n = 74, 59.7%), with high scores on all measured variables (working memory, planning, cognitive flexibility, inhibition and verbal fluency), and a low executive functions profile (n = 50, 40.3%) with low scores on all executive functions‐related variables. The multivariate analysis of variance (MANCOVA) found variations between the different clusters in the intelligence quotient, spelling performance and interpersonal emotional intelligence variables, showing that students with a higher executive functions profile had a higher intelligence quotient, spelling performance and interpersonal emotional intelligence values. This study could help to better define what kind of teaching methodology favours the development of academic performance in schoolchildren and the connections between important educational aspects such as emotional intelligence and executive functions.


| INTRODUCTION
The role of emotional intelligence (EI) in the educational context is one of the lines of investigation that has generated most interest in the past few years (Caballero & Cifuentes, 2017), where recent meta-analysis studies have assessed the role of emotions and their relationship with student academic performance (AP) (MacCann et al., 2020).In this sense, EI can be interpreted from two different points of view, trait and ability.Thus, EI understood as a trait is based on self-report measures and refers to our emotional perceptions: how good we believe we are in terms of perceiving, understanding, managing and utilising our own and other people's emotions (Petrides et al., 2018).This study will focus on trait EI, because it is considered to influence the management of emotional challenges in school situations and therefore affects AP.Eastabrook et al. (2005) were among the first to find a positive relationship between EI and AP in the primary education stage, the students with the highest grades in school being those who obtained the highest scores on global EI in two of its dimensions (interpersonal and adaptability).Another study, specifically focused on children between the ages of 6 and 13, reported that emotional regulation was associated with their AP (Nesayan et al., 2019).Pulido-Acosta and Herrera-Clavero (2019), after analysing children from different cultures, drew the conclusion that there was a strong and direct relationship between EI and AP, each of these factors acting as the main predictor of the other.Therefore, children with emotional problems are more likely to reach situations of failure at school compared with those who enjoy emotional stability (Mundy et al., 2017).Furthermore, following MacCann et al. (2020), emotional content is more relevant in subjects related to the humanities (where the understanding of motivations and emotions can be part of the evaluable content), than science or mathematics.
Focusing attention on subjects related to the humanities, an important factor in learning a language is the ability to show EI, that is, having the ability to recognise, use, comprehend and manage emotions (Perpiñ a et al., 2020).For language acquisition, children need socio-cognitive skills that allow them to understand the communication intentions of others in a wide variety of interactive situations (Tomasello, 2000), where a high correlation between grammar skills and emotional understanding of the children's language was found (Pons et al., 2003).
In this regard, numerous studies show that spelling proficiency is an integral part of students' AP and school success (Schuchardt et al., 2006;Smith, 2015).Associated with AP is the concept of spelling performance (SP), the competence of a person to apply, in an appropriate manner in writing, the orthographic rules governing a language.In accordance with the above, positive correlations between SP and executive functions (EFs) are evidenced by Manso-Luengo (2001), which shows the role played by working memory (WM) in the maintenance and cognitive processing of visual orthographic information.Thus, children with low levels of SP also show deficits in WM (Schuchardt et al., 2006).Furthermore, training in EFs such as WM and inhibition leads to improved SP in students with learning disability (Malekpour & Aghababaei, 2013).
EFs are composed of different behavioural and cognitive elements related to self-control which play a significant role in learning and AP (Perpiñ a et al., 2023).Stelzer and Cervigni (2011), in a systematic review on children and adolescents, conclude that there is an association between AP in children and cognitive processes.Thus, a poor performance of EFs is related to deficits in most school areas including literature (Nouwens et al., 2021) and arithmetic (Gashaj et al., 2019).In relation to language, research has shown a positive association between inhibition, WM and reading in young people (Chung & McBride-Chang, 2011;Foy & Mann, 2013).Likewise, a systematic review has shown that if one of the EFs is worked on in the different curricular subjects, such as prosocial decision-making (which in turn is related to interpersonal intelligence, a principle of EI that explains the socioemotional functioning of young people), then more altruistic behaviour in situations of high acute cognitive activation is encouraged (Duque et al., 2022).
The intelligence quotient (IQ) has been one of the areas most studied, and these different studies concur that the correlation between general intelligence level and academic skills ranges from moderate to strong (Benson et al., 2016;Deary et al., 2007).In primary school students, authors have observed a moderate relationship with IQ (Almeida et al., 2008;Bl azquez Garcés et al., 2015;Ferrando et al., 2011;Luo et al., 2003).Indeed, Laidra et al. (2007) indicate that IQ is the factor that best explains the variability of academic skills in primary and secondary school students.While McCoach et al. (2017) emphasise that IQ is a good predictor of reading and mathematics skills at early ages.However, more recent studies such as those by Martin-Requejo and Santiago-Ramajo (2021), where EI and EFs are included in addition to IQ, state that it is the EFs followed by the EI and finally the IQ that best predict academic ability, although in the case of literature, only the EI has a significant contribution compared with the other variables (IQ and EI).
Although profuse investigation has allowed us to detail the correlates of EFs with other variables such as EI, IQ or SP, as can be seen above, little is known about the EF profiles and their relationships in the primary school context.EFs prove to be a better predictor of AP than IQ levels (Ar an-Filippetti & Richaud, 2017) and explain the variability in school performance and the social behaviour of schoolchildren (Zorza et al., 2016).Their stimulation plays a relevant role in the selfregulation of cognition and emotional behaviour predicting AP (Blair & Razza, 2007;Razza et al., 2015), showing significant differences when students receive an intervention programme (Moreno-G omez et al., 2020;Razza et al., 2015), thus stimulating interest in finding out more about primary student psychosocial characteristics considering different EF profiles.
In view of the previously mentioned, the novelty of the present study lies in the need to expand the field of research by relating variables IQ, EFs, EI and SP in the same study.Hence, the main objectives were to study the correlation of the different variables presented in this study (IQ, SP, EFs and EI) and to analyse the profiles in EFs presented by primary school students and the differences that could exist between the profiles at the level of IQ, SP and EI.It was hypothesised that the four study variables would show significant and positive relationships among them, and young students with a high profile in EFs would obtain higher values in IQ, EI and SP.

| Study design
The design proposed for this study is descriptive, correlational and cross-sectional in which participants were measured once (Montero & Le on, 2007).

| Participants
The final study sample consisted of 124 students in the fifth and sixth years of primary education (57 girls and 67 boys), belonging to an early childhood and primary education school in the autonomous community of Castilla-La Mancha (Spain).The school belonged to a rural environment, with medium-low family economic conditions.The children investigated did not have bilingual competence and their level of language proficiency corresponded to their age group.At the beginning of the study, the age of the students was between 10 and 13 years (M = 10.92;SD = .68)(see Table 1).The school was selected for accessibility and convenience (Azorín & S anchez-Crespo, 1994).
The exclusion criteria were to exclude students with severe behavioural disorders such as ADHD, autism, severe learning disabilities or students who were late entrants to the Spanish educational system with a profound lack of knowledge of the Spanish language, (one participant was excluded).In addition, statistical tests for outlier detection (Mahalanobis distance) were applied, excluding those students with a distance p < .001.One participant was excluded.

| BADyG battery
The BADyG battery (Battery of Differential and General Aptitudes) (Yuste, 1998)  In this present study, the BADyG-E3 Renovado test (Yuste et al., 2000) was used, adapted to the cognitive characteristics of the sample under study (10-13 years old) to measure general intelligence through IQ and SP.A computerised template was used to obtain the scores.
• IQ.The IQ was analysed by means of the tests called 'basic' by the creators of the BADyG battery.This test consists of the application of the following subtests: Analogical Relations (Rv) (finding analogical relationships between concepts; e.g., in Spanish: pez es a _________ como p ajaro es a _________), Numerical Series (Rn) (completer numerical series; e.g., 19 16 13 10 !?), Logical Matrices (Re) (look for the drawing that should go on a figure from which a portion has been removed; e.g., see Figure 1), Sentence Completion (Vv) (find the concept or word that completes the meaning of a sentence; e.g., in Spanish: radio, television y !? son los medios m as importantes de comunicaci on), Numerical Problems (Nn) (comparing quantities resulting from solving numerical problems to see which is larger; e.g., see Figure 1) and Figure Matching (Ge) (find the figure that perfectly completes the part that has been cut out from a surface; e.g., see Figure 1).In its entirety, the test is made up of six basic mental ability tests, with a total duration for its execution of 75 min for children in fifth and sixth grades of primary school.Each subtest is formed by 32 items, and its reliability is Rv α = .87;Rn α = .87;Re α = .92;Vv α = .85;Nn α = .84;Ge α = .74. • SP.The SP was analysed by means of the Visual Spelling Memory (Mv) test.This test consist of selecting the incorrect spelling option from among three options (e.g., in Spanish: baliente hallar arrojar), the remaining two being correct.In its entirety, the test is made up of 32 items, with a total duration for its execution of 5 min for children in fifth and sixth grades of primary school.The reliability of the Mv is α = .98.

| NIH EXAMINER
For the assessment of EFs, version A of the NIH EXAM-INER battery (Kramer et al., 2014) was selected for Spanish children.Specifically, the following tests were applied: the dot counting test for WM.The task is to count and memorise a number of blue circles, which are interfered with by distracting stimuli represented by green circles and blue squares.After a series of stimuli, the person has to add up the memorised quantities.The test is of increasing complexity (see Figure 1); the Flanker task for inhibition consists of pressing the right or left arrow on the computer keyboard according to the image represented on the screen, made up of five arrows in different directions, with the person's answer having to coincide with the one that appears in the centre of the screen (see Figure 1); set shifting for cognitive flexibility (CF) participants have to match the stimulus that appears in the centre of the screen with two other stimuli that appear in the lower corners of the screen.To do so, they have to sort the stimulus presented by the category 'colour' or by the category 'shape'.The colours are red and/or blue, while the shapes are triangles and/or rectangles (see Figure 1); the verbal fluency test consists of two subtests, each lasting 60 s.First, the examinees are asked to name aloud as many words as they can remember from the category 'animals'; and the Unstructured Task test for cognitive planning (CP) participants have a total of 6 min to accumulate as many points as possible.To do this, they need to plan which activities they are going to carry out by calculating their cost-benefit, that is, they have to estimate the time it might take them to do them with the score assigned to each of the 16 tasks of which the test is composed (see Figure 1).

| Procedure
Before starting the research, several meetings were held with the management team and the tutor teachers.They were informed of the purpose, procedure, tests, times and possible benefits of this research.After these meetings, the head of the school informed the Cuenca Provincial Delegation of Education for authorisation.Once this approval was received, the Ethics Committee of the University of Murcia was informed, which authorised the start of all relevant processes (ID: 2036/2018).All students were given a consent form to be signed by their parents and returned to the research team.Finally, a meeting was held with the children's parents before the start of the study to explain its purpose.Data collection took place during the months of February, March and April 2019, during the second school term.All tests were conducted as a group.The BADyG battery and the Bar-On EQ-i: YV questionnaire were administered by means of pen and paper in the normal classrooms, while the EF test (NIH EXAMINER battery) was administered in the IT classroom, which consisted of 26 Lenovo laptops.In this test, each student used an individual computer.For all tests, the start time was 9.00 AM, and the temperature and lighting conditions were identical for each group of participants.These conditions were intended to ensure that attention levels were not influenced by the fatigue of the working day.The BADyG battery and the Bar-On EQ-i: YV questionnaire were administered by the principal investigator, while the NIH EXAMINER battery was administered with the help of the students' tutor teachers.

| Data analysis
First, descriptive statistics and bivariate correlations were calculated for the study variables.The Mahalanobis statistical test was used to detect outliers.Next, we tried to identify different profiles of EFs in the study sample based on the variables measured (WM, inhibition, CF, verbal fluency and CP).For this, a hierarchical cluster analysis with the Ward method was used, using the variables of the EFs.Subsequently, a multivariate analysis of variance (MANCOVA) was performed to identify the differences between the EF profiles according to IQ, EI and SP, using the clusters as independent variables and age factor as a covariate due to its possible effect.This analysis was based on testing the statistical assumptions of univariate normality and multivariate normality and the assumption of homoscedasticity (Hair, 2020).The effect size was calculated to quantify the magnitude of the differences between the two-cluster means through the effect size, which could be considered small (η 2 = .01),medium (η 2 = .06)or large (η 2 = .13)(Cohen, 1977).Finally, statistical power was calculated using post hoc analysis difference between two independent means (two groups) and assuming the minimum power level of .80 following Ellis (2010).For the quantitative analysis, SPSS 25.0 (New York, USA) and G*Power v. 3.1 were used to calculate statistical power.

| Descriptive and bivariate correlation analysis
Table 2 shows the means, standard deviation, skewness, kurtosis and correlation matrix of the study variables.The skewness and kurtosis values were all <3 and <10, respectively, which is within the limits established as normal (Field, 2017).Furthermore, the correlations between most of the variables were significant and with values <.80 which supports the absence of multicollinearity between them (Hair et al., 2019).Specifically, IQ correlates positively with SP, all the EFs and two of the five variables linked to EI (adaptability and interpersonal EI).SP, on the other hand, correlates positively with all dimensions of the EFs and EI, except for intrapersonal EI.The EFs, for their part, were related to some dimensions of EI (CF and verbal fluency with interpersonal IE and inhibition of stress).

| Cluster analysis
After eliminating outliers (Z > ±3 and Mahalanobis distance p < .001), a first hierarchical cluster analysis was performed with the Ward method with half of the randomly selected sample, using the Z scores of the variables WM, planning, CF, inhibition, verbal fluency and age.The dendrogram and clustering coefficients suggested the existence of two clusters.An attempt was then made to confirm the profile solution found, using a K-means cluster analysis with the rest of the randomly selected sample.When the solution was found to be consistent, a K-means cluster analysis with two groups was performed, concluding the existence of two distinct profiles among the students: a high EF profile (n = 74, 59.7%), with high scores on all dimensions, and a low EF profile (n = 50, 40.3%), with low scores on all variables linked to EFs (Table 3).Furthermore, Table 3 shows the differences between variables that were configured with the twocluster solution, with a multivariate effect (Box's value = 50.130,F = 3.187, p < .001),violating the homogeneity of covariances and suggesting the use of the Pillai's trace as a statistical test (Field, 2017) showing a value of .318(F = 6.654).
Figure 2 shows the existence of two profiles in EFs with their respective Z values for each of the variables analysed (IQ, SP and total EI and its dimensions), showing how the profile with high EFs has higher values in IQ, SP, mood and interpersonal EI; similar values in adaptability and total EI compared with the profile with low EFs; and lower values in stress and intrapersonal intelligence.
To examine the characteristics of each EF profile according to their IQ, SP and EI, a MANCOVA was conducted, using the clusters as independent variables and IQ, SP and EI subfactors (mood, adaptability, stress, interpersonal intelligence, intrapersonal intelligence and total EI) as dependent variables and age as a covariate due to the significant effect it might have (Table 4).Tests showed a value for Box's test = 69.356(F = 1.785, p = .003)and for Pillai's trace = .318(F = 6.654, p < .001).
Differences (Wilk's Λ = .682,F(8, 114) = 6.654, p < .001)were found between the different clusters in the variables IQ ( p < .001),SP (p < .001)with a large effect size (η 2 > .13)and the expected power (β > .80),and in interpersonal IE ( p < .05)with a small effect size (η 2 = .03)and a power below the expected (β = .52).It shows that students with a high EF profile have higher values of IQ, SP and interpersonal EI.For the rest of the variables related to EI, no significant differences were found, the effect sizes were small or non-existent and the calculated power was below the expected (β < .80).

| DISCUSSION
In relation to the first of the objectives, to discover the relationships between the study variables, the results of this research show a high correlation between IQ with SP and all the EFs and part of the EI dimensions (adaptability and interpersonal).Very similar results were obtained by Eastabrook et al. (2005) who split a cohort of primary school children into three categories: above average, average and below average, based on the grade average scores (GPA).The above average group scored higher relative to the other two groups on an overall EI scale and its adaptability and interpersonal dimensions.Following Pulido-Acosta and Herrera-Clavero (2019), and their study with a sample of 764 primary students with a 9.41 years average from different cultures, those with a higher EI score presented better AP, identifying that age, status and gender were predictors of the EI.
Focusing on the correlation between EI and the performance in humanities, Jiménez et al. (2019) pointed to a direct relationship between reading competence and EI in a reading intervention programme in high school students lasting two academic years.Another study with F I G U R E 2 Intelligence quotient, spelling performance and emotional intelligence as a function of the executive functions profile.
80 students from 4 to 11 years (Pons et al., 2003) obtained a very high correlation between children's emotional understanding and their language ability, even when the effect of age was controlled.More recently, Perpiñ a et al.
(2020) addressed a study with 180 students from 8 to 11 years to identify which EI components were more related to linguistic competences in primary education, concluding that adaptability and interpersonal were the most outstanding, specially adaptability for reading comprehension.Therefore, the importance of considering and enhancing emotional skills to improve the learning process of students should be highlighted, becoming a support alternative to learning at any of the stages (Pulido-Acosta & Herrera-Clavero, 2019).In this study, SP correlates positively with all dimensions of EFs, IQ and EI dimensions except intrapersonal, being in line with studies assessing the influence of EFs on AP in primary school (Barber et al., 2020;Hoyo et al., 2019;Nesayan et al., 2019;Spiegel et al., 2021).Authors such as Peng and Kievit (2020) analysed the bidirectional relationships between PA, EFs, reasoning, among other cognitive variables, finding a significant positive correlation between reading, mathematics and cognitive skills during children's development.In this same line of study, the research by Zorza et al. (2016) with primary and secondary students from 8 to 13 years observed that EF measures explained 41% of the variability in school performance and 29% of the variance in the social behaviour of schoolchildren.In a study only with primary students from 8 to 11 years old, they analysed EI and EFs and the components more closely related to AP, showing that EFs are better predictors than EI.Inhibition and WM were the EFs most associated with AP while adaptability emerged as the EI dimension most linked to it (Perpiñ a et al., 2023).There was a significant negative correlation between the scores on the intrapersonal EI component and the students' scores in the mathematics task, while showing a positive one between the interpersonal component and language.
With regard to the second objective, to analyse the profiles in EFs presented by the students and the differences in the levels of IQ, SP and EI, it was found that there are two clearly identified profiles, one with high values and one with low values in all the EFs dimensions.Furthermore, the high EF profile showed higher values of IQ, SP and interpersonal EI.To date, no studies have been found that have yielded results in this regard, which makes these initial findings especially valuable.
Studies related to these variables are Rebollo-Goni and de la Pena-Alvarez (2017), who found significant correlations between interpersonal and intrapersonal EI and the EF dimensions (verbal fluency, planning and interference) in a primary students sample of 87 schoolchildren from 6 to 9 years old.In the present study, only interpersonal EI showed differences between the profiles and was seen in the correlations related to CF and verbal fluency.
Regarding the stress EI dimension, it was related to the inhibition of the EF dimension in this study, but no differences were found between profiles.In this sense, Molero-Jurado et al. (2021) analysed the impact of PA with burnout and emotional competences in adolescent students, showing that low PA affected the level of burnout and that stress management and mood in EI acted as These results reinforce the relationship between cognitive abilities, academic competencies and emotional competencies that could have positive consequences on learning (Barkley, 2001;Moreno-G omez et al., 2020).Several studies agree that EF stimulation plays a relevant role predicting AP (Blair & Razza, 2007) in kindergarten students and enhancing self-regulation and school success (Moreno-G omez et al., 2020;Razza et al., 2015).A study on EFs and AP in schoolchildren aged 6-12 years (Porto et al., 2021) evidenced a statistically significant relationship between PA and components of EFs such as phonological fluency and CF.On the other hand, semantic fluency and inhibition were shown to be predictive factors for PA.
Regarding the limitations of the study, it is a descriptive and correlational design study in which cause-effect relationships are not established between the variables analysed.Likewise, it would be necessary to increase the sample size of the study in order to extrapolate results regarding the bulk of the school population.Therefore, quasi-experimental studies are needed to study in depth, the relationship between the application of educational programmes that apply EI techniques and the work of the EFs, with better academic results in schoolchildren.Likewise, randomised controlled trials (RCTs) are needed that intervene through programmes that integrate the development of EFs and emotional competences, with the purpose of finding out, through neuroimaging techniques (e.g., electroencephalogram and functional magnetic resonance imaging), what happens in brain function and how this is related to the behaviour of schoolchildren in and out of the classroom.
In accordance with the proposed objectives, a direct relationship between the variables under study EFs, SP, IQ and some EI dimensions was confirmed.With respect to the profiles in EFs presented by primary school students, there are two clearly differentiated profiles: one with high and another with low values in IQ, SP and interpersonal EI.However, caution should be exercised with the number of profiles obtained based on EFs, as these profiles may vary according to sample size, and covariates such as age and gender.Finally, it was observed that the development of interpersonal EI is related to SP in schoolchildren with a profile of high EFs.Therefore, schoolchildren require the development of executive skills, the practice of emotional regulation, interpersonal communication, the adoption of perspectives and adherence to social rules that can be provided from the educational and social context (Perry et al., 2019).All this, with the ultimate purpose of enhancing their school success (cognitive and academic performance) and their overall well-being.

| Practical application
The results found in the present study, together with the review of numerous studies that link the cognitive and emotional competencies of schoolchildren with AP, show the importance of designing training programmes in interpersonal EI, which stimulate EFs such as metacognition, attention, perception, WM and mental flexibility, in favour of the academic, personal and social performance of this population (Rebollo-Goni & de la Pena-Alvarez, 2017).The need to apply teaching methodologies that focus on the management of personal and interpersonal emotions and cognitive abilities of young students is evident.Authors such as Bisquerra (2016) and Garaigordobil (2018) propose the integration of emotional education throughout the curriculum at different age stages (from kindergarten to secondary school).Thus, classroom activities could be designed (in mathematics, language, languages, physical education, …) to work with emotional management in interpersonal interaction, with reflection on the emotions felt in different situations and with mental flexibility and valuing the opinion of others in order to make 'ecological' decisions (positive decisions for oneself and for the social environment).From this postulate, we also derive the need to train teachers in the implementation of these activities, in order to be able to verify the consequences of these didactic practices on PA, coexistence in the classroom and school absenteeism.
is made up of different subtests for the evaluation of the linguistic factor (it integrates two verbal content tests: Sentence Completion [Vv] and Verbal Analogies [Rv]), the mathematical factor (of general numerical content, assessed by means of two different tests: Numerical Problems [Nn] and Numerical Series [Rn]) and the visuo-spacial factor (composed of the Rotated Figures [Ge] test and the Logic Matrices test [Re]) variables in the school environment, developed in Spain.
Socio-demographic data of the study sample.
T A B L E 1 age = 10.92 (SD = .68) The internal reliability of the battery ranges from .78 to .93 depending on the different tests.Adaptability and Mood subscales.A total EI score is derived from all of them.In this research, we used the Spanish validated version proposed by Ferr andiz et al. (2012), administered in a group setting in the classroom with the help of the tutor teacher.Cronbach's alpha coefficient was used to calculate the internal reliability of each item of the questionnaire used in the IE assessment (Bar-On EQ-i: YVTM).The values were .77for the interpersonal dimension, .73 for the intrapersonal dimension, .77for stress management, .78 for adaptability, .80 for mood and .89for the total scale.
T A B L E 3 Differences in variables according to executive function profile.
T A B L E 4 Differences in IQ, spelling performance and emotional intelligence according to executive functions profile.