School Refusal Behavior Profiles, Optimism/Pessimism, and Personality Traits in Spanish Children

Abstract

The relationship between school refusal behavior (SRB) profiles and personality traits has received little attention from investigators. Identifying the profiles of students with school attendance problems may improve the understanding of the characteristics defining these students. The aim of this study was to identify different SRB profiles and analyze the relationship between these profiles, and optimism/pessimism and personality traits. The School Refusal Assessment Scale-Revised, the Youth Life Orientation Test, and the Big Five Questionnaire were administrated to 739 Spanish students aged 8–11 (M_age = 9.92; SD = 1.12). Pearson’s correlation coefficients revealed a significant association between personality dimensions and SRB. Three distinct profiles were identified: (1) SRB by negative reinforcement (high scores on avoiding school-related stimuli provoking negative affectivity), (2) SRB by positive reinforcement (high scores on pursuing positive tangible reinforcement outside of school), and (3) Low SRB. The SRB profile by positive reinforcement scored higher on Extraversion, Agreeableness, Conscientiousness, Openness, and Optimism, whereas the SRB profile by negative reinforcement scored higher on Neuroticism and Pessimism. More statically significant differences were found between the negative and positive reinforcement profiles. The role of negative personality traits and pessimism as risk factors for students who are truant or refuse to attend school are discussed.

1. Introduction

Regular school attendance and academic success are fundamental to the integral development of children and adolescents [1]. The failure to attend school is considered a breach of the basic right to education and shared social values [2]. School attendance problems refer to a set of distinct types of school absences or difficulties in attending or staying in school [3]. A wide variety of factors may lead to refusal to go to school, including school, personal, family, or social factors. This variety prevents the proposal of a unique theoretical model that combines all of the causes of this behavior [4]. Thus, there appear to be diverse causes of this complex issue in terms of its occurrence and persistence [5,6].

Over recent years, various studies have attempted to identify the profiles or subgroups of students sharing similar traits who engage in school attendance problems [7,8]. Most of the studies are based on the functional model which includes four possible causes of student school rejection [9]. These four conditions are: (1) Avoidance of stimuli that cause negative affectivity, (2) Escaping from social aversion and assessment, (3) Seeking attention from significant individuals, and (4) Obtaining tangible reinforcement outside of the school. In the first two conditions, school rejection is maintained by negative reinforcement (e.g., avoiding oral or written test assessment), and in the last two, by positive reinforcement (e.g., attracting the attention of parents or devoting time to recreational activities such as playing videogames or watching television). The following school rejection behavior profiles have been identified the most frequently: Non-School Refusal Behavior, School Refusal Behavior by Negative Reinforcement, School Refusal Behavior by Positive Reinforcement, and School Refusal Behavior by Multiple or Mixed Reinforcements (characterized by high combined scores on the conditions for both negative and positive reinforcement) [10,11,12,13]. In this study, the term School Refusal Behavior (SRB) has been used to name the profiles of students who are truant or refuse to attend school because this is the term most widely used in this type of study [10,11,12,13].

In addition to identifying profiles, these works have also analyzed the relationship between the distinct groups of students who refuse to go to school and their relationship with other psychological variables. Results suggest that the Mixed SRB profile is the least adaptive group since it is associated with lower scores on variables such as anxiety [12,14] or cyberbullying [10]. On the other hand, students demonstrating refusal behavior by positive reinforcement who engaged in this behavior to attract the attention of other significant individuals tend to suffer from separation anxiety disorders, while those who reject school to obtain tangible reinforcement outside of the school tend to suffer from oppositional defiant disorder and behavior issues [15]. Similarly, students obtaining higher scores on negative reinforcement are considered to be at higher risk when forming a perception of themselves [16].

Prior studies seem to suggest that there are certain profiles of students who refuse to attend school or have difficulties in remaining in school that are at greater risk of suffering from mental health problems and internalizing problems such as depression or anxiety [17,18,19,20,21]. Since emotions such as anxiety and enthusiasm may alter an individual’s emotional regulation [22], feelings associated with happiness, confidence, participation, or enthusiasm may act as preventive variables protecting against refusal to go to school [13]. The term ‘optimism’ refers to the tendency to highlight or consider the more positive and likable aspects of an individual or object, while pessimism refers to the expectation of the occurrence of negative events [23]. While over recent years, the constructs of optimism and pessimism have been studied more frequently, most works have examined these variables in the labor environment, with few education-based studies [24]. However, prior studies have suggested that optimism is linked to successful academic performance [25]. Fernández and Doldán [26] suggested that a positive attitude is related to good academic performance, low levels of illness, and expectations to achieve the anticipated. Pessimism, on the other hand, is associated with at-risk behavior, a decline in health, or poor academic performance [27]. As for students who refuse to go to school, only one study suggested that pessimism is a significant predictor of high school refusal scores, except in cases in which students based their refusal to go to school on the obtaining of tangible reinforcement outside of the school, in which case, the results were the opposite [13]. Therefore, schools should attempt to offer students the necessary emotional training to encourage their positive development, since optimism and pessimism are related to the emotional wellbeing of youth [28].

On the other hand, past studies have revealed that personality plays a major role in the academic context, with personality traits negatively affecting the student’s attendance due to the relationship existing between school attendance issues and emotional instability [29,30,31,32,33,34]. However, the role that these personality traits may play in students who are truant or refuse to attend school has yet to be extensively studied [29]. Personality refers to the stable characteristics of an individual and is revealed by their behavior [35]. In the study of personality, the Big Five Model is a benchmark that examines the structure of the same, based on five descriptive elements or personality traits. It is one of the most frequently used theoretical models to define and measure an individual’s personality [36,37]. Based on this approach, the following personality dimensions have been proposed: (1) Extraversion, referring to sociable, communicative, and secure individuals, (2) Agreeableness, referring to the capacity for empathy and reflecting interpersonal trends, (3) Conscientiousness, referring to the ability to control impulses and to organize the tasks at hand, (4) Neuroticism, referring to emotional instability and related to negative emotions such as anger or guilt, and (5) Openness, referring to an individual’s ability to tackle new situations.

The objectives of this study are intended to offer additional information on the personality traits of students who are truant or refuse to attend school. On the one hand, the work attempts to identify the school refusal behavior profiles by analyzing latent profiles, combining high and low scores on the first and fourth factors of the SRAS-R. On the other hand, it attempts to analyze the relationship between the identified profiles and the optimism/pessimism variables, and the personality traits. According to past studies, the following hypotheses have been proposed:

Hypothesis 1.

Based on prior studies which have identified three distinct profiles of school refusal behavior [10,12,13], the following groups are expected to be found: a low school refusal profile, school refusal behavior by negative reinforcement to avoid negative affectivity, school refusal behavior by positive reinforcement to obtain tangible reinforcement outside of the school, and a mixed school refusal behavior profile.

Hypothesis 2.

Given that pessimism has been related to poorer academic results [25,26], it is anticipated that the profile of students with school refusal behavior by mixed or negative reinforcement will have higher mean scores on pessimism.

Hypothesis 3.

Since refusal to go to school is associated with a more internalizing character and mental health problems [17,18,19,20], it is expected that the student profile with school refusal behavior by mixed or negative reinforcement will have higher mean scores on the personality trait associated with neuroticism.

2. Materials and Methods

2.1. Participants

A total of 739 participants were recruited, aged 8 to 11 (M = 9.2; SD = 1.12). The sample distribution based on sex and age was as follows: 150 8-year-old participants (69 boys and 81 girls), 205 9-year-old participants (95 boys and 110 girls), 162 10-year-old participants (85 boys and 77 girls), and 222 11-year-old participants (118 boys and 104 girls). No significant differences were found in the sample distribution based on sex and age (χ² = 3.30; p = 0.35).

2.2. Instruments

School Refusal Behavior Assessment Scale-Revised (SRAS-R) [38]. The SRAS-R is a self-report measure consisting of 24 items with a 7-point response scale (0 = never; 6 = always). This instrument assesses school refusal behavior in students aged 8 to 17. Specifically, it permits the assessment of the causes of school refusal, according to the functional model that proposes four factors contributing to school refusal: I. Avoiding stimuli that provoke negative affectivity (e.g., “How many times have you tried not to go to school because if you go, you will feel sad or depressed?”), II. Avoid aversive social situations (e.g., “If it were easier for you to make new friends, would it be easier for you to go to school?”), III. Attracting the attention of significant individuals (e.g., “How often would you prefer that your parents taught you at home instead of your teacher at school?”), and IV. Obtaining tangible reinforcement outside of the school (e.g., “How often do you reject going to school because you would rather have fun outside of school?”) [38]. Appropriate internal consistency indices were identified for each of the previously mentioned factors: 0.70, 0.79, 0.87, and 0.72, respectively [13]. In this study, the Spanish validation of this instrument proposed by Gonzálvez et al. [30] was used. It consists of the four cited dimensions, but with 16 items. The internal consistency of the SRAS-R in this study was 0.72 (Factor I) and 0.70 (Factor IV).

Youth Life Orientation Test (YLOT) [39]. The YLOT is an assessment tool consisting of 16 items with a 4-point Likert-type response format (0 = not true for me; 3 = true for me). It is intended for children and adolescents aged 8 to 16 and is intended to assess positive expectations (e.g., “I usually expect to have a good day”) and negative expectations (e.g., “When things are good, I expect something to go wrong”). The measure was designed to yield three score types: an optimism score, a pessimism score, and a total score. Internal consistency for this instrument has been reported to have adequate reliability indices: 0.78 (pessimism), 0.79 (optimism), and 0.78 (total score) [39]. In this study, the Spanish version of the instrument was used, which continues to use 16 items [40]. Internal consistency indices for this work were 0.80 (optimism) and 0.78 (pessimism).

The Big Five Questionnaire for Children (BFQ) [36]. This instrument contains 65 items, 13 per scale in a 5-point Likert-type response format (1 = almost never; 5 = almost always) with 5 dimensions for personality assessment resulting from its analysis: I. Extraversion-introversion: aspects related to creativity, enthusiasm, assertiveness, and self-confidence (e.g., “I like talking to others”); II. Agreeableness-hostility: concern and sensitivity towards others (e.g., “I treat others very kindly”); III. Conscientiousness: assessing the autonomy, order, precision, and compliance with rules and commitments (e.g., “I respect the rules and order”); IV. Neuroticism-stability: refers to anxiety, depression, or anger (e.g., “I get angry easily”), and V. Intelligence or openness to experience: assesses intellectual aspects, cultural interests, creativity, fantasy, and interest in other cultures (e.g., “When the teacher explains something, I understand it quickly”). The levels of internal consistency for this instrument range from 0.80 (Extraversion) to 0.94 (Agreeableness) [36]. In this study, the Spanish validation proposed by Carrasco et al. [37] was used. The following internal consistency indices were found: 0.78 (Extraversion), 0.92 (Agreeableness), 0.89 (Conscientiousness), 0.82 (Neuroticism), and 0.84 (Openness).

2.3. Procedure

First, the school principals were contacted and the informed consent of the parents was requested to authorize the participation of their children. During this process, the school principals and families were informed of the study objectives and protocol of data usage, in accordance with the ethics committee of the Universidad de Alicante (UA-2017-09-05) which approved the study. The informed consent was offered in writing for participation in the study. It was granted by the legal guardian or parent of each participant, or by their closest relative. Instruments were administered in approximately 45 min. Instructions were read out loud and all participants were verified to be acting in accordance with their own voluntary will. In all of the sessions, at least one of the researchers was present, in addition to the classroom teacher. Finally, all members of the educational community were thanked for their participation.

2.4. Statistical Analysis

First, a Latent Profile Analysis (LPA) was performed to establish the profiles of school refusal behavior. To determine the number of latent profiles that best agree with the data from this work, distinct profile analysis models were used, as well as a variety of adjustment indices. The Bayesian Information Criteria (BIC), Akaike Information Criterion (AIC), Vuong–Lo–Mendell–Rubin Likelihood Ratio Test (LRT), Bootstrap Likelihood Ratio Test (BLRT), entropy, and size index were used as measures of goodness of fit. The lowest values of AIC and BIC were preferred, as were the values closer to 0.05 for the LRT and BLRT likelihood tests, and scores closer to 1 for entropy and closer to 0 for size index.

In this study, version 8 of the MPlus was used, as well as the social sciences statistics package, SPSS-25. A multivariate analysis of variance (MANOVA) was performed, in addition to post hoc tests (the Bonferroni correction), to identify the statistically significant differences between groups. In addition, the size of the differences was calculated using Cohen’s d (effect size), with resulting values between 0.20 and 0.49 indicating a low effect size, those between 0.50 and 0.79, a moderate effect size, and those above 0.80, a high effect size.

3. Results

3.1. Correlations between School Refusal Behavior, Personality, and Optimism/Pessimism

As for the personality dimensions, Factor I of the SRAS-R correlated negatively and significantly with Extraversion (−0.14, p < 0.001), Agreeableness (−0.17, p < 0.001), Conscientiousness (−0.20, p < 0.001), and Openness (−0.20, p < 0.001), and it correlated positively and significantly with Neuroticism (0.20, p < 0.001). The fourth factor of the SRAS-R correlated positively and statistically significantly with the dimensions of Extraversion (0.18, p < 0.05), Agreeableness (0.09, p < 0.05), Conscientiousness (0.08, p < 0.001), and Openness (0.13, p < 0.001).

As for Optimism, Factor I of the SRAS-R was negatively and significantly correlated with this factor (−0.13, p < 0.001), although it was positively and significantly correlated with Pessimism (0.34, p < 0.001). The fourth factor of the SRAS-R correlated positively and significantly with Optimism (0.23, p < 0.001) and negatively and significantly with Pessimism (−0.20, p < 0.001).

3.2. School Refusal Behavior Profiles

Table 1. Data fit of all models.

Models	AIC	BIC	BIC-Adjusted	LRT p	LRT-Adjusted	BLRT	Entropy	Size
2 profiles	4093.67	4125.90	4103.68	<0.001	<0.001	<0.001	0.85	0
3 profiles	4052.18	4098.23	4066.48	0.01	0.01	<0.001	0.70	0
4 profiles	4014.92	4074.79	4033.51	0.05	0.05	<0.001	0.73	1
5 profiles	3996.62	4070.30	4019.49	0.11	0.12	<0.001	0.72	1
6 profiles	3981.23	4068.73	4008.39	0.12	0.13	<0.001	0.79	2

Note: AIC = Akaike Information Criterion; BIC = Bayesian Information Criteria; LRT = Vuong–Lo–Mendell–Rubin Likelihood Ratio Test; BLRT = Bootstrap Likelihood Ratio Test.

shows the adjustment indices of the 6 examined models. The lowest BIC and AIC were obtained in the model of 6 classes; however, this model was rejected since, as with the models of 4 and 5 profiles, the adjusted LRT index was higher than 0.05. Furthermore, the representativeness of the index size was above 0 for models 4, 5, and 6, indicating that one or more groups were formed by a reduced number of individuals and, therefore, representativeness was not achieved.

Combining all of the criteria, the three-profile model was selected as the best model in terms of adjustment, having p < 0.01 for the Vuong–Lo–Mendell–Rubin Likelihood Ratio Test, the second-lowest scores for the AIC and BIC as compared to the other models, and a lower entropy value than the other models.

Figure 1 presents the three identified profiles. On the one hand, the SRB by negative reinforcement profile was characterized by high scores (1.95) on the first factor of the SRAS-R (avoiding school-related stimuli provoking negative affectivity) and the lowest scores for the fourth factor (0.12); the Low SRB had low scores on the fourth factor (−1.16) and a negative trend (−0.16) for the first factor of the SRAS-R, and the SRB by positive reinforcement was characterized by high scores (0.51) on the fourth factor of the SRAS-R (pursuing positive tangible reinforcement outside of the school) and the lowest scores on the first factor (−0.27).

3.3. Differences between the Profiles of School Refusal Behavior in Personality and Optimism/Pessimism

As for the differences in scores of the personality dimensions between the three clusters, the results of the MANOVA suggest the existence of statistically significant differences in the five personality variables and in optimism/pessimism (Wilks’ lambda = 0.83, F(14,739) = 10.04, p < 0.001, η_p² = 0.09) (see

Table 2. Means and standard deviations obtained by the three profiles of school refusal behavior in personality traits.

		SRB by Negative Reinforcement (n = 73)		Low SRB (n = 184)		SRB by Positive Reinforcement (n = 482)		Statistical Significance
		M	SD	M	SD	M	SD	F(2, 736)	η2
Dimensions BFQ	Extraversion	43.39	9.62	42.49	8.77	45.64	9.63	8.12 **	0.02
	Agreeableness	44.63	10.77	46.47	11.39	48.16	12.44	3.43 *	0.01
	Conscientiousness	43.65	10.28	46.20	10.82	47.95	11.52	5.41 *	0.02
	Neuroticism	37.12	7.94	34.28	10.33	33.61	11.06	3.48 *	0.01
Dimensions YLOT	Openness	42.79	10.47	43.58	10.01	46.60	10.03	8.86 **	0.03
	Optimism	14.43	2.73	14.25	3.14	15.33	2.42	12.57 **	0.03
	Pessimism	8.25	4.43	5.83	3.46	4.11	3.34	51.32 **	0.12

Note: SRB = School Refusal Behavior; BFQ = Big Five Questionnaire; YLOT = Youth Life Orientation Test. * p < 0.05; ** p < 0.001.

). These differences were statistically significant for all of the personality dimensions and the optimism/pessimism variables. The highest mean scores were obtained in the profile of SRB by pursuing tangible reinforcement outside of the school for the following personality dimensions: Extraversion, Agreeableness, Conscientiousness, and Openness. On the other hand, for the dimension of neuroticism, the profile of SRB by avoiding negative affectivity obtained the highest mean scores. As for the dimensions of Optimism and Pessimism, the profile of SRB by tangible reinforcement outside of the school received the highest scores in Optimism while the profile of SRB by avoidance of negative affectivity generated the highest scores for Pessimism.

Table 3. Cohen’s d value for post hoc contrasts between profiles on BFQ and YLOT dimensions.

		SRB by Negative Reinforcement vs. Low SRB	SRB by Negative Reinforcement vs. SRB by Positive Reinforcement	Low SRB vs. SRB by Positive Reinforcement
Dimensions BFQ	Extraversion	−	−	−0.34
	Agreeableness	−	−0.29	−
	Conscience	−	−0.38	−
	Neuroticism	0.29	0.33	−
	Opening	−	−0.38	−0.30
Dimensions YLOT	Optimism	−	−0.37	−0.41
	Pessimism	0.64	1.18	0.51

Note: SRB = School Refusal Behavior; BFQ = Big Five Questionnaire; YLOT = Youth Life Orientation Test.

shows the post hoc comparisons made to identify the effect size of the statistically significant differences found between the groups. When comparing the profile of SRB by negative reinforcement and the Low SRB profile, low effect sizes were found for the dimension of Neuroticism (0.29) and moderate effect sizes were found for the Pessimism (0.64) dimension. On the other hand, between the profiles of SRB by negative reinforcement and SRB by positive reinforcement, low effect sizes were found for the dimensions of Agreeableness (−0.29), Conscientiousness (−0.38), Neuroticism (0.33), Openness (−0.38), and Optimism (−0.37). High effect sizes were found, however, for Pessimism (1.18). Finally, when comparing the profiles of Low SRB and SRB by positive reinforcement, low effect sizes were found for the dimensions of Extraversion (−0.34), Openness (−0.30), and Optimism (−0.41), and moderate effect sizes were found for the Pessimism (0.51) dimension, when comparing these profiles.

4. Discussion

The objective of this study was to identify the profiles of students who are truant or refuse to attend school through the analysis of latent profiles combining high and low scores on the first and fourth factor of the SRAS-R and to analyze the relationship between the different profiles identified and the optimism/pessimism variables as well as the five personality traits.

In this work, three school refusal behavior profiles have been identified: (1) SRB by negative reinforcement (avoiding school-related stimuli provoking negative affectivity), (2) Low SRB, and (3) SRB by positive reinforcement (pursuing positive tangible reinforcement outside of the school). Taking into account the first hypothesis created, these profiles coincide with the anticipated results [10,11,13]. However, one of the profiles proposed in this hypothesis has not been identified: the mixed profile, characterized by the combination of high scores on functions of school refusal behavior by negative and positive reinforcement. This may be because, in this study, the creation of profiles has been carried out by only assessing Factor 1 (Avoidance of Stimuli that Provoke Negative Affectivity) and Factor 4 (Obtaining Tangible Reinforcement outside of the School) of the SRAS-R. Despite the fact that these two factors represent causes for not going to school by negative reinforcement (Factor 1) and by positive reinforcement (Factor 4), prior studies that have identified the mixed profile have characterized it by high scores on negative reinforcement, and in the case of positive reinforcement, it was not the fourth factor, but rather, the third, in which it was identified (Seeking the Attention of Significant Individuals). This functional condition of refusal to go to school was not assessed in this work.

In the identified profiles, it was found that the SRB profile by negative reinforcement was the most maladaptive, obtaining the highest scores on Pessimism and being associated with the highest mean scores on Neuroticism. Thus, Hypotheses 2 and 3 were confirmed. The results suggest that students with a profile of school refusal behavior by negative reinforcement scored higher on pessimism, in agreement with past studies that related school refusal by this condition with more internalizing problems such as anxiety, sadness, or apathy [17,18,19]. The results obtained in this work reveal significant differences in personality variables based on the high and low scores on school refusal behavior, as seen in past studies which suggested that personality plays a major role in the academic context [29,31,32,33,34] and that those positive emotions are more closely related to optimistic attitudes [23].

Given the growing interest in determining which factors act as risk variables and which may serve as protective factors against school attendance problems, this study reveals that the personality variables of Neuroticism and Pessimism are related to students whose refusal to go to school is based on negative reinforcement, as suggested by past studies [13]. It should be noted, however, that future studies are necessary to determine if the personality variables of Optimism, Extraversion, Agreeableness, and Openness serve as protective factors against refusal to go to school, since in this study, the highest scores were received on these variables, as compared to students in the Low School Refusal Behavior group. This may be due to the fact that profiles for tangible reinforcement (Factor IV) have been associated with variables such as empathy, happiness, or confidence [16]. Therefore, it may be interesting to determine whether or not these students have an adjusted view of the reality of their impact on their peers.

Although this study offers numerous contributions, it also has certain limitations. First, despite the fact that almost 1000 students participated, their age range was between 8 and 11, making it impossible to generalize the study results to other age ranges or cultural groups. The study was conducted with students of primary education; therefore, future studies may be conducted in subsequent phases on the reference population used in this work, as well as in students from distinct countries. On the other hand, it would be interesting to carry out longitudinal studies that permit the observation of the causal relationship between the mentioned variables and the determination of changes taking place over time in the studied variables. Furthermore, the instruments used in the study were based on questionnaires completed by the students, so it is expected that the use of self-reporting as the sole collection instrument may be complemented by other multi-source tools, including the participation of families and teachers.

Despite these limitations, this work offers a representative image of the relationship between the profiles of refusal to go to school and optimism/pessimism and the distinct personality dimensions. It increases the understanding of the heterogeneity of students who refuse to attend school and offers relevant empirical evidence on the relationship between profiles of refusal to go to school and the optimism/pessimism dimensions with respect to other past studies. Given that school attendance is a major pillar of the educational, social, and personal development of children and adolescents, efforts should be made to determine the specific characteristics and needs of these students. Further study is necessary to determine which variables serve as protective factors against refusal to go to school.

5. Conclusions

This study offers empirical evidence on the influence of personality traits (Extraversion, Agreeableness, Openness, Conscientiousness, Neuroticism, Optimism/Pessimism) on distinct groups of students who reject school. Although this work offers interesting information on the significant differences in the personality variables based on the high and low scores on refusal to go to school, additional studies are likely to be carried out in this area. The results suggest that not all adolescents who refuse to go to school have the same behavior patterns, with the SRB by negative reinforcement profile group being associated with the highest scores on pessimism and neuroticism. Therefore, prevention programs are proposed to emotionally and psychologically promote the mental health development of children and adolescents, such as the FORTIUS program [41], which seeks the personal development and the promotion of wellbeing; the Training Program of Optimistic Explanatory Style and Skills in Adolescent Communication [42], aimed to increase optimism and reduce depressive symptoms or the Student´s Optimistic Attitudes, and Resiliency Program, designed to improve levels of optimism and resilience [43]. Programs should be carried out in centers to promote personality traits that understand and encourage positivity in everyday life, leading to better attitudes towards school and thereby avoiding scholastic absence problems. In addition to promote mental health programs, school environment can also be important. Schools should be a place where everyone—students, staff, parents, and the community—feels safe, welcome, and respected. An inclusive and accepting school environment is essential for student wellbeing and achievement. Lack of secured school environment may bring school attendance problems [44]. Creating a warm and supportive classroom climate where educational inclusive practices are implemented will help to improve students’ expectations and results, making it easier for their adaptation to school [45].