Exploring Mexican lower secondary school students’ perceptions of inclusion

Abstract

Emotional well-being, social inclusion, and academic self-concept are crucial for students’ academic and socio-emotional development, and are seen as key factors for successful inclusion. Considering this relevance, the Perception of Inclusion Questionnaire (PIQ) has been developed as a short screening instrument to assess these three aspects—not only for research, but also for teachers to enhance students’ well-being in school and the selection of intervention goals. The present study aimed to evaluate the psychometric properties of the Spanish PIQ student version. Participants were 673 Mexican lower secondary school students (52% female, M_age = 13.25 years) who either attended inclusive classes (33%) or special schools (67%). 15% have been diagnosed as having special educational needs (SEN). Factorial and convergent validity and measurement invariance as well as mean differences based on students’ gender, diagnosis of SEN, and classroom setting were analyzed. The three-factor structure of the PIQ was confirmed, and the three scales (emotional well-being, social inclusion, academic self-concept) showed adequate internal consistency. Further, some evidence for convergent validity was shown. Strong measurement invariance was established, allowing for comparison across gender, SEN status, and classroom setting. Regarding gender, boys indicated higher levels of social inclusion and academic self-concept than girls. Students with SEN scored lower on all three subscales compared to students without SEN. Further, students with SEN in inclusive classrooms showed lower academic self-concepts compared to those attending special schools. Although more research is needed, the psychometric properties of the Spanish PIQ student version are encouraging for its potential for use in research and practice.

Keywords:

• perceptions of inclusion questionnaire
• emotional well-being
• social inclusion
• academic self-concept
• student perceptions

1. Introduction

In recent decades, schools have increasingly emphasized their students’ emotional and social development. Especially within the recent COVID-19 crisis, children’s psycho-social risks have come into the focus of attention (e.g. Flack et al., 2020; Mælan et al., 2021; Styck et al., 2021). For instance, as a result of the COVID-19 school related closures, the social inclusion of students has been investigated in numerous of studies (e.g. Drane et al., 2021; Loades et al., 2020). Thus, with the critical impact that the COVID-19 pandemic has had, and continues to have, inclusive education has become even more important worldwide (OECD, 2023). According to UNESCO (2018), inclusive education is the “process of strengthening the capacity of the education system to reach out to all learners” (p. 7). In line with this, Schwab (2020) further explains that inclusive education is focused on ensuring equity in education, increasing participation, and providing high-quality educational opportunities that support all learners. As a result, inclusive education aims to support every student’s academic achievement outcomes as well as non-achievement outcomes, such as socio-emotional variables (European Agency for Special Needs and Inclusive Education, 2017). Thus, assessing the socio-emotional outcomes of students can provide important information concerning the quality of inclusive schooling (DeVries et al., 2018). However, perceptions of inclusion in the form of socio-emotional variables, such as students’ emotional well-being, social inclusion, and academic self-concept, can vary across gender, educational settings (such as inclusive or non-inclusive classroom settings), and special educational needs (SEN) status (DeVries et al., 2018; Pozas, Letzel, et al., 2021; Savolainen et al., 2018). With this background, appropriate assessment tools that allow for the measurement of such socio-emotional variables are necessary, especially from the perspective of students, as they are key stakeholders (De Leeuw et al., 2018). Assessing students’ perceptions of inclusion within schools is not only important for research and school effectiveness, but also for developing appropriate interventions within the classroom. To this aim, the Perceptions of Inclusion Questionnaire (PIQ), which explores students’ emotional well-being, social inclusion, and academic self-concept in a simple and easily understandable way, has been recently developed. Although the PIQ is currently available in different languages, and previous research on the PIQ has indicated that the instrument holds high psychometric properties (DeVries et al., 2018; Guillemot & Hessels, 2022; Zurbriggen et al., 2019), up to now, the PIQ’s psychometric properties have not been explored within a Spanish-speaking country, and thus a comprehensive evaluation of this instrument in Mexico is lacking. This is particularly significant, given that even though “certain system level policies and practices in Mexico […] can favor equity, such as delayed tracking and limited ability grouping” (OECD, 2018, p. 8), Mexico still faces significant challenges when it comes to developing a more inclusive education system (Romero-Contreras & García-Cedillo, 2013; Romero-Contreras et al., 2013).

Considering these challenges and knowledge gaps, this study’s objective was to validate the Spanish PIQ student version. To this end, a sample of lower secondary school students in Mexico was used. First, we will briefly describe the Mexican educational system as well as its inclusive education regulations. We will then elaborate on the relevance of emotional well-being, social inclusion, and academic self-concept, and its assessment, as well as the available (international and national) research on the topic.

1.1. Inclusive education in Mexico

The Mexican education system caters to the educational needs of a large and highly diverse population (Forlin et al., 2010), in terms of for instance languages and dialects (OECD, 2019). Students with SEN attend basic, mainstream schools or specialized institutions (I. García-Cedillo et al., 2015). Although certain policies and practices such as delayed tracking and limited ability grouping are implemented, reports from the OECD (2018) indicate that Mexico ranks amongst the countries with the lowest level of social inclusion. In 2012, a constitutional reform established quality education as a right for all Mexicans (OECD, 2018), ensuring full participation access to quality learning (OECD, 2023). In addition, Mexican law states that private schools cannot deny admission to students with SEN (Official Gazette of the Federation Diario Oficial de la Federación, & DOF, 2011). Even though Mexico has established educational support such as curricular adaptations and accommodations to students with disabilities in inclusive settings (SEP, 2010), Romero-Contreras et al. (2013) argue that inclusive education in Mexico is still unsatisfactory. In fact, out of the 15% of students who have a disability, only 2.85% receive inclusive education (DOF, 2017). Moreover, approximately 90% of students attend public schools (OECD, 2019), having access to free education. In contrast, private schools are not publicly subsidized and are thus tuition-based, operating without any government funding (Santiago et al., 2012). Most private schools are attended by students of middle and high socioeconomic status (I. García-Cedillo et al., 2015). Even though private schools cannot deny admittance based on disability (DOF, 2011; Lavin et al., 2022), many students with SEN are left without appropriate support for their specific learning needs, as they would need to pay private school tuition as well as for the additional support required within these school settings (I. García-Cedillo et al., 2015). However, despite these barriers and challenges, empirical research has shown that students with SEN have greatly benefited from attending inclusive classrooms with regards to their emotional, social, and academic development (Contreras & Cedillo, 2013). For instance, a study by Márquez et al. (2021) reported that students with SEN learning together with students without SEN in inclusive classrooms indicated the ability to develop friendships with their classmates. Additionally, students with SEN expressed feeling accepted and participated actively with their peers. Considering the fact that the key objectives of inclusive education are to foster students’ emotional, social, and academic development and the impact these might have on their quality of life, it is necessary to empirically explore students’ perceptions of inclusion.

1.2. Emotional well-being, social inclusion, and academic self-concept in the context of inclusion

Considering the significant amount of time students spend at school, they are not only a place of development but also a living space. Positive relations between students’ school well-being as well as their academic achievement and development have already been demonstrated in numerous studies (e.g. Bücker et al., 2018; López et al., 2017; Taylor et al., 2017). Therefore, the school well-being of students is seen as a crucial factor for successful learning. Emotional as well as psychological well-being is now actively addressed by the educational policy (e.g. Ecclestone, 2012). According to Hascher’s (2004) six-component model of school-related well-being, in addition to positive attitudes towards school, social situations and students’ self-esteem also play significant roles for students’ well-being in school. Students who have lower levels of emotional well-being (e.g. Knickenberg et al., 2022; Schwab, 2018b) and those who struggle with their social inclusion (Schwab, 2018b) tend to leave school earlier. In the context of inclusion, those three aspects of school well-being can be considered important outcome variables (Schwab, 2020). Therefore, it is of major importance for research but especially for teachers to have instruments available to assess those variables, yet these are missing for Spanish-speaking students.

Social participation was used as a primary argument for why inclusive education is needed. It cannot be regarded as legitimate that some students are socially excluded from settings of others because of individual characteristics, such as having a diagnosis of SEN. However, while the inclusion of students with SEN in mainstream classes can be fostered by law (see e.g. Salamanca Statement or the United Nations Convention on the Rights of Persons with Disabilities), building friendships cannot be fostered by legal guidelines. In this context, previous research has indicated that students with SEN in inclusive education are at risk of lower social inclusion compared to their peers without SEN (see e.g. reviews of Bossaert et al., 2013; Schwab, 2018a). Similarly, lower scores for students with SEN in academic self-concepts have been shown in previous studies (e.g. DeVries et al., 2021; Savolainen et al., 2018; Weber & Freund, 2017). Only some studies have found lower levels of well-being for students with SEN (e.Skrzypiec et al., 2016), however, while others have not (Schwab et al., 2018; Zurbriggen et al., 2018).

Taking the school setting into account, a lower academic self-concept has been shown in students attending inclusive classes as opposed to special schools (e.g. Bear et al., 2002; Marsh et al., 2006). For emotional well-being, the results are unclear. Some studies indicate that students with SEN have higher levels of emotional well-being compared to their peers in special schools (Rossmann et al., 2011; Venetz et al., 2010). However, Schwab (2014) did not find an effect of the classroom setting on emotional well-being, though he did find higher levels of social inclusion for students with SEN attending inclusive classes. Group differences not only in terms of SEN but also considering gender variables have been investigated intensively. In the study of Schwab et al. (2019), girls experienced higher levels of emotional well-being compared to boys. For school well-being (Schwab et al., 2018) and academic-self-concept (e.g. DeVries et al., 2018, 2022), no group differences or rather inconsistent results have been found.

1.2.1. Emotional well-being, social inclusion, and academic self-concept: their relation to students’ intentions to quit school and general self-efficacy

When exploring students’ emotional well-being, social inclusion, and academic self-concept, it is important to consider other variables that could be related to their experiences in inclusive schooling. Such variables are students’ general self-efficacy (Briones et al., 2005; Klassen, 2010; Mælan et al., 2021) and their intentions to quit school (Frostad et al., 2015; Schwab, 2018b). For the case of emotional well-being, several studies have identified a close link between such constructs and general self-efficacy (Gómez et al., 2007; Malo et al., 2011). For instance, a study by De Caroli and Sagone (2014) revealed positive relationships between secondary school students’ general self-efficacy and emotional well-being. Likewise, Céspedes et al. (2021) explain that students with higher levels of self-efficacy also report greater degrees of social inclusion. Lastly, research has suggested that both general self-efficacy and academic self-concept are positively and substantially correlated (Bong & Skaalvik, 2003; Scherer, 2013). On the other hand, though, previous research has demonstrated a negative relationship between students’ emotional well-being (Schwab, 2018b), social inclusion (Pijl et al., 2014; Schwab, 2018b), academic self-concept, and their intentions to quit school. In particular, according to De Witte et al. (2013), a substantial high negative correlation can be expected between emotional well-being and an intention to quit school.

1.3. The perceptions of inclusion questionnaire

The PIQ (Venetz et al., 2015) is a short, simple, easily understandable questionnaire consisting of 12 items and designed to measure the domains of emotional well-being, social inclusion in class, and the academic self-concept. The PIQ measures these three constructs with four items each and on a 4-point Likert scale (1 = not at all true to 4 = certainly true). The PIQ has been designed to assess three different perspectives and can thus be self-administered or responded to by a child’s teacher, parent or primary caregiver. The student version is designed for children and adolescents from ages 8 to 16.

The PIQ is based on the questionnaire by Haeberlin et al. (1989; originally in German as the “Fragenbogen zur Erfassung von Dimensionen der Integration von Schülern [FDI]”). The FDI consisted of 45 items across 3 factors and was one of the first and only measurement instruments available to assess the levels of perceived inclusion by students at schools (DeVries et al., 2018). The FDI was revised in order to lower its language complexity, shortened to 12 items (Venetz et al., 2014), and renamed the Perceptions of Inclusion Questionnaire (Venetz et al., 2015). Additionally, the PIQ student version was expanded to a parent and a teacher version. The three PIQ versions are available in 24 languages (e.g. Spanish, English, German, Arabic, French) and are open access (www.piqinfo.ch).

The psychometric properties of the PIQ student version have been previously documented in several different countries. For instance, in a study with 823 German-speaking Swiss students (including 190 students with SEN) demonstrated it to be a reliable tool also on the individual item level (Zurbriggen et al., 2019). Knickenberg et al. (2020) have validated the PIQ for students with learning disabilities as well, however, and negatively worded items appeared to be somewhat problematic. The teacher version has been tested as well, and the consistency between teachers and students evaluated (Schwab & Alnahdi, 2020; Venetz et al., 2019). A study by Schwab et al. (2020) has also shown good psychometric qualities for the parent version. More recently, the PIQ student version has been validated in non-German speaking countries, for instance in Saudi Arabia (Alnahdi & Schwab, 2021), Slovenia (Schmidt et al., 2021), Poland (Zwierzchowska et al., 2022), Sweden (DeVries et al., 2022), and France (Guillemot & Hessels, 2022), indicating that the instrument has good psychometric properties in several language versions. To our knowledge, only one small-scale study (Pozas et al., 2023) has used the PIQ in the Latin American context. In that study, however, the Spanish (Mexican) PIQ student version was not validated. Analyses were conducted with a sample of 101 primary and lower secondary school students in Mexico. Results from this small-scale study indicated that students in general perceive themselves to be included at school. However, students in private schools experience less school well-being.

1.4. The present study

Based on the aforementioned (validation) studies, it can be concluded that the PIQ has high psychometric qualities. Nevertheless, the PIQ’s properties “cannot be considered universally valid” (Guillemot & Hessels, 2022, p. 3). Consequently, it is important to examine the instrument’s characteristics in each language and each context in which it is applied. Even though the PIQ is available in Spanish, its psychometric properties have until now not been explored, and as mentioned, never before within the Latin American (Spanish speaking) context. With this background in mind, the present study is guided by the following research questions:

(1) Does the PIQ have a three-factorial structure for the Mexican sample as indicated by previous theoretical and empirical literature?

It is hypothesized that the Spanish PIQ student version will perform similarly within the Mexican context such as it has in other countries, thus confirming the validity of the instrument.

(2) Does the PIQ hold appropriate reliability and convergent validity within the Mexican student sample?

In light of the aforementioned theoretical and empirical background, it is expected that the subscales of emotional well-being, social inclusion, and academic self-concept are positively related to students’ general self-efficacy, whereas these are negatively correlated to their intention to quit school. Especially for school well-being, a high negative correlation with the intention to quit school is hypothesized.

(3) Does the PIQ demonstrate sufficient measurement invariance across gender, SEN status, and classroom setting (inclusive vs. not inclusive)?

Given that it is expected that the instrument will be validated within the Mexican context, it is therefore hypothesized that a sufficient measurement environment will be obtained in order to make comparisons across the target subpopulations.

(4) Are there differences in Mexican lower secondary students’ emotional well-being, social inclusion, and academic self-concept based on gender, SEN status, and classroom settings (inclusive vs. not inclusive?

It is hypothesized that students with SEN will have lower scores in social inclusion as well as in academic self-concept compared to their peers without SEN. For the classroom setting, it is expected that students with SEN attending inclusive classes will have lower levels of academic self-concept compared to students attending special schools.

2. Methods

2.1. Participants

A total of 673 Mexican lower secondary school students (52% female) with a mean age of 13.25 years (SD = 1.96 years) participated voluntarily in the study. While 33% of the students were taught in an inclusive classroom, 67% were educated in a non-inclusive classroom. Of this sample, 98% of the students attended private schools and 15% were diagnosed as having SEN. Data collection took place between February and March of 2022. The participants completed a voluntary online survey during regular school hours, which took approximately 15 to 20 minutes. During the assessments, three assistants monitored and supported those students who might have encountered difficulties filling out the questionnaire. Given that students were underage, informed consent from their parents or tutors was obtained from all participants prior to their completing the questionnaire. The research project was approved by the [blinded] Ethical Committee.

2.2. Instruments

2.2.1. PIQ student version

The PIQ includes the subscales social inclusion (e.g., “I have a lot of friends in my class”), emotional well-being (e.g., “I like going to school”), and academic self-concept (e.g., “I am a fast learner”), with four items each. Students were asked to indicate their agreement using a 4-point Likert scale (1 = not at all true to 4 = certainly true). One statement in each of the subscales is negatively worded (e.g., “I have no desire to go to school”). Responses were recorded for the analyses.

2.2.2. General self-efficacy

Students’ general self-efficacy was measured using the Spanish version (Bläßler & Schwarzer, 1996) of the General Self-Efficacy Scale from Schwarzer and Jerusalem (1999). The scale consists of 10 items (e.g., “I can always manage to solve difficult problems if I try hard enough”) and comprises a 4-point Likert scale ranging from 1 = not at all true to 4 = exactly true (Cronbach’s α = .90).

2.2.3. Intention to quit

Students’ intentions to leave education early was measured using the Intention to Quit Scale by Pijl et al. (2014), which was translated into Spanish following a back-translation process (Bundgaard & Brøgger, 2019). The scale consists of 3 items (e.g., “I really feel that I am wasting my time in school.”) and comprises a 5-point Likert scale ranging from 1 = never to 5 = always (α = .86).

2.3. Data analysis

The collected data were analyzed using SPSS 28, Mplus 8.3 and R software (packages lavaan and equaltestMI). As a first initial step, the Kaiser-Meyer-Olkin (KMO) test and Bartlett’s test of sphericity were conducted to verify the adequacy of the data. Second, confirmatory factor analyses (CFA) were conducted in order to test the PIQ’s factor structure. Given that the items are categorical and were treated as such, the weighted least-square means and variances estimator (WLSMV) were used in order to consider the data’s non-normal distribution (Liang & Yang, 2014). To assess the quality of the model, different fit indices were employed: χ²/df ratio, where a ratio≤3 indicates a good fit with the data (Byrne, 2013); root mean square error of approximation (RMSEA), where a value≤0.08 suggests a well-fitting model; standardized root mean square residual (SRMR), with acceptable values considered to be ≤ 0.08 (Hu & Bentler, 1999); and comparative fit index (CFI) and Tucker-Lewis index (TLI), where values≥0.9 demonstrate adequate fit (Bentler, 1990). Third, the internal consistency of the factors was verified by means of the Cronbach’s α and McDonald’s ω. Fourth, the PIQ’s construct validity was explored by means of correlation analyses between self-efficacy, intention to quit, and academic achievement. Lastly, to test mean differences between groups, t-tests were performed. Effect sizes were assessed by Cohen’s d. Students’ emotional well-being, social inclusion, and academic self-concept were considered dependent variables whereas gender, SEN status, and classroom setting were entered as independent variables. Given that strong or scalar measurement invariance is a prerequisite for calculating meaningful mean comparisons across groups (e.g., gender, SEN status or classroom setting), the PIQ was first checked for measurement invariance. Configural, metric, scalar, and finally strict scalar invariance were tested.

3. Results

3.1. Factor structure, reliability, and convergent validity

The Kaiser-Meyer-Olkin measure and Bartlett’s test of sphericity demonstrated strong relationships among the items (KMO = 0.868; χ² = 3117.82, p ≤ .001), indicating appropriateness to perform a CFA. The PIQ’s three-factor structure was modeled with each of the four corresponding indicators. Except for the RMSEA, the fit indices for the three-factorial model indicated an appropriate model fit (χ² = 384.702, df = 51, p ≤ .001; RMSEA = 0.10, CFI = 0.98, TLI = 0.98, SRMR = 0.07). The standardized factor loadings ranged between .477 and .878 (emotional well-being: .59 ≤ λ ≤ .78; social inclusion: .56 ≤ λ ≤ .84; academic self-concept: .52 ≤ λ ≤ .81).

As indicated by the Cronbach’s α and McDonald’s ω, the three factors’ internal consistencies were of an acceptable level (emotional well-being α = .80, ω = .88; social inclusion α = .79, ω = .87; academic self-concept α = .72, ω = .80). To verify the evidence on convergent validity for the PIQ, correlation analyses were conducted between the variables of general self-efficacy and intention to quit. Prior to these analyses, the internal consistencies for both variables were examined. Both the general self-efficacy (α = .90, ω = .90) and intention to quit scales (α = .86, ω = .86) showed good internal consistencies for the present sample.

As seen in Table 1 and 2, the correlations were all moderate to high, and significant, which indicates a good convergent validity between the three PIQ scales and the corresponding scales. Students’ emotional well-being was moderately correlated with their general self-efficacy (r = .29, p ≤ .001) and negatively related to their intentions to quit school (r = −.31, p ≤ .001). With regards to students’ social inclusion, there was a positive correlation with general self-efficacy (r = .30, p ≤ .001), and a slightly negative correlation with intention to quit school (r = −.11, p ≤ .05). Lastly, there was a strong positive correlation between academic self-concept and general self-efficacy (r = .46, p ≤ .001), but a negative correlation to students’ intentions to quit (r = −.17, p ≤ .001).

Table 1. Descriptive statistics of the measures

	Emotional Well-being	Social Inclusion	Academic Self-concept	General Self-Efficacy	Intention to Quit
N	673	673	673	585	585
M	3.10	3.16	2.89	3.08	1.79
SD	.63	.61	.60	.61	1.02
Min	1	1	1	1	1
Max	4	4	4	4	5

Table 2. Correlations between the scales

	(1)	(2)	(3)	(4)
Emotional Well-being	-
(2) Social Inclusion	.54**	-
(3) Academic Self-concept	.47**	.39**	-
(4) General Self-Efficacy	.29**	.30**	.46**	-
(5) Intention to Quit	−.31**	−.11	−.17**	.04

** p < .01.

3.2. Descriptive statistics

Table 1 shows the means and standard deviations of the PIQ subscales as well as the other measures used within this study. The three PIQ means were high and with homogeneous standard deviations. As the theoretical mean of the scales was 2.5, the scores were significantly positive: emotional well-being (t(672) = 24.66, p ≤ .001, d = .63), social inclusion (t(672) = 28.39, p ≤ .001, d = .61), and academic self-concept (t(672) = 18.05, p ≤ .001, d = .56). The general self-efficacy mean score was also significantly positive: (t(584) = 22.77, p ≤ .001, d = .62). For a scale ranging from 1 to 5, the intention to quit scale had a low mean, (t(584) = −28.65, p ≤ .001, d = 1.02), thus indicating that students have a low level of intention to leave education early.

In sum, these results indicate that the majority of the participating students in this sample have a rather positive perception of inclusion. Nevertheless, when revising the results in detail, it is possible to identify that a total of 10% of the students who participated in this study could be considered to be at risk, given their significantly low levels of emotional well-being (4%), social inclusion (3%), and academic self-efficacy (3%).

3.3. Measurement invariance across gender, SEN status, and classroom setting

To ensure meaningful interpretation of group mean differences, measurement invariance for the three PIQ scales across groups (male vs. female students; students with SEN vs. students without SEN; inclusive classroom setting vs. non-inclusive classroom setting) was explored. Measurement invariance analyses revealed that the PIQ shows partially strong measurement invariance (Putnick & Bornstein, 2016) across gender (Table 3), classroom setting (i.e., inclusive vs. non-inclusive classrooms; Table 4), and across students with vs. without SEN (Table 5).

Table 3. Fit statistics of the CFA models and for the group comparison across gender

Model	χ^2wlsmv	df	p	CFI	RMSEA	SRMR	Δχ^2	Δdf	p
Baseline models
Overall	357.858	51	<.001	.981	.096	.073
Male	189.952	51	<.001	.982	.093	.079
Female	222.431	51	<.001	.981	.100	.079
Group comparison
Configural invariance	412.383	102	<.001	.982	.097	.079
Metric invariance	459.444	111	<.001	.979	.098	.083	19.074	9	.025
Scalar invariance	464.286	132	<.001	.980	.088	.079	12.027	21	.007

n = 339 male students, n = 314 female students. WLSMV = weighted least-square means and variances estimator; CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual. Δχ² = Chi-square difference tests (with reference to the previous model).

Table 4. Fit statistics of the CFA models for the group comparison across classroom settings

Model	χ^2wlsmv	df	p	CFI	RMSEA	SRMR	Δχ^2	Δdf	p
Baseline models
Overall	385.388	51	<.001	.980	.099	.074
Inclusive	203.005	51	<.001	.975	.116	.089
Non-inclusive	233.156	51	<.001	.984	.089	.073
Group comparison
Configural invariance	436.161	102	<.001	.981	.099	.078
Metric invariance	453.444	111	<.001	.981	.096	.079	6.433	9	.696
Scalar invariance	476.744	132	<.001	.980	.088	.079	80.565	21	<.001

n = 448 students in inclusive classrooms, n = 224 students in non-inclusive classrooms. WLSMV = weighted least-square means and variances estimator; CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual. Δχ² = Chi-square difference tests (with reference to the previous model).

Table 5. Fit statistics of the CFA models for the group comparison across SEN status

Model	χ^2wlsmv	df	p	CFI	RMSEA	SRMR	Δχ^2	Δdf	p
Baseline models
Overall	384.702	51	<.001	.980	.099	.074
Non-SEN	307.312	51	<.001	.982	.094	.072
SEN	154.425	51	<.001	.942	.142	.131
Group comparison
Configural invariance	461.738	102	<.001	.978	.103	.081
Metric invariance	503.543	111	<.001	.976	.103	.084	15.937	9	.068
Scalar invariance	509.683	132	<.001	.977	.092	.082	8.408	21	.993

n = 572 students without SEN, n = 101 students with SEN. WLSMV = weighted least-square means and variances estimator; CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual. Δχ² = Chi-square difference tests (with reference to the previous model).

3.4. Differential analyses

In order to explore group mean differences in terms of emotional well-being, social inclusion, and academic self-concept, independent sample t-tests were conducted. Results from an independent samples t-tests indicate that boys stated higher values on social inclusion (M_SI = 3.25, SD_SI = .62; t(651) = 2.98, p ≤ .01, d = .59) and academic self-concept (M_ASC = 2.95, SD_ASC = .56; t(651) = 2.98, p ≤ .05, d = .55) than girls (M_SI = 3.12, SD_SI = .57; M_ASC = 2.86, SD_ASC = .54). Male and female participants do not significantly differ regarding emotional well-being, t(651) = .127, p = n.s.

Concerning differences among students with and without SEN, the t-tests reveal that students with SEN indicated lower values for emotional well-being (M_SWB = 2.84, SD_SWB = .68; t(671) = 4.48, p ≤ .001, d = .62), social inclusion (M_SI = 2.82, SD_SI = .62; t(671) = 6.42, p ≤ .001, d = .59), and academic self-concept (M_ASC = 2.58, SD_ASC = .55; t(671) = 6.26, p ≤ .001, d = .54) than students without SEN (M_ASC = 3.15, SD_ASC = .61; M_SI = 3.23, SD_SI = .58; M_ASC = 2.94, SD_ASC = .54). The results of the final t-tests showed that students learning in inclusive classrooms stated lower values for academic self-concept (M_ASC = 2.83, SD_ASC = .58; t(670) = −2.05, p ≤ .05, d = .56) than students in non-inclusive classrooms (M_ASC = 2.92, SD_ASC = .54). There were no significant group differences regarding students’ emotional well-being and social inclusion.

4. Discussion

Even before COVID-19, especially in the context of inclusive education, students’ school well-being, their sense of social inclusion, and their academic self-concept have been in the focus of research. This might be because is it not easy to legitimate why students with SEN should attend separate classes or even schools—especially because this causes social isolation from their peers without SEN. However, on the other hand, in inclusive classes, students with SEN might feel overwhelmed and therefore might develop a rather negative academic self-concept.

As there is no short screening instrument available in Spanish to assess students’ emotional well-being, social inclusion or academic self-concept, the primary purpose of the current study was to examine the psychometric properties of the Spanish student version of the PIQ. Further, first insights in the status quo of Mexican students’ perceptions were given—focusing on comparisons between gender, students with and without SEN, and on classroom setting.

Confirmatory factorial analyses confirmed the three-factor structure as proposed by Zurbriggen et al. (2019). In addition, reliability was within an acceptable range for all three subscales. Important to highlight is that, as in other previous studies, the lowest reliability was found for the subscale of academic self-concept (e.g., Knickenberg et al., 2020; Zurbriggen et al., 2019). This could indicate that the construct is rather broadly assessed with the PIQ. Nevertheless, an internal consistency of .80 can still be considered good.

With regard to convergent validity, the findings can only be partially interpreted as evidence. All correlations were in the expected direction, indicating that more positive perceptions of inclusion are linked with a lesser intention to quit school and a higher level of general self-efficacy. For the subscale of students’ intention to quit, results indicate a high correlation with emotional well-being, whereas the correlation is rather moderate for the subscale of social inclusion. This finding seems to be somewhat different from that of the study of Knickenberg et al. (2022), in which a much higher correlation between emotional inclusion and students’ intentions to quit was found. Additionally, general self-efficacy was also found to be significantly associated with all three scales. However, the highest correlation was found between general self-efficacy and academic self-concept. These results are in line with previous research, which has identified strong associations between both constructs (Bong & Skaalvik, 2003; Ferla et al., 2009). Furthermore, measurement invariance analyses revealed (partially) strong measurement equivalence across gender, students with and without SEN status, and students attending inclusive and non-inclusive classrooms. These results are in line with other studies that have explored measurement invariance of the PIQ’s subscales, e.g., in the German version (Knickenberg et al., 2020) and the French version (Guillemot & Hessels, 2022) version.

Other important findings were that the means of the PIQ were significantly high: all were above the theoretical mean of the scale (2.5). Hence, these results indicate that the majority of the participating students in this sample have a rather positive perception of inclusion. Moreover, this confirms results from other studies which have explored students’ perceptions of inclusion using the PIQ, such as those done in Germany (Schwab & Alnahdi, 2020) and in France (Guillemot & Hessels, 2022). However, emphasis must not only be placed on the overall class mean, as high mean scores do not automatically imply that all students perceive high levels of inclusion. In detail, within the present sample, a total of 10% of the students would be considered to be at risk given their significantly low levels of emotional well-being (4%), social inclusion (3%), and academic self-efficacy (3%). Therefore, it is important to focus on those at-risk students in particular, as well as to implement prevention and intervention strategies to support them.

Differential analysis showed no significant differences between gender. On the other hand, with regards to social inclusion and academic self-concept, results revealed gender differences. Surprisingly, it appears that boys perceived significantly higher levels of social inclusion and academic self-concept than girls. Comparing the present study’s results with previous empirical evidence, they seem to be somewhat contradictory. For instance, in the study of Schwab et al. (2020), girls felt higher levels of emotional well-being than boys, whereas no gender effects were found for the subscales of social inclusion or academic self-concept. Other studies, such as those by Schneekloth and Andresen (2013) and Venetz et al. (2019), revealed that females reported higher levels of emotional well-being. Similarly, results from Ato et al. (2014) and Krull et al. (2018) indicated that girls felt higher levels of social inclusion compared to boys. Lastly, Venetz et al. (2019) reported that boys have higher levels of academic self-concept. Furthermore, the present study’s findings also seem to be contradictory to previous research in Mexico, which has shown for instance that females have higher academic self-concepts (Valdés-Cuervo et al., 2015), as well as other important socio-emotional variables (Blanco et al., 2012). However, it is important to highlight that all of these studies refer to pre-pandemic times. Hence, it is possible that these contradictory results could be due to the impact of the COVID-19 pandemic on the educational landscape and the changes and challenges brought to students, with which they were not prepared to deal and cope (Wyse et al., 2020). For instance, a study by Styck et al. (2021) showed that female secondary school students experienced, in general, higher levels of stress (e.g. concerning school workload) and social isolation compared to their male counterparts. Thus, a possible explanation for the present study’s results can be that girls had a greater need for physical social contact during the COVID-19 period. It is potentially possible that a post-pandemic effect is that girls struggled far more with the feeling of being socially included, with their emotional well-being, and with their academic competency beliefs.

For students with SEN, results of the current study draw a rather negative picture: they have lower levels of emotional well-being, social inclusion, and academic self-concept compared to their peers without SEN. This might indicate that the inclusive education practices in Mexican schools are still poor. In this context, several authors have argued that, despite attempts to promote inclusive education policies with the new educational model from 2017, inclusive education in Mexico still faces a long, challenging road ahead (I. G. García-Cedillo, 2018; Zhizhko, 2020). However, the results do not necessarily indicate that special education schools are the better option as, for instance, no group differences in emotional well-being or social inclusion were found between students with SEN attending inclusive classrooms and special schools. On the other hand, in line with other studies, findings indicate that students in inclusive classrooms have lower levels of academic self-concept (Bear et al., 2002; Marsh et al., 2006). This might be partially explained with the big-fish-little-pond effect (Marsh et al., 2008; Seaton et al., 2010). In this context, a study by Nusser and Wolter (2016) revealed that, in special schools, the average class competence did not influence academic self-concept when individual competence was controlled. For inclusive classrooms, however, the authors found evidence for the big-fish-little-pond effect. Lower academic self-concepts of students with SEN in inclusive settings (compared to those in special classes) might also be partially explained by the reference standards teachers use when evaluating academic achievement (see e.g. Lüdtke et al., 2005). For instance, if social reference standards or criterion reference standards are used, lower academic self-concepts for students with SEN compared to those without SEN might occur, as was found in this study.

Based on these outcomes, and given the fact that comparative research on inclusive education (in which Mexico is also considered) is still scarce (I. García-Cedillo et al., 2015), this paper makes an urgent call for educational researchers to conduct cross-cultural investigations on students’ experiences and perceptions of inclusive schooling. Such data would not only extend the existing knowledge base, but also contribute toward informing countries about intervention measures that can support their efforts in the development of inclusive education.

4.1. Limitations and future research

The present study underlines several limitations. First, the student sample in this study consists of mainly private students. Likewise, the sample also is composed of only lower secondary students from a city in northern Mexico. Considering the extreme economic, social, and educational disparities across the Mexican states, as well as between public and private schools (Lavín et al., 2020; Pozas et al., 2021), the results of this study must be interpreted very carefully. Investigating Mexican students’ perceptions of inclusion while taking into consideration the large sociodemographic differences that exist throughout the country is necessary for future research. A second and very important limitation to consider is that the data collection process for this study took place during the first trimester of 2022, right when schools in the Mexico reopened after the two-year COVID-19 school lockdowns. Thus, given that the effects of the pandemic have manifested in a widespread from (Hevia et al., 2022), the findings of this study must be considered with caution. Considering this context, this study thus calls for further studies to follow a longitudinal design in order to investigate the development of the PIQ dimensions as well as their longitudinal measurement invariance (Knickenberg et al., 2022). Lastly, psychometric properties of the Spanish PIQ version should also be explored in younger age groups, for instance, with Mexican primary school students.

5. Conclusion

Students’ emotional well-being, social inclusion, and academic self-concept are considered key factors as well as outcomes of successful inclusive schooling. Thus, it is of the utmost importance to listen to students’ own voices and their experiences of being taught in inclusive classrooms. To this end, the present study validated the Spanish PIQ student version in the Mexican context. The results revealed that the Spanish PIQ is a reliable, valid, economical instrument that holds high psychometric standards, and thus seems to be an appropriate tool to assess emotional well-being, social inclusion, and the academic self-concept of Mexican students.

Disclosure statement

No potential conflict of interest was reported by the authors.