The inclusive approach refers to the type of education that allows students with diverse needs to find their own place and to receive their education in regular schools and classrooms (Ebersold and Watkins, 2011). Countries from all around the world have decided that, by 2030, inclusive and quality education will be guaranteed for all, thus reducing the negative impact on the emotional well-being of the students (Tardi, 2012).

Traditionally, persons with disabilities (Groce et al., 2011; Schneider et al., 2016) have been subject to discrimination, school segregation (Kim et al., 2016), as well as deep-seated forms of exclusion (Tenenbaum et al., 2017). Among the main consequences of the phenomenon of exclusion is the erosion of the well-being and health of the adults and children who suffer from it (UNICEF, 2013), profoundly affecting the quality of life of individuals as well as the equity and cohesion of society as a whole (Bodvin et al., 2018). Currently, however, this paradigm is changing: there is a shift toward understanding diversity from a social perspective (Díaz Velázquez, 2010) of human rights or inclusion as leading to respectful responses toward people both with and without disabilities (Tardi, 2012).

As the theory of planned behavior affirms (Ajzen, 2005), the attitudes of individuals constitute the best predictor of their behavior. In this sense, one of the aspects that most frequently seems to positively affect inclusive education (IE) in the classroom is related to the attitudes and beliefs of teachers, whether they are professionally in-service teachers (de Boer et al., 2011; Chen, 2018; Glock et al., 2018) or future teachers (Sharma et al., 2014). In addition, attitudes are not innate but learned (Ajzen, 2005). Therefore, they should be taught to teachers because they, together with families, have the responsibility of educating and training children (Schneider and Coleman, 2018).

In this regard, the attitude of the different education professionals is a key factor in the development of inclusion (de Boer et al., 2011; Vaz et al., 2015; Gutentag et al., 2018). Attitudes toward inclusion have been studied in in-service teachers (Boyle et al., 2013; Collins, 2013) and in future teachers (teaching students) (Beacham and Rouse, 2012; Kuyini et al., 2018; Saéz-Gallego et al., 2019). In order to achieve inclusion, it is not enough that teachers (Messiou, 2017) or future teachers (Loreman et al., 2014; Sharma and Sokal, 2015) have good knowledge of the educational needs of students. They must also have certain attitudes, abilities, skills and strategies that, applied in their educational praxis, enable quality education for all (Loreman et al., 2014; Novo-Corti et al., 2015). Therefore, as Gale et al. (2017) argue, it is important to understand and monitor teachers’ attitudes toward diversity (Loreman et al., 2014). Thus, the type of response, positive or negative, that the teacher (Booth and Ainscow, 2015) or future teacher (Novo-Corti et al., 2015) has toward the object of attitude—in this case diversity—will depend on the combination of the three integrating components of the attitude: cognitive (positive or negative beliefs toward diversity), affective (evaluation and positive or negative valuation toward diversity), which translates into an acceptance (inclusion) or rejection (exclusion), and conative-behavioral responses (disposition toward diversity depending on the other two components) (Ajzen, 2005; Fishbein and Ajzen, 2011).

Thus, various investigations have attempted to analyze the attitudes of teachers (de Boer et al., 2011; Booth and Ainscow, 2015) and future teachers in the educational field (Novo-Corti et al., 2015). The studies explain that factors, such as gender, age, experience, training on attention to people with disabilities (Forlin and Chambers, 2011; Loreman et al., 2014; Schmidt and Vrhovnik, 2015) as well as types and levels of disability in children (Forlin and Chambers, 2011), could be influenced in the IE.

In the case of teachers, studies suggest that both work experience and contact positively affect teachers’ inclusive attitudes (Tsakiridou and Polyzopoulou, 2014), while some older and more experienced teachers appear to tend toward the opposite (Loreman et al., 2007). On the other hand, university students’ contact with people with disabilities is crucial for inclusion. In this sense, it improves and modifies beliefs and attitudes toward them (Mellado et al., 2017) at the same time as it enables the students to form comprehensive conceptions of them (Novo-Corti et al., 2015). Regarding gender, its influence on these types of attitudes seems to be controversial. While studies, such as those by Efilti and Arslan (2017), found no significant effect of gender on teachers’ attitudes toward inclusion, some studies have indicated that women have more positive attitudes toward educational inclusion (Tsakiridou and Polyzopoulou, 2014; Shatri, 2017), whereas other studies (de Boer et al., 2011) found more positive attitudes among male teachers. All this controversy may be due to the very measurement instruments used to measure attitudes toward inclusion—instruments that have rarely been adapted and validated for the specific context in which they are being used. In the case of future students, the influence of these factors seems clearer, with studies suggesting that women tend to have more positive attitudes toward students with disabilities than men (Novo-Corti et al., 2015). It seems that beliefs and attitudes toward diversity can facilitate or hinder inclusion in the educational context (Booth and Ainscow, 2015; Messiou et al., 2016); therefore, their adequate measurement is highly relevant.

Moreover, there are other socials factors (Irby and Clough, 2015), such as the culture (Armstrong et al., 2016), policy, and practices in schools; i.e., the implementation of inclusion in schools, resources and their distribution, the support of school administration and teachers themselves, and the organizational framework (teacher role or supports available in the classroom) (Galović et al., 2014; Schmidt and Vrhovnik, 2015). In addition, such attitudes may also be affected by the contact that individuals have had with people with special needs (Crowson and Brandes, 2014; Gonzalez et al., 2015). However, few studies have analyzed the differences in attitudes toward the inclusion of teachers and teacher training students as a function of these type of variables.

On the other hand, one of the variables that can influence the development of IE is whether one is a teacher training student or an in-service teacher. Some studies about future teachers (teacher training students) affirm that they feel that they do not receive adequate training on educational inclusion (Tsakiridou and Polyzopoulou, 2014), which leads to a decrease in confidence in the design of inclusive curricular activities and programs (de Boer et al., 2011), although it seems that students do have a positive attitude toward the concept of inclusion (Lambe et al., 2013). Other studies have found that younger teachers tend to adopt more favorable views toward the inclusion of students with special educational needs (SENs) (Forlin et al., 2011). However, we have not been able to observe studies that simultaneously analyze such inclusive attitudes in both groups in the same study, and we therefore cannot determine which group has more positive attitudes toward inclusion: in-service teachers or students.

During the last few years, there has been a proliferation of tools to measure attitudes toward the inclusion of students with educational needs in the Spanish context (Loreman et al., 2014; Abellán et al., 2017). Some examples of these include the Teacher Education for Inclusion Assessment Questionnaire (CEFI) (González-Gil et al., 2013), the Questionnaire on Attitudes Toward IE for Teachers in Training (Vélez, 2013), or the most recent Questionnaire for Future Teachers of Secondary Education on the Perceptions About Attention to Diversity (Ruiz and Palomino, 2015). Although these scales have been used in the Spanish context, they do not focus on specifically measuring attitudes, feelings, and concerns toward inclusive education; they also focus on other aspects, such as training or inclusion needs. Internationally, the probably most common scale in use is the Sentiments, Attitudes, and Concerns about Inclusive Education (SACIE) scale, created in 2007 (Loreman et al., 2007), which, revised 4 years later, was subsequently named SACIE-R (Forlin et al., 2011).

This scale consists of 15 items grouped into three factors. The first factor (Sentiments) evaluates the sentiments in the interaction or contact with students with disabilities or with students cataloged by the educational system with SENs; the second factor (Attitudes) focuses on the acceptance of these students; and the last factor (Concerns) evaluates concerns about IE education (Loreman et al., 2007). The scale has been used with working teachers and teachers in training (Loreman et al., 2007; Forlin et al., 2011; Cansiz and Cansiz, 2018; Murdaca et al., 2018), although neither of groups have been compared in the same study before.

Since its inception, SACIE-R has been adapted and validated to different cultural contexts, such as Italian (Murdaca et al., 2018), Turkish (Cansiz and Cansiz, 2018), and Portuguese (Santos and César, 2010). However, so far, it has never been adapted and validated for the Spanish context. Likewise, the existing studies have been carried out mostly with teachers, and the scale has thus never been analyzed simultaneously in both in-service teachers and future teachers (teachers in training).

Therefore, the objective of this study was twofold: on the one hand, the aim was first to adapt and validate the SACIE-R scale to Spanish in a sample of university students and teachers; on the other hand, the aim was to observe the attitudes toward inclusion of in-service teachers and teacher training students in Spain while at the same time analyzing the effects of gender, training, or the level of contact have those attitudes.

The participants were recruited using convenience sampling methods. The inclusion criteria for the case of the teachers were to be active at the time of the data collection, to exercise the profession at non-university levels, to not have any type of disability or illness preventing them from completing the questionnaire on their own, and to accept participation in the study. In the case of the students, the criteria were belonging to the last course of the degree in teaching or being a post-graduate student, not working as teachers at the time of the data collection, not having any type of disability or illness that would prevent them from completing the questionnaire on their own, and accepting participation in the study. The sample consisted of 647 subjects where 323 were university students studying teaching (49.92%). The students were in the last year of their degree, and most of them were studying for a degree in Primary Education (52.56%), followed by those studying for a double degree in Pre-school and Primary Education (31.73%) and those studying for a degree in Pre-school Education (14.74%). All of the students belonged to a medium socioeconomic level. The other 324 (50.08%) were teachers of the various educational stages (non-university levels): pre-elementary, primary, and secondary education. According to the educational stages in which they teach: 29% teach in pre-school education, 73.1% in primary education, 19.4% in secondary education, and 7.1% in high school education. The average age of the group of respondents was 32.42 years (SD = 12.32), with a minimum age of 18 and a maximum of 64 years. The percentage of women was 71.60%. The students’ subsample (323), consisted of individuals aged between 18 and 45 years (M = 21.20, SD = 6.37), with 77.2% being women (n = 249). In comparison, ages in the active teacher subsample (324) ranged between 35 and 58 years (M = 41.79, SD = 10.10), and the percentage of women was 66% (n = 214). The teachers’ teaching experience ranged from 0 to 39 years (M = 19.72, SD = 90.12). The number of teachers that claimed to have received training on diversity attention quite or very often during their career was 30.09%, while 20.38% said they had occasionally received this type of training, and 49.53% stated that they had never or rarely received it. Regarding the students, nearly half of those interviewed (47.95%) stated they had received training on diversity quite or very often during their studies, while 30.48% had occasionally received this type of training, and 21.56% said that they had never or rarely received it.

In terms of the frequency with which teachers had had their diversity training updated, 36.99% said they had often or very often been updated, 36.36% stated they had occasionally been updated, while 26.65% claimed they had never or rarely been updated. As for the students, 36% had often or very often been updated, 30% had occasionally been updated, while 34% had never or rarely been updated.

Regarding the frequency with which teachers had taught students with disabilities, 40.44% had often or very frequently taught these types of students, 32.29% indicated that they had occasionally taught them, and 27.27% indicated that they had never or rarely taught students with disabilities. In the case of students, 20.09% had often or very frequently taught this type of pupils, and the same percentage indicated that they had occasionally taught them, while 59.82% stated that they had never or rarely taught pupils with disabilities.

Finally, about half of the students (50.32%) claimed to have had peers with SENs during their schooling, while 49.68% reported not having had peers with SENs. Furthermore, 48.25% indicated that they had a family member or acquaintance with some type of disability, while 51.75% stated that they did not.

SACIE-R: adaptation of the SACIE-R (Forlin et al., 2011) scale to the Spanish context was carried out by the research team. The scale assesses three dimensions of IE: Sentiments, Attitudes, and Concerns. This questionnaire consists of 24 items to be answered with a five-level Likert item (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree). The scale showed an acceptable internal consistency (α > 0.74) within previous studies (Forlin et al., 2011).

An ad hoc questionnaire was carried out to collect the variables of gender, age, specific training in the face of diversity, frequency of updating educational training, any experience or contact with students with diversity, and the presence of a relationship with peers or family members with diversity.

The SACIE-R scale was adapted according to the guidelines of the International Testing Commission (ITC) (Muñiz et al., 2013). For the language contextual adaptation of the instrument, two native linguists translated and back-translated (English–Spanish–English) the items of the scale; this process was complemented with the review of the resulting items by experts in IE, which ensured the adequacy of the statements for the Spanish context (Table 1). Next, the participants were selected: both the university students and the teachers of the different educational stages from the Valencian Community. The collection of both samples was carried out over 3 months. In the case of the student sample, the information was collected during school hours and during the time allowed by the university professors (approximately 20 min) who voluntarily facilitated the interruption of the class session in order to explain and proceed to collect the sample voluntarily from the student teachers who placed their questionnaire in a blank envelope. This sample was of convenience to students of teaching at the Catholic University of Valencia. The student response rate was 100%. In the case of the sample of teachers, 10 centers that voluntarily wanted to participate were provided with the total number of questionnaires requested and some hermetically sealed boxes in which the questionnaire could be deposited, thus guaranteeing anonymity among the participating subjects. The response rate from teachers was 45%. Due to the nature of the study, participants answered voluntarily, and the questionnaires were anonymous. The study was approved by the bioethics committee of the Catholic University of Valencia (PRUCV/2015/660). All participants received detailed information about the aims and procedures and were informed of confidentiality.

The statistical analysis was carried out with the programs SPSS Statistics (version 24), EQS (version 6.3) (Satorra and Bentler, 2001; Crawford and Henry, 2003) and FACTOR (Lorenzo-Seva and Ferrando, 2006). First, the descriptive statistics for each item were calculated. Second, the reliability and validity of the scale for the whole sample and for each subsample was analyzed. Third, Pearson correlations were performed to determine the relationship between the different dimensions of the questionnaire. The instrument’s invariance was calculated, and then descriptives comparing both samples were calculated. Finally, the influence of gender, training, and contact played on the attitude toward inclusion was evaluated by means of t-test and Pearson correlations.

In order to examine the reliability of the scale, its internal consistency was calculated using Cronbach’s alpha. However, since this index does not consider the influence on the other reliability construction, both the composite reliability coefficient (CRC) and the average variance extracted (AVE) (Fornell and Larcker, 1981) were calculated. The minimum CRC value was 0.70, which was considered adequate, as (Nunally, 1978) values higher than 0.50 are recommended for the AVE (Bagozzi and Yi, 1988), although others studies have considered that values of AVE higher than 0.40 are adequate too (Vila et al., 2000).

The questionnaire showed acceptable scores for overall reliability for the whole sample and for the different subsamples (in-service teachers and university students). The values of the different dimensions of the SACIE-R ranged between α = 0.64–0.83, CRC = 0.63–0.84 and AVE = 0.37–0.52 for the different dimensions, fairly acceptable values as indicated by the literature (Fornell and Larcker, 1981; Vila et al., 2000). The analysis of internal consistency suggested the elimination of item 1 in the Sentiments dimension and item 10 in the concerns dimension because the factorials loads were below 0.40 (Table 2).

The internal validity was checked by means of several exploratory factor analyses (EFA) and confirmatory factor analyses (CFA). First, the adequacy of the data for these analyses was tested using the Kaiser Meyer Olkin index (KMO) and the Bartlett sphericity test. In all cases both were adequate: for the whole sample (KMO = 0.77, χ² = 2523, df = 91, p ≤ 0.001) for the students (KMO = 0.74, χ² = 1324.1, df = 91, p ≤ 0.001) and, finally, for the teachers (KMO = 0.78, χ² = 1418.8, df = 91, p ≤ 0.001).

Next, an EFA was performed using the unweighted least squares (ULSs) method. Likewise, the normalized direct oblimin rotation was used. However, the analysis was repeated using the normalized Varimax rotation since the correlations between the factors proposed by the EFA, with oblimin rotation, were quite low (<0.30) and did not allow for the relationship between the factors proposed by the EFA to be verified (Lloret et al., 2014).

First by performing the EFA without any restriction on the number of factors, the EFA recommended the grouping of items into four common factors. However, this solution did not offer a good theoretical interpretation so it was decided to check the adjustment of the factorial structure fixed in three dimensions (attitudes, sentiments, and concerns) derived from the proposal of the original authors of SACIE-R (Forlin et al., 2011). After the application of the EFA fixed to three factors in the whole sample, an indicator (item 1) with a saturation lower than 0.40 was eliminated. The factorial solution coincided with the whole of the sample and the subsample of students; however, in the teacher sample, some indicators saturated in a different factor and two indicators presented a factorial load lower than 0.40 (items 10 and 15). For the sample, the variance explained by the three factors was 58.55%; in the students it was 59.75% and in teachers 54.52%.

Once the EFAs had been performed, several CFAs were also carried out with the aim of, on the one hand, verifying the factorial structure extracted by the EFA (14 items grouped into three dimensions) and, on the other hand, contrasting the adjustment of the original proposal considering the 15 items grouped into three dimensions (Sentiments, Attitudes, and Concerns). The indicators of adjustment of the different proposed models are shown in Table 3. The estimation was made using Maximum likelihood (ML) estimation with Satorra-Bentler’s robust correction (Satorra and Bentler, 2001) in order to correct the absence of multivariate normality. A factorial solution with a good fit was obtained after eliminating four items from two dimensions—items eight and 10 from Concerns and items 1, 13, and 14 from Sentiments—because their factorial loads were lower than 0.40.

Based on the significance of the χ² statistic (p < 0.05), an adequate model fit could not be ensured in any of the cases. However, given that this analysis is very sensitive to the sample size, other goodness-of-fit indices were applied: the χ²/df, showing adequate fit with values lower than 5, the Non-Normed Fit Index (NNFI), the Comparative Fit Index (CFI), the Incremental and the McDonald’s Fit Indices (IFI and MFI, respectively), values above 0.90 being indicators of good fit (Wallin and Ahlstrom, 2006), and the Root Mean-Square Error of Approximation (RMSEA) in which adequate fit scores are equal to or below 0.08. Table 3 presents a summary showing these indicators for each of the CFA.

Furthermore, in order to increase the empirical evidence for the construct validity, on the one hand, we tested the convergent validity of the scale on the basis of the results obtained in the CFA, and the items that make up the SACIE-R were significantly and strongly correlated with the latent variable they assumed to measure, oscillating from 8.49 to 12.31 (t ≥ 1.96) for the whole sample, from 5.70 to 9.00 for the students, and from 5.34 to 8.63 for the teachers. In all cases, the t-values were above 3.29 and failed to improve when new loads were included. On the other hand, discriminant validity was evaluated by means of the AVE (Fornell and Larcker, 1981; Netemeyer et al., 1990). To determine the existence of discriminant validity, we found that all correlations between the various factors were less than 0.85, and the AVE square root was higher than the correlation among the pairs of factors or dimensions considered (Fornell and Larcker, 1981; Netemeyer et al., 1990; Vila et al., 2000). The results, displayed in Table 4, suggest an acceptable discriminant validity.

The analysis of the descriptives, the mean differences, and the correlations were carried out with the final model obtained from the previous AFCs.

Based on the results obtained (Table 5), it appeared that both in-service teachers and teacher training students had positive Attitudes toward inclusive education. Specifically, high scores were observed in the attitudes dimension of the SACIE-R (whole sample: M = 4.02; SD = 0.82; teacher training students; M = 3.95; SD = 0.89; and in-service teachers M = 4.09; SD = 0.73), while they present medium-low levels of Sentiments (whole sample: M = 2.81; SD = 0.98; teacher training students; M = 2.69; SD = 1.06; and in-service teachers M = 2.94; SD = 0.88), and Concerns (whole sample: M = 2.67; SD = 0.89; teacher training students; M = 2.42; SD = 0.83; and in-service teachers M = 2.92; SD = 0.88).

Comparing in-service teacher and teacher training students, the highest scores for all three dimensions were found in the in-service teacher sample in Attitudes [t(360) = −2.18; p = 0.03], although the effect size was small (d = 0.17); Concerns [t(636) = −3.36; p ≤ 0.001], where the effect size was (d = 0.15); and Sentiments [t(630) = −2.36; p = 0.02], where the effect size was also small (d = 0.19).

Regarding gender, statistically significant differences (p ≤ 0.05) were found for the Attitudes dimension in the student sample but not in the others [t(298) = −2.63; p ≤ 0.01], although the effect size was small (d = 0.15). In general, women seemed to show more positive Attitudes toward IE than men (M_women=4.02, SD_women=0.88, M_men = 3.70, SD_men=0.91).

Regarding differences according to contact during schooling with peers with SENs or having friends or family with some type of disability, no statistically significant differences were found (p > 0.05). Nevertheless, slightly higher scores were observed in Attitudes and Concerns for those who did have peers with SEN or those with friends or family with some type of disability.

Likewise, considering the relationship of age with attitudes toward inclusion, a low but significant positive correlation was observed (p ≤ 0.01) with the dimensions of the SACIE-R (Attitudes: r = 0.11; Sentiments r = 0.12; Concerns r = 0.22). On the other hand, when considering experience or training related to SEN, low significant correlations were observed (p ≤ 0.01) between the frequency of receiving training during the career on attention to diversity, frequency of updating training on attention to diversity, and frequency of teaching to students with disabilities with the dimensions of the SACIE-R. In all cases there was a positive relationship with the Attitudes dimension and negative relationships with Sentiments and Concerns, apart from the frequency with which they had taught students with disabilities and in the Concerns dimension whose correlation was not significant. When considering each sample separately, similar correlations were observed, with slightly higher levels in the case of teachers than students and with the exception of the correlations of the dimensions of the SACIE-R, with respect to the frequency of receiving training on attention to diversity during the career/studies; here, low values were also observed, but they were slightly higher in the case of the students (Table 6).