Participants

A total of 237 teachers participated in the study who were selected through intentional, non-probabilistic sampling. The teachers mainly taught in the north of Spain, specifically Asturias and Cantabria. Just over three-quarters (185; 78.1%) were women. The mean age was 39.22 years old (SD = 10.59).

Half of the teachers (119, 50.2%) taught in secondary-school (ages 11-16), while 118 (49.8%) taught primary education. A fifth of the primary-school teachers (23; 19.4%) had little or no experience of students with SLD-R, 37 (31.3%) had moderate experience, and 58 (49.1%) had a lot of experience of these students. Just over a third of the secondary-school teachers (43; 36.1%) had little to no experience with SLD-R students, 43 (36.1%) reported moderate levels of experience, and 33 (27.7%) reported a lot of experience. The differences in distribution of experience levels were statistically significant in primary-school teachers χ²(2) = 15.780, p < .001; but not in secondary-school teachers χ²(2) = 1.681, p = .432.

Measuring instruments

The survey was split into three sections. The first covered personal data, including sex, age, educational stage taught (primary or secondary), educational specialties, type of school (state-funded, private, or concertado [partly state funded, but educationally independent schools]), and experience with SLD-R students. This experience was recorded on a scale from 1 to 5 (1 = no experience; 2 = little experience; 3 = moderate experience; 4 = quite a lot of experience; 5 = a lot of experience). To make subsequent analysis simpler, we grouped the experience with SLD-R students variable into three: none-little experience (n = 66), moderate experience (n = 80), and quite a lot-a lot of experience (n = 91).

The second section included the Knowledge of SLDs Scale (^{Guzmán et al., 2015}), made up of 38 items in three dimensions (Appendix 1): General Information (11 items), Symptoms/Diagnosis (14 items), and Intervention (12 items). One of those original items was not included (meaning there were 37 in total) as it dealt with learning difficulties with writing, which the present study did not address. The scale was used to produce results in terms of teachers' real knowledge, incorrect ideas, and gaps in knowledge related to SLD-R for each of the three dimensions (General Information. Symptoms/Diagnosis, and Intervention) and overall. Each item has three possible answers: true, false, or don't know. Analysis of the psychometric properties of the scale gave a Cronbach α of .88 and McDonald's ω of .88 (^{McDonald, 1999}).

The third and final section included the teachers' causal attributions questionnaire from ^{Navarro and Rueda (2011}). This questionnaire aims to examine the causal attributions exhibited by the teachers for the success or failure of their students with and without SLD-R in reading tasks (Appendix 2). It has four items. Two of the items ask teachers' about the key factors they attribute success or failure to in reading tasks for their students with SLD-R. The other two items ask about students without SLD-R. Each item has six possible responses and the option to provide an open, individual answer. Some of the alternatives are associated with attributions linked to internal factors (ability, interest, motivation), while others are associated with attributions linked to external factors (teacher or family support or the difficulty of the tasks). Participants can only choose one answer, which they feel most strongly identified with. This produces four attribution variables: success of students without SLD-R, failure of students without SLD-R, success of students with SLD-R, and failure of students with SLD-R. The variables were scored as 1 if the teachers' causal attributions dealt with internal factors, 2 if they were external factors, and 3 if the responses indicated a combination of the two. Analysis of the questionnaire's psychometric properties gave a Cronbach α of .85 and McDonald's ω of .85 (^{McDonald, 1999}).

Procedure

Data was collected online. The data-collection process began on the 16th of October 2020 and ran until the 30th of November the same year. We contacted 321 schools in Asturias and 173 in Cantabria by email to distribute the survey to the teaching staff. Email addresses were obtained from the websites of the respective regional government education departments. The survey was also spread through social networks, using non-probabilistic snowball sampling (^{Elfil & Negida, 2017}).

Before completing the survey, participants were informed that their responses would be anonymous and asked for their agreement to use the results in the study, ensuring informed consent from each participant.

Design and data analysis

The study followed a survey strategy, described above. This meant that the study was a transversal, comparative study, analysing a set of variables, which in this case related to teachers' knowledge of SLD-R and the type of causal attributions made by the teachers about their students with SLD-R.

Data analysis was done using SPSS 27.0, with statistical significance set at = 0.05. First, a correlation matrix was calculated and the distribution of the variables was examined. In pursuit of the first objective and to analyse the differences in knowledge about SLD-R based on educational level (primary or secondary; independent variable), various univariate analyses of variance (ANOVA) were performed with overall knowledge of SLD-R and knowledge in each of the three dimensions (General Information, Symptoms/Diagnosis, and Intervention) taken as dependent variables.

For the second objective, ANOVA were again carried out to determine the differences in knowledge about SLD-R both generally and in each of the three dimensions (dependent variables) according to teachers' experience with SLD-R students (none-little, moderate, a lot; independent variable). Post hoc analyses (using Bonferroni correction) were performed for the dimensions where statistically significant differences were found.

Effect sizes were estimated using partial Eta squared (η_p ²), with ^{Cohen's (1988}) classic criteria, by which there is a small effect when η_p ² < 0.01, a moderate effect when η_p ² ≥ 0.059 and a large effect when η_p ² ≥ 0.138.

We also performed a descriptive analysis of the different causal attributions exhibited by teachers' about the success and failure of their students with and without SLD-R based on educational level. Multidimensional contingency tables were produced in order to produce the total frequencies of attributions of success or failure to internal or external factors or a combination of the two for students with and without SLD-R.

Preliminary Analysis

Table 1 shows the descriptive statistics for the variables, along with the Spearman correlation matrix. The values for asymmetry and kurtosis indicate that the study variables complied with the criteria for normality (values below three for asymmetry and below ten for kurtosis; ^{Kline, 2011}). The results of the correlation analysis indicate that real knowledge of the General Information, Symptoms/Diagnosis, and Intervention dimensions, and experience of students with SLD-R, were positively and significantly correlated, with greater experience being associated with greater knowledge.

Teachers' knowledge of SLD-R according to educational profile

Teachers' knowledge about SLD-R considered real knowledge, incorrect ideas, and gaps in knowledge via the Knowledge of SLD Scale.

There were statistically significant differences between primary- and secondary-school teachers in real knowledge about SLD-R, F(1, 235) = 37.07, p < .001, η_p ² = 0.136, with a large effect size. Primary-school teachers exhibited greater real knowledge (M = 25.18, SD = 4.99) than secondary-school teachers (M = 20.98, SD = 5.6). The differences between primary- and secondary-school teachers' scores in the three dimensions were also statistically significant, with a small effect size for General Information, moderate for Intervention, and large for Symptoms/Diagnosis (see Table 2), primary-school teachers scoring higher.

There were also statistically significant differences in incorrect ideas based on educational level, F(1, 235) = 15.41, p < .001, η_p ² = 0.062, with a moderate effect size. Primary-school teachers demonstrated more incorrect ideas (M = 4.75, SD = 3.15) than secondary-school teachers (M = 3.28, SD = 2.58). The differences between the two were also statistically significant for the three dimensions, with a small effect size for General Information and Symptoms/Diagnosis, and a moderate effect size for Intervention (see Table 2), primary-school teachers demonstrating more incorrect ideas.

Lastly, there were also statistically significant differences between primary- and secondary-school teachers in gaps in knowledge, F(1, 235) = 53.69, p < .001, η_p ² = 0.186, with a large effect size. In this case, secondary-school teachers exhibited greater gaps in knowledge (M = 10.73, SD = 6.48) than primary-school teachers (M = 5.08, SD = 5.32). The differences between the two in gaps in knowledge in the three dimensions were also significant, with a moderate effect size for General Information and a large effect size for Symptoms/Diagnosis and for Intervention (see Table 2). Secondary-school teachers exhibited more gaps in their knowledge in the three dimensions.

In summary, in the Knowledge of SLD Scale, secondary-school teachers exhibited lower levels of real knowledge and more gaps in knowledge than the primary-school teachers. Although the primary-school teachers demonstrated greater real knowledge, they also demonstrated more incorrect ideas.

Teachers' real knowledge of SLD-R according to experience with SLD-R students

The next step was to examine whether experience of students with SLD-R was related to knowledge about it. The results of the ANOVA indicated statistically significant differences in levels of knowledge about SLD-R by the amount of experience teachers had with it (see Table 3), with a moderate effect size. The results showed that the differences between the different levels of experience (little-none, moderate, a lot) were evident in the three dimensions of the scale, with a small effect in the General Information dimension, a large effect in the Symptoms/Diagnosis dimension, and a large effect in the Intervention dimension (see Table 3).

Via multiple comparison tests, it was possible to determine that the differences in the General Information dimension were between the little-none and moderate experience groups (p = .014). In the Symptoms/Diagnosis dimension, the differences were between the three groups - in other words, between little-no experience and moderate experience (p = .001), between little-no experience and a lot of experience (p < .001), and between moderate experience and a lot of experience (p = .047). In the Intervention dimension, there were statistically significant differences between the little-no experience and the moderate experience groups (p = .024) and between the little-no experience and the a lot of experience groups (p < .001). Teachers who reported having a lot of experience with these types of students scored higher in real knowledge both overall and in the three scale dimensions.

Teachers' causal attributions of the success and failure of their students with and without SLD-R in reading tasks

The final step was to perform a descriptive analysis, with multidimensional contingency tables, to determine what the teachers (primary and secondary) attributed the successes and failures of their students (with and without SLD-R) to in reading tasks. As Table 4 shows, both primary- and secondary-school teachers mainly attributed successes and failures of students without SLD-R and successes of students with SLD-R to intrinsic factors. In contrast, teachers attributed failure of students with SLD-R to both intrinsic and extrinsic factors. Primary-school teachers felt that failures of students with SLD-R was mainly due to extrinsic factors, whereas secondary-school teachers attributed it to intrinsic factors.

In summary, the most common teacher attributions were related to intrinsic factors, however, for students with SLD-R, there were more attributions of failure to extrinsic factors than for students without SLD-R (see Table 4).

Practical implications

The most immediate practical implication of our results concerns the need for education and training programs for teachers of all educational stages, both in their initial training and throughout their careers. This training should emphasize improving teachers' skills for identifying and dealing with students who have specific educational support needs, such as those with SLD-R. Tutorial support is implicit and inherent to teaching, and is the first level of educational guidance (^{González-Benito, 2018}). Good teaching and tutoring practices require broad pedagogical mastery allowing the adoption of the best methods and pathways to encourage the development of students with difficulties (^{Martín & Villanueva, 2018}). Greater awareness and knowledge of neurodevelopmental disorders pays dividends in prevention, detection, and intervention for students that present them (^{Cueli et al., 2022}; ^{Fuchs & Fuchs, 2006}), as well as being positively related to teachers' inclusive attitudes in the classroom (^{Thomas & Uthaman, 2019}).

At the same time, new teachers' initial training must consider the importance of establishing an approach to students with SLD-R to provide them with an educational experience guided by professionals who are experts in working with such students.

Limitations

Lastly, it is important to note some of the limitations of our study. One of the main limitations was the small proportion of respondents compared to the number of schools we contacted and had access to. The teachers who did respond would probably be those who were more motivated or interested in SLD-R. In addition, the geographical reach of the study was limited to the north of Spain. It is also worth noting the possibility of response bias. As noted above, the primary-school teachers exhibited greater knowledge than the secondary-school teachers, but also exhibited more incorrect ideas. These results may be explained by a response pattern with a liberal bent from the primary teachers, which would encourage them to give an answer rather than omit an item, leading to greater levels of knowledge and incorrect ideas.

In any case, and despite those limitations, the study underscores the need to increase teacher training related to SLD-R, given that greater awareness may benefit prevention, detection, and intervention for these students and promote inclusive attitudes in the classroom.