Assessing the Psychometric Properties of the Practice and Product Inventory of Supporting Students with ASD (PPI-SSA): A Concise Assessment Tool for Teachers in Inclusive Classrooms

Abstract

Globally prevalent, Autism Spectrum Disorders (ASDs) have negative consequences on students’ social, intellectual, emotional, and post-school transition results. While assessment tools exist, the majority of them were developed from the perspective of clinical psychology and/or healthcare. Some are lengthy and commercialized, and some have weak factorial validity. A feasible and practical assessment method, from the perspective of educational psychology and assessment, is required to better accurately and effectively assess the social and academic performance of ASD students in inclusive settings. With an emphasis on teachers’ practice and output in helping students with ASD in inclusive education, this study evaluated the Practice and Product Inventory of Supporting Students with ASD (PPI-SSA) psychometric qualities. PPI-SSA was designed to be practical, concise, and especially suited for quantitative research on ASD. The inventory was administered to 411 teachers in Hong Kong, and exploratory and confirmatory factor analyses confirmed the PPI-SSA’s reliability and validity. A follow-up path analysis examined the concurrent validity of the PPI-SSA, revealing significant pathways between teachers’ intentions to implement inclusive education and each of the respective dimensions within the PPI-SSA. Measurement invariance results showed that the PPI-SSA demonstrated configural, metric, scalar, and residual invariance across primary and secondary teachers. The PPI-SSA, as a non-commercial instrument, has meaningful implications. It can facilitate the efficient collection of more empirical data on the social and academic achievements of students with ASD, informing evidence-based improvement of inclusive practices. It can assist in teachers’ practice to identify the diverse social and academic needs of students with ASD to align with their corresponding adjustments of teaching strategies.

1. Introduction

Autism Spectrum Disorders (hereafter abbreviated as ASDs) are prevalent worldwide. In the United States, the prevalence of ASD is 1 in 36 children [1], while in Mainland China, it is approximately 1%. In Hong Kong, ASD is the third most common type of special educational need, with over 13,000 autistic students studying in mainstream primary and secondary schools [2]. Students with ASD are characterized by persistent deficits in social interaction and communication, repetitive patterns of behavior, and restricted interests or activities [3]. Despite their potential for exceptional talents and strengths, such as in math and reading skills [4], they are more likely to have co-occurring conditions, including intellectual disabilities [5,6], ADHD [7,8], and epilepsy [9,10], compared to typically developing peers. Students with ASD are also at risk of poor outcomes, including poor mental health conditions [11] and underachievement in schools [12]. Previous studies have reported higher rates of academic underachievement among students with ASD than typically developing peers [13] and a lower likelihood of maintaining competitive employment or postsecondary education [14]. Therefore, it is important to focus on examining their academic and social performance in inclusive settings.

While there are some instruments available for assessing the educational experiences and outcomes of students with Autism Spectrum Disorder (ASD) in inclusive settings, such as social skills, adjustment, emotion recognition skills, classroom engagement, and communication skills, the majority of these instruments, including the Strengths and Difficulties Questionnaires (SDQ) [15], Social Responsiveness Scale [16], and Social Communication Questionnaire [17], as well as standardized academic tests such as the WJ-III test of achievement [18], Assessment of Basic Language and Learning Skills (ABLLS) [19], Child Behavior Checklist (CBCL) [20], and Social Skills Improvement System Rating Scales (SSIS-RS) [21], tend to primarily function as diagnostic screeners with limited connections to assessing social and academic aspects. The ABLLS, CBCL, and SSIS-RS stand out for containing both social and academic indicators. However, they have several limitations. Firstly, they are relatively lengthy (with over 60 items) and may pose difficulties for teachers who already have high levels of workload. Secondly, their fees can be a challenge, especially in low-resource settings such as schools in developing countries. Thirdly, previous studies have shown poor factorial validity of the CBCL among children with ASD aged 6 to 18 years [22] or with concurrent intellectual disabilities [23] and limited information regarding the psychometric properties of the ABLLS and SSIS-RS in the ASD context. We also summarized key features of these measures and their limitations to assess ASD students’ social and academic achievements from the teachers’ perspective.

Based on our review of existing instruments, the objective of this study is to validate a specialized assessment tool to evaluate the effectiveness of teaching practices and educational outcomes for students with Autism Spectrum Disorders (ASDs) in inclusive classroom settings. Our instrument, named the Practices and Products Inventory to Support Students with ASD (PPI-SSA), aims to provide a practical, reliable, and quantitative method for assessing both the practices employed by teachers and the academic and social outcomes experienced by students with ASD. Diverging from existing assessment tools, which are predominantly rooted in clinical psychology or healthcare and often lengthy or commercialized, the PPI-SSA grounds on the fields of educational psychology and assessment and with a perspective of teachers’ assessment. By incorporating these measures, our goal is to enable a more comprehensive and evidence-based understanding of how to effectively support students with ASD in inclusive educational environments. The PPI-SSA offers meaningful implications for educational practice by facilitating the efficient collection of empirical data, thus informing the tailored improvement of inclusive teaching strategies to meet the diverse needs of students with ASD in inclusive educational settings.

1.1. Literature Review

1.1.1. Achievements of ASD Students: Social/Academic Aspects

In recent years, there has been a growing interest in understanding the relationship between social skills and academic achievement in children with Autism Spectrum Disorder (ASD). Among numerous studies on ASD students, we found several relevant studies on this topic that have contributed to this area of research, shedding light on the factors that influence the social and academic outcomes of children with ASD. These nine studies generated from our review of high-quality publications explored a range of aspects of the relationship between social skills and academic achievement in children with Autism Spectrum Disorder (ASD).

Several cross-sectional studies have explored the relationship between academic achievement and intellectual ability among children with ASD. For example, Estes et al. (2011) found significant discrepancies between actual achievement levels and levels predicted by intellectual ability among higher-functioning children with ASD [24]. Rosen et al. (2019) conducted a cross-sectional study to examine the association between school services and parent, teacher, and clinician ratings of ASD and co-occurring psychiatric symptom severity and intellectual functioning [25]. Focusing on the stakeholder of teachers, Milgramm et al. (2021) investigated whether teachers could discern differences among students with ASD concerning their social skills, problem behaviors, and academic competence. In their study, 61 children ranging from kindergarten to fifth grade underwent extensive psychological evaluations for Autism Spectrum Disorder (ASD) at a university-affiliated center. The results of the study indicated that, on average, children who were referred for ASD evaluation displayed lower levels of social skills and exhibited more pronounced problem behaviors compared to their peers. Additionally, the study revealed that cognitive ability, social skills, and problem behaviors each independently contributed to predicting academic competence. This suggests that students perceived as having impaired social skills and an elevated incidence of problem behaviors tended to be viewed as less academically competent, regardless of their ASD diagnosis [26]. However, Milgramm et al. (2021) did not consider teachers’ instructional practices in inclusive education in examining their predictive effects on these social and academic outcomes [26].

Some observational studies have evaluated tools designed to measure behaviors associated with positive educational outcomes in students with ASD. For instance, Sparapani et al. (2016) evaluated the Classroom Measure of Active Engagement (CMAE) to measure active engagement in students with ASD and found it to be a reliable and valid tool [27]. For collecting process data on ASD students’ learning, Sparapani et al.’s (2016) CMAE would be a useful assessment tool. Consistently, Leifler et al. (2022) investigated the feasibility of social skills group training (SSGT) in naturalistic settings and found its feasibility and validity [28]. There are also some randomized controlled trials that have investigated the efficacy of specific interventions for improving social and academic outcomes for children with ASD. For example, Lopata et al. (2018) evaluated the efficacy of a cognitive-behavioral school-based intervention (schoolMAX) in ratings of ASD symptoms and social communication skills in children who received the intervention and found significant improvements in emotion recognition skills [29]. From the assessment perspective, SchoolMAX could be considered a helpful tool for assessing cognitive and behavioral interventions for promoting ASD students’ emotional domains. Recently, considering the social skills of ASD students, Temkin et al. (2022) evaluated the effectiveness of the Secret Agent Society (SAS) in improving social-emotional functioning for youth with ASD, ADHD, and/or Anxiety and found SAS to be effective in improving social skills and emotion regulation [30]. In search of predictors of school success for ASD students, researchers also used qualitative and quasi-experimental studies to explore the perspectives of educational professionals. Noteworthy, Carter et al.’s (2019) study considered teachers as important assessors of students’ academic performance, and they found that teacher-rated academic skills predicted child social skills, engagement, and adjustment, while child problem behavior negatively predicted parent and teacher ratings of placement success [31]. Adaptive behavior predicted teacher and principal ratings of placement success. Taking account of a whole-school approach to support ASD students, Van Der Steen et al. (2020) found that educational professionals need collaboration within the school, practical teaching suggestions, and confidence to teach students with ASD, and the ability to enhance students’ social and communication skills to provide optimal support to ASD students [32].

The studies reviewed above revealed the complex and multifaceted challenges associated with educating children with ASD. They provide valuable insights into the academic and social achievements of students with ASD as they highlight the complexity of assessing social and academic achievements in students with ASD. Despite contributions, none of these studies provided a measure of assessing the social and academic domains of ASD students in inclusive education (we summarized their designs in

Table A4. The complex and multifaceted challenges associated with educating children with ASD: Snapshot.

Studies	Research Design	Snapshot of Key Findings
Estes et al. (2011) [24]	Cross-sectional	Significant discrepancies were found between actual achievement levels and levels predicted by intellectual ability among the majority of higher-functioning children with ASD.
Carter et al. (2019) [31]	Quasi-experimental	Teacher-rated academic skills predicted child social skills, engagement, and adjustment, while child problem behavior negatively predicted parent and teacher ratings of placement success. Adaptive behavior predicted teacher and principal ratings of placement success.
Leifler et al. (2022) [28]	Mixed-methods	Social skills group training was found to be largely feasible and socially valid, and broader implementation of social skills group training in school settings appeared meaningful.
Lopata et al. (2018) [29]	Randomized controlled trial	Children with ASD who received a cognitive-behavioral school-based intervention (schoolMAX) exhibited significantly greater improvements in emotion recognition skills, ratings of ASD symptoms, and social communication skills relative to children in schools with treatment as usual.
Milgramm et al. (2021) [26]	Cross-sectional	Children with and without ASD did not differ significantly in terms of overall academic competence, social skills, or problem behaviors as rated by their teachers.
Rosen et al. (2019) [25]	Cross-sectional	School services were associated with ASD severity and IQ, but no significant associations were found between internalizing/externalizing symptoms and school service presence/frequency.
Sparapani et al. (2016) [27]	Observational	The Classroom Measure of Active Engagement (CMAE) was found to be a reliable and valid tool for measuring behaviors associated with positive educational outcomes in students with ASD.
Temkin et al. (2022) [30]	Randomized controlled trial	The Secret Agent Society was superior to treatment as usual in improving social skills and emotion regulation for youth aged 8–12 with ADHD, ASD, and/or anxiety.
Van Der Steen et al. (2020) [32]	Qualitative	Educational professionals need collaboration within the school, practical teaching suggestions, confidence to teach students with ASD, and the ability to enhance students’ social and communication skills to provide optimal support to ASD students.

Note: These relevant studies were screened from the database of the WoS.

). These studies also highlight the importance of collaboration between various stakeholders, the need for specialized interventions tailored to the unique needs of children with ASD, and the importance of reliable and valid tools for measuring academic and social outcomes. These findings could inform the development of effective educational interventions as well as the validation of measures to assess the social and academic achievements of ASD students in inclusive education.

1.1.2. Measurement of ASD Students’ Achievements in Inclusive Education

In the realm of assessing Autism Spectrum Disorders (ASDs), Constantino and Gruber (2005) developed the SRS to measure social communication, interaction, and repetitive behaviors in individuals with ASD, which was found to be reliable and valid [16]. Similarly, Goodman (1997) created the SDQ, a reliable and valid measure of emotional and behavioral problems in various populations, including children with ASD [15]. Gresham et al. (2011) conducted a comparison between the SSRS and SSIS measures, discovering that they were highly correlated despite some differences in content and psychometric properties [21]. Medeiros et al. (2017) investigated the factor structure of the CBCL and found that it was a valid measure of emotional and behavioral problems in children with ASD, though with a different factor structure from the general population. Lastly, Dovgan et al. (2019) examined the measurement invariance of the CBCL and concluded that it is an effective instrument for assessing emotional and behavioral problems in children with ASD, regardless of whether or not they have intellectual disabilities [22]. The tools mentioned in

Table 1. Measurements related to ASD students’ achievements in a range of aspects.

Measures	Source	Dimensions	How Many Items	Social Indicator	Academic Indicator	School- Indicators (others)	Teacher-Form	Fees
Social Responsiveness Scale (SRS)	Constantino and Gruber, 2005 [16]	5: Social Awareness, Social Cognition, Social Communication, Social Motivation, and Autistic Mannerisms/Restricted Interests and Repetitive Behavior	65	Yes	No	No	Yes	Yes
Woodcock-Johnson III Tests of Cognitive Abilities.	Woodcock et al. 2001 [18]	5: Reading, Mathematics, Writing, Oral Language Abilities, and Academic Knowledge.	22 subtests	No	Yes	No	No	Yes
Social Communication Questionnaire (SCQ)	Chandler et al., 2007 [17]	4: Reciprocal Social Interaction, Language and Communication, and Repetitive and Stereotyped Patterns of Behavior	40	Yes	No	No	Yes	Yes
Strength and Difficulties Questionnaires (SDQ)	Goodman, 1997 [15]	5: Emotional Symptoms, Conduct Problems, Hyperactivity/Inattention, Peer Relationship Problems, and Prosocial Behavior	25	Yes	No	No	Yes	Free
Assessment of Basic Language and Learning Skills (ABLLS)	Partington, 2010 [19]	25: Cooperation and Reinforcer Effectiveness, Visual Performance, Receptive Language, Motor Imitation, Vocal Imitation, Requests, Labelling, Intraverbals, Spontaneous Vocalizations, Syntax and Grammar, Play and Leisure, Social Interaction, Group Instruction, Classroom Routines, Generalized Responding, Reading, Math, Writing, Spelling, Dressing, Eating, Grooming, Toileting, Gross Motor Skills, Fine Motor Skills	544 tasks	Yes	Yes	Yes	Yes	Yes
Social Skills Improvement System Rating Scales (SSIS-RS)	Gresham and Elliott, 2008 [21]	3: Social Skills, Problem Behaviors, Academic Competence	67	Yes	Yes	Yes	Yes	Yes
Child Behavior Checklist (CBCL)	Achenbach and Rescorla, 2001 [20]	8: Anxious/Depressed, Withdrawn, Somatic Complaints, Social Problems, Thought Problems, Attention Problems, Rule-Breaking Behavior, and Aggressive Behavior	118 syndrome items 4–7 academic items	Yes	Yes	Yes	Yes	Yes

, such as the SRS, SDQ, SSRS, SSIS, and CBCL, are reliable and valid measures for assessing emotional, behavioral, and social communication aspects that are often associated with ASD. However, none of the instruments listed in the table explicitly assess the social and academic achievements of ASD students by taking into account teachers as important stakeholders in inclusive education. Instead, they focus on measuring emotional, behavioral, and social communication aspects that are often associated with ASD. Thus, it is crucial for educators and practitioners to consider additional assessment methods and strategies to ensure that individuals with ASD receive comprehensive support and opportunities for success in academic and social domains.

We also used Table 1 to highlight the limited availability of measurement tools that adequately assess both social and academic indicators in children, particularly those with autism. While three scales (ABLLS, SSIS-RS, and CBCL) integrate both types of indicators and provide additional school-specific information, they have significant limitations, such as being lengthy, expensive, and lacking psychometric validation for this population. Thus, there is a clear need for the development of a new, validated, and cost-effective measurement tool that can effectively capture the complex social and academic needs of children with autism. Given the difficulties of collecting data from ASD students themselves, it is meaningful to consider their significant others in school settings (e.g., teachers) in designing this assessment tool. Such a tool would significantly enhance our understanding of the unique needs of this population of students with ASD, which in turn could guide more effective interventions and support to them.

1.2. The Present Study

The assessment of Autism Spectrum Disorders is a critical step in supporting individuals with ASD in inclusive educational settings. Therefore, the present study aims to develop and validate a novel instrument for assessing teachers’ practice and product of supporting ASD students integrated in inclusive classrooms. Our assessment tool, called the Practice and Product Inventory of Supporting Students with ASD (PPI-SSA), is designed to be concise, reliable, and easily accessible, providing educators with valuable information about the school outcomes of students with ASD in inclusive settings. Given the crucial role that teachers play in the success of inclusive education for children with ASD [33,34], we also included indicators of teachers’ inclusive practices regarding ASD in the PPI-SSA. By incorporating these measures, we aim to provide a comprehensive assessment of both the product and process of supporting students with ASD in inclusive educational settings. The PPI-SSA has the potential to be a valuable tool for educators and researchers seeking to improve outcomes for students with ASD in inclusive settings. By providing a reliable and easily accessible means of measuring instructional practices and student outcomes, the PPI-SSA can help support the development of more effective and inclusive educational environments for all students. This validation study is guided by three fundamental research questions (RQs):

• RQ 1: What are the psychometric properties of the Practice and Product Inventory of Supporting Students with ASD (PPI-SSA)?
• RQ 2: To what extent might teachers’ intentions to implement inclusive education predict their self-assessed inclusive practice and social and academic products of ASD students measured by the PPI-SSA?
• RQ 3: What is the extent to which the PPI-SSA measures the same constructs across primary and secondary school teachers?

2. Methods

2.1. Participants and Procedure

Our current study involved local schools and teachers who educated ASD students. We sought and obtained ethical approval from our academic institution before undertaking the survey study, and we meticulously verified that our research practices were in accordance with established ethical guidelines. Our data collection team gathered physical copies of the questionnaire from the participating schools, and prior to their completion, teachers were provided with a comprehensive explanation regarding the rationale and objectives of the research. They were also reassured that their confidential information would be handled exclusively by our research group. Participation was voluntary, and teachers retained the right to withdraw their involvement from the study at any time. We would like to extend our sincere gratitude to the schools and teachers who generously participated in our research, and we remain dedicated to continued community engagement as we progress with our investigations.

A total of 411 teachers participated in this study, including 194 teachers from 61 secondary schools and 217 teachers from 76 primary schools. Among them, 23 teachers had missing data but still provided information on at least 9 out of 14 (64.2%) PPI items. Approximately 23.1% of the teachers were Special Educational Needs Coordinators (SENCO). All participants confirmed that their schools had students with ASD.

Table 2. Demographic information of these teachers.

		Overall n = 411		Secondary n = 194		Primary n = 217
Gender
	Male	107	26.6%	70	36.1%	37	17.1%
	Female	295	71.8%	121	62.4%	174	80.2%
	Missing	9	2.2%	3	1.5%	6	2.8%
Age
	20–29	87	21.2%	47	24.2%	40	18.4%
	30–39	162	39.4%	74	38.1%	88	40.6%
	40–49	106	25.8%	46	23.7%	60	27.6%
	50 or above	51	12.4%	24	12.4%	27	12.4%
	missing	5	1.2%	3	1.5%	2	0.9%
Current positions
	Special Educational Needs Coordinator (SENCO)	95	23.1%	51	26.3%	44	20.3%
	Non-SENCO teacher with more than 10 years teaching experience	161	39.2%	66	34.0%	95	43.8%
	Non-SENCO Teacher with less than 10 years teaching experience	145	35.3%	72	37.1%	73	33.7%
	missing	10	2.4%	5	2.6%	5	2.3%
Previous teaching experience with a student with ASD
	Yes	345	72.0%	143	72.2%	197	90.8%
	No/missing	74	18.0%	54	27.8%	20	9.2%

presents the demographic information for the sample.

2.2. Measures

Practice and Product Inventory of Supporting Students with ASD (PPI-SSA): The Practice and Product Inventory of Supporting Students with ASD (PPI-SSA) is a measure prepared by Centre for Special Educational Needs and Inclusive Education, The Education University of Hong Kong to assess teacher’s evaluation of their inclusive practice and product in supporting students with ASD in inclusive settings. It was developed by referring to a previous study conducted by the research team in inclusive education, but with a focus on over five hundred parents of students with special educational needs including ASD [35]. In this study, Practice refers to teachers’ evaluation of their inclusive practice of adopting suitable instructional strategies to support ASD students’ social and academic achievements in inclusive education. As for the importance of considering instructional strategies, we echo Stahmer et al.’s (2023) argument that “Delivery of high-quality educational services [to ASD students] requires contextual supports, skilled providers, and effective use of evidence-based autism intervention and instructional strategies” (p. 32) [36].

Consistently, Product focused on assessing two essential aspects of ASD students’ gains from inclusive education (social and academic) but measured from the teachers’ perspective. The measure consists of 14 items, with 12 items (PRA1-PRA4, PA1-PA4, and PS1-PS4) adapted from the Parents’ evaluation of their Child’s School Life survey, including their evaluation of their children’s social and academic achievements in inclusive education, which showed good reliability and factorial validity [35]. Two additional items, PS5 (“general students with ASD get on well with non-SEN students”) and PRA5 (“school lessons enhance all students in understanding individual differences”), were added to capture the interaction between students with and without ASD and teaching practices to improve students’ understanding of individual differences, respectively. Teachers rated the PPI-SSA using a 4-point scale (1 = “strongly disagree” and 4 = “strongly agree”). All PPI-SSA items are presented in Table 3 for the reference of researchers and practitioners. We also enclosed a Chinese version of the PPI-SSA in Supplementary Materials for researchers’ and practitioners’ use under the terms and conditions of the Creative Commons Attribution (CC BY) license.

Teachers’ intentions to implement inclusive education: To evaluate the concurrent validity of the PPI-SSA, we assessed teachers’ intentions to implement inclusive education in mainstream schools using Mahat’s (2008) behavioral intention scale [37]. This scale comprised six items (e.g., “I am willing to modify assessments for individual students to enable inclusive education”) rated on a 4-point scale ranging from 1 (“strongly disagree”) to 4 (“strongly agree”). A higher score reflected a greater readiness to implement inclusive education. The Cronbach’s alpha for this scale was 0.816, and CFA with a robust MLR estimator confirmed the appropriateness of a single-factor model (CFI = 0.955, TLI = 0.925, RMSEA = 0.081, SRMR = 0.044).

2.3. Data Analysis

Data analysis was performed using version 27 of SPSS and version 7.2 of MPLUS. Missing data were handled using maximum likelihood methods. We randomly divided the database into two samples (A and B).

2.3.1. Sample A

In this study, Exploratory Factor Analysis (EFA) was conducted using a random half of the sample (n = 206) to investigate the underlying structure or patterns within the data collected through the Practice and Product Inventory for Supporting Students with ASD (PPI-SSA). EFA was needed to identify and extract the key factors or dimensions that contribute to teachers’ practices to support students with ASD and their assessment of ASD students’ social and academic gains from inclusive educational settings. This analysis was essential to ensure that the PPI-SSA accurately captured the multidimensional nature of inclusive education practices for students with ASD, as well as teachers’ evaluation of these students’ social and academic gains, providing a solid foundation for subsequent analyses and instrument validation.

2.3.2. Sample B

Confirmatory Factor Analysis (CFA) using another half of the data (n = 205) was conducted in this study primarily for the validation of the factor structure identified through Exploratory Factor Analysis (EFA). This step aims to assess the consistency and stability of the identified factors and their relationships across an independent sample. By verifying that the proposed factor structure holds true for the entire dataset, CFA further establishes the validity of the instrument, ensuring that it accurately represents the underlying dimensions of teachers’ practices and products in supporting students with ASD in inclusive educational settings. Additionally, CFA provides statistical indices to assess the fit of the model to the data, further confirming the appropriateness of the factor structure for capturing the nuances of inclusive education practices for ASD students.

2.4. Model Fit Assessment

In performing CFA, we used the MLR estimator when conducting CFA as it adjusts the standard errors to be robust against non-normality and heteroscedasticity, making it a robust alternative for obtaining more accurate parameter estimates and fit indices by using real-world datasets. Values of root mean square error of approximation (RMSEA) and standardized root mean square residual (SRMR) less than 0.08, and comparative fit index (CFI) and Tucker–Lewis index (TLI) greater than 0.90 indicated acceptable model fit [38]. Values of RMSEA and SRMR less than 0.06 and CFI and TLI greater than 0.95 indicated excellent model fit [39]. A factor loading higher than 0.42 was considered significant [40,41]. It is a common convention in factor analysis to use this threshold, suggesting that the identified factors are robust and interpretable [40,41]. Table 3 presents detailed items of the PPI-SSA and EFA factor loadings.

Table A3. CFA Factor loadings for the four revised models (n = 205).

	Single Factor	Three Correlated Factors			Second-Order Product and First-Order Practice			Higher Order (PP)
Factor Loading		PRA	PA	PS	PRA	PA	PS	PRA	PA	PS
PRA1	0.402	0.674			0.674			0.674
PRA2	0.513	0.720			0.720			0.720
PRA3	0.430	0.629			0.629			0.629
PRA4	0.516	0.723			0.723			0.723
PRA5	0.569	0.774			0.774			0.774
PA1	0.789		0.812			0.812			0.812
PA2	0.774		0.837			0.837			0.837
PA3	0.796		0.890			0.890			0.890
PA4	0.792		0.825			0.825			0.825
PS1	0.648			0.632			0.632			0.632
PS2	0.649			0.658			0.658			0.658
PS3	0.667			0.822			0.822			0.822
PS4	0.775			0.887			0.887			0.887
PS5	0.737			0.799			0.799			0.799
Second-order product					0.649	0.787	0.904
Higher order factor								0.649	0.787	0.904

Note: PRA = Practice; PA = Academic product; PS = Social product.

shows detailed factor loadings of the four alternative CFA models.

2.5. Factorial Invariance Analysis

Factorial invariance testing was conducted to examine whether measures of the PPI-SSA were equivalent between primary and secondary teachers. First, a baseline model with the same latent structure defined across the two groups of teachers (from primary and secondary schools) was estimated to establish configural invariance. After determining the appropriate factor solution, multiple-group CFA was employed to test invariance across primary and secondary teachers, as well as across SENCO and non-SENCO teachers. Four invariance models were tested: configural (free factor loadings and free intercepts), metric (equal factor loadings), scalar (equal factor loadings and intercepts), and residual (equal factor loadings, intercepts, and item residuals). Invariance was supported when the change in CFI was less than 0.010 and the change in RMSEA was less than 0.015 [42].

2.6. Concurrent Validity Analysis

Finally, a separate latent variable path analysis was conducted to examine the impact of teachers’ intentions to implement inclusive education on the inclusive practice and product dimensions in the PPI-SSA. Significance was determined at a p-value threshold of less than 0.005, confirming the concurrent validity of the instrument.

3. Results

We organized the results of our study in relation to the three research questions we aimed to answer.

3.1. RQ 1: What Are the Psychometric Properties of the PPI-SSA?

Exploratory and Confirmatory Factor Analyses were conducted in two phases. In the first phase, an exploratory factor analysis (EFA) was performed on half of the sample (n = 206), revealing three distinct factors: inclusive practice, academic product, and social product. These factors were identified based on the five most salient items in each category (Table 3). Together, these three factors explained 59.24% of the total variance, with no significant cross-factor loadings observed.

Table 3. The Practice and Product Inventory of Supporting Students with ASD (PPI-SSA) and its Exploratory factor analysis results.

	Teachers’ Perceived Inclusive Practice to Support ASD Students and ASD Students’ Social and Academic Products		EFA Factors and Loadings
		F1 Practice	F2 Academic Product	F3 Social Product
	In my school…
PRA1	Staff modify the curriculum to meet the needs of students with ASD	0.793	−0.146	0.025
PRA2	Lessons are planned in response to the diversity of student with ASD	0.780	−0.001	−0.077
PRA3	Teachers are concerned to support the learning of students with ASD	0.613	−0.065	0.128
PRA4	Staff have sufficient professional knowledge to support the learning of students with ASD	0.667	0.110	0.009
PRA5	Lessons enhance all students (especially considering ASD) in understanding individual differences	0.761	0.079	−0.087
	In my school, ASD students…
PA1	Grasp a range of learning skills (e.g., note-taking, problem-solving)	0.055	0.671	0.042
PA2	Understand what the teacher is teaching in the classroom	0.049	0.836	0.033
PA3	Learn on their own	−0.070	0.926	−0.114
PA4	Are motivated to learn	−0.073	0.734	0.109
PS1	Participate in extracurricular activities	0.158	0.134	0.537
PS2	Participate in inter-school activities	0.177	0.122	0.573
PS3	Have a social circle of friends	−0.046	−0.038	0.839
PS4	Socialize with students without ASD	−0.124	−0.025	0.946
PS5	Get on well with students without ASD	0.023	−0.028	0.747
Factor correlation
F1		-
F2		0.534	-
F3		0.535	0.616	-

Notes: Extraction Method: Principal Axis Factoring; Rotation Method: Promax with Kaiser Normalization; Rotation converged in 4 iterations. All items were rated on a 4-point scale. Factor loadings for the corresponding three dimensions and larger than 0.42 are bolded for easy reference.

Table 3. The Practice and Product Inventory of Supporting Students with ASD (PPI-SSA) and its Exploratory factor analysis results.

	Teachers’ Perceived Inclusive Practice to Support ASD Students and ASD Students’ Social and Academic Products		EFA Factors and Loadings
		F1 Practice	F2 Academic Product	F3 Social Product
	In my school…
PRA1	Staff modify the curriculum to meet the needs of students with ASD	0.793	−0.146	0.025
PRA2	Lessons are planned in response to the diversity of student with ASD	0.780	−0.001	−0.077
PRA3	Teachers are concerned to support the learning of students with ASD	0.613	−0.065	0.128
PRA4	Staff have sufficient professional knowledge to support the learning of students with ASD	0.667	0.110	0.009
PRA5	Lessons enhance all students (especially considering ASD) in understanding individual differences	0.761	0.079	−0.087
	In my school, ASD students…
PA1	Grasp a range of learning skills (e.g., note-taking, problem-solving)	0.055	0.671	0.042
PA2	Understand what the teacher is teaching in the classroom	0.049	0.836	0.033
PA3	Learn on their own	−0.070	0.926	−0.114
PA4	Are motivated to learn	−0.073	0.734	0.109
PS1	Participate in extracurricular activities	0.158	0.134	0.537
PS2	Participate in inter-school activities	0.177	0.122	0.573
PS3	Have a social circle of friends	−0.046	−0.038	0.839
PS4	Socialize with students without ASD	−0.124	−0.025	0.946
PS5	Get on well with students without ASD	0.023	−0.028	0.747
Factor correlation
F1		-
F2		0.534	-
F3		0.535	0.616	-

Notes: Extraction Method: Principal Axis Factoring; Rotation Method: Promax with Kaiser Normalization; Rotation converged in 4 iterations. All items were rated on a 4-point scale. Factor loadings for the corresponding three dimensions and larger than 0.42 are bolded for easy reference.

In the second phase, a Confirmatory Factor Analysis (CFA) was conducted on the other half of the sample (n = 206) using maximum likelihood estimation with robust standard errors (MLR). The results indicated that the three-factor solution from EFA closely fit the data (χ² = 226.096, df = 74, CFI = 0.887, TLI = 0.861, RMSEA = 0.100 [90% CI = 0.087, 0.114], SRMR = 0.053). Modification indices suggested that correlating the residuals of two items (PS1 “Participate in extracurricular activities” and PS2 “Participate in inter-school activities”) could improve the model fit, as these two items had overlapping meanings. After making this modification, the revised model demonstrated excellent fit (χ² = 126.758, df = 73, CFI = 0.960, TLI = 0.950, RMSEA = 0.060 [90% CI = 0.044, 0.076], SRMR = 0.051).

To further confirm the appropriateness of our factor solution, alternative models were tested, including a single-factor model, a model combining one second-order conduct factor and one practice factor, and a second-order product and practice factor (see Figure 1), with and without correlating residuals (see

Table 4. Model fit indices of the four alternative models of the PPI-SSA.

The Remaining Half Sample (n = 205)		χ²	df	CFI	TLI	RMSEA [90% CI]			SRMR
Not correlating the residuals of ps1 and ps2
1	single factor	498.642	77	0.686	0.629	0.163	[0.151	0.176]	0.105
2	three correlated factors	226.096	74	0.887	0.861	0.100	[0.093	0.125]	0.053
3	second-order product and first-order practice	226.096	74	0.887	0.861	0.100	[0.093	0.125]	0.053
4	Higher order (PP)	226.096	74	0.887	0.861	0.100	[0.093	0.125]	0.053
Revised models (by correlating the residuals of ps1 and ps2)
1.1	Revised single factor	407.356	76	0.753	0.705	0.146	[0.132	0.160]	0.102
2.1	Revised three correlated factors	126.758	73	0.960	0.950	0.060	[0.044	0.076]	0.051
3.1	Revised second-order product and first-order practice	126.758	73	0.960	0.950	0.060	[0.044	0.076]	0.051
4.1	Revised higher order (PP)	126.758	73	0.960	0.950	0.060	[0.044	0.076]	0.051

for details). While all three models produced acceptable fit indices with factor loadings higher than 0.42, they had poor fit for the SENCO (n = 95) and secondary (n = 194) samples (e.g., RMSEAs = 0.086, 0.092; TLIs = 0.876, 0.897). The revised correlated factor model had excellent fit across all subgroups of teachers, including overall, SENCO, non-SENCO, primary, and secondary samples, and reliability analysis based on this factor solution produced acceptable Cronbach’s alphas ranging from 0.836 to 0.887 (see

Table 5. Measurement invariance of the revised correlated factors model *.

MLR	χ²	df	RMSEA [90% CI]			CFI	TLI	SRMR	AIC	BIC	Δχ²	Δdf	p	ΔCFI	ΔRMSEA
Across primary and secondary
Primary (n = 217)	118.653	73	0.054	0.038	0.069	0.959	0.949	0.063	4251.022	4406.497
Secondary (n = 194)	117.759	73	0.056	0.039	0.073	0.963	0.954	0.047	3957.037	4107.359
Configural	236.425	146	0.055	0.043	0.066	0.961	0.952	0.056	8208.059	8577.770
Metric	246.423	157	0.053	0.041	0.064	0.962	0.956	0.060	8196.878	8522.384	9.998	11	0.531	0.001	0.002
Scalar	264.735	168	0.053	0.042	0.064	0.958	0.955	0.060	819.074	8474.376	18.312	11	0.075	0.004	0.000
Residual	286.032	182	0.053	0.042	0.063	0.955	0.955	0.061	8195.558	8420.600	21.297	14	0.094	0.003	0.000
Across SENCO and NON-SENCO
SENCO (n = 95)	85.498	73	0.042	0.000	0.074	0.974	0.968	0.061	2054.289	2171.767
NON-SENCO (n = 306)	165.474	73	0.064	0.053	0.076	0.952	0.940	0.053	6009.312	6180.597
Configural	255.193	146	0.061	0.050	0.072	0.956	0.945	0.055	8063.601	8431.046
Metric	272.558	157	0.061	0.049	0.071	0.953	0.945	0.069	8061.313	8384.824	17.365	11	0.098	0.004	0.001
Scalar	298.615	168	0.062	0.052	0.073	0.947	0.942	0.071	8064.731	8344.308	26.057	11	0.006	0.006	0.003
Residual	311.949	182	0.060	0.049	0.070	0.947	0.947	0.071	8059.262	8282.924	13.334	14	0.500	0.000	0.003

* The results were similar for the revised model of first-order practice and second-order product (see Table A1 and Table A2 in the Appendix A)

for detailed results). Therefore, we concluded that the revised correlated factor model was appropriate and used it for subsequent analyses.

3.2. RQ 2: To What Extent Might Teachers’ Intention to Implement Inclusive Education Predict Inclusive Practice and Products Assessed by the PPI-SSA?

Establishing a relationship between teachers’ intention to implement inclusive education and their self-assessed inclusive practice and the product of ASD students through the PPI-SSA is important to build the intention–behavior connection. Since the PPI-SSA is centered around inclusive education practices for ASD students and the evaluation of ASD students’ social and academic gains from inclusive education, it is reasonable to assume a connection between the subscale scores of the PPI-SSA and teachers’ intention to implement inclusive education. Teachers who express a strong intention to embrace inclusive education are likely to be actively engaged in practices that support ASD students’ social and academic development within the inclusive classroom setting. With this rationale, concurrent validity was assessed through a latent variable path analysis. The analysis revealed significant path coefficients of 0.346, 0.258, and 0.248, and p-values less than 0.001. These results suggest that a teacher’s intention to implement inclusive education would be a significant predictor of the inclusive practices and products (social and academic gains) of students with ASD within inclusive settings. These empirical findings are consistent with existing literature, such as studies by Alkeraida (2021) and Low et al. (2018) [31,34], which emphasized the role of teachers’ proactive and inclusive behaviors in facilitating the successful inclusion of students with ASD. Our study adds to the existing body of evidence supporting the concurrent validity of the PPI-SSA as an effective tool.

3.3. RQ 3: What Is the Extent to Which the PPI-SSA Measures the Same Constructs across Primary and Secondary School Teachers?

Measurement invariance was tested using multiple-group Confirmatory Factor Analysis (CFA) to examine the equivalence of the PPI-SSA across different groups of teachers, including primary and secondary teachers, as well as SENCO and non-SENCO teachers. The results of the CFA, as shown in Table 5, indicated that the PPI-SSA demonstrated configural, metric, scalar, and residual invariance across primary and secondary teachers. This means that the factor structure, factor loadings, intercepts, and item residuals of the PPI-SSA were equivalent across the two different teacher groups.

4. Discussion

This study investigated the psychometric properties of the Practice and Product Inventory of Supporting Students with ASD (PPI-SSA) in relation to three research questions and with a sample of Hong Kong teachers. The analysis procedure in this study revealed several notable results. Exploratory Factor Analysis (EFA) identified three distinct factors within the Practice and Product Inventory for Supporting Students with ASD (PPI-SSA), explaining around 59.24% of the total variance. Confirmatory Factor Analysis (CFA) validated this three-factor model, with indices such as CFI (0.960), TLI (0.950), RMSEA (0.060), and SRMR (0.051) supporting the appropriateness of the proposed model and demonstrating a good fit to the data. Concurrent validity was established as teachers’ intentions to implement inclusive education positively predicted their PPI-SSA scores. Additionally, measurement invariance was confirmed across different teacher subgroups. Similarly, the measurement invariance across SENCO and non-SENCO teachers suggests that the PPI-SSA is applicable to teachers with and without special education responsibilities, making it a flexible tool for assessing inclusive practices in diverse educational settings. These findings collectively underscore the instrument’s validity and reliability, offering a valuable tool for assessing teachers’ practices and products in supporting students with Autism Spectrum Disorder (ASD) in inclusive educational settings.

The findings of the study highlight the usefulness of the PPI-SSA in supplementing previous research for assessing teachers’ inclusive teaching practices to support ASD students. The study’s conclusions show the value of the PPI-SSA in enhancing empirical studies by evaluating instructors’ use of inclusive teaching strategies to benefit children with Autism Spectrum Disorder (ASD) in both their social and academic achievements. The study’s findings could be compared and contrasted with previous research, such as the study by Estes et al. (2011) [24], which highlighted the need to consider individual differences in academic performance among ASD students in inclusive education settings. The instructional practices identified through the PPI-SSA can provide insights into how teachers can effectively support academic outcomes for ASD students, taking into account ASD students’ unique abilities and needs. In addition, the findings of this study could be related to the influence of school services on adjusting the severity of ASD symptoms in children with ASD, as highlighted by Rosen et al. (2019) [25], examining how certain instructional practices positively adjust symptom severity could highlight the benefits of inclusive education for ASD students.

The results of the study could also be compared to those of Milgramm et al.’s (2021) study, which found no appreciable differences in academic competence overall, social skills, or problem behaviors between children with and without ASD [26]. In this study, we found a positive association between inclusive instructional support and interpersonal competencies (academic and social domains) among students with ASD in inclusive education. This comparison may shed light on further examining of comparative studies between ASD and non-ASD students by taking into account inclusive instructional support between the two groups of students. Additionally, the reliability and validity of the PPI-SSA as a tool for assessing instructional practices in supporting ASD students could be related to those of the Classroom Measure of Active Engagement (CMAE) as a tool Sparapani et al. (2016) for measuring active engagement in ASD students [27]. Empirically, further studies can explore the extent to which engagement of ASD students measured by CMAE relates to social and academic outcomes by using the PPI-SSA. The psychometric properties of the PPI-SSA highlight the rigor of the study and the validity of the current findings, further strengthening its usefulness to the field. The study’s findings could also be related to the feasibility and effectiveness of social skills group training (SSGT) in naturalistic settings [28] for improving social skills in students with ASD and the effectiveness of cognitive-behavioral school-based interventions, such as schoolMAX, in improving emotion recognition skills in children with ASD [29]. Recently, in their pioneering study, Stahmer et al. (2023) found teacher participation in Classroom Pivotal Response Teaching (CPRT) training had a significant positive effect on the engagement of ASD students during CPRT instruction, significantly reducing reported approach/withdrawal problems [36]. Discussing the alignment between the identified instructional practices and evidence-based interventions (e.g., SSGT, schoolMAX, CPRT) could provide support for the significance of our design of the PPI-SSA to include assessing the effectiveness of the instructional practices identified in the current study. The findings of the study align with established interventions and practices in the field of ASD research, which aim to enhance the academic, social, and emotional skills of students with ASD in inclusive education settings. The study provides valuable insights into the mechanisms that underlie the impact of instructional practices on the social and academic aspects of ASD students’ development in inclusive education.

In summary, in an era where sustainability underscores our educational and societal endeavors, it becomes paramount to consider the inclusivity and comprehensiveness of our educational systems to cater to diverse students’ needs. This includes ensuring that every student, regardless of their individual challenges, receives a quality education that prepares them for a sustainable and socio-economically viable future. This study offers significant insights that resonate with the broader theme of sustainability by addressing the inclusion of students with Autism Spectrum Disorder (ASD) in mainstream educational settings. The challenges of assessing learning outcomes for students with special educational needs intersect with the broader challenges of ensuring sustainable, inclusive, and equitable education for all. Our findings highlight the importance and efficacy of the PPI-SSA instrument in supplementing existing research on assessing teachers’ inclusive teaching practices to support students with ASD. By identifying evidence-based instructional practices tailored to the unique needs of ASD students, our research contributes to crafting a more integrated and inclusive approach to education. Such inclusivity, when considered through the lens of sustainability, suggests that by catering to the unique needs of all students, researchers and practitioners contribute to preparing a future generation that is more resilient, adaptable, and equipped to face socio-economic challenges. Furthermore, by linking the outcomes of our research with prior studies on ASD interventions and outcomes, we reinforce the idea that sustainable educational policies should encompass a comprehensive understanding of student needs, including those with various types of special educational needs. With a focus on ASD students, as discussed above, the findings of this study can be related to other previous research on ASD symptomatology and associated interventions (e.g., problem behaviors, engagement, social skills group training, cognitive-behavioral interventions, teacher-rated academic skills, and social relationships establishment). These connections with previous research provide valuable evidence for the relevance and effectiveness of the instructional practices identified through the PPI-SSA in promoting positive outcomes for students with ASD in inclusive education settings. This study adds depth to the literature on educational practices for supporting ASD students but also emphasizes the need to consider these inclusive practices as part of a supportive framework for creating sustainable and inclusive learning environments.

Limitations and Future Directions

One limitation of this study is that the measures of inclusive education practice to support ASD students and ASD students’ social and academic products were based on self-reported data from teachers. Despite the contribution of the self-report survey design, it also has limitations [43] due to its potential personal bias from teachers. Future research could incorporate additional sources of data, such as observations or student school portfolios and reports from other stakeholders of inclusive education (e.g., ASD students’ parents and educational psychologists), to provide a more comprehensive understanding of inclusive education practices and outcomes for ASD students. Another limitation is that the study only assessed the concurrent validity of the PPI-SSA, and future research could examine its predictive validity over time. Additionally, while the measurement invariance analysis demonstrated the equivalence of the factor structure and item parameters across different groups of teachers, it is possible that other sources of bias, such as response style bias, could still be present. Future directions could include using the PPI-SSA in longitudinal studies to investigate how teachers’ inclusive intentions and behaviors impact the outcomes of students with ASD over time. The PPI-SSA could also be used in intervention studies to assess the effectiveness of interventions designed to increase teachers’ productive, inclusive practice to support ASD students. Additionally, future research could explore the use of the PPI-SSA with other populations, such as students with other types of disabilities or in different cultural contexts, to determine the generalizability of the factor structure and item parameters. Furthermore, it may be worthwhile to examine the construct validity of the PPI-SSA by exploring its relationship with other measures of teacher inclusiveness and student outcomes. Finally, it is essential to acknowledge the notable limitations associated with the use of self-reports. This method may lack the capacity to capture nuanced details that might be more comprehensively explored through interviews or observations. In order to enhance the comprehensiveness and robustness of the instrument, future research directions could involve supplementing self-reports with complementary methods such as interviews or observations. These qualitative approaches would be especially valuable when aiming to gain a deeper and more nuanced understanding of teachers’ experiences and practices in supporting students with ASD within inclusive educational settings.

5. Conclusions

Students with ASD are vulnerable to poor outcomes, including poor mental health conditions and underachievement in schools. Therefore, identifying key dimensions of teachers’ professional practice related to inclusive education can play a critical role in promoting positive outcomes for these students. In this study, we evaluated the Practice and Product Inventory for Supporting Students with ASD (PPI-SSA) to assess its suitability for examining teachers’ inclusive practices and self-assessed products in terms of ASD students’ social and academic gains from inclusive education. Our analysis, comprising exploratory and confirmatory factor analyses, identified three key factors within the PPI-SSA: inclusive practice, social product, and academic product. These factors explained a substantial portion of the variance, indicating the instrument’s ability to capture essential dimensions of inclusive education practices for ASD students. Concurrent validity was also confirmed, as a positive relationship was observed between teachers’ intentions to implement inclusive education and their PPI-SSA scores. This suggests alignment between teachers’ intentions and their inclusive behaviors. Additionally, measurement invariance across different teacher subgroups, such as primary and secondary teachers and SENCO and non-SENCO teachers, further supports the instrument’s adaptability to diverse educational contexts. The PPI-SSA thus holds promise as a useful and concise tool for assessing and understanding inclusive education practices for students with ASD, potentially informing targeted interventions to enhance their educational experiences and outcomes in inclusive classrooms.