The conceptual understanding of IE in this study was framed by the IEM (Mitchell 2015), as it covered all key aspects of IE. The IEM (Figure 1) (Mitchell 2015) illustrates the different elements of IE, hence its adoption as a conceptual frame that guided the researchers’ perspective of IE. The IEM was adapted by the researchers into four key themes, guided by relevant literature and the EWP6, including guidance and intentions, adaptations, access and acceptance and support. Although the IEM is a model based on international principles, it encompasses all of the necessary aspects to implement a successful IE programme, thus justifying its use in this study.

Guidance and intentions involve the common, philosophical vision across leading stakeholders towards providing a conducive environment suited for learners with and without special educational needs. Access and acceptance involve fostering equitable, physical access to education depending on the learner’s special needs, while optimising academic and social skills development and mutual acceptance among differently abled learners (Lorger et al. 2015). Support entails the provision of necessary academic resources and other support, with intention to ensure equitable access for learners with differing disabilities. Adaptations entails the modifications made to the curriculum (teaching, learning and assessment) to accommodate all learners with differing capabilities (Mitchell 2015). These themes, and the categories within each, were used to organise and analyse the data.

Within an interpretive paradigm, a phenomenology design was adopted for this study (Irarrázaval 2020). Two key phenomenological tenets framed the researcher’s perspective on the educators’ experiences. Firstly, the world exists as experienced, and secondly, no two individuals can experience the same phenomenon completely in the same way (Neubauer, Witkop & Varpio 2019). The current study adopted a qualitative, descriptive, explorative approach to explore and describe the experiences of educator’s teaching of learners with HL in an inclusive context (Lambert & Lambert 2012). This design best suited the need to explore the first-hand experiences of the educators that participated in the study as it was not the phenomenon of teaching learners with HL, but the experiences thereof, as lived by the participants, that the researchers were interested in.

The study was conducted in fulfilment of a postgraduate degree, and the data was collected in public schools that provide IE, within the KZN Province, South Africa. A total of 25 full-service and ordinary schools that catered to the learner with HL were selected and contacted, with the intention of identifying 15 appropriate participants. The number of participants was deemed appropriate as it would most likely allow for the saturation of data (Latham 2020). Out of the 25 schools that were contacted, two schools participated in the study. The KZN Province was selected for the research project as it encompassed the multicultural tapestry of South Africa and allowed the researcher to obtain data from educators of various settings (rural and urban). Thus, the research findings were, arguably, a fair representation of the current inclusive teaching practices for learners with HL.

Data collection was deemed complete after six purposively sampled educators had participated in the study (Table 1), as data saturation was reached, where further data collection did not produce themes that were not already present in the data (Saunders et al. 2017). Each participant had to be currently working in an inclusive school and exposed to learners with HL. All participants had been teaching learners with HL within an inclusive context for at least a minimum of 2 years, with the most experienced in this regard being 10 years. An in-depth interview with each participant indicated that they each had been teaching for many years prior to teaching learners with HL within an inclusive context (see Table 1). Therefore, the cumulative experience of teaching learners with HL, within an inclusive context, was at least 29 years across the participants, which was copious exposure for them to provide in-depth reflections. The participants were exposed to a variety of learners with HL throughout their teaching experience, thus providing them with experience across varying degrees of HL, loss of functionality and amplification needs. However, it must be noted that at the time the study was conducted, none of the educators were teaching learners with HL who had any assistive or amplification devices currently in use.

Participants were all female and mostly homogeneous, in being from Quintile 3 schools within the urban area. The Quintile 3 schools are not considered affluent, yet not poor enough to be prioritised for maximum funding, thus presenting a vulnerable context (Mestry, & Bisschoff, 2009), worthy of the attention in research projects like the current one. Of the participants, three of them only used verbal cues when communicating with their learners; however, one participant supplemented verbal instructions with written cues. However, all participants were exposed to at least one or more learners with a significant HL of moderate level (35 dBHL) or more in at least one ear.

A pilot study was conducted with two participants, following the same recruitment and data collection process as the main study (Leon, Davis & Kraemer 2011). No changes were needed to the data collection process and methods prior to conducting the main study. The participants who took part in the pilot and the data gathered from them were excluded from the main study. For the main study, semi-structured interviews were conducted with each participant, guided by the interview schedule (developed by the researcher and guided by the IEM), via an online platform (Zoom). In addition to closely following the interview schedule, the author was able to probe further, with intention to elicit rich reflection from each participant where needed. Thus, despite other participants being less forthcoming with their experiences than others, this process allowed for all interviews to be relatively similar in depth. All audio data were automatically recorded onto a password-protected computer that belonged to the researcher.

The raw data was transcribed twice. Firstly, verbatim transcription was manually conducted by the researcher, transforming audio files into text files (Duranti 2009). Secondly, the audio files were transcribed using the NVivo 12 software, separately, as a parallel process and a way of double checking the accuracy of the transcription (Zamawe 2015). Where differences existed in the transcription for each interview, the researchers reassessed the audio at specific timestamps to correct the errors. Where necessary, data cleaning included the removal of redundant speech, which did not alter the intention of the speaker, such as ‘Uhm’ and ‘yeah’, along with aspects of speech that did not contribute to the topic. The transcripts were then sent to the participants to observe member checking, thus increasing credibility and the richness of the data (Honorene 2017).

Through content analysis, the data was coded and organised into the preexisting subthemes of the IEM, as shown in Figure 1, following a deductive approach (Elo & Kyngäs 2008). Subsequently, emergent themes were formulated by collating the organising subthemes, through thematic analysis following the steps recommended by Braun and Clarke (2006:87). Each of the two authors generated the emerging themes separately, prior to meeting to compare and agree on the final emerging themes. Therefore, related excerpts within each organising subtheme were placed into themes by the principal researcher. Both researchers cross-checked and agreed on the emergent themes prior to finalising the current article.

Methodological rigour and transparency about the processes followed in the study enhanced dependability (Kyngäs, Kääriäinen & Elo 2020). Data saturation was duly achieved when there were no longer new codes generated from the new data (Urquhart 2013). This was evident between the fifth and sixth interviews, as the sixth did not generate any new codes. This was used to judge the completeness and richness of the generated data, to ensure that the data represented, as closely as possible, the experiences of educators in the study’s context, thus enhancing dependability and rigour before terminating data collection (Guest, Namey & Chen 2020). Transferability was ensured through a rigorous methodological application, including representative sampling and bracketing of the researcher’s own experiences ensuring that the findings could be representative of and transferable to similar contexts (Nowell et al. 2017). Confirmability was maintained throughout by a rigorous verbatim transcription and retranscription to correct the errors identified from the first transcription. Bracketing out the researchers’ biases also contributed positively in confirmability, as this ensures that the conclusions made were founded in the study findings. Member checking further contributed in enhancing confirmability of the findings as participants reviewed their transcripts (Johnson, Adkins & Chauvin 2020). The use of direct quotes, over and above sharing the rigorous processes in the study, further enhanced credibility (FitzPatrick 2019).

Ethical clearances were obtained from the Humanities and Social Science Research Ethics Committee (HSSREC/00002351/2021) of the University of KwaZulu-Natal and the KZN Department of Education (Ref: 2/4/8/7042), to ensure that the study is ethically and scientifically sound.

Gatekeeper permission was obtained from the school management of every participating school and all participants were required to provide a written form of consent via an informed consent form. Participants were ensured that their participation was voluntary with autonomy to withdraw if the participants wished to do so. To maintain confidentiality, all identifying factors were removed from the final report.

Three preexisting organising subthemes under guidance and intention included leadership, vision and placement. The emerging themes are depicted in Table 2.

The adaptation of the curriculum content, assessment and teaching practices play a critical role in the success of the implementation of IE (Dalton et al. 2012; Mitchell 2015). Practically, curriculum adaptation presented the participants with challenges, which are shared in the emerging themes (Table 3).

The EWP6 specified the Curriculum Assessment Policy Statement (CAPS) as a suitable syllabus for all learners. With appropriate guidelines, educators should be confident adapting the curriculum content (Dalton et al. 2012). However, this did not seem to be the case with the current study participants for various reasons. Participant 3 noted that educators may feel that they do not have the knowledge to sufficiently differentiate the curriculum as a means to accommodate all learners:

As some learners may struggle to complete assessment tasks on time, their assessment time was increased. Two participants (Participants 1 and 4) reported use of this practice:

A further adaptation is made in the presentation of the assessment tasks as indicated by Participant 3, who would include the auditory presentation to supplement the visually presented paper, by reading out the questions to the learners. Participant 4 noted that they presented the learners with different difficulty levels so as to allow learners with disabilities to answer what they were able to:

Despite the aforementioned adaptations, learners are essentially assessed on the same exams, which leave some participants feeling that this sets the learners up for failure. Another challenge is that the participants are clearly following different, individualistic approaches to adapting their assessment, which leaves room for some level of standardisation in this regard.

Four participants (Participants 1, 2, 4 and 5) optimised the learners’ access to visual and auditory stimuli by positioning specific learners closer to the front, based on their disability and strategically grouping the learners for the similar purpose, so that peer learning would be optimised between all learners:

Similarly, to the assessments, the mode of presentation was adapted during teaching. While three of the participants (Participants 1, 2 and 6) only used verbal communication when providing instructions to learners with HL, Participant 2 supplemented the verbal with written instructions:

The above excerpt indicates how teachers enforced inclusivity in classroom, in a way that would provide opportunity for a diverse range of interaction. The learners without disability not only assisted those with disabilities, but they are likely to also learn and appreciate opportunities to interact with their peers.

This study was conducted within one province of South Africa, where circumstances might have elements of uniqueness from other provinces. Yet, where conditions match those of the study settings, the findings remain applicable, even though the aim was not to generalise the findings to other contexts. The current study has indicated that IE is perceived differently by different educators, with some foregrounding the educational development while others foreground the social aspect. These perspective differences suggested a possibility that the implementation of IE is also different among the educators. This was further supported by the findings that indicated a predominantly institution-based implementation of IE, as opposed to one that follows a standardised approach, despite the availability of a policy and related guidelines for the implementation of the IE (DoBE 2010; DoE 2001).

However, the findings indicated that the study participants did not feel clearly guided by the EWP6 although being the main policy on IE for over 20 years. The same concern was previously noted by Storbeck and Moodley (2011) who noted the poor guidance towards the implementation of IE policies by educators. Likewise, Donohue and Bornman (2014) argued that the guidance at a policy level is simply insufficient for educators. For this and other reasons, the curriculum in IE remains largely suitable for learners without disability, despite some adaptations (Störbeck et al. 2010). Therefore, the room for different interpretations of the EWP6 remains wide enough for IE to be implemented differently at each school, which, we argue, needs to be addressed urgently. A further concern is the lack of specialised training that educators reported, while the current IE system, as it is set out, relies on their expertise in teaching learners with disabilities. The lack of training is a major concern among educators as indicated in literature (Mpu & Adu 2021; Nel et al. 2014) and supported by the current study.

Among many challenges, the correct placement of learners was mentioned where the administration process was open to all learners with disability, regardless of whether they met inclusive school criteria. Therefore, educators’ challenges started with the access and placement process. These were then followed by tedious administration processes once the learners had been admitted, including when they had to access support from other professionals, or needed to be referred to a more suitable school. This potentially contributes to the educators’ silent frustrations (Nel et al. 2014), suggesting that the administration that should ideally be carried out by administrators (Flem, Moen & Gudmundsdottir 2004) becomes their responsibility.

The lack of support, or delayed access to other stakeholders including parents, resulted in learners placed with educators who struggled to accommodate them. The findings suggested that the parents have a differing conception and attitude about the importance of their role in their children’s learning; the reasons for their lack of involvement in the context studied were not provided. However, a study conducted in Germany found that the parental involvement depended largely on their attitude towards IE, and those with a positive attitude tended to be much more involved in their children’s education (Paseka & Schwab 2020).

On the contrary, the participants received the needed support from school management and colleagues through frequent communication, where concerns about teaching would be addressed. This support helped with institutional-based IE implementation, fostering positive experiences in this regard. These findings are similar to the findings reported by Nel et al. (2016) who found that educators had a positive experience with institutional-based support as opposed to the insufficient support they received outside their specific schools towards implementing IE. The support from internal structures alone, mentioned earlier, was clearly insufficient in closing the gap of the needed support from other stakeholders. The educators reporting the lack of stakeholder support as a challenge in IE suggests that such stakeholders, including parents, have a significant role in the effective implementation of IE.

All the above challenges put the educators in a position where they needed resources, including teaching aids and personnel. These were not always available, potentially because of the education funding system that does not seem to prioritise IE as it should, at least as experienced by the educators in this study. Instead, the current funding model (Quintile system) prioritises funding for schools who are classified as Quintile 1, whereas most of the participants in the study teach in Quintile 3 schools. Yet, their needs could be equal to or exceed Quintile 1 schools because of having learners with different disabilities. These findings are consistent with a study conducted by Makoelle and Burmistrova (2020) in a South African province, which concluded that the current funding system and practices are insufficient to support IE, leading to resource constraints in those schools. Furthermore, Van Dyk and White (2019) revealed the same, especially in low-quintile schools, and further add that there is a need to reconsider the funding approach to a more holistic one, if learners with HL and other disabilities are to be afforded IE experiences.

Despite the general challenges in the participants’ experiences, they still managed to accommodate the learners with HL and other disabilities through adjusting how they teach and assessed, with the CAPS curriculum keeping the content relatively unchanged. Thus, the participant followed the model of differentiated instruction and assessment, as opposed to a content-based model of accommodating learners with HL and other disabilities.

The teachers typically used their creativity and experience to implement IE, as there are no standardized set of steps they could follow in adjusting the curriculum to suit learners with HL within IE settings. Some participants used ungraded reading books, which was typically easier for the learners to cope with, thus encouraging the learners (Scruggs, Mastropieri & Marshak 2012), while accommodating the different reading skills in the same classroom, as recommended in the EWP6. Other teaching adaptations included the adjustments to the modalities of communication, where some educators supplemented visual information with auditory stimuli. They further adjusted how learners sit in class in order to optimise their exposure to stimuli, depending on their needs. The strategic grouping of learners with different capabilities was also used to facilitate cooperation and peer learning between the learners, as recommended in the Guidelines for Responding to Learner Diversity in the classroom through Curriculum and Assessment Policy Statements (DoBE 2011).

The assessment adaptations were similar to the aforementioned teaching adaptations. Adapting assessments to suit the needs of a learner provides them with opportunities to thrive to the best of their abilities (Gardner 1983). Findings in the current study indicated that the educators adapted the assessments by adjusting the length of examination times, presentation of exam content to their learners and by varying the difficulty of questions throughout the article. Again, there was no standardised method of doing this, which presented a challenge to the educators. The reported lack of guidance in this regard suggests that the educators were either not aware of the Guidelines for Responding to Learner Diversity in the classroom through Curriculum and Assessment Policy Statements (DoBE 2011) or they did not find them to be sufficient in guiding their curriculum adaptation practices. As a result of the lack of guidance, learners with HL may experience difficulties completing assessments, thus affecting their academic progression.

The study findings suggest that the extent to which the IEM and the EWP6 were adopted by educator of learners with HL in IE was very limited, if at all. The curriculum adaptations, staff support, admissions and other practices were typically unique to each school, with each participant having unique methods of implementing IE.

Therefore, it is possible that the lack of awareness of the details of the EWP6 and the IEM, or other suitable models of implementing IE, contributed to the diversity of IE practices among educators. Those who were aware of the EWP6 had different interpretations of its key intentions. This, along with the vague nature of this policy (DoE 2001) and related guidelines (DoBE 2005, 2010, 2011), could have further contributed to the limited extent to which they were adopted by the educators in this study.