Diagnostic language assessment is central to speech-language therapy, underpinning the identification of language disorders and guiding intervention planning (Norbury et al., 2016). Clinicians must make complex diagnostic decisions through the integration of multiple sources of information, including standardised test scores, caregiver report, observations and clinical judgement (Pascoe & Singh, 2023). However, diagnostic decision-making is shaped by the nature and quality of available data. Evidence suggests that clinicians demonstrate variability in their judgements, particularly when relying on informal or perceptual measures (Bawayan & Brown, 2022). In South Africa, these challenges are compounded by extensive linguistic diversity and longstanding concerns regarding the validity of assessment tools that have not been developed or normed for multilingual populations (Khoza-Shangase & Mophosho, 2021; Southwood et al., 2025; Van Dulm & Southwood, 2008). A key component of diagnostic practice is the selection and use of assessment tools. In South Africa, speech-language pathologists (SLPs) commonly use standardised instruments such as the Clinical Evaluation of Language Fundamentals (CELF), Peabody Picture Vocabulary Test (PPVT) and Renfrew Action Picture Test, which are primarily developed and normed in Western, monolingual populations (Van Dulm, 2013; Van Dulm & Southwood, 2008). The use of these tools in linguistically diverse populations raises concerns regarding cultural and linguistic validity, as normative samples do not reflect the multilingual realities of South African children (Khoza-Shangase & Mophosho, 2018; Mdlalo, 2016). This mismatch introduces a risk of misdiagnosis, including both over-identification and under-identification of language disorders (Bedore & Peña, 2008; Paradis, 2010). Despite these concerns, standardised assessments continue to play a dominant role in diagnostic decision-making. Globally, clinicians often prioritise standardised test scores when determining eligibility for services (Fulcher-Rood et al., 2018). At the same time, clinicians use informal measures such as language sampling, caregiver interviews and observations to capture functional language use in naturalistic contexts (Verdon et al., 2015). However, these measures are often weighted less heavily in diagnostic decisions and may be applied inconsistently across clinicians (Bawayan & Brown, 2022; Fulcher-Rood et al., 2018), creating a tension between norm-referenced data and more ecologically valid sources of information. This tension may be particularly pronounced in South Africa, where clinicians frequently assess children whose linguistic and cultural backgrounds differ substantially from those represented in available normative datasets. In multilingual contexts such as South Africa, where no single assessment approach adequately captures children’s linguistic repertoires, the complexity of diagnostic assessment has led to recommendations for multi-method approaches. Research emphasises that no single assessment tool is sufficient for diagnosing language disorders, particularly in multilingual populations (Bedore & Peña, 2008). Instead, clinicians integrate multiple data sources to build a comprehensive profile of a child’s language abilities. Approaches such as dynamic assessment have been proposed as more equitable alternatives, as they focus on learning potential rather than static performance (Peña et al., 2001). Assessment practices are shaped by systemic and contextual constraints. South African SLPs often work in resource-constrained environments characterised by high caseloads, limited assessment time and restricted access to culturally appropriate tools (Pascoe & Singh, 2023). Similar challenges have been reported internationally (Hunt et al., 2025), contributing to more streamlined assessment processes and reliance on readily available tools. International evidence further suggests that assessment practices vary across service settings (Fulcher-Rood & Castilla-Earls, 2023), highlighting the importance of understanding how South African clinicians negotiate diagnostic decision-making within their own service contexts. An additional challenge lies in the interpretation of assessment outcomes. Standardised scores are often used as primary indicators of impairment; however, their interpretation is problematic when applied to children whose linguistic and cultural backgrounds differ from those of the normative sample (Carter et al., 2005; Thordardottir, 2015). Clinicians may therefore draw on alternative strategies, including clinical judgement and integration of multiple data sources. Diagnostic decisions typically emerge from the synthesis of diverse and sometimes conflicting information, requiring significant clinical expertise (Hunt et al., 2025; Thomas et al., 2019). Clinician confidence is closely linked to these complexities. In contexts where assessment tools are not well suited to the populations being assessed, clinicians may experience uncertainty in their diagnostic decisions (Khoza-Shangase & Mophosho, 2021). International findings similarly indicate reduced confidence when assessing multilingual children, particularly in the absence of appropriate training and resources (Hunt et al., 2025). Despite growing recognition of the mismatch between available assessment tools and the multilingual realities of South African children, there remains limited empirical evidence describing how South African SLPs conduct diagnostic language assessments in practice. Existing research has largely focused on the mismatch between available tools and linguistic diversity (Mdlalo et al., 2016; Van Dulm, 2013), with less attention given to assessment workflows, including screening practices, areas assessed, time and cost considerations, and decision-making processes. International evidence further suggests that assessment practices may vary across service settings (Fulcher-Rood & Castilla-Earls, 2023). This study aims to provide a comprehensive overview of diagnostic language assessment practices among South African SLPs. Specifically, it examines the tools used, screening and assessment workflows, areas of language assessed, time and resource considerations, scoring and interpretation practices, and clinician confidence in diagnostic decision-making. The study sought to answer the following question: How do South African SLPs conduct diagnostic language assessments in multilingual and resource-constrained contexts? By situating these practices within multilingual and resource-constrained contexts, the study contributes to a more contextually grounded understanding of language assessment in South Africa.

The findings of this study illustrate a diagnostic landscape characterised by variability, fragmentation and professional adaptation, reflecting the structural realities of practising in a linguistically diverse and resource-constrained country. These patterns reflect systemic gaps in the availability and cultural relevance of diagnostic tools in South Africa (Pascoe & Singh, 2023; Salako, 2025; Southwood et al., 2025).

A central interpretive insight is that SLPs are navigating two competing demands: the requirement to provide accurate, defensible diagnostic decisions and the lack of validated, culturally fair tools. The high reliance on tools such as the CELF, TACL and Renfrew tasks – none of which are normed on South African populations – reflects the absence of viable, locally normed alternatives. This suggests that SLPs are working within a diagnostic ecology shaped by historical inequities in the development of assessment tools and the dominance of monolingual English norms (Pascoe & Singh, 2023; Van Dulm, 2013; Van Dulm & Southwood, 2008). This reflects broader structural constraints identified across African contexts (Salako, 2025). The pervasive use of informal measures and clinician-developed tools reflects adaptive, context-responsive practice. Speech-language pathologists appear to draw on linguistically attuned approaches such as narrative assessment, observation and language sampling, which align more closely with multilingual development than standardised tools (Higgins & Lefebvre, 2024; Verdon et al., 2015). However, informal measures may introduce variability when not supported by systematic frameworks (Bawayan & Brown, 2022). Moderate clinician confidence reflects the limits of relying on individual expertise within constrained systems (Thomas et al., 2019). The screening practices reveal a similar tension. Screening practices varied, reflecting tension between standardisation and contextual responsiveness. Speech-language pathologists appear to favour flexible approaches, suggesting that imported screening models may not align with multilingual contexts. This reflects broader concerns regarding the cultural validity of standardised frameworks (Bedore & Peña, 2008; Carter et al., 2005). Interpretation practices provide an important insight into how clinicians navigate diagnostic uncertainty in multilingual contexts. Rather than representing a weakness, the dominance of clinical judgement may reflect an appropriate and necessary strategy when standardised measures are not culturally or linguistically valid. Multilingual development, variable language exposure and cultural communication norms challenge traditional psychometric models (Bedore & Peña, 2008; Bortz, 2023; Paradis et al., 2021), requiring clinicians to integrate information from multiple sources rather than rely on test scores alone. The low use of monolingual comparison groups further suggests awareness of their limited validity and supports recommendations that diagnostic decisions should be based on the synthesis of standardised measures, language samples, caregiver report, observations and informed clinical judgement (Thomas et al., 2019; Verdon et al., 2015). The challenges identified by clinicians – particularly linguistic bias, a lack of adapted tools and difficulties in interpreting bilingual children’s results – reflect interconnected systemic inequities. These findings support arguments regarding the Eurocentricity of Speech-Language and Hearing training and tools (Khoza-Shangase & Mophosho, 2018, 2021). Finally, the data provide insight into the relationship between multilingualism and diagnostic certainty. Clinician confidence was generally modest, reflecting broader systemic constraints. Similar patterns have been reported in South African contexts (Mdlalo et al., 2016; Pascoe et al., 2010). Taken together, these findings indicate that improving diagnostic language assessment in South Africa will require more than the development of new tools. It will necessitate a broader reframing of assessment theory, training and policy to recognise multilingualism as the norm. This aligns with broader calls for culturally grounded, contextually responsive frameworks (Salako, 2025).

This study reveals a diagnostic landscape characterised not by inconsistency, but by clinicians’ adaptive responses to systemic constraints and the enduring inadequacy of monolingual assessment models in a multilingual society. SLPs in South Africa navigate a complex intersection of linguistic diversity, limited access to culturally valid tools and structural inequities that shape both their assessment choices and their diagnostic confidence. The findings underscore the need for a fundamental reorientation of assessment theory, training and policy – towards culturally grounded, multilingual frameworks (International Expert Panel, 2017) that reflect the realities of South African clients. Achieving equitable diagnostic language assessment will require sustained investment in the development, validation and implementation of locally relevant assessment tools and multilingual assessment practices. Meaningful progress will depend not only on creating new tools, but also on transforming the epistemological foundations of diagnostic language assessment to recognise multilingualism as central to, rather than peripheral in, clinical practice.