Bias in Language Assessment: What Speech-Language Pathologists Needs to Know

Speech-Language Pathologists (SLPs) rely on language assessment to make informed and ethical decisions about children’s language abilities. However, assessment tools and procedures are not neutral by default. When a child comes from a culturally and linguistically diverse background, including many Indigenous children, multilingual, newcomers, the risk of bias increases, and that bias can affect everything from test performance to score interpretation, and to referral decisions.

Assessment bias matters because it can make a child appear less capable than they truly are. In language assessment, this can mean overidentifying language disorders such as Developmental Language Disorder when what we are really seeing is cultural difference, unfamiliarity, or a mismatch between the child’s experiences and the expectations built into the assessment process (Laing & Kamhi, 2003; Laurie & Pesco, 2023; Reynolds et al., 2021). Taylor and Payne (1983) described several types of bias that remain highly relevant today: examiner bias, situational bias, value bias, and linguistic bias. These forms of bias can occur even when clinicians have the best of intentions.

They are useful to know because they help us answer an essential question:

Is this child struggling with language, or are they being asked to perform in a way that does not reflect who they are and how they learn? 

1. Examiner Bias

Examiner bias happens when the person administering or interpreting the assessment brings assumptions, often unintentionally, into the process. These assumptions can shape how responses are judged, how much prompting is given, and what is considered “typical” or “concerning.”

This can happen in very subtle ways. For example, if an assessor expects a child to answer quickly, maintain eye contact, volunteer details, or use a certain type of vocabulary, they may interpret hesitation or different communication styles as a deficit. But some children are processing carefully, waiting to be sure, or communicating in ways that are more aligned with their home or community experiences.

For example, if a clinician expects a child’s story to follow a mainstream, goal-directed narrative structure, they may interpret a different storytelling style as disorganized or immature. But in some cultural contexts, storytelling may emphasize relationships, environment, shared knowledge, or emotional significance over the kind of linear structure often rewarded in school-based tasks. A child may not be demonstrating a deficit. They may be demonstrating a different narrative style (Westby et al., 2002).

Another example is that some Indigenous children may end a story at the high point or climax rather than include a tidy resolution. Research with Algonquin children found that some children told personal narratives that ended at the high point or climax, rather than including the type of resolution expected in mainstream story structure (Pesco & Crago, 1996). In a standardized framework, this might look like an incomplete or poorly organized story. But culturally, it may be a valid and meaningful way to tell good story.

Examiner bias can also show up when clinicians assume that standardized scores are equally valid for all children, even when the test was not normed on populations that reflect the child being assessed. If a child’s cultural or linguistic background is not well represented in the norming sample, the score may look objective while still being misleading. This is where clinical judgment matters deeply, but only if it is informed by cultural humility and awareness.

2. Situational Bias

Situational bias refers to how the testing context itself can disadvantage some children. This includes the physical setting, the testing format, the materials, and the social expectations embedded in the task.

Many language assessments are designed around one-on-one interactions with an unfamiliar adult in a quiet room, often using tasks that are highly decontextualized or school-like. For some children, especially those whose communication and learning are shaped by more relational, community-based, or observational ways of participating, this context may not reflect how they communicate best.

Narrative assessment is a strong example. A child may be asked to tell a story from a picture sequence showing experiences that are unfamiliar to them, such as a trip to a pool, a fast-food restaurant, or a carnival. If the child has little connection to the context or has not had opportunities to engage with that type of storytelling task, their performance may reflect unfamiliarity rather than language difficulty.

For example:

• A vocabulary task may include words tied to experiences more familiar to some socioeconomic or cultural groups than others.

• A listening comprehension task may assume background knowledge that a child has never had access to.

• A sentence repetition task may be influenced by how familiar the syntax or rhythm is to the child’s language variety.

• A child may appear less capable simply because the task format itself is unfamiliar.

  
  

Another important example is the one-on-one testing format itself. In some Indigenous communities, storytelling and communication may happen in more relational, group-based, or elder-led ways. Asking a child to independently generate language for an unfamiliar adult in a formal setting may not reflect how they communicate best (Ball, 2007; Westby et al., 2002).

Sometimes the child is not “failing the task.” The task is failing to reflect the child.

3. Value Bias

Value bias occurs when an assessment privileges one way of communicating over another. In other words, the test is built around what the dominant culture considers important, appropriate, or “good” language.

This is where we need to be especially careful, because many assessments are built around what mainstream school systems value.

In narrative assessment, that often means rewarding:

• clear story grammar

• linear sequencing

• cause-and-effect links

• internal state language

• conflict and resolution

  
  

These are useful school-based skills. But they are not the only signs of language competence.

Research has shown that some Indigenous children’s stories may place greater emphasis on community, relationships, humour, environmental context, and harmony rather than the highly goal-driven, cause-and-effect structure often expected in mainstream assessments (Peltier, 2014; Westby et al., 2002). These are rich and meaningful narrative features, but they may not score highly if the assessment only values one kind of “good story.”

And again, this is not just about narratives. Value bias can also show up when we:

• expect children to answer in complete sentences rather than accept shorter but meaningful responses

• reward children who explain their thinking solely in an academic manner

• assume that “good language” always looks like the discourse style most familiar in mainstream educational settings

  
  

A child may have strong ideas, solid comprehension, and excellent learning potential, but still underperform if the assessment only values one communication style. That is a huge issue in school-based decision-making, because children are often judged not just on what they know, but on whether they show it in the way the system expects them to.

4. Linguistic Bias

Linguistic bias happens when there is a mismatch between the language or language variety expected by the assessment and the language variety used by the child.

This can include differences in vocabulary, syntax, discourse style, or dialect. If a child uses a non-mainstream variety of English, a heritage language, or culturally influenced discourse patterns, they may be marked as incorrect even when their responses are linguistically appropriate within their own language system.

This is one of the most dangerous forms of bias because it can make normal variation look like impairment. A child may have strong underlying language abilities, but if the assessment only recognizes one “correct” way of speaking or structuring meaning, those strengths can be missed (Laing & Kamhi, 2003; Gutiérrez-Clellen & Peña, 2001). When a child’s language does not match the one embedded in the assessment, their performance can be underestimated in ways that seem objective but are actually deeply shaped by cultural and linguistic assumptions. That can affect:

• eligibility decisions,

• intervention planning,

• school support,

• and how adults understand the child’s learning potential.

----

This is why it is so important not to rely on one score, one task, or one interpretation. Language should always be understood in context. Bias cannot always be eliminated, but it can be reduced. That starts with asking better questions:

What Questions Should SLPs Ask?

• Is this child’s performance reflecting a disorder, or a difference?

• How familiar is this task, context, or topic for them?

• What cultural assumptions are built into this measure?

• Am I evaluating their language, or their fit with the test?

What Can SLPs do?

• practice cultural humility

• use multiple sources of information

• use dynamic assessment that measures learning potential (which is consistent across language learning)

• gather language samples and naturalistic observations

• discuss with families and educators about communication styles and expectations

• consult with other professionals and individuals who come from the same linguistic and cultural background as the child.

• consider how culture and context may be shaping performance

Because the goal of assessment should never be to see whether a child fits the test. The goal should be to understand the child as clearly and fairly as possible and making sure every child has a real opportunity to show what they know. This is how we provide the most appropriate support.

Are you interested in being certified on a culturally and linguistically diverse narrative assessment? Click here for training information and dates.

a Speech-Language Pathologist working with a child on a computer.

References

Ball, J. (2007). Aboriginal young children’s language and literacy development: Research evaluating progress, promising practices, and needs. Canadian Language and Literacy Networked Centre of Excellence.

Gutiérrez-Clellen, V., & Peña, E. (2001). Dynamic assessment of diverse children: A tutorial. Language, Speech, and Hearing Services in Schools, 32(4), 212–224. https://doi.org/10.1044/0161-1461(2001/019) 

Laing, S. P., & Kamhi, A. (2003). Alternative assessment of language and literacy in culturally and linguistically diverse populations. Language, Speech, and Hearing Services in Schools, 34(1), 44–55. https://doi.org/10.1044/0161-1461(2003/005)

Laurie, A., Pesco, D. (2023). Curriculum-based dynamic assessments of narratives for bilingual Filipino children. Language, Speech, and Hearing Services in Schools, 54(2), 489–503. https://doi.org/10.1044/2022_LSHSS-22-00117

Peltier, S. (2014). Assessing Anishinaabe children’s narratives: An ethnographic exploration of elder’s perspectives. Canadian Journal of Speech-Language Pathology and Audiology, 38(2), 174–193. https://cjslpa.ca/files/2014_CJSLPA_Vol_38/No_02/CJSLPA_Summer_2014_Vol_38_No_2_Paper_3_Peltier.pdf 

Pesco, D., & Crago, M. (1996). " We Went Home, Told the Whole Story to Our Friends": Narratives by Children in an Algonquin Community. Journal of Narrative and life history, 6(4), 293-321.

Reynolds, C. R., Altmann, R. A., Allen, D. N. (2021). The problem of bias in psychological assessment in mastering modern psychological testing. Springer, Cham. https://doi.org/10.1007/978-3-030-59455-8_15

Taylor, O. L., & Payne, K. T. (1983). Culturally valid testing: A proactive approach. Topics in Language Disorders, 3(3), 8–20. https://doi.org/10.1097/00011363-198306000-00005

Westby, C., Moore, C., & Roman, R. (2002). Reinventing the enemy’s language: Developing narratives in Native American children. Linguistics and Education, 13(2), 235–269. https://doi.org/10.1016/S0898-5898(01)00063-8