Language Sample Analysis Notes (SLP)

Milestones and Expected Language Development

By 12 months (one year), expected to have roughly between 1 and 3 words in expressive production. $ext{Expected words at }12 ext{ months} <br /> ightarrow [1,3]$
By 24 months (two years), expected to produce two-word utterances. If a child at this age has an MOU (mean utterance length) of less than two words, milestone concerns arise.
By 30 months, expected to have two to three word utterances.
If a 24-month-old uses only one-word utterances, you should consider whether they have missed the milestone for combining words.
For older children, two-year-olds using two to three word utterances reflect typical development; a discrepancy prompts a different assessment plan.
General rule of thumb used in practice: assess language development relative to age norms and milestones to determine the focus of the assessment and intervention plan.
Mean Utterance Length (MOU) mention: when describing an individual with MOU = 1, there are no bound morphemes to analyze, so syntax analysis is limited in that moment; semantics and pragmatics can still be analyzed.
When you encounter a child with a single-word utterance, you can still collect the language sample but interpret it with a focus on semantics and pragmatics; syntax may be limited due to lack of multiword utterances.

Purpose and Context of Language Sample Analysis (LSA)

Language sample analysis (LSA) is used to analyze semantic, syntactic, and pragmatic aspects of language. The authors recommend looking at semantic information first, followed by syntax, then pragmatics (though this order is a recommendation, not a strict requirement).
LSA captures language in a naturalistic setting and is considered informal testing because there is no single normative sample or standardized procedures for collecting and analyzing the data.
Formalized tests (e.g., OWLS, PLS) have standardized procedures and norms; LSA provides complementary information and can be used when normative data are lacking or when culturally/linguistically diverse children are involved.
Federal law requires multiple data sources to determine a child’s eligibility for special education services; LSA can serve as one data source alongside standardized tests.
LSA is useful for measuring how a child’s language abilities have generalized in therapy, as a benchmark, or when language samples are needed for CLD populations where standardized tests may be inappropriate.

Informal vs Formal Testing in Language Evaluation

Informal testing (language sampling) is flexible and context-dependent; clinicians decide how to collect the language sample, which settings, and what analysis tools to use.
Formal testing involves standardized measures and normative data; however, it may not always capture a child’s functional language in daily life.
LSA is typically part of a broader assessment; it is rarely used in isolation.
In culturally/linguistically diverse populations, LSA may be particularly valuable because standardized tests may not have representative normative samples.

When to Use Language Sample Analysis

In comprehensive evaluations to supplement formal tests.
As a benchmark after therapy to see if skills have generalized.
When an individual’s first language is not well represented in available standardized tests; LSA can provide usable data for that language.
To inform intervention goals and objectives by revealing which areas are not meeting typical developmental trajectories.
To support decisions about eligibility for special education services by contributing multiple data points across modalities.

How LSA Addresses Semantic, Syntactic, and Pragmatic Language

Semantic analysis: what is being said, vocabulary, word meanings, and conceptual content.
Syntactic analysis: word order, sentence structure, and morphological markers (bound morphemes).
Pragmatic analysis: language use in social contexts, topic maintenance, turn-taking, repair of breakdowns, and intent.
The analysis often proceeds with an interpretation framework that includes developmental norms and clinician expertise; results are synthesized to determine typical, delayed, or disordered language.

Practical Considerations for LSA Data Collection

LSA uses naturalistic communication to reveal language production in context, not just elicited responses in a test situation.
A language sample is typically not collected in isolation; it is part of a broader assessment.
In some cases, a child’s MLU (mean length of utterance) or morphophonological development can be limited if only one-word utterances are produced.
For CLD children, LSA can help circumvent limitations of standardized tests that may not capture the child’s dominant language.

Language Sample: Data Sources and Norms

Norm-referenced tests provide comparison to normative groups but have limited opportunities and contexts for exploring a child’s full linguistic repertoire.
Language samples provide rich qualitative data and may reveal skills not captured by standardized tests (e.g., use of specific word classes, pragmatics, discourse abilities).
A typical approach is to combine LSA with formal tests to form a comprehensive view of language abilities.

Collecting a Representative Language Sample

A representative sample reflects how the child typically uses language across contexts and may include language below, at, or slightly above their typical level.
The sample should reflect daily communication and include a range of language use, not just high-level performance.
A representative sample helps avoid misinterpretation due to an unrepresentative snapshot of language.

Factors That Impact Representativeness

Nature of interaction: who the child is interacting with (clinician, parent, peer) and whether the partner engages the child in conversation or just asks questions.
Setting: clinic, home, school, or other environments; differences across settings can change language use.
Materials used: familiar vs unfamiliar toys; single-object vs multi-object pictures; scenario complexity in picture books or toys.
Sample size: number of utterances; more utterances generally yield a more representative sample.
Recording method: video vs audio; video captures nonverbal context and turn-taking, which are important for pragmatics; video may distract some children.
Context specification: the situation, prompts, and the level of support provided during sample collection.

Interaction Partners and Flexibility in Sampling

The partner can be the clinician, a parent, a peer, etc.; the child’s language can vary significantly with different partners.
Collect multiple samples with different partners to capture a broader range of language abilities.
Sample collection often involves interactive play to maximize language output; however, questions limit spontaneity and may narrow the range of language.
Training parents to avoid overly directive questions improves sample representativeness when parents are involved.

Using Play and Structured Prompts for Language Sampling

Structured play activities can help elicit a broader range of language, but if a child refuses to engage or only engages in one activity, the sample may be limited.
Use a variety of play activities and prompts to maximize language production (e.g., farm toys, city setups, grocery/play scenarios).
If a child’s preferred activity is not the planned prompt, follow the child’s lead to maintain naturalistic discourse and maximize language output.
Book readings or wordless picture books can be used, but may not suit all children; familiarity with reading practices at home influences success.
For older children, personal narratives, story retells, and expositions may yield a more representative sample.

Recommended Settings and Prompts

Use multiple locations: home, school, clinic, and residential facilities when possible.
Obtain a sample of roughly fifteen to twenty minutes per setting, with the goal of collecting 50–100 utterances total across sessions.
If time-limited, aim for 50 utterances; if possible, collect 50–100 utterances for a representative sample.
If a sample falls short of target utterances in one session, consider adding another sampling session or combining samples from multiple settings.
Use multiple prompts to elicit varied discourse: e.g., “Tell me something fun you did this weekend,” “Tell me about your favorite vacation,” etc.
Obtain parent-provided samples (e.g., at home) with clear guidance on when the child is most communicative and what activities generate language.
Record the context of the language sample: who was present, what activity occurred, the setting, and any external events that influenced language use.

Recording and Transcription Methods

Videotaping is preferred because it provides audio and visual context; audio alone misses nonverbal cues and contextual information.
If video is impractical, high-quality audio recording is acceptable; ensure proximity to the microphone and minimal movement that disrupts recording.
You must obtain written permission from parents to record; many institutions require consent forms as part of the assessment package.
Do not rely on personal devices for official recording in many schools; use school-issued devices (e.g., iPads) to store recordings.
Transcription should be done with a combination of real-time notes and later playback to fill in gaps; use online transcription when possible but verify against the recording.
If a word or utterance is unintelligible on the recording, note it as such and do not count it toward utterance counts; where possible, fill in from real-time notes or context.
The quality of the language sample analysis is heavily dependent on the quality of the sample and transcription; accurate coding is essential for reliable results.
Keep a detailed log of what occurred in the room (nonverbal cues, context, and any interruptions) to support interpretation of the language sample.

Transcription Rules and Utterance Coding (Key Rules Covered in Class)

Only fully intelligible child utterances are numbered and analyzed. Partially intelligible or totally unintelligible utterances are not numbered or analyzed.
A language sample should contain 50 or more fully intelligible utterances to be analyzable.
If a child’s utterance is interrupted or overlapped by another speaker, the interrupted portion may not be counted; context is critical.
Repetitions: if a child repeats an utterance with no intervening activity or no other speaker’s utterance, it is considered a repetition and not counted as a separate utterance.
- Example: If the child says, “The cow’s on the barn” and then continues with the same phrase after a pause and without another speaker’s interruption, that repetition does not count as a new utterance.
- If the child repeats after an intervening utterance (e.g., an adult adds “the horse is on the barn”), the repetition could be counted as a separate utterance depending on context.
Utterance boundaries are often marked with a slash at the end of each child utterance; boundaries are influenced by pause length (commonly a two-second pause) and contour.
Present progressives should be used for simultaneous action (e.g., “is eating” while doing something); simple present is used for actions that occur before or after an utterance.
Boundaries between child utterances are determined by pauses and the observed conversational turn-taking in the sample.
Emphasis or exclamations: an exclamation by an adult after a child utterance is marked with an exclamation point; if the child’s utterance is explanatory and followed by an exclamation point, a slash is added to indicate the utterance boundary.
Morphology and tense marking are important for syntactic analysis: count morphemes to compute MLU; for example, counting -ed, -ing, plural -s, possessives, and auxiliary forms.
The decision to count an utterance for MLU depends on whether the child uses bound morphemes; a single-word utterance contains one morpheme per word (approximately) unless it includes bound morphemes (which would be counted if present).
When calculating MLU, the numerator is the total number of morphemes in fully intelligible utterances and the denominator is the total number of utterances; for a child with one-word utterances, the MLU will reflect the number of morphemes per utterance accordingly.
C-units (communication units) are not used for very young children; for younger children, rely on pause-based boundaries, actions in play, and immediate contextual cues to delineate utterances.

Morpheme and Syntax Considerations (Brown’s Morphemes and Beyond)

Brown’s morphemes provide a framework for typical morpheme acquisition; however, in early stages, ING tends to appear before ED in some cases (the transcript notes that ING often emerges early in development).
If a child uses present tense verbs (e.g., walk, talk, drink) but not -ing, -ed, or other morphemes, you would plan intervention to introduce the next developmental step (e.g., -ing) before addressing more complex tense forms.
The sequencing for intervention often follows typical development: begin with -ing (present participle), then -ed (past tense), then other morphemes (e.g., plurals, possessives, third person -s).
Dynamic assessment can inform whether a child is likely to acquire certain morphemes with minimal facilitating support, but these decisions must be tested and documented via formal assessment data and analysis.

Interpreting Language Sample Results and Report-writing Considerations

The clinician is responsible for evaluation and interpretation; supervisors must review your analysis and help ensure the client’s best care.
When used in a school setting, you should be prepared to justify your analysis and describe how you administered the test, how you collected and analyzed the language sample, and why you chose specific methods.
Language sample findings are often used to set intervention goals and objectives; however, norm-referenced tests are not ideal for directly deriving goals, so LSA data are critical to informing goals.
If a child’s norm-referenced results show potential deficits, LSA helps determine whether the child uses past tense, completes verb tenses, uses appropriate syntax, and demonstrates pragmatic skills, enabling precise goal setting.
Interpretation must consider the child’s developmental level, norms, and clinical expertise; a child can present typical language development in some contexts and delays in others.

Using LSA with AAC and Multilingual Contexts

When a child uses AAC (augmentative and alternative communication) devices, you can assess expressive use, including requests and sentence formulation, but the specifics of how to implement and interpret AAC-based language sampling are covered in a separate class.
LSA data can inform AAC goals (e.g., increasing use of subject-verb-object constructions) but the AAC-specific assessment and intervention strategies require specialized training.
For bilingual or multilingual children, LSA can provide valuable data when standardized tests may not be representative; however, it also requires careful interpretation of linguistic features across languages and consideration of cross-language transfer and influence.

Practical Tips for Building Competence as a Clinician

Be prepared to answer supervisor questions about how you will administer tests, collect language samples, and analyze results; the supervisor ensures client care and adherence to ethics.
Practice collecting representative samples across multiple scenarios, partners, and settings to develop flexibility and competence.
Plan to perform the transcription accurately, including tagging intonation, pauses, and pragmatic cues; consider using software (e.g., SALT) but be prepared to adjust or correct automatically generated data.
Ensure you have a backup plan for sample collection (e.g., multiple sessions) in case you fail to reach target utterance counts in a single session.
Clinicians should avoid giving a diagnosis before completing a full analysis and consulting with supervisors; even when results strongly suggest a diagnosis, final determinations require complete data synthesis.
Understand that the language sample collection process can be challenging; be ready to adapt to a child’s needs and use multiple prompts to elicit meaningful language data.

Myth Busting: Language Sample Analysis Is Not Too Hard to Learn

Breakout activity: discuss common myths about LSA (e.g., "LSA is too hard to learn").
Takeaways from group discussions: LSA is learnable with practice, and the benefits of a well-collected language sample far outweigh the initial effort required to master procedures.
Emphasize the value of practice with real transcripts and hands-on transcription to build confidence and skill.

Practical Considerations for Class and Practice

Rules for transcription and analysis (pages 13–14 of the week one materials) will be reviewed in the next session; you should review pages 13–14 ahead of class.
The class will cover more on the rules for transcription (8 through 18) and related topics; come prepared with questions and ideas for how to apply to your own future practice.
In this session, we highlighted: timing, sample size, partner variety, settings, material variety, and recording/transcription considerations as foundational to high-quality language sample analysis.

Final Notes and Next Steps

Remember: the quality of the language sample and its transcription directly impacts the accuracy of analysis and interpretation.
Strive for a representative sample by sampling across partners, settings, and tasks; aim for 50–100 fully intelligible utterances, or a time-based target (e.g., 15–20 minutes per session) when possible.
Prepare to discuss your sampling plan with supervisors and be ready to adjust based on the child’s responsiveness and the clinical setting.
Upcoming topics: continuing rules for transcription, c-units and their relevance for older or younger children, and more on how to apply LSA findings to intervention planning.

References to Upcoming Sessions

SALT software details and limitations; differences between manual vs computer-assisted analysis; accuracy considerations when using automatic counting of morphemes and utterances.
The role of LSA within a comprehensive evaluation (05/16 content); how to integrate LSA data into formulation and goals.
The ethics and professional practice guidelines (ASHA) that govern recording, sampling, and reporting.

Quick Reference Formulas and Rules

Mean Utterance Length (MLU): ext{MLU} = rac{ ext{Total morphemes}}{ ext{Total utterances}}
Utterance boundary cue: typically a pause > 2 seconds; boundaries may also be indicated by changes in activity or intonation.
Representation criteria: a representative sample may include utterances below and above the child’s typical level; a focus on naturalistic, everyday language.
Target utterance counts: aim for 50–100 fully intelligible utterances; 30-minute samples often yield 100–200 utterances in toddlers, but 50 utterances is a common practical minimum.
Repetition rule: repetition without intervening activity or another speaker is not counted as a new utterance; repetition with intervening activity can be counted as a separate utterance depending on context.