A practical guide to language sampling in telepractice

A language sample gives an SLP a view of how a child communicates in everyday interaction. Standardized tests can show whether a student performs within expected limits on specific tasks, while a spontaneous sample reveals vocabulary, grammar, narrative organization, turn-taking, and repair strategies in connected speech.

Telepractice changes how that sample is collected. The clinician may be working through a screen, relying on a caregiver to manage materials, and navigating audio delays or limited access to the child’s natural environment. A well-chosen language sampling tool can make the process more consistent without making the interaction feel artificial.

The most useful workflow balances clinical judgment with efficient data collection. The tool should support—not replace—the SLP’s observation of communicative intent, participation, cultural-linguistic context, and the child’s comfort during a virtual session.

Choose a tool that fits the clinical question

Before opening a platform, spreadsheet, or digital form, define what you need to learn. A preschool language sample might focus on spontaneous utterances, grammatical morphemes, vocabulary diversity, and conversational reciprocity. A school-age sample may target narrative structure, complex syntax, inferencing, cohesion, or classroom discourse.

Different tools offer different levels of support. Some allow audio or video uploads and automatically segment speech. Others provide fields for transcription, coding, mean length of utterance, words per sentence, type-token ratio, or number of different words. A simple clinician-created template may be sufficient when the goal is to track a small set of individualized targets over several sessions.

Consider accessibility and workflow before selecting software. Check whether the tool works with your browser, supports secure data handling, allows manual corrections, and produces reports that are understandable to families and educational teams. Automated analysis can save time, but it may misinterpret unintelligible speech, dialect features, overlapping talk, or speech generated with augmentative and alternative communication.

Prepare the virtual environment

A strong telepractice sample begins with technical preparation. Test the microphone, camera, screen-sharing function, and recording settings before the appointment. Confirm whether the platform permits recording and whether your workplace requires separate written consent. Families should understand what will be recorded, where files will be stored, who can access them, and when they will be deleted.

Reduce distractions without trying to create a perfectly controlled laboratory setting. Ask the caregiver to choose a quiet area with good lighting and to place the camera so the child’s face, hands, and communication materials are visible. If the child uses a device, picture board, sign language, or speech-generating device, make sure the communication system is within reach and visible in the video frame.

Plan a backup if the connection becomes unstable. You might continue with audio only, collect a shorter sample, or ask the caregiver to complete a brief home interaction using a secure method approved by your organization. Avoid asking families to send recordings through personal messaging apps or unapproved cloud services. Telepractice efficiency should never come at the expense of confidentiality.

Elicit natural communication without overprompting

The clinician’s language shapes the sample. Too many direct questions can produce short, test-like answers, while too little support may leave a hesitant child unsure how to participate. Begin with familiar, motivating topics such as pets, favorite games, a recent activity, or a picture book. Follow the child’s lead and use comments that invite elaboration: “That looks exciting,” “Tell me what happened next,” or “I noticed something interesting in your picture.”

Use a consistent elicitation routine when you need to compare samples over time. For example, start with five minutes of play or conversation, include a shared picture sequence, and finish with a personal narrative. Keep prompts open-ended and allow enough wait time for processing. In a virtual session, pauses can feel longer because of transmission delay, so count silently before repeating or rephrasing.

Caregiver involvement should be planned rather than improvised. A parent or teacher can position materials, model access to AAC, or help the child remain engaged, but frequent corrections and answer-giving may reduce the validity of the sample. Explain that the goal is to observe communication, not to coach every utterance. Record contextual notes about who was present, what prompts were used, and any technical interruptions.

Clinical purpose	Useful sample activity	Measures or observations	Telepractice note
Early expressive language	Toy play, pretend routine, or shared book	Utterance length, vocabulary, communication functions, grammatical forms	Keep preferred objects near the camera
Narrative language	Wordless picture sequence or personal event	Story grammar, sequencing, cohesive ties, temporal words	Share images digitally or mail a simple picture set
Social communication	Conversation, game, or collaborative problem-solving	Topic maintenance, turn-taking, repair, perspective-taking	Watch for latency caused by audio delay
AAC-supported communication	Familiar activity using the child’s system	Mode selection, message length, initiation, partner response	Keep the device visible and avoid speaking for the child
Progress monitoring	Repeated structured sample	Change in target forms, vocabulary, participation, independence	Use similar prompts while documenting differences

Capture and code the sample carefully

A language sample tool is most useful when the transcript reflects what actually happened. Mark unintelligible segments, abandoned utterances, immediate repetitions, scripts, sound effects, and communication produced through AAC according to your established conventions. Do not silently “clean up” a child’s grammar. The errors and repairs are part of the clinical information.

Separate the child’s language from adult models and prompts. If the clinician says, “What is the dog doing?” and the child responds, “He running,” code the child’s production rather than treating the full exchange as one language unit. Note whether the utterance was spontaneous, imitated, prompted, or selected from a communication system. These distinctions help prevent inflated estimates of independent language ability.

Automated metrics should be interpreted alongside qualitative observations. Mean length of utterance can show growth in sentence complexity, but it does not describe pragmatic effectiveness or the communicative value of a short utterance. Number of different words may be influenced by topic familiarity, cultural experience, bilingual development, and the length of the sample. A brief sample can produce unstable results, especially for children who speak infrequently or need additional response time.

When analyzing speech sound production within a language sample, connect the findings to the child’s broader phonological profile rather than treating every error as an isolated target. A resource on evidence-based articulation can help guide decisions about assessment, treatment selection, and outcome measurement when connected speech reveals patterns that warrant closer examination.

Interpret results within the child’s context

The numbers generated by a digital language analysis tool are descriptive, not diagnostic by themselves. Compare performance with developmental expectations that match the child’s age, language exposure, dialect, communication mode, and participation demands. For multilingual children, gather samples in each relevant language when possible and work with an appropriately qualified interpreter or bilingual professional.

Telepractice may change the sample in ways that have nothing to do with language ability. A child may speak less because the camera feels unfamiliar, the audio cuts out, the caregiver is nearby, or the session occurs at the end of a demanding school day. Conversely, a favorite online game may produce unusually high engagement and vocabulary. Document these conditions instead of presenting a single score as a complete representation of functioning.

Look for patterns across several sources: the virtual language sample, caregiver report, teacher observations, classroom work, standardized assessment, and therapy data. If a child uses short utterances during the session but produces detailed narratives in a familiar home routine, that contrast is clinically meaningful. It may suggest differences in confidence, topic knowledge, partner support, or task demands.

Turn analysis into useful therapy decisions

After reviewing the sample, choose a small number of functional targets. A child who omits grammatical markers may benefit from recasts and focused stimulation during play. A student with weak narrative organization may need visual story maps, explicit teaching of causal language, and repeated retells. A child who struggles with conversational repair may practice requesting clarification, explaining a message in a new way, or using an AAC support when misunderstood.

Share findings in language that families and educators can use. Instead of reporting only that a student’s mean length of utterance increased, explain that the student is combining ideas more often during shared activities or giving more complete explanations in class. Include a short transcript excerpt when appropriate, with identifying information removed and the context clearly described.

Repeat sampling at meaningful intervals rather than every session. A consistent monthly or quarterly routine may provide a clearer progress picture than frequent brief checks that interrupt therapy. Use the same broad activity when comparison is important, while also collecting varied samples to see whether new skills generalize beyond the treatment context.

Build a repeatable telepractice routine

A dependable process reduces preparation time and improves data quality. Create a digital checklist that covers consent, equipment, activity selection, caregiver instructions, file naming, transcription conventions, and follow-up. Keep a small set of screen-shareable books, picture sequences, visual schedules, and motivating conversation prompts ready for different age groups.

Protect the child’s privacy throughout the workflow. Use initials or assigned identifiers in filenames, store recordings only in approved systems, and limit access to team members who need the information. If a student is discussed in a school meeting, share the clinical interpretation and relevant excerpts rather than distributing an entire recording without a clear purpose.

Use the following practices to keep virtual language sampling efficient and clinically sound:

• Define the communication behaviors you want to observe before choosing the activity or tool.
• Give the child processing time and use comments, visual supports, and open prompts instead of rapid-fire questions.
• Record technical issues, caregiver assistance, prompt levels, and changes in the child’s engagement.
• Review automated transcripts manually, especially for AAC, bilingual speakers, dialect differences, unintelligible speech, and overlapping conversation.
• Compare results across sessions and settings before changing goals or making eligibility decisions.

A language sample collected through telepractice can be both practical and meaningful when the technology remains in service of the clinical question. With careful preparation, respectful elicitation, transparent coding, and context-sensitive interpretation, the SLP can gather evidence that informs therapy while preserving the natural, interactive quality of communication.

Choose one familiar activity for your next virtual session, define two or three behaviors to monitor, and use a secure language sampling tool to capture the interaction. Then review the results with the family or educational team and turn the clearest pattern into an actionable therapy target.