Home Private Practice The How and Why of Collecting a Language Sample

The How and Why of Collecting a Language Sample

by Scott Prath
written by
Male teacher assists elementary age girl with homework in school classroom or library setting.

SLPs reading this essay: I have a favor to ask you. Please suspend judgment when reading this post.

I’m sharing a time-saving and accurate way to complete one of our most valued activities. And also one of the most despised. The dreaded language sample! I know. Many of us still need counseling from the language sampling rituals we endured in grad school. But things change, and technology and techniques improve, so please read on.

Language samples provide some of the most useful information we can gather about a child’s communication because it’s an immediate snapshot of:

  • Utterance length
  • Complexity
  • Articulation abilities
  • Narrative skills
  • Perspective-taking
  • Comprehension
  • Imitation
  • Direction-following abilities

Here’s what a language sample looks like:

“There once was a boy and a frog.” [initiation / character identification]

“Then, he jumped inside the box” [cohesive element / past tense / preposition use/ object name]

“They ran behind the trees” [ pronoun / past tense / preposition / article / plural use]

With a mean-length-of-utterance of 6.3. Beat that for data-taking!

As a bilingual SLP, I also find that language samples offer bonus details for the diverse populations I—and many of you—serve. No formal assessment can provide information I get from language samples, such as:

  • Second-language influence
  • Cultural differences
  • Low language experience

Yet, a language sample helps us sort this out.

4 Steps to quickly and accurately collect a language sample

Choose a wordless picture book. We can’t assume any level of literacy and it’s not what we’re testing. So choose a book with no words. The standard in our field is the Mercer Meyer Frog Books. Salt Software offers free, downloadable audio elicitations and transcripts in Spanish and English. Its database also contains thousands of children for you to compare against if you really want to get serious.

Some of my favorites include:

  • Frog on His Own by Mercer Mayer
  • Frog Goes to Dinner by Mercer Mayer
  • The Arrival by Shaun Tan
  • Tuesday by David Wiesner

Set up your phone or computer to record the language sample and type at the same time. While the child is telling the story, type it! I know this sounds unorthodox. However, you can get all or the majority of the sample typed while the child is talking. Record as well to check for lines you missed. You might surprise yourself by collecting enough detail from just the typed utterances. In many cases, 25 to 50 utterances can give you sufficient information.

Technology wise, download Audacity to your computer for free or use your phone’s voice recorder. Speech-to-text programs constantly get better.

Have a child tell the story and maybe retell a story. The difference between tell and retell is important. Let the child try to tell the story without any help.

If the story seems incomplete, we can’t assume there’s an issue yet. What if the child just doesn’t have experience telling stories? This is where the retell comes in.

Say: “That was a great story. Now I’m going to tell you a story and then you tell it back to me.”

The child’s retelling might be twice as long.

Easily analyze your language sample. You collected the language sample, so now what? Now you analyze the macro (big picture) and micro (details) of the sample. I helped create a cheat sheet (Assessment of Fictional Narratives) of the different aspects of language and did research on when they are expected to develop.

  • Begin by looking for the narrative components then compare with expected language development.
  • On the second page, check for expected grammatical components.
  • For articulation, count all words, count all unintelligible words, and divide to get a percentage.
  • Calculate average utterance length.

I posted a detailed case study and example of language sampling for those who want to see exactly what this looks like. So, here’s to effective language sampling!

This July, I will present a and lead a lab on this topic at ASHA Connect in Baltimore: “Difference or Disorder? Language Development in Bilingual English Learners” (SC04). Stop by and say “hi” if you are there.

Scott Prath, MA, CCC-SLP, is vice president of Bilinguistics in Austin, Texas, and serves a diverse caseload in school and early-childhood settings. He is also the author of “Literacy-Based Speech Language Therapy Intervention Activities,” and writes for The Speech Therapy Blog. He is vice president of public relations for ASHA’s Hispanic Caucus.  scott.prath@bilinguistics.com


Related Articles


Hamza Bashir Ahmad June 22, 2018 - 5:54 am


Frances Weldon August 30, 2018 - 7:54 am

Very Timely! 🙂

Comments are closed.