Get 25% off on lifetime Praxis 5331 prep with code PASS25

00Days
:
00Hrs
:
00Min
:
00Sec
← All articles

Deep Dive · 7 min read

Language Sample Analysis: MLU, TTR, and Why the Praxis Loves It

When scores look fine but classroom language doesn't, the defensible answer is a language sample.

Written by The Praxis Path Editorial TeamLast verified Editorially reviewed

Cross-checked against the current ETS Praxis 5331 Speech-Language Pathology Test at a Glance and ASHA CCC-SLP standards.

Language sample analysis is the assessment tool the Praxis 5331 keeps rewarding. When a stem describes a child whose standardized scores look fine but whose classroom language does not, or a bilingual child with no valid norms, the defensible answer is usually a language sample.

Collecting a usable sample

  • 50–100 utterances is the common clinical target; shorter samples inflate variability.
  • Use more than one context — play, narrative retell, conversation — because syntax complexity shifts with task.
  • Follow the child's lead and avoid yes/no questions, which suppress utterance length.
  • Record and transcribe; live tallying loses the errors you most need.

MLU and what it actually measures

Mean length of utterance is total morphemes divided by total utterances. Count free and bound morphemes: "the dogs jumped" is 5. Do not count disfluencies, fillers, or the parts of an unanalyzed chunk like "gonna" when the child does not use "going to" productively. MLU is most valid roughly between 1.0 and 4.0 morphemes, which maps onto Brown's stages. Beyond that, syntactic complexity measures serve better.

Other measures the exam likes

  • Type-token ratio (TTR): different words divided by total words — a lexical diversity index sensitive to sample length.
  • NDW: number of different words, often more stable than TTR.
  • Percent grammatical utterances and finite verb morphology composite — sensitive markers for developmental language disorder.
  • Narrative measures: story grammar elements, cohesion, episode structure.

Why it beats a test score in some cases

Sampling produces authentic, functional, culturally fair data. For a child from a dialect or language background without representative norms, a standardized score can misclassify difference as disorder — the core issue in our bilingual assessment guide. Dynamic assessment and language sampling are the two answers the Praxis wants in those stems.

Turning analysis into goals

Pick targets that are emerging rather than absent, high frequency in the child's environment, and functionally useful. If the sample shows omitted past tense -ed and third person -s with intact MLU, target finite verb morphology in narrative contexts rather than drilling isolated words.

Practice applying this to case stems in our practice question set.

More Praxis 5331 prep from the Praxis Path library.

Free download

Grab the free Praxis 5331 cheat sheet

The 2-page high-yield PDF top-scorers actually use — cranial nerves, aphasia matrix, dysphagia stages, dysarthrias, and Brown's morphemes. Instant download.

No spam. Unsubscribe anytime.