Examqo › Examqo library › Speech-language pathology and audiology › Child language
Language assessment: standardised tests, scores, percentiles and measurement error
A standardised test compares a child with peers on the assumption that performance is normally distributed; a standard score of 85 sits one SD under the mean (about the 16th percentile), and every score carries measurement error — so a test is combined with a free language sample and information from the child's environment.
Why it matters in the exam
Questions ask you to compute or interpret: a score of 85 on a test with mean 100 and standard deviation 15 matches about the 16th percentile; a score of 70 is two standard deviations below the mean; "language age" is the chronological age that matches the score obtained; and what to do when a girl scored 82, the standard error of measurement is 4 and the cut-off is 85 — the confidence interval includes the cut-off, so you cannot say for certain that she is below it and you add further information.
Also asked: screening is a quick first look that only answers whether a full assessment should follow, and it should have high sensitivity and specificity; a standardised test allows comparison with same-age norms, and a structured test and a free sample each add unique information; MPU (morphemes per utterance) is a measure of morphosyntactic command; and for a 9-year-old with average scores whose teacher reports retells that are hard to follow, discourse is assessed across genres. Genetic testing is not part of a language assessment.
Key points
- A standardised test assumes a normal distribution and compares with peers.
- 100 mean, 85 = 16th percentile, 70 = 2nd percentile.
- Measurement error: a confidence interval around the score.
- Screening only decides whether a full assessment should follow.
- A structured test and a free sample complement each other.
Summary table
| Concept | Definition | How to interpret |
|---|---|---|
| Standardised test | Comparison with same-age norms, assuming a normal distribution | A score reflects only the areas tested |
| Standard score (mean 100, SD 15) | 85 = one SD below the mean; 70 = two | 85 ≈ 16th percentile; 70 ≈ 2nd percentile; 115 ≈ 84th percentile |
| Measurement error and confidence interval | The true score lies in a range around the obtained score | If the range includes the cut-off — do not decide on the score alone |
| Language age | The chronological age matching the score | The gap from the child's age gives an idea of severity, but is less precise than a standard score |
| Screening | A quick first look that decides if a full assessment should follow | High sensitivity and specificity |
| Language sample and discourse | MPU, different genres (narrative, expository), conversation | Adds information the structured test does not give |
In the full version for subscribers
- A simple algorithm
- Memory aid
- Exam traps
- 5 review cards
- What matters for the exam
The full version is open to subscribers of: Speech and hearing (communication disorders).
Subscriber? Read the full topic Plans and prices 25 free questions
Sources
- McCauley, Assessment of Language Disorders in Children, 2001 — Basic measurement concepts: norms, reliability and validity
- ASHA Practice Portal, 2024 — Spoken Language Disorders: assessment
- Paul, Norbury & Gosse, Language Disorders from Infancy through Adolescence, 5th edition (2018) — Assessment principles
Sources are given for further reading; the text is written afresh and not copied from them.
Reviewed by the Examqo AI Professor · 2026-11-02. The text may contain errors; when in doubt, check the source.