How to Calculate Mean Length of Utterance (MLU)
Accurately determine the Mean Length of Utterance (MLU) for speech-language assessments. Enter the total morphemes and utterances below to calculate the score, identify Brown’s Stage, and compare against normative age data.
Fig 1. Comparison of calculated MLU against Brown’s Normative Stages.
| Stage | MLU Range | Approximate Age (Months) | Key Milestones |
|---|---|---|---|
| I | 1.0 – 2.0 | 12 – 26 | Basic Semantic Roles |
| II | 2.0 – 2.5 | 27 – 30 | Grammatical Morphemes |
| III | 2.5 – 3.0 | 31 – 34 | Simple Sentences |
| IV | 3.0 – 3.75 | 35 – 40 | Embedded Sentences |
| V | 3.75 – 4.5 | 41 – 46 | Compound Sentences |
| V+ | 4.5+ | 47+ | Advanced Syntax |
What is How to Calculate Mean Length of Utterance?
Learning how to calculate mean length of utterance (MLU) is a fundamental skill for Speech-Language Pathologists (SLPs), linguists, and researchers focused on child development. MLU is a quantitative measure of linguistic productivity. It represents the average number of meaningful units of language (morphemes) a child produces per utterance.
The concept was popularized by Roger Brown in the 1970s as a reliable index of syntactic development in early childhood. Unlike simple word counts, understanding how to calculate mean length of utterance allows professionals to gauge the grammatical complexity of a child’s speech. A higher MLU typically indicates more advanced grammatical skills, such as the use of prefixes, suffixes, and complex sentence structures.
This metric is most accurate for children developing language normally between the ages of 18 months and 5 years, or up to an MLU of approximately 4.0 to 5.0. Beyond this point, sentence complexity grows without necessarily adding length, making MLU less reliable for older children.
How to Calculate Mean Length of Utterance: Formula and Logic
The core mathematics behind how to calculate mean length of utterance is straightforward, yet the rules for counting can be intricate. The basic formula is a simple ratio:
To perform this calculation correctly, one must understand the variables involved:
| Variable | Definition | Example |
|---|---|---|
| Total Utterances (N) | A complete spoken thought or sentence, marked by a pause or breath. | “I want cookie” (1 utterance) |
| Total Morphemes (M) | The smallest units of meaning (words + grammatical markers). | “Walked” = 2 morphemes (walk + ed) |
| MLU (m) | The average length calculated from the sample. | 3.5 (Mid-Stage IV) |
Key Counting Rules (Brown’s Rules)
When learning how to calculate mean length of utterance, you must adhere to counting protocols to ensure validity:
- Count as 1 Morpheme: Compound words (birthday), proper names (Uncle Bob), reduplications (choo-choo), irregular past tense (went, saw).
- Count as 2 Morphemes: Possessives (‘s), plural nouns (s), third-person singular (s), regular past tense (ed), present progressive (ing).
- Do Not Count: Fillers (um, ah), false starts, or unintelligible speech.
Practical Examples of How to Calculate Mean Length of Utterance
To better understand how to calculate mean length of utterance, let’s look at two realistic scenarios encountered in clinical practice.
Example 1: Early Language Learner
An SLP collects a 50-utterance sample from a 24-month-old child. After transcribing the audio, the data shows:
- Total Utterances: 50
- Total Morphemes: 85
Calculation: 85 / 50 = 1.70
Interpretation: This places the child in Brown’s Stage I (1.0 – 2.0), which is age-appropriate for a 24-month-old. The child is primarily using single words and simple two-word combinations.
Example 2: Advanced Preschooler
A sample is taken from a 4-year-old (48 months) suspected of language delay. The transcript reveals:
- Total Utterances: 100
- Total Morphemes: 320
Calculation: 320 / 100 = 3.20
Interpretation: An MLU of 3.20 corresponds to Brown’s Stage IV. While Stage IV is typical for children around 35-40 months, a 48-month-old would typically be expected to be in Stage V or V+ (MLU > 3.75). This result suggests a potential delay in syntactic development, prompting further assessment.
How to Use This MLU Calculator
This tool simplifies the process of how to calculate mean length of utterance by automating the division and stage-matching. Follow these steps:
- Collect Sample: Record a language sample of at least 50-100 intelligible utterances.
- Transcribe & Count: Use standard protocols (like Brown’s or SALT) to count total utterances and total morphemes manually.
- Enter Data: Input the “Total Number of Utterances” and “Total Number of Morphemes” into the calculator fields above.
- Optional Age Comparison: Enter the child’s age in months to visualize their performance against the standard developmental curve on the chart.
- Analyze: Review the calculated MLU, the identified Brown’s Stage, and the predicted age range to inform your clinical decision-making.
Key Factors That Affect MLU Results
When mastering how to calculate mean length of utterance, it is crucial to recognize that the number is influenced by more than just linguistic ability. Several external factors can skew results:
- Sample Size: While 50 utterances is the minimum standard, 100 utterances provide a more reliable representation of the child’s true ability. Small samples may miss complex structures.
- Context & Setting: A child may produce shorter sentences in a rigid testing room compared to free play at home. Context greatly impacts linguistic output.
- Interlocutor Style: If the adult asks many “Yes/No” questions, the child’s MLU will artificially drop (e.g., answering “Yes” is MLU 1.0). Open-ended questions encourage longer responses.
- Dialectal Differences: Certain dialects (like AAE) have different morphological rules (e.g., zero copula). Standard counting rules may penalize these dialects if not adjusted.
- Child’s Temperament: A shy or tired child may speak less or use simpler sentences, resulting in a score that underestimates their competence.
- Counting Rules Used: Variations exists (e.g., counting “gonna” as 1 vs. 2 morphemes). Consistency in applying counting rules is vital for longitudinal comparison.
Frequently Asked Questions (FAQ)
A typical 3-year-old (36 months) usually falls within Brown’s Stage IV, with an MLU ranging roughly between 3.0 and 3.75. However, individual variation is common.
It provides an objective baseline for syntactic development. It helps SLPs diagnose language disorders, set therapy goals, and monitor progress over time.
Generally, no. MLU is most valid up to age 5 or an MLU of 5.0. For older children and adults, measures like T-units or C-units are more appropriate as sentence length stops correlating directly with complexity.
In standard Brown’s rules, catenatives and negatives like “don’t”, “won’t”, and “can’t” are often counted as single morphemes initially until the child demonstrates they can separate the segments (e.g., using “do” and “not” separately).
Not necessarily. A low MLU can result from shyness, disinterest in the topic, or a bad day. It should always be interpreted alongside other assessments and observations.
Strictly speaking, MLU-M (Morphemes) is the standard for syntactic development. MLU-W (Words) exists but is less sensitive to grammatical growth (e.g., “cats” is 1 word but 2 morphemes).
Research suggests 50 is the minimum for a usable sample, but 100 is preferred for clinical reliability when determining how to calculate mean length of utterance.
Unintelligible utterances should be excluded entirely from the count. Including them would skew the ratio since you cannot accurately count the morphemes within them.
Related Tools and Internal Resources
Expand your SLP toolkit with these related resources found on our site:
- Speech Therapy Assessment Tools – Comprehensive list of digital tools for modern SLPs.
- Child Language Milestones Guide – Detailed breakdown of communication skills by age.
- Brown’s Stages Printable Chart – A quick-reference PDF for clinical settings.
- Guide to Morpheme Counting Rules – Deep dive into the specific rules for bound vs. free morphemes.
- Early Signs of Language Disorders – What to look for when MLU is low.
- Data Collection Templates for SLPs – downloadable sheets for tracking session data.