How To Calculate Mean Length Of Utterance

Calculate the average length of utterances in morphemes for language development analysis

Comprehensive Guide: How to Calculate Mean Length of Utterance (MLU)

Mean Length of Utterance (MLU) is a fundamental measure in child language development that quantifies the average length of a child’s spoken utterances. First introduced by Roger Brown in 1973, MLU has become one of the most reliable indicators of syntactic development in children aged 2-5 years.

Why MLU Matters in Language Development

• Developmental Milestone: MLU correlates strongly with grammatical complexity and vocabulary growth
• Clinical Utility: Used by speech-language pathologists to identify potential language delays
• Research Standard: Common metric in child language acquisition studies worldwide
• Cross-linguistic Application: Adaptable to multiple languages with proper morpheme analysis

Step-by-Step Guide to Calculating MLU

1. Collect Language Sample:

   Record 50-100 utterances from the child in a natural setting. The sample should be:

   • Spontaneous (not elicited through direct questioning)
   • Representative of typical communication
   • Recorded during play or daily activities
2. Transcribe Utterances:

   Write down exactly what the child says, preserving:

   • All words and word attempts
   • False starts and repetitions
   • Filler words (“um”, “uh”)

   Example: “I want… I want cookie please” should be transcribed exactly as spoken.

3. Segment into Utterances:

   Define utterances as:

   • Complete thoughts separated by pauses
   • Independent clauses
   • Not separated by conjunctions (“and”, “but”)

   Example: “I want cookie and milk” = 1 utterance (6 morphemes in English)

4. Count Morphemes:

   Analyze each utterance for morpheme count following language-specific rules:

   English Morpheme Rules	Example	Morpheme Count
   Regular plurals (-s, -es)	“dogs”	2 (dog + s)
   Past tense (-ed)	“walked”	2 (walk + ed)
   Present progressive (-ing)	“running”	2 (run + ing)
   Possessive (‘s)	“Mommy’s”	2 (Mommy + ‘s)
   Articles (a, the)	“the ball”	2 (the + ball)
   Contracted forms	“I’m”	2 (I + am)

5. Calculate the Average:

   Sum all morphemes and divide by total utterances:

   MLU = Total Morphemes ÷ Total Utterances

   Example: 5 utterances with morpheme counts of 3, 2, 4, 3, 5 = (3+2+4+3+5)÷5 = 3.4 MLU

MLU Norms by Age (English-Speaking Children)

Age (months)	Average MLU	Expected Range	Developmental Stage
18-20	1.5	1.0-2.0	Single words, some two-word combinations
24	2.2	1.8-2.7	Simple sentences, emerging grammar
30	2.8	2.3-3.3	Complex sentences, questions
36	3.5	3.0-4.0	Compound sentences, conjunctions
42	4.1	3.6-4.6	Advanced syntax, narratives
48+	4.5+	4.0-5.0+	Adult-like sentence structures

Common Mistakes in MLU Calculation

• Oversegmenting utterances:

  Incorrectly splitting compound sentences. Wrong: “I want cookie and milk” as 2 utterances. Correct: 1 utterance (6 morphemes).

• Undercounting morphemes:

  Missing bound morphemes like plural -s or past tense -ed. Wrong: “dogs” as 1 morpheme. Correct: 2 morphemes (dog + s).

• Including adult speech:

  MLU measures child utterances only. Adult prompts or repetitions should be excluded.

• Small sample size:

  Analyzing fewer than 50 utterances may not be representative. Aim for 50-100 utterances for reliability.

• Ignoring language specifics:

  Morpheme rules vary by language. Spanish, for example, counts verb conjugations differently than English.

MLU vs. Other Language Measures

Measure	What It Assesses	Advantages	Limitations
MLU	Syntactic complexity	Objective, quantifiable, cross-linguistic	Requires morpheme analysis, time-consuming
Vocabulary Size	Lexical development	Easy to measure, correlates with cognition	Doesn’t assess grammar or syntax
Type-Token Ratio (TTR)	Lexical diversity	Shows vocabulary variety	Sensitive to sample size
Developmental Sentence Score (DSS)	Grammatical structures	Detailed grammatical analysis	Complex scoring, less standardized
Index of Productive Syntax (IPSyn)	Syntactic structures	Comprehensive grammatical assessment	Time-intensive, requires training

Clinical Applications of MLU

Speech-language pathologists (SLPs) use MLU in several key ways:

1. Language Delay Identification:

   Children with MLU scores below age expectations may warrant further evaluation. Research shows that MLU below 1.5 at 24 months correlates with later language disorders (Paul & Smith, 1993).

2. Progress Monitoring:

   MLU can track intervention effectiveness. A study by Hadley & Holt (2006) found that children with specific language impairment (SLI) show MLU growth of 0.1-0.3 per month during intervention, compared to 0.4-0.6 in typically developing peers.

3. Treatment Planning:

   MLU guides therapy targets. For example:

   • MLU < 2.0: Focus on two-word combinations
   • MLU 2.0-3.0: Target simple sentences and question forms
   • MLU > 3.0: Work on complex sentences and conjunctions
4. Differential Diagnosis:

   MLU patterns help distinguish between:

   • Language delay (consistently low MLU across contexts)
   • Language difference (MLU appropriate in first language but not second)
   • Intellectual disability (slow MLU growth with other cognitive delays)

Expert Resources on MLU:

• National Institute on Deafness and Other Communication Disorders (NIDCD) – Speech and Language Development
• American Speech-Language-Hearing Association (ASHA) – Spoken Language Disorders
• Vanderbilt University – MLU Calculation Guide (PDF)

Advanced Considerations in MLU Analysis

For researchers and clinicians seeking deeper insights:

1. MLU in Bilingual Children:

   Bilingual MLU should be calculated separately for each language. Research by Gutierrez-Clellen & Kreiter (2003) shows that:

   • Bilingual children’s MLU in each language may be 0.3-0.5 lower than monolingual peers
   • Combined MLU (both languages) often matches monolingual norms
   • Language exposure percentage correlates with MLU in each language
2. MLU in Children with Disorders:

   Disorder	MLU Characteristics	Typical Pattern
   Specific Language Impairment (SLI)	Lower than age expectations	Slow growth rate (0.1-0.2/month vs. 0.4-0.6 typical)
   Autism Spectrum Disorder (ASD)	Variable, often lower	May show spikes in MLU with echolalia included
   Down Syndrome	Delayed but parallel growth	MLU trajectories similar to typical but shifted later
   Hearing Impairment	Depends on intervention	With early intervention, MLU can approach typical norms

3. MLU in Different Languages:

   Morpheme counting varies significantly across languages:

   • Spanish: Verb conjugations add multiple morphemes (e.g., “había comido” = 4 morphemes)
   • Mandarin: Tonal differences and classifiers affect counting
   • Finnish: Extensive case system increases morpheme count
   • Arabic: Root-and-pattern morphology requires specialized analysis
4. Automated MLU Calculation:

   Emerging technologies in natural language processing (NLP) are enabling:

   • Automatic transcription from audio samples
   • Morpheme tagging with 90%+ accuracy in some languages
   • Longitudinal tracking through mobile apps

   Tools like TalkBank provide platforms for automated MLU analysis in research settings.

Frequently Asked Questions About MLU

1. How many utterances are needed for reliable MLU?

   Research suggests:

   • 50 utterances: Minimum for clinical use (Miller & Chapman, 1981)
   • 100 utterances: Preferred for research reliability
   • 200+ utterances: For detailed longitudinal studies
2. Should I count repetitions in MLU?

   Standard practice is to:

   • Count immediate repetitions once (e.g., “I want I want cookie” = “I want cookie”)
   • Count delayed repetitions as new utterances if separated by other speech
   • Always count self-repetitions of single words (e.g., “more more” = 2 morphemes)
3. How does MLU relate to intelligence?

   MLU correlates with cognitive development but isn’t a direct IQ measure:

   • Typical correlation with nonverbal IQ: r = 0.3-0.5 (Fenson et al., 1994)
   • Children with intellectual disabilities often show delayed MLU trajectories
   • Some children with high MLU may have cognitive strengths in verbal domains
4. Can MLU be used for adults with aphasia?

   Modified MLU (mMLU) is sometimes used in aphasia assessment:

   • Focuses on grammatical complexity rather than length
   • Often combined with other measures like correct information units (CIUs)
   • Less standardized than child MLU but clinically useful

Future Directions in MLU Research

Emerging areas of study include:

• Neural Correlates:

  fMRI studies linking MLU growth to brain development in Broca’s and Wernicke’s areas

• Genetic Factors:

  Twin studies investigating heritability of MLU growth rates (estimates: 40-60%)

• Environmental Influences:

  Longitudinal studies on how parent child-directed speech affects MLU trajectories

• Cross-linguistic Databases:

  International collaborations to establish MLU norms across 50+ languages

• Early Identification:

  Machine learning models using MLU to predict later literacy outcomes with 85%+ accuracy