Language Samples (Exam 2)

What is a Language Sample?

A language sample is a “snapshot” of a child’s language skills in a more natural environment.

What is Included in a Language Sample?

• Transcribed utterances from observation, interactions of child’s language (does not have to be in IPA)

• Calculation of MLU or other measures

• Analysis of purposes of communication (no standardized test)

• Analysis of language components (5 components)

• Examples of language components

• Analysis of speech sound production in connected speech

Why do we need a Language Sample?

• Standardized tests do not provide all the information needed to assess language skills (no functional language skills).

• Some children cannot complete standardized testing (e.g., can’t follow directions, test anxiety), making results not representative.

• Language samples allow us to analyze the components and purposes of communication.

What kinds of information do we get from a Language Sample?

• Utterance length

• Complexity

• Articulation skills

• Narrative skills (pragmatics)

• Perspective-taking

• Comprehension (e.g., “put baby under blanket,” knowing the word "under")

• Imitation (a way to track new skills)

• Direction-following abilities

Language Sample Analysis - Mean Length Utterance (MLU)

Average number of morphemes in an utterance (most used and easy; minimum of 50 utterances or more i s better).

Language Sample Analysis - Number of Different Words (NDW)

How many types of words are used (e.g., nouns, verbs, negatives, conjunctions)

Language Sample Analysis - Total Number of Words (TNW):

How many total words are in the sample?

Language Sample Analysis - Type-Token Ratio (TTR):

NDW/TNW (type of words/total number of words = percentage of different types of words); measures lexical diversity.

Language Sample Analysis - Percent Grammatical Utterance (PGU): 

How many utterances are grammatically correct in a sample? Calculated as (total utterances-ungrammatical utterances)/(total utterances) * 100%

Components of Language in a Sample: Phonology

• Sound errors

• Phonological processes (phonological patterns to simplify words)

Components of Language in a Sample: Morphology

• Grammatical/inflectional morphemes (Brown's morphemes)

• Derivational morphemes (prefixes and suffixes)

Components of Language in a Sample: Syntax

• Length of utterances

• Grammatical complexity of utterances

• Note: Infants do not demonstrate syntax; they don't make sentences.

Components of Language in a Sample: Semantics 

• Semantic feature analysis

• Word knowledge (receptive and expressive lexicon)

Components of Language in a Sample: Pragmatics

• (conversational skills, narrative skills)

• Inability to tell a story in the right order (or unable to tell at all)

• Purposes of communication

Semantic Categories (Schwartz & Chapman): Agent-Object

• e.g. "Boy hat", "Daddy coat", "Mommy doll"

Semantic Categories (Schwartz & Chapman): Agent-Action

• e.g. "She ate", "He jumped", "Dad fell", "Mom ran"

Semantic Categories (Schwartz & Chapman): Action-Object

• e.g. "Eat pie", "Toss football", "turn wheel", "fasten seatbelt", "add sugar"

Semantic Categories (Schwartz & Chapman): Entity-Attribute

• e.g. "Daddy tall", "Baby happy"

Semantic Categories (Schwartz & Chapman): Entity-Possession

• e.g. "Mommy cookie", "Daddy coat"

Semantic Categories (Schwartz & Chapman): Entity-Locative:

• e.g. "Doggie bed"

Rules for Counting Morphemes: RULE #1:

• Only fully transcribed utterances are used. Blanks that could not be understood are not counted (if you can understand the word, write it down).

  • e.g. “He X here.” = 2 Morphemes

Rules for Counting Morphemes: RULE #2:

• Include all exact utterance repetitions with a plus sign. If stuttering occurs, only count the attempt at the single word once in the most completed form.

  • e.g. “I (w) (w) want this” = 3 Morphemes

Rules for Counting Morphemes: RULE #3:

• Do not count fillers (e.g., mhm, uh, hm, um, ah, etc.), BUT do count "hi," "yeah," and "no."

Rules for Counting Morphemes: RULE #4:

• Count compound words, proper names, and ritualized reduplications as one single word.

  • e.g. "basketball", "lifetime", "airport", "birthday" (compound)

  • e.g. "choo-choo", "night-night", "quack-quack" (ritualized reduplication)

  • e.g. "John Smith", "Mary Miller" (proper nouns)

Rules for Counting Morphemes: RULE #5:

• Count ALL irregular past tense forms of verbs as one morpheme (e.g., swam, got, did, said, knew, saw, caught, won). Count ALL irregular plurals as one morpheme (e.g., men, feet, sheep, mice).

Rules for Counting Morphemes: RULE #6:

• Count all diminutives as one morpheme.

  • e.g. "doggie", "mommie", "birdie", "fishy"

Rules for Counting Morphemes: RULE #7:

• Count auxiliaries as separate morphemes (helper verbs like is, will, have, can, would, be). All catenatives are counted as single morphemes (e.g., gonna, wanna, gotta). All inflections are counted as separate morphemes (Brown's morphemes).

Inflection Examples:

• Possessive (-'s): e.g. "Dad’s hat", "Tina’s cookie", "the teachers’ lounge"

• Plural (-s): e.g. "dogs", "horses", "computers", "books"

• Third person singular (-s): e.g. "He runs", "she jumps", "it goes"

• Regular past (-ed): e.g. "hopped", "dragged", "loved", "stopped", "watched"

• Progressive (-ing): e.g. "jumping", "turning", "going", "running"

Rules for Counting Morphemes: RULE #8:

• The range count follows all these rules, but it should always be calculated for the total transcription, NOT for 100 utterances.