Download SUGAR Language Sample Analysis Made Sweet and more Study notes History in PDF only on Docsity!
SUGAR
Language Sample
Analysis Made Sweet
Dr. Stacey Pavelko, CCC-SLP
How we define a problem usually determines how we
analyze it. It sends us in a particular direction. And how
we analyze a problem — the direction we take —
absolutely determines whether we find a solution and
what the quality of that solution is.
Jones, 1998
Advance Organizer
A BRIEF LSA History
LSA use
Changes to LSA
What is robust sampling and how do you do it?
What is SUGAR
SUGAR QUICK Analysis
Roger Brown
Seminal study of early language development, entitled A First Language **(1973) *** Revolutionized the study of child language language development studies Changed the assessment of communication abilities of children with language impairment (LI). SLP’s have changed the basic methodology very little in the intervening four decades. *Taken from: Brown, R. (1973). A first language the early stages. Cambridge, MA: Harvard University Press.
Brown
Collected extensive language samples
Each sample consisted of at least 700 utterances per month Sample size only included 3 children
Documented information on early language form between ages 2 and 4
Different grammatical stages described by mean length of utterance (MLU)
Brown’s Participants
Three typically developing (TD) children Adam: 27 months old; parent’s education level- college Eve: 18 months old; father’s education- college ; mother’s education- high school Sarah: 27 months old; parent’s education level- high school Children were typically developing (TD) All from Cambridge, Massachusetts Parent occupations: Minister, Harvard graduate student, and clerk
LSA Research Questions
Do school-based SLPs use LSA in their clinical practice
What are the characteristics of the language samples SLPs collect?
What are the methods of analysis?
What are the barriers that school-based SLPs identify in using LSA?
Do SLPs use LSA?
Not so much…
33% did not use LSA AT ALL the 2013-2013 school year…
Frequency of LSA
10 or fewer Samples 55%
11-20 Samples 23%
21-30 Samples 9%
> 30 Samples 11%
* Of the 67% who reported using LSA
during the 2012-2013 school year
Length of Samples Collected
1-5 Minutes 27%
6-10 Minutes 46%
11 minutes or longer 27%
Number of Utterances Collected
Less than 25 utterances 11%
26-50 utterances 42%
51-100 utterances 35%
More than 100 utterances 12%
Recording Sample
Audio 43%
Video 17%
Neither* 52%
Transcribe what the child
is saying while the child is talking*
An Alternative Method
Best general indication of language use
More naturalistic than testing
Correlates well with results from many test formats
Only way to elicit some language features
Potentially represents everyday use
So… how do we
make LSA more
clinician friendly?
SUGAR!!
SUGAR changes
Robust Sample Transcription Rules Analysis Rules
Robust Sampling Study Can we change the quality of samples?
22 students each collected a language sample from a child ( = 51.36 months, SD = 12.14) Small group of student trainers Prepared handout on collecting sample emphasizing narrative elicitation Trained same 22 students via role-playing in elicitation techniques Six months after 1 st^ sample, same 22 students each collected a second language sample from a different child ( = 57.81, SD = 13.2)
Conversational Strategies
Use narrative elicitations instead of yes/no questions Build on what the child says or on what you know Begin with… Your mom says you…. That sounds like fun. Tell me what happened. I know that you…. Tell me what happened. Did you ever…. Tell me what you did.
Example of Robust Sample
- Listen to the following sample (transcript of examiner’s utterances will be on the next slide)
- What do you notice about how the adult interacts with the child?
- How long does she wait?
- How many comments does she make?
- How many examiner utterances?
Example of Robust Sample
- Listen to the following sample (transcript of examiner’s utterances will be on the next slide)
- What do you notice about how the adult interacts with the child?
- How long does she wait?
- How many comments does she make?
- How many examiner utterances?
What is a more robust sample?
- Engages the child in a conversation that gives the child the opportunity to produce complex language
- Allows the child to control the interaction
- Children are more communicative and use more complex language (Lapadat,1983; Norris & Hoffman, 1990)
Conversational Strategies
Turnabouts = Comment + Cue for child to talk Process Questions How did… What happened… Tell me… I wonder what you… Why did… More than one-word “why” questions Not appropriate for kids below 4.5 yrs
Conversational Strategies
Use narrative elicitations instead of yes/no questions Build on what the child says or on what you know Begin with… Your mom says you…. That sounds like fun. Tell me what happened. I know that you…. Tell me what happened. Did you ever…. Tell me what you did.
Rules for
Transcribing
- Type in plain English as spoken.
- Omit punctuation to save time.
- Do NOT embellish the student’s utterance. In
other words, don’t add morphemes that are
missing.
SUGAR Quick Analysis: Transcription
- Type words in full even when pronunciation omits
portions as follows:
- Talkin’ should be transcribed as “Talking”
- Gonna, wanna, gotta, hafta should be transcribed
as “going to, want to, got to, have to”.
- Type contractions as is. In other words, don’t should
be typed as “don’t” and I’m as “I’m”.
SUGAR Quick Analysis: Transcription
SUGAR Quick Analysis: Transcription
- Do NOT include fillers (uhhhh, ummm, like, you know).
- Do NOT include disfluences. Only include the fullest form of what the student actually said. Example: “He said…he says…he tell me secrets” becomes “He tell me secrets.”
- Do NOT include repeated words unless it is for emphasis, as in “He went down down down in the cave.”
SUGAR Quick Analysis: Transcription
- If the entire utterance is unintelligible, omit
it. If a word is unintelligible, type nonsense,
such as “XXX” in place of the word.
- Transcribe the sample directly into your
computer.
- Only type the student’s utterances, NOT
yours.
- Do NOT include identifying data.
SUGAR Quick Analysis: Transcription
- Set “Numbering”, found on the tool bar in the “Paragraph” section, to ensure that you only type 50 utterances. - Remember that an utterance is a sentence or less, separated by a pause, drop in voice, inhalation or combination of these.
Example
HOW TO ANALYZE
FOUR METRICS
- Quick Analysis Includes
- Total Number of Words
- Mean Length of Utterance (MLUS)
- Words per Sentence
- Clauses per Sentence
Quick Analysis Metrics
Let’s Review the
SUGAR Method
Handout
On copy 1
- Before doing any actual analysis, make sure the “Numbering” function is off. - Turn “off” the “Numbering function by highlighting the entire document(Control-A or Apple-A), going to the “Paragraph” section of the Toolbar and clicking on “Numbering”.
Total Number of Words (TNW)
On copy 1
- Nothing else should be on the page except the child’s 50 utterances.
- Word count is on the tool bar at the bottom of your screen.
- Record the number of words before moving on.
Total Number of Words (TNW)
SUGAR Quick Analysis: Words Per Sentence
- Make sure the “Numbering” function is off.
- Delete all utterances that are NOT sentences.
- Both a sentence and a clause contain a subject and a verb, as in Mommy walked. A sentence can have more than one clause, as in “ Mommy walked but I ran”. (2 clauses, 1 sentence).
- The critical element in a sentence is a verb.
SUGAR Quick Analysis: Words Per Sentence
- Once you have only sentences represented, record the total words from the word count section on the toolbar at the bottom of the screen.
- Switch on the “Numbering” function again.
- With this on, you can tell how many sentences you have. - Record this value for future use.
SUGAR Quick Analysis: Words Per Sentence
- Divide the number of words by the number
of sentences to get the mean
words/sentence. Record this value
somewhere. This is the number of words
per sentence
Words = 150
Sentences = 50
Words per sentence (150/50)=
Time is of the essence. No one is looking over your shoulder to see if each clause is exactly correct or that the remainder may be a partial clause.
Example: “ The boy who’s in my class is yukky ” consists of two clauses, “ The boy is yukky ” and “ Who is in my class .” Separate these as follow: The boy who’s in my class is yukky
SUGAR Quick Analysis:
Clauses Per Sentence
- When you have separated the clauses, note the number and divide it by the number of sentences from the previous step. Example:
Number of Clauses: 56 Number of sentences: 50
Clauses per sentence= 56/50=1.
SUGAR Quick Analysis:
Clauses Per Sentence
Questions?
