Overview of Language Assessment Flashcards
goals of language assessment
- determine if child has significant impairment in form, content, and/or use
- describe deficit relative to typical developmental sequence of language acquisition
- determine how deficit will affect child’s ADLs
- use multiple methods
- describe child’s strengths and weakness in all areas (the whole child)
review of referral/background information
- what is the problem in medical terms (if known)?
- how does the family describe the deficit?
- how do other professionals describe the deficit?
- how does the problem impact the client academically?
- does the problem vary in severity based on the setting?
- how does the problem impact the client socially?
case history
- should highlight family’s and teacher’s primary concern
- pre-, peri-, post-natal development
- risk factors related to language
- family history of speech, language, literacy delays
- health + medical history
- education history
- previous assessment + treatment
informal observations
- free play + observations made throughout the session
- unstructured conversation
- interaction with parents, siblings
- various settings, situations
- build rapport
- note interests
free play + observations made throughout assessment
- child’s preferences, motivations
- child’s response to failed attempts to communicate
- examples of child’s communicative
standardized assessments
specified procedure for administration, analysis to achieve uniformity
norm-referenced assessments
provide means of comparing child’s core to standard/norm
NR assessments: potential pros
- reliable, valid
- compares child to peers that completed same assessment
- can be used to diagnose language disorder (but not alone)
- more easily understood by other, related professionals, public
NR assessments: cons
- decontextualized
- formal, structured, inauthentic
- can be difficult for some kids to complete
- not useful in writing treatment objectives
- can’t always be used alone to identify a language disorder
- underestimate abilities of children who are culturally and linguistically diverse
standardized, NR tests should include the following properties
- clear administration and scoring criteria
- reliability
- validity
- characteristics of the normative sample
- diagnostic accuracy
reliability
how consistent and accurate are the results of the instruments?
validity
does it measure what it’s supposed to measure?
characteristics of the normative sample
- overall size of the sample
- number of children who are like the subject who will be given the test
- represents or includes children who are like the subject wo will be given the test (ethnicity, race, gender, SES language status, geographic location, rural/urban/suburban)
diagnostic accuracy
- sensitivity
- specificity
- 90% accuracy = good
- 80%-90% = fair
- <80% = the test should not be used
sensitivity
does the test accurately identify individuals with disorders s/l skills?
specificity
does the test accurately identify individuals with typical s/l skills?
goal of cut scores
balanced sensitivity and specificity (and/or PPP and NPP–probability that the test will accurately ID those with/without LD)
standard scores
- derived from the raw score
- a type of norm-referenced score
standard scores: range of scores
- mean of 100
- SD of 15
- 115-85 = WNL
- 84-70 = mild
- 69-55 = moderate
- 54 or < = severe
percentile ranks
tells what proportion of the normative population scored lower than the child who took the test
equivalent scores (age, grade)
- don’t take into consideration the range of normal performance for individuals whose scores fall within the average range
- compare child’s performance to age group or specific grade level that has an average score that is in the same range (performance isn’t compared with others in a similar population)
- AE scores are unevenly distributed + skewed, and half of the examinees of a given age will earn an AE score that is above their CA, and half below
criterion-referenced tests
- don’t compare an individual’s performance to anyone else’s
- not norm-referenced, but most are standardized
- student performance reported based on number and percent correct, speed, and quality
- obtain baseline score and then re-administer at a later date
- more helpful in setting treatment goals than NR tests
- can more easily be modified for cultural/dialectal differences
- less structured and less decontextualized compared to NR tests
- can also be clinician-developed, informal
purpose of criterion-referenced tests
- ID what a child can and cannot do compared to a predefined criterion
- how does the child compare to an expected level of performance?
- compared to own performance at different times
dynamic assessment
- ZPD
- allows you to measure language-learning ability
- typical development
- atypical development (disorder)
- allows us to determine language difference vs. disorder
dynamic assessment: allows you to measure language-learning ability
- determine the size of the child’s ZPD
- does the child perform poorly because they’re unfamiliar with the tasks?
- would just a little bit of extra cueing/support allow the student to experience success on the task?
dynamic assessment: typical development
- performs below what is expected during testing
- performs within normal limits after mediated learning
dynamic assessment: atypical development (disorder)
- performs below what is expected during testing
- continues to perform below normal limits even after mediated learning
