ACQ Vol 13 no 2 2011

41 / 52

ACQ Volume 13, Number 2 2011 95
These include assessment of children from a range of
cultural and linguistic backgrounds and children with severe
disabilities. In fact, for assessment to be balanced, it should
always include information collected in different ways. We
need to use standardised but not norm-referenced tests,
criterion referenced assessments, non-standardised
protocols, dynamic assessment, observation, and language
sampling. We need to collect our data from a wide range of
people involved in the process, including significant others,
through interviews, observation, and questionnaires.
9 Discourse based sampling methods
and analyses
normal distribution and the use of scores such as standard
scores/percentile ranks allow us to determine where a child’s
score fits in the normal distribution. While they should be
used with caution, they do allow us to interpret a score. A
chart of the normal distribution is a good visual reference to
explain standardised scores to parents/carers (see left).
In contrast, other scores such as age equivalents can be
dangerous. They do not tell us the degree of impairment/
difference. They do not say a child is performing the same
as another child at that age equivalent (if a child aged 12
years obtains an age equivalent score of 5 years, it does
not mean he or she is performing the same as a 5-year-old)
as the score is usually calculated on a total raw score and
the pattern of responses can be completely different. Age
equivalent scores do not measure change and should not
be used to state, for example: “he made 6 months progress
in 3 months”.
5 Criteria to evaluate a test against,
which allow us to make an informed
decision when we purchase a test
When we plan to buy a standardised test, we need to
ensure it has strong psychometric properties. The books
outlined above (see no. 2) cover these concepts in detail. In
brief, we want a standardised test to provide information
about the standardisation sample (so we know if it is valid
to compare our client to the norms), as well as reliability and
validity data – so we know that the scores will be consistent
and reflect the areas the test aims to test, and that the test
administration and scoring are clear. It is tempting to skip
through the manual when considering purchase but this is
the most important part of evaluating a test!
6 A well-developed standardised test
I have found the Clinical Evaluation of Language
Fundamentals (4th ed.) (CELF-4) and the Clinical Evaluation
of Language Fundamentals, Preschool (CELF-P) to be good
value for money and they have strong psychometric
properties (as well as the benefit of local “norms”).
These tests are available from Pearson at http://www.
and
7 Access to evidence in the research
literature
Evidence based practice is as important to assessment as
it is for intervention. When we come to develop and use
non-standardised approaches to collect assessment data,
we must still make sure they are reliable and valid as well as
driven by theory.
Speech Pathology Australia has provided a position
statement that makes a great starting point: http://
statements/EBP_Position_Statement.pdf
And my favourite resources include:
• Dollaghan, C. (2007). The handbook for evidence based
practice in communication disorders. Baltimore: Paul H.
Brookes
• SpeechBite website:
•
path/
8 A broad range of assessments in the
“tool kit”
There are many times when it is not appropriate to use
standardised tests – whether broad spectrum or specific.
Dr Suze Leitão is a clinician, researcher, and lecturer in speech
pathology at Curtin University in Western Australia. She has been
teaching a unit about assessment for a few years and it has given her
much to think about in her own clinical practice! Her top 10 is a mix
of her thoughts and resources that have had an influence on her
assessments of paediatric clients.
Correspondence to:
Dr Suze Leitão
email:
We are seeing increasing use of language sampling in a
range of text and discourse genres. One method of
sampling that is very practical and has been field-tested
with hundreds of children is the Westerveld and Gillon
language sampling protocol available for free from http://
The systematic analysis of language transcripts (SALT) is
also increasingly used in clinical practice and a lot of useful
information on this approach is available on their recently
updated website:
10 Caroline Bowen’s website
I don’t think I could survive without this rich source of
information. I often go to the website for information and
the listserv discussions – a new “nugget” pops up regularly
and I have found out about many new assessments,
assessment protocols, and resources this way!