You are on page 1of 4

Phonological Assessment: An Overview

a. Strength-based assessment and child characteristics


A thorough assessment is required to fully describe a child’s sound system. This includes
performance-based tasks that identify the child’s strengths, needs and personal
characteristics. Oftentimes, parents comment that reports are esoteric and focus on what
their children cannot do; consequently, clinicians should strive to objectively characterize the
child’s sound system in parent-friendly, yet positive terms. In particular, the clinician may
highlight the fact that a child is stimulable for many sounds, demonstrated willingness to take
risks or that he or she showed interest in various activities. SLPs may wish to write reports
that feature parent-friendly sections in bullet points that identify strengths, describe the
family or teacher concerns, characterize the sound system and offer facilitative strategies to
support the child as he or she acquires sounds.

According to Barlow, Storkel and Taps (2007), the assessment process is marked by three
phases: characterization of the sound system, the predicted reorganization of the system
based on the proposed treatment targets and ongoing monitoring. The next few sections
focus on the characterization of the child’s sound system and the predicted results of
targeting sounds for treatment.

b. Thorough assessment
i. Limitations of standardized tests
A complete assessment allows the examiner to identify what sounds are in the child’s
repertoire and what sounds are not. Frequently, standardized tests, such as the Goldman-
Fristoe Test of Articulation-2 (GFTA-2) (Goldman & Fristoe, 2000) and the Assessment of
Phonological Processes-Revised (APP-R) (Hodson, 1986), are employed for this purpose.
Such instruments provide a quick “snapshot” of a child’s system as compared to their same
age peers and yield standard scores and percentiles. However, these tests usually sample
sounds once in each word position. It may be the case that a child has partially acquired a
sound in his or her system, but that he or she did not use it in that particular target. For
example, a child might substitute /j/ for /l/ (as in [jaIt]) even though he or she produces /l/
with 60% accuracy in words. A clinician may recommend that /l/ be targeted even though
this sound does not warrant treatment. Consequently, SLPs cannot use this information to
fully represent the phonological system.

In 2005, Kollia and Eisenberg compared the performance of children on five standardized
tests, including the GFTA-2, the Photo Articulation Test-Third Edition (PAT-3) (Lippke,
Dickey, Selmar & Soder, 1997), Bankson-Bernthal Test of Phonology (BBTOP) (Bankson &
Bernthal, 1990), the Clinical Assessment of Articulation and Phonology (CAAP) (Secord &
Donohue, 2002) and Smit-Hand Articulation and Phonology Evaluation (SHAPE) (Smit &
Hand, 1997). They wanted to determine if these tests accurately identify articulation or
phonological disorders. Their preliminary findings suggested that the CAAP was the most
stringent while SHAPE, GFTA-2 and PAT-3 were the most lenient (regarding error scores).
Overall, they concluded that the content from these tests provided insufficient information to
determine appropriate goals because they focused on relational analyses. That is, each test

© 2010 Jennifer Taps, M.A., CCCSLP and SLPath.com


compared the child’s productions to adult-like productions rather than determining what
sounds are present or absent in the inventory (independent of adult targets). To accomplish
this goal, Kollia and Eisenberg recommended using independent probes to fully describe a
child’s phonological system.

ii. Independent probes of singletons and clusters


Independent probes sample sound production in various positions and contexts (next to
different vowels and consonants), allow clinicians to assess the level of knowledge for each
sound (from “least” to “most”) and ultimately enable the identification of sounds out of
inventory that may be targeted to create maximal change in the sound system. Of course, the
elicitation of a large sample and subsequent analysis require more time to complete.
Masterson and Apel (2006) argued that this time in analysis leads to more effective treatment
approaches. Similarly, Dyson and Robinson (1987) stated that “Despite differences, the most
appropriate treatment for a given child emerges directly from the results of the diagnostic
assessment of the phonological disorder.”

Ideally, independent probes sample singleton sounds across word positions at least five times
each. The sounds should also be elicited in clusters to determine which clusters are
represented in the child’s repertoire. Due to the limited number of picturable clusters, it may
be most feasible to sample each cluster two to three times (instead of five as mentioned for
singletons). Elbert and Gierut (1986) suggested that an effective independent probe should
be easy to administer, easy to analyze and that it should measure sound production in the
context of treatment (i.e., at the word level). They also emphasized that these words should
remain untreated so that the probe measures change throughout the treatment process. The
same probe should be readministered frequently to monitor the effectiveness of treatment
and/or treatment target selections. Moreover, these samples reveal generalization not just for
target sound(s), but also within-class and across-class generalization. For example, a child
might present with a sound system that includes all sounds, except for /s/, /z/, /ʃ/, /ʧ/, /ʤ/, /θ/,
/ð/, /l/ and /r/ and most clusters. Based on the results of the assessment, the SLP may
recommend targeting /sl/ and /θr/ words. During the treatment process, changes in the entire
sound system should be monitored (not just the treatment sounds) to determine progress and
whether or not the predictions for change came to fruition. The SLP may have predicted that
some of the missing fricatives and affricates would emerge and that other clusters would be
added to the system.

The contents of Gierut’s Phonological Knowledge Protocol (independent probe) are


available in Dinnsen and Gierut’s book, Optimality Theory, Phonological Acquisition and
Disorders (Advances in Optimality Theory) (2008). In addition, McLeod and Hand created an
independent clusters probe accessible at:
http://athene.riv.csu.edu.au/~smcleod/Consonantclustertest.pdf.

iii. Stimulability tasks


Research suggests that children will likely acquire stimulable sounds without intervention
(Powell, Dinnsen & Elbert, 1991; Gierut, 2007). Consequently, to complete a full
assessment, clinicians must determine the stimulability of sounds missing from the
phonological system. For years, the most common method of determining stimulability

© 2010 Jennifer Taps, M.A., CCCSLP and SLPath.com


consisted of asking the child to produce the sound in isolation given a model. In the last few
years, researchers have used more in-depth tasks to determine the modifiability of a child’s
production given various contexts, articulatory cues and modeling. Glaspey and Stoel-
Gammon (2005) developed a task that samples target sounds across word positions and in the
context of various vowels. The article abstract describing this dynamic assessment process is
accessible here:
http://journals.lww.com/topicsinlanguagedisorders/Abstract/2005/07000/Dynamic_Assessme
nt_in_Phonological_Disorders__The.5.aspx. For example, if a child does not produce /l/
during the independent probes, the examiner determines if the child can produce /l/ given
various levels of support. Given cues and modeling, the SLP asks the child to produce /l/ in
isolation, with the /a/ vowel (as in /la/, /ala/ and /al/), with the /i/ vowel (as in /li/, /ili/ and /il/)
and with the /u/ vowel (as in /lu/, /ulu/ and /ul/). This provides the child with ten
opportunities to produce the target sound rather than the one opportunity described above.
Barlow, Storkel and Taps (2007) stipulated that any sound produced with 30% accuracy or
greater be deemed stimulable and anything below this accuracy nonstimulable. Like
sampling sounds more thoroughly in the independent probes, this provides the clinician with
a more complete picture of what a child can do.

iv. Oral mechanism examination


A full phonological assessment includes a thorough oral mechanism examination. The
purpose of this task is to note any structural or organic issues that may impact sound
production. A child may present with a high palate or poor coordination of the jaw and lips.
It is critical to note, however, that other children may possess these characteristics and yet
produce sounds accurately.

The purpose is not to identify areas of need for oral-motor exercises. These practices lack
empirical support (Lof, 2006). Moreover, Lof and Green et al (2000) have argued that
“while differentiated movement is necessary for speech production, speed and strength of
oral movements have little impact on speech sound production.” SLPs may access an oral
motor examination resource (Addington, n.d.) here:
http://www.speechfriends.com/BAoral_mechanism.pdf

During these tasks, SLPs may discover nonspeech apraxia, which may accompany childhood
apraxia of speech. Other motor speech disorders may also be evident. SLPs are directed to
this current summary of motor speech disorder assessment and treatment (Kent, 2000):
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6T85-41PP1TJ-
2&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&_docanchor=&view=c&_acct=C00
0050221&_version=1&_urlVersion=0&_userid=10&md5=280e4a86f69e304404737a39df55
4a65

c. Conversation samples vs. single-word samples


Masterson, Bernhardt and Hofheinz (2005) addressed the results generated by conversation
samples as opposed to the information provided by single word samples. They established
that conversation samples took three times as long to elicit and transcribe as compared to
single-word tasks. Fortunately, Masterson et al concluded that single-word tasks yielded the
same information and fully represented a child’s system. Both tasks resulted in comparable

© 2010 Jennifer Taps, M.A., CCCSLP and SLPath.com


Percentage Consonant Correct (PCC) scores. Moreover, word-level samples can be
transcribed during administration and they provide the “big picture” of the phonological
system. However, Masterson and her colleagues still suggested that conversation samples be
collected to assess a child’s prosody, overall intelligibility, syntax, morphology, etc. Such
tasks enable clinicians to rule in or out the possibility of childhood apraxia of speech (CAS)
or other disorders. McCauley and Velleman (2008) reviewed the ASHA CAS technical
documents. Based on existing evidence, the panel concluded that there are three hallmarks of
CAS: 1) inconsistent error patterns, 2) disordered vowel productions and 3) prosodic
difficulties. Children with CAS demonstrate one or more of these characteristics. For more
information, readers may visit:
http://convention.asha.org/handouts/1420_1850McCauley_Rebecca_071142_Nov18_2008_T
ime_102700PM.pdf.

d. Phonetic and phonemic inventories


i. Phonetic inventory
As stated earlier, an independent analysis reveals the essence of a child’s phonological
system rather than just comparing the child’s productions to those of adults. Once an in-
depth independent probe has been administered, the results yield information about a child’s
phonetic and phonemic inventories (Elbert & Gierut, 1986). The phonetic inventory consists
of phones that occur at least two times in the probe sample (Stoel-Gammon, 1990) and
encompasses all sounds that a child produced regardless of whether these sounds are correct
or incorrect relative to the target. (For instance, a child may produce [θ] as a substitute for /s/
targets ([θʌn] for /sʌn/), but not for /θ/ targets. If the /θ/ sound occurs twice in the sample, it
is considered to be part of the child’s phonetic inventory because there is evidence of sound
production. Alternatively, relational analysis would not represent knowledge of this sound
because it is a substitution rather than a “correct” target.) Sounds in the phonetic inventory
are those which the child can physically produce. This implies at least some knowledge of
the target sound.

ii. Phonemic inventory


A more complete assessment also provides data regarding the child’s phonemic inventory.
Phonemes in the child’s phonemic inventory show a contrast in meaning. Such a contrast
may be shown in initial (“hat” vs. “bat”), medial (“beg“ vs. “bag“) or final (“flag” vs. “flap”)
positions. In these examples, /h/ and /b/ would be part of the child’s phonemic inventory as
would /ɛ/, /æ/, /g/ and /p/. SLPs identify minimal pairs from the results of the independent
probes. A smaller sample (like most standardized tests) would not yield enough
opportunities for a child to demonstrate this kind of knowledge. Inclusion in the phonemic
inventory implies greater knowledge of a sound because there is evidence that the child has
both the motoric ability to produce the sound and the linguistic knowledge that its purpose is
to denote a different meaning.

© 2010 Jennifer Taps, M.A., CCCSLP and SLPath.com

You might also like