A nonverbal phoneme deletion task administered in a dynamic assessment format

Abstract

Purpose

The purpose of the project was to design a nonverbal dynamic assessment of phoneme deletion that may prove useful with individuals who demonstrate complex communication needs (CCN) and are unable to communicate using natural speech or who present with moderate–severe speech impairments.

Method

A nonverbal dynamic assessment of phoneme deletion with a systematic prompting and scoring system was developed and compared to the same measure administered using a traditional “static” format. Both versions were administered to 64 typically developing children ranging in age from 6 to 8;5. The dynamic phoneme deletion task was designed so that no verbal response was required. Children were asked to point to pictures representing each stimulus item. The same stimulus items were administered to children in a traditional static format, without the use of pictures. For example, in the dynamic assessment task, children were asked to point to 1 of 4 pictures (e.g., ice) to indicate their response when asked to delete “m” from “mice.” In the static format, children were asked to “say” the real word left over after deleting “m” from “mice.” Correlations between phoneme deletion tasks and word-level reading tasks were calculated.

Results

Internal reliability for the nonverbal dynamic phoneme deletion task was α = .88 for the total sample. Correlations between the dynamic and static formats were high (r = .84) as were correlations between the deletion tasks and the measure of word-level reading (dynamic; r = .54, static; r = .53).

Conclusions

These preliminary findings suggest that the nonverbal dynamic phoneme deletion task holds promise for use with children for whom speech output is compromised. Implications for use with children who use AAC or who have significant speech impairments and recommendations for future research are discussed.

Learning outcomes: Readers will learn about a nonverbal dynamic assessment of phoneme deletion that may prove useful for measuring phoneme awareness for children who are unable to communicate using natural speech or who present with moderate–severe speech impairments. Readers will learn about the importance of accurate measurement of phoneme awareness for children with complex communication needs. Readers will also learn about how to develop and utilize reliable and valid measures of phoneme awareness for this population of children with communication impairments.

Introduction

Children with complex communication needs (CCN) sometimes demonstrate unintelligible speech, limited vocabulary, or impaired or grammatically incorrect sentence structures and often experience difficulty learning to read (Iacono & Johnson, 2004). In order to be understood, many children with CCN must depend on augmentative and alternative communication (AAC) systems. AAC systems offer a wide array of communication options including the use of pointing to icons or words, spelling, and/or the use of drawings or speech generating devices that produce words, phrases or sentences for children (Beukelman & Mirenda, 1998). Communication for children with CCN is dependent to a large extent on the type of AAC system that is being used. That is, children may only communicate those needs or wants that they have access to through the AAC system or device. If a thought is not represented in some way on the device, it cannot be expressed. For example, children may want to talk about a giraffe seen at the zoo. If a picture, icon, or word is not available in the AAC device to represent “giraffe,” then children are at a loss to talk about it. Devices that allow for spelling and “generation” of new thoughts and ideas may afford the best opportunity to express individual thoughts for optimal independence in daily communication (Blackstone & Cassatt-James, 1988).

In addition to the difficulties that children with CCN may experience in intelligibility and in mastering a generative language, 50–90% have significant difficulty acquiring age-appropriate literacy skills (Dahlgren Sandberg, 2001, Koppenhaver and Yoder, 1992, Zascavage and Keefe, 2004). Unfortunately, these difficulties often persist into adulthood. The achievement of literacy is critical for individuals with CCN because the attainment of communicative independence is dependent on literacy skills and the use of a generative AAC device (Blackstone and Cassatt-James, 1988, Blischak, 1994).

Several language skills have been identified as critical to the development of skilled reading and writing such as phonological and phonemic awareness. Phonological awareness is a broad term describing a conscious awareness of the sounds of the language. Phonemic awareness is the ability to deal explicitly and “segmentally” with sound units smaller than the syllable (Stanovich, 1994). For example, phonemic awareness allows children to identify the position of a sound in a word (Where is the /s/ in /sun/?), or to isolate a sound from a word to discover the “real word” that is left (cold without /k/ is “old”). Phonemic awareness skills measured in blending (/s/, /o/, /p/ is soap), segmenting (the sounds in cat are /k/, /□/, /t/) and deletion or elision tasks (ski without /k/ is see) have been shown to be the most closely tied to the development of early word reading and spelling in children (Perfetti et al., 1987, Rosner and Simon, 1971, Schuele and Boudreau, 2008). One of the greatest difficulties involved in the assessment of children with CCN is the fact that most standardized, norm-referenced tests, such as those measuring phonological awareness require verbal responses.

Traditionally, clinicians and researchers have “adapted” assessments for children with CCN by replacing verbal responses with pointing or scanning responses (Beukelman and Mirenda, 1998, Zascavage and Keefe, 2004). There is limited research on the impact that adaptations may have on response interpretation. That is, it is possible that fundamental differences in modality (auditory-verbal, visual-pointing) may influence the results that are obtained from the assessment (Ginns, 2005). However, it is generally accepted that the use of direct selection methods (recognition tasks) place less burden on memory than free-recall or visual scanning responses (Light and Lindsay, 1991, Wagner and Jackson, 2006). In the past, the tasks that have required the least adaptation such as word/sound matching or counting tasks have been recommended as the most effective and reliable ways to assess phonological awareness in children with CCN (Blischak, 1994). However, as was discussed earlier, these basic-level phonological awareness tasks may not predict reading as well as tasks that require phoneme manipulation (e.g., deletion, segmentation).

Standardized, norm-referenced assessments can be particularly problematic for use with children with CCN because they are designed to yield diagnostic information to the exclusion of prescriptive information, are lengthy to administer, and often yield floor effects (Iacono & Cupples, 2004). Dynamic assessments that allow clinicians to gather diagnostic and prescriptive information simultaneously may be time and cost-effective for gathering pre-literacy information for children with CCN without producing floor effects, which may be a result of task requirements rather than difficulty with the actual skill (Blischak, 1994). Dynamic assessment of phonemic awareness, specifically segmentation, has also been shown to reliably predict reading performance in children without CCN (Laing and Espeland, 2004, Laing and Kamhi, 1998, Spector, 1992).

One type of dynamic assessment involves the use of graduated prompting or “cueing.” Typically, graduated prompts are designed to assist children incrementally in their efforts to perform a specific task. For example, children may have difficulty blending sounds in words. To design a graduated prompting task, the clinician can think about the steps involved in blending sounds in words and arrange a testing/teaching session based on successively more supportive cueing. The least supportive prompt might be to ask children, “Tell me what word the sounds /k/, /u/, /p/ make?” The most supportive cue might be to have children imitate the sounds one at a time after the clinician and then to say the word. The clinician ascertains the level of prompting that is required to enable children to be successful and begin instruction at that level. After the dynamic assessment session, the clinician knows whether or children have a problem performing the skill (diagnostic information) and the prompts/cues that are appropriate for moving the individual toward independence in performing the skill (prescriptive information).

The purpose of this project was to design a nonverbal dynamic assessment measure of phoneme deletion with a systematic prompting and scoring system and compare it to the same measure administered using a verbal, traditional “static” format. As in previous research on the development of a nonverbal word recognition task for children with CCNs, we sought to determine if the nonverbal dynamic assessment measure was a valid measure for typical children before administering it to children with CCNs (Erickson et al., 2008). Internal consistency of each measure was calculated. Correlations between the 2 formats (dynamic, static) were computed. Correlations were also calculated among the phoneme deletion tasks (dynamic, static) and a measure of word-level reading.