Phonological awareness:
Investigating the phonological awareness knowledge of New Zealand primary schools’ educators

Jane Carroll

“What is the first sound in the word ‘black’?” “What is the last sound in the word ‘though’?” “How many syllables in the word ….” These are all things that teachers say as part of their instructional literacy programmes. If the roles are reversed can the teachers answer these questions themselves? Recent studies have highlighted the need for teachers to have in-depth knowledge of the sound structures and corresponding phoneme/grapheme relationship.¹

This paper explores the results of a Phonological Awareness Test administered to a sample of final year college of education students, teachers, teacher aides, Resource Teachers of Learning and Behaviour, and Resource Teachers of Literacy in order to establish their personal phonological awareness understandings. The results are similar to other overseas studies and suggested a wide variance of teachers’ understandings and use of the concept of “a sound” and how they then used this knowledge to segment words into sounds. This variance may have quite profound implications for children in our classrooms.

Typically, teachers are heard to give instructions such as “Write down all the sounds that you can hear”, “What’s the next sound you can hear?”, “What is the last sound you can hear?” This assumes that the child (no matter what their age) can accurately segment the word into individual sounds; retain the phoneme sequence in their short-term memory; and (if writing the word down) match the phonemes one-to-one with graphemes and write the graphemes on to the paper. If the word is multisyllabic, then it is also assumed that the child can segment the word into syllables and retain the phonemic sequence within the syllabic sequence. All this occurs while the child maintains the memory of the sentence he or she is writing within the sequence of the piece of writing, whether that be, for instance, a story during writing or a set of instructions for a science experiment. This is a huge cognitive task for those children with the necessary skill level let alone those having difficulties.

Research is increasingly documenting the importance of phonological awareness in children’s literacy development. Phonological awareness is the ability to identify and manipulate the sounds of language, and is not to be confused with phonics. The definition of phonological awareness used for this research is the “awareness of the phonological structure of a word at the syllable level, onset-rime level² and the phoneme level” (Gillon, 2004, p. 9). Phonological awareness is one of the key concepts of learning that underpins learning to read (Dickinson, McCabe, Anastasopoulos, Peisner-Feinbergt, & Poe, 2003; Gillon, 2000, 2002; Jenkins & Bowen, 1994; Richey, 2004).

Like all language acquisition, phonological awareness is hierarchical in nature and not all learners need to be actively taught its component skills. Children “tune in” to sentences, words, and syllables before being able to hear and isolate groups of sounds and individual sounds within them. This, in turn, allows children to develop skills that enable them to manipulate sounds and sound patterns in more complex ways (Munro, 1998; Gillon, 2004). Stainthorp (2004) suggests that the “awareness of phonemes enables children to identify the building blocks of language, which can then be paired with letters to make possible the building up of grapheme–phoneme correspondences” (p. 754).

In the course of the author’s daily work and during several oral language/literacy professional development days throughout New Zealand, discussions with participants led the researcher to wonder whether the teachers themselves had the pedagogical knowledge and personal phonological awareness to successfully assess and teach children. Cunningham, Perry, Stanovich, and Stanovich (2004) show in their study of teachers knowledge of reading that they “demonstrated limited knowledge of … phonological awareness ” and yet the teachers self-evaluated their knowledge positively. Cunningham et al. continue by stating that their findings “suggest that teachers tend to overestimate their reading related subject matter knowledge, and are often unaware of what they know and do not know” (p. 6). Of the 722 teachers in their study, less than 1 percent of the sample were able to identify correctly all the phonemes in the 11 test words, and 20 percent were not able to identify correctly the number of phonemes in any of the 11 words presented to them. Only 30 percent of the teachers correctly identified the number of phonemes in half of the 11 words. However, having low phonological awareness does not imply that the teacher is not a literate individual; “rather, it points out that they lack a degree of technical knowledge that is relevant and that many consider fundamental to the teaching of reading” (Cunningham et al., 2004, p. 14).

In order to write and read (encode and decode text), children must be aware of sounds (phonemes) and know how to manipulate them into words. It is recognised that children who struggle to segment orally and blend sounds within words will struggle to learn to decode text (Pogorzelski & Wheldall, 2005). A child needs to master the (oral) phonemic skills before being taught the letter names and forms (Richey, 2004).

Blending and segmentation of speech sounds in oral language provides an essential foundation for reading and writing. Some of the children who are struggling with learning to read have difficulties processing phonological information. It is for these children, that the “underpinnings” of the relationship between sound, phoneme and letter needs to be secure before they can become effective at using the information (Stainthorp, 2004). Low levels of phonemic awareness have been shown to hinder a child’s reading ability as well as their progress through the developmental spelling levels (Munro, 1998).

Method

The Teachers’ Phonological Awareness Test (Love & Reilly, 1995) was adapted, with the authors’ permission, to allow for the slight vowel variation between Australian English and New Zealand English.

Participants (including classroom teachers, specialist teachers, teacher aides, and college of education students) at various oral language/literacy professional development days were asked to complete a short Teachers’ Phonological Awareness Screening Test. Teachers were asked to specify which professional group they identified with—classroom teacher, Resource Teacher of Learning and Behaviour (RTLB), Resource Teacher of Literacy (RTLit), Speech Language Therapist (SLT), teacher aide, college of education student, or “other”. They were also asked to identify if they had been trained in or used any “phonic or phonological awareness programme” in their classroom (e.g., Jolly Phonics) or had any specialist training (e.g., Reading Recovery). The college of education students were in their final year. A total of 212 people completed the test.

The test took approximately 15 to 20 minutes to administer and each of the five sections evaluated a different aspect of phonological awareness. All instructions were delivered in a manner similar to that heard in junior school classrooms. No definitions of any terminology used in the instructions were given, even if requested.

In sections 1 to 4, the speech–language therapist read out the words and in sections 5 and 6 the words were written on the answer sheet. The instructions for each section were written on the answer sheet as well as given verbally. Section 1 was read aloud and then marked cooperatively. Sections 2, 3, and 4 were administered as a block and then, again, marked co-operatively. Sections 4 and 5 were completed by the teachers and the teachers marked their own sheets during discussion following the test’s completion. This administration differs from both the Cunningham et al. (2004) and Stainthorp (2004) studies where both studies used solely pen and paper formats with the participants reading the words silently to themselves.

There was discussion at the end of the test where the participants were encouraged to reflect on their personal scores. The facilitator asked questions generally of the group so individuals were not directly asked, and answers were provided by participants who chose to. Discussions generally continued for 15 to 30 minutes.

Results

The test ceiling was 40 with scores of 38 or more considered a “pass”; scores of 34 to 37 considered “ look at error patterns before proceeding”; and scores of below 34 a “fail”. Love and Reilly’s pass rate of 38 or more is “arbitrary” but the test’s author Elizabeth Love stated that these were skills that she would expect “teachers should be able to do” (personal communication, May 2005).

The individual results were collated into the occupational groups. Each answer sheet was checked for marking accuracy by the writer with a marking key. Forty randomly selected papers were blind marked by an independent speech–language therapist. When compared to the original scores allocated, there was 98 percent agreement with the researcher’s marking.

As shown in Table 1, there was a wide variation in average scores. There was also a greater than expected variance between occupational groups. The teacher aides had a range of 18 between the lowest and highest score, which was slightly smaller than the range for teachers (range of 19), RTLBs (range of 11), RTLits (range of 13), and the control group (range of 14). It is also interesting, and maybe a little disturbing, that teachers who identified that they used or had used phonological awareness or phonics-based programmes such as Jolly Phonics had the same range of scores as those with no specialist training. It can, therefore, be concluded that use of these programmes did not improve the teacher’s personal phonological awareness or proficiency. Those who identified that they were Reading Recovery-trained had a range of 12 between the lowest and highest score, with a percentage of correct answers of 83; this was lower than expected as Reading Recovery uses Elkonin boxes³ for identifying individual sounds in target words. However, the number of individuals in each of these subgroups is low and further investigation would be required to establish any causal link between their higher average scores and any specialist training. This would be an interesting area to pursue in more depth in subsequent research.

The results of sections 1 to 4 will to be discussed more fully below and sections 5 and 6 will not be covered in this article.

Table 2 shows the average score, by professional group, for each of the individual sections.

Section 2 (phoneme segmentation) had the widest variance of all the sections. On average, teachers were able to accurately segment words into phonemes for just over a third of the 10 target words while college of education students were able to segment words accurately for just under a third of the words. Teacher aides had most difficulties and their scores were positively altered by one teacher aide scoring 36 out of 40. If her scores are removed, the average for teacher aide phoneme segmentation drops from 19 percent to 14 percent. For section 3 (second sound identification), the college of education students had difficulties, identifying just over half of the required phonemes correctly. If the high-scoring teacher aide is removed, the average accuracy for second sound identification drops from 40 percent to 35 percent.

Teachers wrote a number of unsolicited comments and explanations for their answers on the test forms about why or how they had segmented the words in the way they did. Some participants also wrote explanations during the discussion time. Most considered blends to be “a sound” or “chunked sounds”. The average accuracy of all groups improved for section 4, identifying the last sound; however, some had difficulties separating the sound from the “orthographic” spelling. During the discussion, this theme was frequently raised, with the facilitator describing word origins and how factors such as a word’s origin influence its pronunciation; for example, the digraph “ch” can be pronounced “ch” as in chips or “k” as chemist (from Greek origins) or “sh” as in champagne (from French).

The range for each section was used to identify which groups had the widest variance in ability. Teacher aides and teachers had the greatest variance across all sections. Of the 212 participants, 7.5 percent segmented all words in section 2 correctly which is significantly higher than the 1 percent in the Cunningham et al. study. At the individual participant level, a high score in one section did not always translate to high scores in subsequent sections; for example, one teacher scored 10/10 in section 1 (syllables) then scored 0/10 in section 2 (segmentation), 2/10 in section 3 (second sound), but 6/6 in section 4 (last sound). This suggests that some participants were not always consistent in the rules that they applied across subtests; for example, they considered blends as one sound unit in some situations but then segmented a blend into smaller sound units in another. During the discussion some participants made comments that they had not thought about blends being at the ends of words so always segmented them into individual sounds.

The simple three-sound word “bought” had the highest correct responses. The words which contained consonant blends were less likely to have their second sound identified accurately, with many teachers identifying the blend as being a single sound unit. This is consistent with what Cunningham et al. found. In the word /skrim/ the consonant blend was segmented into smaller phonemic units by 62 percent of the teachers; with 13 percent identifying /sk/ as the first phoneme and /r/ as the second; and 53 percent correctly segmenting the blend into three phonemes. Thirty-three percent of the teachers identified /skr/ as a single sound unit. For more complex sounds which do not follow the orthographic pattern; for example,. “qu”=/kw/ in the word queen /kwin/ only 17 percent of teachers identified that /w/ was the second sound. The diagraph /Ɵ/ in thrive was correctly identified as a single sound by 61 percent of teachers.

As shown in Table 4, the word “thought”, with the least number of phonemes and with no consonant blends, was segmented into “sounds” with the most phonemic accuracy with 79 percent correctly identifying it has three phonemes. This is significantly higher than the participants in the Cunningham et al. study, where only 67 percent could correctly segment “sun” into its three phonemes. However, in this current study, 14 percent of participants allocated more sounds; that is, they believed that “thought” had more than three sounds. This is possibly due to the participant “stretching” the sounds and adding a vowel to the consonant phonemes (e.g., “t” becomes “ta”) when segmenting the word and “hearing” more sounds rather than segmenting phoneme by phoneme.

Words with consonant blends were often allocated fewer phonemes because blends were counted as one sound; for example, “flag” was allocated three phonemes instead of four (“f-l-a-g”). The blend segmentation was inconsistent with the word rust with the “st” blend at the end of the word being segmented into the individual phonemes by more participants than the blends at the beginning of words. This is also evident by the identification of “p” as the last sound in “crisp”, in section 4, by the majority of participants. “Clump” with two consonant blends had only 27 percent of participants correctly identifying that it had five phonemes. This again is consistent with the Cunningham et al. (2004) study which found that “with the introduction of increasingly complex letter–sound patterns such as consonant blends, performance declined further” (p. 9).

“Instrument” had the largest variance of phoneme identification and the lowest accuracy of correct identification at 16 percent. “Flag” and “rust” had the smallest range of variance at three phonemes. For “rust”, “flag”, and “clump”, none of the teachers allocated more phonemes than the word contained. These words have direct mapping of phoneme-to-letter; however, words such as “change”, “thought”, and “chemist” all have phonemes that are represented by more than one letter—for example, “ch” is said /k/. Cunningham et al. (2004) suggest that this is evidence of teachers who “operate more on the level of orthographic patterns (i.e., the spelling of “x” instead of hearing the sounds /k/and /s/ in “x”) when attempting to dissect a word, and fail to shift their attention to the sound stream within words (p. 9).”

There was evidence of several differing patterns of segmentation through comments on the answer sheets and issues debated in the discussion time. For example, some segmented “flag” into “f-l-a-g” (four phonemes); others “fl-ag” (two phonemes) or “fl-a-g” (three phonemes) or “f-l-ag” (three phonemes). During the discussion, some teachers talked about breaking words into chunks before breaking them into smaller sounds and there was some evidence of use of onset and rime; that is, “fl-ag”. When asked how they knew how to segment words, the most typical comment was they had “copied” how other teachers did it rather than learning about how to do from their training. The above results are consistent with Munro’s (1998) Australian study, Cunningham et al.’s (2004) North American study and Stainthorp’s (2004) British study.

Discussion

The author is not being critical of the individual teachers or teacher aides involved in this study. The results are indicative of a broader level of teacher understanding that has led to be some “confusion” over what a sound is and how this concept should be taught. When asked “What is a sound or phoneme?” in discussions, few teachers gave a comprehensive or accurate explanation. In the classroom, children are commonly asked “What is the first sound you can hear?” If teachers do not have adequate understanding themselves and therefore give conflicting and contradictory models, this will only add to the confusion that some children are experiencing. Examples of confusion shown by teachers during the testing include blends being identified as one sound and their identification as a “single phoneme”—blends are, in fact, made up of two or more phonemes. Teacher aides are often asked to work with those who require the most in-depth level of expertise and yet have the least training, and, as shown in this study, understanding,

Teachers, on average, only scored 30 out of a possible 40 which, as a group, is a “fail”. Only 12 percent of the teachers achieved over the 95 percent pass criteria that Love and Reilly set for this test and a further 13 percent attained the proceed with caution threshold. As a phonological awareness task, syllable counting was the most accurate across all groups, with anecdotal evidence suggesting that subjects use strategies such as tapping out the beats to aid identification. Again, this is consistent with the findings of Stainthorp’s (2004) study.

As groups, teachers, teacher aides, college of education students and RTLBs had the most difficulties in sections 2 (phoneme segmentation) and 3 (second sound), and therefore at the most in-depth level of segmentation of dividing spoken words into sounds. Participants wrote (unsolicited) reasons why they had difficulties. These included identifying consonant blends as a single sound: difficulties identifying the difference between blends and digraphs; having difficulties doing it in their heads (and wanting to write down the word); and not listening to the instructions (the instructions were also written above each section of the answer sheet).

Some participants wrote the words on the answer sheets and then grouped the written letters into sound groups, thus using a visual strategy. This also may reflect their reliance on written prompts to support their metalinguistic skills and the use of orthographic knowledge to aide their segmenting. Unfortunately, because of the anonymous nature of the study, it was not possible to further question participants about this. Other participants were observed to use strategies such as counting on their fingers, closing their eyes and mouthing the words, and using tally marks to record sounds as they segmented the words.

This has implications for classroom practice because if the adults are unable to model consistent segmentation of words into sounds then their ability to teach children this skill effectively will be compromised. We can confidently assume that as a group, most teachers did not struggle to learn to read and therefore may not be as aware of the need for the explicit teaching of segmentation and blending of the phonemes within words as they might be. Scarborough et al. (1998, cited in Pogorzelski & Wheldhall, 2005) suggest that as we become more proficient readers, our use of phonological awareness skills becomes more implicit and therefore not recognised as part of our internal reading processes. There must also be consideration of other phonological awareness segmentation techniques; for example onset and rime, and how teachers scaffold children’s awareness of rime unit whose spelling is more consistent “than the spelling of phonemes so knowledge of these larger orthographic units may bring an added bonus” (Goswami, 1998, cited Stainthorp, 2004, p. 755).

The other strategies that the adults in this study used to aide their segmenting of the words also have implications for classroom practice. It would be interesting to be able to evaluate how many of the strategies teachers explicitly model within their classroom programmes, such as tally marks, using fingers to mark each sound, and so on. For children to use these strategies, other skills are required; for example, one-to-one matching of phoneme to material and then counting with numerals, working memory skills, and so on. Many adults appeared to close their eyes and “visualise” the words, thus making the task into a letter-to-phoneme task.

During discussions, some of the teachers suggested that in their training in the teaching of reading and writing they were told the questions to ask children to elicit appropriate responses and facilitate the child’s learning but this did not equip them with underlying theoretical knowledge in the area of breaking the spoken word into phonemes. Can it be assumed that the reverse skill of blending phonemes into word is also not taught? It may be over simplistic to conclude that by introducing explicit phonological awareness and in particular as shown by this study, phonemic awareness, into teacher training that teachers’ ability will improve as a direct result because “… phonological awareness skills will only become fluent if training is followed by regular practice” (Stainthorp, 2004, p. 763).

There is confusion and a lack of common understandings within the teaching community around several aspects of the phonemic complexity of our language and its importance in literacy learning. This may be part of the answer to Tunmer, Chapman, and Prochnow’s suggestion that Year 5 teachers in New Zealand are placing a greater emphasis on teaching decoding skills in comparison to other PIRLS (Progress in International Reading Literacy) countries (2004, p. 138).

It can reasonably be assumed that all participants in this study were literate, with the large majority having at least one tertiary qualification. However, it does not necessarily follow that they have a correspondingly high level of phonological awareness, especially at the level of segmentation of words. Not all teachers were able to correctly identify syllables—this is often a skill that teachers assume that new entrants are able to do before starting school.

If teachers are able to segment words explicitly into individual sounds rather than chunks or blends, will this have an impact on the classroom teaching of literacy? If they incorporate phonological awareness activities (at an oral level) and then introduce phoneme to grapheme links then the research suggests an overwhelming yes! (Gillon, 2004). Many teachers are working from the grapheme to phoneme level, therefore making it explicit that letters have sounds. This requires children to use more working memory and retain the “rules” that accompany this phonics-based approach so children are starting with teachers expecting children to be competent at a higher skill level rather than teachers having the knowledge to scaffold those children who require earlier phonological awareness skills to this level.

During the post-test discussion, some teachers suggested that they had been “put on the spot” with the test which caused their poor results. However, it could be argued that teachers put children “on the spot” all the time during literacy time when they ask children to do the same skills that were asked of them as part of this study.

Conclusion

From this study, it appears that the New Zealand teaching profession has widely varying personal phonological awareness abilities and professional linguistic knowledge. The research (both within New Zealand and overseas) strongly suggests that the phoneme level of phonological awareness is a very important skill for children to master as part of their literacy development. As Stanovich (1996) states, “Impaired language segmentation skills lead to difficulties in phonological awareness which in turn impede the word recognition process which underpins reading comprehension” (cited Tunmer, et al., 2004, p. 131). However, if, as this study suggests, there is a wide variation within the educational sector in understandings about the definition of what a sound is, then the teaching in this area will be inconsistent between teachers and within schools.

Phonological awareness is a critical aspect of oral language and developing metalinguistic skills is an essential part of a child’s literacy development. It is well-documented that those with low phonological awareness abilities are more susceptible to literacy difficulties than those children who have developed higher levels at school entry. They are, therefore, also able to profit more from reading instruction in the classroom (Tunmer et al., 2004).

Are teachers further compounding these children’s difficulties by not explicitly (or incorrectly or inconsistently) teaching phoneme level strategies because of their own misunderstandings? Tunmer et al.’s (2004) PIRLS review suggests that “relatively small changes to the predominant approach … increase the overall effectiveness of beginning literacy instruction” (p. 132).

If, as Tunmer et al. suggest, small changes in practice can have big impacts on literacy acquisition then this area of teachers’ professional abilities must be explored further and on a larger scale. However, if we do examine teachers’ phonological awareness on a larger scale and the results are similar to this pilot, then the implications for children as readers and writers, and for teacher practice in our schools, are profound and in need of urgent attention by all those involved in our education system.

Acknowledgements

Thanks to Dr John Smith, Dunedin College of Education; Maryanne Pease and June Laverty, Ministry of Education; and Greg Carroll, Principal Pine Hill School, for providing mentoring and support for the writing of this paper.

References

Cunningham, A., Perry, K., Stanovich, K., & Stanovich, P. (2004). Disciplinary knowledge of K–3 teachers and their knowledge calibration in the domain of early literacy. Annals of Dyslexia, 54(1), 1–18.

Dickinson, D. K., McCabe, A., Anastasopoulos, L., Peisner-Feinbergt, E. S., & Poe, M. P. (2003). The comprehensive language approach to early literacy: The interrelationships among vocabulary, phonological sensitivity, and print knowledge among preschool-aged children. Journal of Educational Psychology, 95(3), 465–481.

Gillon, G. T. (2000). The efficacy of phonological awareness intervention for children with spoken language impairment. Language, Speech and Hearing Services in Schools, 31, 126–141.

Gillon, G. T. (2002). Phonological awareness intervention for children: From the research laboratory to the clinic. Retrieved 1 March 2005, from http://www.asha.org/about/publications/leader-online/archives/2002/q4/021203a.htm

Gillon, G. (2004). Phonological awareness: From awareness to practice. New York: Guilford Press.

Jenkins, R., & Bowen, L. (1994). Facilitating development of preliterate children’s phonological abilities. Topics in Language Disorders, 14(2), 26–39.

Love, E., & Reilly, S. (1995). A sound way: Phonological awareness activities for early literacy. Melbourne: Longman.

Ministry of Education. (2003). Effective literacy practice in Years 1–4. Wellington: Learning Media.

Munro, J. (1998). Phonological awareness: Their impact on learning to read prose and to spell. Australian Journal of Learning Disabilities, 3(2), 15–21.

Pogorzelski, S., & Wheldall, K. (2005). The importance of phonological processing skills for older low progress readers. Educational Psychology in Practice, 21(1), 1–22.

Ritchey, K. (2004) From letter names to word reading: The development of reading in kindergarten. Reading Research Quarterly, 39(4), 374–376.

Stainthorp, R. (2004). W(h)ither phonological awareness? Literate trainee teachers’ lack of stable knowledge about the sound structure of words. Educational Psychology 24(6), 753–765.

Tunmer, W. E., Chapman, J. W., & Prochnow, J. E. (2004). Why the reading achievement gap in New Zealand won’t go away: Evidence from the PIRLS International Study of Reading Achievement. New Zealand Journal of Educational Studies, 39(1), 137–145.

Notes

1.&&“A phoneme is the smallest unit of sound in a word” (Ministry of Education, 2003, p. 32). A grapheme is the graphic representation of a sound, for example, the sound /f/ can be represented by the letter “f” or the digraph “ph”.

2.&&Awareness at the onset-rime level is the awareness that words and syllables can be divided at the intrasyllabic level (e.g., “sit” can be divided into “s-it” and “split” into “spl-it”). It is the recognition that words can have different beginnings but can share a common ending (Gillon, 2004).

3.&&Elkonin boxes or “sound boxes” are visual activities designed to help the child think about sounds in spoken words.