Introduction
Phonological disorder is a subset of speech sound disorders (SSD) in which the increase in the complexity of speech leads to increased errors [1, 2]. The articulation of polysyllabic words (with three or more syllables [3]) requires the planning of chain and hyperchain parts of speech, which usually does not happen in producing simpler words. The correct production of polysyllabic words depends upon appropriate stress patterns and mastery of vowels and consonants in collocations [4]. Therefore, the analysis of polysyllabic word production can offer a sound knowledge of the phonology system development in children [5]. 
To date, studies have been conducted on polysyllabic word production in children speaking English, Italian, Spanish, and Malay [3, 7-11]. In 2006, the production of 50 polysyllabic words was examined in 6 typical 4-year-old English-speaking children. Despite the correct production of all vowels and most consonants in monosyllabic and bisyllabic words, children encountered challenges producing the same sounds in polysyllabic words [7]. To describe the acquisition of consonants in the Malay language in 2014, a tool containing 11 polysyllabic words was used. Four- and six-year-old children produced consonants more accurately in monosyllabic and bisyllabic words than in polysyllabic words [12].
SSD is the most common speech disorder in children, affecting 1 out of every 20 preschool children and constituting 30% to 75% of the clients of speech therapy clinics [12]. Most children diagnosed with SSD recover spontaneously or with treatment. However, 1 in every 1000 SSD children will experience a more permanent disorder, such as childhood apraxia of speech (CAS), which may persist into later ages [17]. Studies have shown that the error in the production of polysyllabic words can present one of the speech characteristics of school children with CAS that has extended since childhood [4, 13]. Also, low production accuracy in polysyllabic words is a crucial prognosis for poorer developmental conditions in literacy [13], limited improvement in production accuracy over time [14], and weaker phonological processing abilities [15]. The possibility of secondary psychological-social effects and problems in finding a job in adulthood have also been reported [16, 17]. Overall, polysyllabic word production provides critical information about children’s phonological systems; however, it has received less attention in related studies. 
A review of literature in other languages showed that although children usually achieve the correct production of vowels by the age of three [19] and the correct production of consonants correctly by the age of four, incorrect articulation of acquired sounds in polysyllabic words continues until older age [7]. Zarifian and Fotuhi reported that Persian-speaking children master all vowels before age three, but consonant mastery extends up to four years and six months [20]. However, the phonetic test used in this study included only a small number of polysyllabic words. So far, no suitable test has been designed to investigate the production of polysyllabic words by Persian-speaking children and the process of producing speech sounds in these words. Due to the extensive research of these words in other languages and the minimal presence of these words in the current production tests, despite their practicality, it is necessary to develop a tool to evaluate these words in the Persian language. 
Given the high prevalence of SSD and the significance of early diagnosis and prognosis determination, this research was conducted to build a valid and reliable native instrument in the Persian language. The purpose of the first phase was to prepare a test of polysyllabic words. The second phase aims to determine the psychometric characteristics of the test to examine the ability to produce polysyllabic words in Persian-speaking children aged 4 to 7 years.

Materials and Methods 
Study design and participants 
A total of 120 typically-developed children (60 girls, 60 boys) and 20 SSD children (10 girls and 10 boys) aged 4 to 7 years (in six groups with age intervals of six months and an equal number of girls and boys in each group) participated in the research. Children with normal growth were selected from three kindergartens in Tehran City, Iran, and one kindergarten and one school in Karaj City, Iran, using a purposive convenience method. The inclusion criteria included being monolingual, not suffering from SSD, and lacking language impairment. The exclusion criteria included children with chromosomal abnormalities, autism disorders, obvious neurological disorders, and hearing, visual, movement, cognitive, or mental impairment based on the information in the kindergarten file or the individual questionnaire the parents completed. 
Twenty children with SSD (group one: Four girls and three boys, group two: Three girls and two boys, group three: One girl and two boys, group four: One boy, group five: Two girls, and group six: Two boys) were selected among the clients of speech therapy clinics by a purposive convenience method. They were matched with the normal group in terms of gender distribution and socio-economic level. They were Persian-speaking children diagnosed with SSD by a speech therapist without known sensory damage or physical disorders, such as cleft palate, cognitive language disorder, chromosomal abnormalities, and brain damage. The inclusion criteria were determined using the information in the kindergarten file and parental reports. In both groups, the child’s or parents’ non-cooperation and withdrawal from the study led to the child’s exclusion from the study groups. 

Production of test items
To prepare test items, an item pool comprised 532 words with three, four, five, and six syllables. These words were collected from the literature review using phonetic tests, a book of basic Persian vocabulary from the language of Iranian children, theses, research projects, and an illustrated dictionary of consonants [21-25]. 
In the first stage, the items were reduced to 66 items through three panels, each consisting of three speech therapists who were experts in sound disorders. Their decision criteria comprised the words’ high frequency, significance, familiarity, visual representability of the word, and inclusion of all Persian vowels and consonants in the two positions at the beginning and end of the syllables [26]. Also, the chosen words were culturally aligned with the popular dictionary of Iranian children [19]. Three images were selected for each word, and the experts selected the most appropriate image based on the research objectives. Finally, the test images file was prepared in the form of PowerPoint.

Content validity 
Five SSD experts quantitatively examined the content validity of the items. The experts evaluated 66 items in four areas: Suitability for evaluating production at the level of polysyllabic words, suitability for children aged 4 to 7 years, cultural appropriateness, and imageability of the items. Each word was rated a score from 1 to 100. If words received a mean score of less than 80, they were candidates for deletion. 

Face validity 
The face validity of the images of 66 items was pilot-tested on 8 normal 4- to 7-year-old children. Items that more than 3 children did not identify without clues were subjected to deletion.
Among the 36 items subjected to deletion, 10 items were kept due to their unique syllabic form and the requirement to have all consonants at the beginning and end of the syllable. The items with non-compliant syllabic forms were deleted. Finally, a test containing 40 words (20 three-syllable words, 15 four-syllable words, and 5 five-syllable words) was compiled.

Construct validity of the test 
The test was administered on 120 typically-developed children and 20 SSD children to determine the construct validity. The data of the subjects who participated in the face validity stage were not used in this stage.

 Reliability 
Three methods were employed to determine the test reliability. In the test re-test method, 20 children with normal growth participating in the test were re-tested after one week. To determine the inter-rater reliability, two speech and language pathologists, unaware of the research objectives, phonetically transcribed and scored the speech sample of 20 children with normal growth. Internal consistency was calculated by calculating the correlation between test items. 

Study procedure 
The diagnostic screening subtest of Zarifian et al.’s [19] phonetic and phonemic evaluation test was initially performed during the session. Then, a 5-minute speech sample was recorded using a voice recorder (Creative 30c model). Then, the image file of the production of the polysyllabic word test was shown to the subject. They were asked to identify the names of the images. If the child cannot answer, he or she receives a semantic clue in the form of a description of the target word. In cases of non-response, a phonological clue was given by stating the first syllable of the word. If the child still could not respond, the examiner would try imitation, prompting the child to say the target word. After phonetically transcribing the child’s productions, a score of 1 was given for the correct production of each word (matching adult production) and a score of 0 for incorrect production. The test’s total score (range: 1-40) represents the general ability to produce polysyllabic words. Also, the phonetic accuracy of polysyllabic words was analyzed by calculating the following quantitative measures of their production ability. The percentage of consonants correct (PCC) was obtained by calculating the correct consonants produced divided by the total consonants, the percentage of vowels correct (PVC) was obtained by calculating the correct vowels produced divided by the total vowels, and the percentage of phonemes correct (PPC) was obtained by calculating the correct phonemes produced divided by the total number of phonemes multiplied by 100.

Data analysis 
The obtained data were entered into the SPSS software, version 26. The statistical methods include a one-way analysis of variance (ANOVA) parametric test to compare the research variables across the 6 age groups. Also, an independent t-test was performed to compare the research variables between two genders and two groups of children with normal growth and children with SSD at a significance level of 0.05. The Tuckey test was used to affirm homogeneous age groups. The intraclass correlation coefficient was used to check reliability. 

Results
Construct validity 
In the second phase, construct validity was investigated by comparing between groups. A total of 140 four- to seven-year-old children (120 children with normal growth and 20 children with SSD) participated in the test. The total scores and PCC, PVC, and PPC measures were calculated for them and presented in Table 1. 


The total scores of the test and the phonological parameters of PCC, PVC, and PPC in children with normal growth in different age groups were compared using a one-way ANOVA. Regarding the total score of the test, significant differences were observed between the three groups of 6-5.5, 5-4, and 6.5-7 years old (F(5, 114)=12.5, P<0.001), also regarding PCC, between the three groups of 4-4.5, 4.5-6.5 and 7-6.5 years (F(5, 114)=9.08, P<0.001) and in PPC, between the three groups of 4-4.5, 4.5-6.5, and 6.5-7 years (F(5, 114)=9, P<0.001). No significant difference was found in vowels between different age groups (F(5, 114)=9, P=0.23). 
No significant difference was observed among boys and girls regarding the total score of the test (t118=0.31, P=0.76), the PCC (t118=0.97, P=0.33), the PPC (t118=0.60, P=0.55), and the PVC (t118=-1.06, P=0.29).
We further evaluated the construct validity by comparing the mean scores of polysyllabic word production tests in children with normal growth and children with SSD. In addition to the mean score of the whole test, the PCC, PPC, and PVC scores in these two groups were compared using an independent t-test. As a result, significant differences were observed between the two groups in the mean total scores (t138=14.54, P<0.001), the PCC (t20.22=8.37, P<0.001), the PPC (t20.70=9.60, P<0.001) and PVC (t22.13=2.55, P=0.02). Table 2 presents the results of the test on SSD children.



Test re-test reliability 
The test re-test reliability correlation coefficient was obtained at 0.98 as an intraclass correlation by performing the test twice in 15% of the samples. 

Inter-rater reliability 
The inter-rater reliability was determined by calculating the correlation coefficient between two speech and language pathologists (as evaluators) who independently calculated the total score of the polysyllabic word test in 15% of the samples, resulting in an intraclass correlation of 0.79. 

Internal consistency 
The internal consistency of the test items, which were coded as binary variables (0 and 1), was calculated in 140 children using intraclass correlation. Internal consistency was found to be 0.75 in three-syllable words, 0.65 in four-syllable words, and 0.63 in five-syllable words. 

Discussion
This study was conducted to design a valid and reliable test for assessing the production of polysyllabic words in preschool children. To obtain the best results from such a test, its design should consider special features. The selected features should cover choosing words with varying syllable numbers, phonological forms, and stress patterns of the target language. Also, the word frequency should be considered. In this study, three-syllable words formed a larger number of test words because the frequency of these words in the Persian language is more than one- and two-syllable words. In Bernhardt’s study, the ratios were different, and the ratio of the syllabic numbers of the words in the target language was considered [10, 27]. Besides having a high visual representation, the test words should also reflect the most frequent vocabulary of preschool children. Also, the test should allow common error patterns of the target language to appear. Among the common error patterns in the Persian language [28], cluster reduction, final consonant deletion, assimilation, voicing, and syllable deletion were frequently seen in the production of children in this study. Baker et al.’s test [3] has such a feature among the international studies. The test prepared in this study covers most of the above cases consistent with Baker’s test.
In this study, the criteria of picturability, syllabic and phonetic construction analyses, and the requirement to sample all sounds were considered, like Chavez-Peon's study [8]. However, unlike Chavez-Peon and Phoon's research, we did not use the words with simple syllables because this test was conducted to investigate the ability to produce polysyllabic words in children, not to prepare a phonetic list. The appropriate context to determine the phonetic list of children includes the words that have a simpler form [6]. Hence, this test is not suitable for obtaining a phonetic list. Also, unlike the tests containing shorter words, this study did not require two for each sound in each position [8, 11]. However, to ensure diversity in syllabic construction and coverage of all phonemes, the number of words in this research was increased to 40, even though Masso considered the final number of 30 words sufficient for testing polysyllabic words in English [3]. 
Like Gozzard's study [7], this study incorporated words containing different sounds and syllabic forms, such as clusters. However, unlike the present study, Gozzard used the extraction of polysyllabic words in continuous speech and naming the picture. In this research, like James's study, we sampled the words with different sounds, different positions, structures, stress, and the number of syllables [5]. Like Arciuli's study, all the words were selected from real and familiar words and contained all consonants of the language. However, a specific syllabic structure in the target language was not required [29].
To examine the test’s construct validity, we tried to comply with the results of similar valid studies as closely as possible [3]. In terms of scores, the difference between age groups was significant, which was consistent with the study conducted by Bernhardt et al. [9]. In terms of production ability indices, the differences between age groups were significant in PCC and PPC. In contrast, no significant difference was observed in PVC. This finding was consistent with the study conducted by Bernhardt [10] et al, and Zarifian et al. [28]. James believed that the PCC median values increase with age, which was consistent with the present study, but the increase in the PVC with age in James’ research [5] was not confirmed in this study. The possible reason is that in Farsi, the accuracy of vowels reaches 100% at the age of three [20], and because the children in this study were older than three years, no difference was found between different age groups.
Regarding the effect of gender on the production ability indices (PVC, PCC and PPC), no significant difference was observed between girls and boys, which agrees with some similar articles in this age group, such as Wellman et al. [34] and Zarifian and Fatuhi [20]. However, it was not consistent with some other studies. Hyde and Linn stated that girls performed better in speech production [35]. Kenny and Prather showed that the gender difference is significant up to 4 years [36], and Phoon stated that girls have a higher accuracy [11]. The possible reason for the lack of significant difference between boys and girls was that the children were older than 4 years, and according to Kenny et al., the difference in the production ability between boys and girls is significant only up to 4 years old [36].
Regarding the production ability index (PVC, PCC, and PPC), significant differences were observed between children with normal growth and SSD children, consistent with the results of Bernhardt et al. [10]. 

Conclusion 
The polysyllabic word production test was prepared to obtain a valid and reliable tool to complement the results of the usual tests of speech therapists. The results of this test have applications for treatment planning. For example, in some approaches, simpler sounds and, in some, more complex sounds are the target of treatment. Therefore, depending on the selected approach and comparing the results with the results of the usual production tests, this test can help therapists identify mastered sounds produced correctly in more challenging contexts and which sounds are still developing. 
According to the obtained results, the test of polysyllabic word has good validity and reliability. This study serves as a preliminary investigation. It lays a foundation for further studies in the natural development of sounds and the age of their mastery in polysyllabic words. It also paves the way for a more comprehensive assessment of speech sound disorders in Persian. 
It seems that this test can better address the needs of clinical experts for an accurate tool to evaluate children's speech problems. It allows the therapist to use valid and reliable evidence presented earlier in decision-making in test selection and clinic.

Ethical Considerations
Compliance with ethical guidelines
This study was approved by the Ethics Committee of the University of Social Welfare and Rehabilitation Sciences (Code: IR.USWR.REC.1400.128). Only children whose parents signed the consent form participated in the study. In the consent form, complete information was given about the test and its implementation process, and if the parents were present, more explanations were given, and their questions were answered. Children's information, including their voices and images, remained confidential, and the child and parents had the right to withdraw from the research at any point. 

Funding
This study was extracted from master's thesis of Hossein Abbasi, approved by the Department of Speech Therapy, University of Social Welfare and Rehabilitation Sciences, and finantially supported by the Research Vice-Chancellor of University of Social Welfare and Rehabilitation Sciences (Project Code: A-10-3632-1). 

Authors' contributions
Conceptualization, methodology, validation, editing, and finalization: Talieh Zarifian and Mersedeh Imani; Analysis, research, review, and sources: Talieh Zarifian, Mersedah Imani and Hossein Abbasi; Writing the initial draft: Hossein Abbasi.

Conflict of interest
The authors declared no conflict of interest.

Acknowledgments
Special thanks to Mr and Mrs Yamininejad, who were the managers of Ghoghnoos kindergartens in Tehran and Karaj, and Mrs Hosseinzadeh, the manager of Alama kindergarten and Mrs Khorsandi, the manager of Shadamooz kindergarten and Mrs Mohammadyari, the assistant manager of Farhang School, and all of the employees of those centers.


References

1. Bernthal JE, Bankson NW, Flipsen P. Articulation and phonological disorders: Speech sound disorders in children. London: Pearson; 2013. [Link]
2. Dodd B. Differential diagnosis and treatment of children with speech disorder. New Jersey: John Wiley & Sons; 2013. [Link]
3. Masso S, McLeod S, Baker E. Tutorial: Assessment and analysis of polysyllables in young children. Language, Speech, and Hearing Services in Schools. 2018; 49(1):42-58. [DOI:10.1044/2017_LSHSS-16-0047] [PMID]
4. Benway NR, Preston JL. Differences between school-age children with apraxia of speech and other speech sound disorders on multisyllable repetition. Perspectives of the ASHA Special Interest Groups. 2020; 5(4):794-808. [DOI:10.1044/2020_PERSP-19-00086] [PMID]
5. James DG. Hippopotamus is so hard to say: Children’s acquisition of polysyllabic words [PhD dissertation]. Sydney: The University of Sydney; 2006. [Link]
6. Eisenberg SL, Hitchcock ER. Using standardized tests to inventory consonant and vowel production: A comparison of 11 tests of articulation and phonology. Language, Speech, and Hearing Services in Schools. 2010; 41(4):488-503. [DOI:10.1044/0161-1461(2009/08-0125)] [PMID]
7. Gozzard H, Baker E, McCabe P. Children’s productions of polysyllabic words. ACQuiring Knowledge in Speech, Language and Hearing. 2006; 8(3):113-6. [Link]
8. Chávez-Peón ME, Bernhardt BM, Adler-Bock M, Avila C, Carballo G, Fresneda D, et al. A Spanish pilot investigation for a crosslinguistic study in protracted phonological development. Clinical Linguistics & Phonetics. 2012; 26(3):255-72. [DOI:10.3109/02699206.2011.608463] [PMID]
9. Goldstein BA, Iglesias A. Phonological patterns in Puerto Rican Spanish-speaking children with phonological disorders. Journal of Communication Disorders. 1996; 29(5):367-87. [DOI:10.1016/0021-9924(95)00022-4] [PMID]
10. Bernhardt BM, Hanson R, Perez D, Ávila C, Lleó C, Stemberger JP, et al. Word structures of Granada Spanish-speaking preschoolers with typical versus protracted phonological development. International Journal of Language & Communication Disorders. 2015; 50(3):298-311. [DOI:10.1111/1460-6984.12133] [PMID]
11. Phoon HS, Abdullah AC, Lee LW, Murugaiah P. Consonant acquisition in the Malay language: A cross-sectional study of preschool aged Malay children. Clinical Linguistics & Phonetics. 2014; 28(5):329-45. [DOI:10.3109/02699206.2013.868517] [PMID]
12. Murray E, McCabe P, Heard R, Ballard KJ. Differential diagnosis of children with suspected childhood apraxia of speech. Journal of Speech, Language, and Hearing Research. 2015; 58(1):43-60. [DOI:10.1044/2014_JSLHR-S-12-0358] [PMID]
13. Larrivee LS, Catts HW. Early reading achievement in children with expressive phonological disorders. American Journal of Speech-Language Pathology. 1999; 8(2):118-28. [DOI:10.1044/1058-0360.0802.118]
14. Masso S, McLeod S, Wang C, Baker E, McCormack J. Longitudinal changes in polysyllable maturity of preschool children with phonologically-based speech sound disorders. Clinical Linguistics & Phonetics. 2017; 31(6):424-39. [DOI:10.1080/02699206.2017.1305450] [PMID]
15. Masso S, Baker E, McLeod S, Wang C. Polysyllable speech accuracy and predictors of later literacy development in preschool children with speech sound disorders. Journal of Speech, Language, and Hearing Research. 2017; 60(7):1877-90. [DOI:10.1044/2017_JSLHR-S-16-0171] [PMID]
16. Reilly S, McKean C, Morgan A, Wake M. Identifying and managing common childhood language and speech impairments. BMJ. 2015; 350:h2318. [DOI:10.1136/bmj.h2318] [PMID]
17. Hildebrand MS, Jackson VE, Scerri TS, Van Reyk O, Coleman M, Braden RO, et al. Severe childhood speech disorder: Gene discovery highlights transcriptional dysregulation. Neurology. 2020; 94(20):e2148-e67. [DOI:10.1212/WNL.0000000000009441] [PMID]
18. Murray E, Iuzzini-Seigel J, Maas E, Terband H, Ballard KJ. Differential diagnosis of childhood apraxia of speech compared to other speech sound disorders: A systematic review. American Journal of Speech-Language Pathology. 2021; 30(1):279-300. [DOI:10.1044/2020_AJSLP-20-00063] [PMID]
19. Lieberman P. On the development of vowel production in young children. Child Phonology. 1980; 113-42. [DOI:10.1016/B978-0-12-770601-6.50012-7]
20. Zarifian T, Fotuhi M. Phonological development in Persian-speaking children: A cross-sectional study. International Journal of Speech-Language Pathology. 2020; 22(6):614-25. [DOI:10.1080/17549507.2020.1758209] [PMID]
21. Zarifian T, Ahmadi A, Ebadi A. Development and measurement of psychometric properties of the Persian test of speech consistency in children with typical development. Applied Neuropsychology: Child. 2022; 11(3):226-34. [Link]
22. Damerchi Z, Jalilehvand N, Mahmoudi Bakhtiari B, Keyhani MR. [Development of Phonetic inventory in 2-to-6 year-old Farsi speaking children (Persian)]. Journal of Research in Rehabilitation Sciences. 2010; 5(1):42-7. [Link]
23. Ghasisin L, Ghayoumi Z, Mohamadi Bakhtiari B. [Study of developmental consonant cluster in 2-4 years old farsi speaking children (Persian)]. Journal of Research in Rehabilitation Sciences. 2011; 7(3):391-8. [Link]
24. Nematzade Sh, Dadres M, Kazemi M, Mansoorizade M. [Basic Persian vocabulary from Iranian children (Persian)]. Journal of Curriculum Studies. 2012; 7(27):175-6. [Link]
25. Daryanavard M. [Pictorial inventory of consonants (Persian)]. Shiraz: Moalefan-e- Farhikhteh; 2015. [Link]
26. Masso S, McLeod S, Baker E, McCormack J. Polysyllable productions in preschool children with speech sound disorders: Error categories and the Framework of Polysyllable Maturity. International Journal of Speech-Language Pathology. 2016; 18(3):272-87. [DOI:10.3109/17549507.2016.1168483] [PMID]
27. Eslami M, Rahimi A, Eslami S. [Phonostatistics of Persian phonological system (Persian)]. Language and Linguistics. 2013; 9(18):65-90. [Link]
28. Zarifian T, Modarresi Y, Tehrani LG, Kazemi MD. Phonetic and phonological acquisition in Persian speaking children. Proceedings of the International Symposium on Monolingual and Bilingual Speech. 2015; 430-8. [Link]
29. Arciuli J, Colombo L. An acoustic investigation of the developmental trajectory of lexical stress contrastivity in Italian. Speech Communication. 2016; 80:22-33. [DOI:10.1016/j.specom.2016.03.002]
30. Anderson C, Cohen W. Measuring word complexity in speech screening: single-word sampling to identify phonological delay/disorder in preschool children. International Journal of Language & Communication Disorders. 2012; 47(5):534-41. [DOI:10.1111/j.1460-6984.2012.00163.x] [PMID]
31. Chiat S, Roy P. The preschool repetition test: An evaluation of performance in typically developing and clinically referred children. Journal of Speech, Language, and Hearing Research. 2007; 50(2):429-43. [DOI:10.1044/1092-4388(2007/030)] [PMID]
32. Kehoe MM. An investigation of rhythmic processes in English-speaking children’s word productions [PhD dissertation]. Washington: University of Washington; 1996. [Link]
33. Sutherland DE. Phonological representations, phonological awareness, and print decoding ability in children with moderate to severe speech impairment [PhD dissertation]. Christchurch: University of Canterbury; 2006. [Link]
34. Wellman B, Case I, Mengert I, Bradbury D. Speech sounds of young children. University of Iowa studies in child welfare. Iowa City: University of Iowa Press; 1931. [Link]
35. Hyde JS, Linn MC. Gender differences in verbal ability: A meta-analysis. Psychological Bulletin. 1988; 104(1):53-69. [DOI:10.1037/0033-2909.104.1.53]
36. Kenney KW, Prather EM. Articulation development in preschool children: Consistency of productions. Journal of Speech, Language, and Hearing Research. 1986; 29(1):29-36. [DOI:10.1044/jshr.2901.29] [PMID]
37. Gierut JA. Complexity in phonological treatment: Clinical factors. Language, Speech, and Hearing Services in Schools 2001; 32(4):229-41. [DOI:10.1044/0161-1461(2001/021)] [PMID]