Assessing phonological short-term memory in Greek: Reliability and validity of a non-word repetition test

This study explores the reliability and validity of a NWR task in a large cohort of 387 TD Greek-speaking children aged 7-13 years attending elementary (Grades 2-6) and secondary school (Grade 1), divided into six age groups. Further, the relationship between NWR and reading fluency skills as well as the predictive value of the NWR on reading fluency skills in TD children are examined. To investigate the external reliability of the NWR task, test-retest reliability was performed, and excellent test-retest reliability was found. Internal reliability was explored with Cronbach's alpha coefficient and good reliability was found. To explore convergent validity, correlation analysis between NWR and reading fluency was conducted and significant and strong correlations were found for all age groups excepted 2 (ages 9-10 and 12-13). To examine predictive validity, regression analysis was conducted between these two variables and showed that performance on NWR contributed significantly to reading fluency skills, suggesting that NWR skills are a good predictor of reading skills. Finally, it was explored whether the relevant scores increase as a function of age and found significant differences between groups that differed in 2 years or more, while this difference was no longer significant after 10 years. This finding suggests that phonological STM increases in capacity along with age, but only until the age of 10, where it seems to reach a ceiling. In addition, linear regression analysis showed that age contributed significantly to performance on NWR test. To sum up, the present study provides normative data of a NWR test for a wide age range, which does not exist in the Greek language (particularly for ages over 9 years) and it can be concluded that the present NWR test can be successfully used as a reliable and valid measure of phonological STM in the age range that was examined in this study.