Language sample analysis: development of a valid language assessment tool and determining the reliability of outcome measures for Farsi-speaking children.

The present study determined how to elicit language samples from Farsi-speaking children, which language measures should be analysed, and whether these analyses are reliable. Two valid sets of picture stories were developed to elicit the language samples. Language measures were chosen by a panel of experts and the reliability of the measures was verified by test–retest reliability. The subjects were children 5–6 years of age (N = 30) who told stories twice at a 7–10 day interval. The results of inter-rater reliability showed that consistency of measurement was high for the transcription and analysis of the stories. The results of test–retest reliability showed there was a correlation between most variables in the longer samples (p < .05). This study demonstrates that language ability can be more reliably assessed when longer language samples are collected. [ABSTRACT FROM PUBLISHER]