Linguistic and emotional-valence characteristics of reading passages for clinical use and research

• Published in: Journal of fluency disorders 2016-09, Vol.49, p.1-12
• Main Authors: Ben-David, Boaz M., Moral, Maroof I., Namasivayam, Aravind K., Erel, Hadas, van Lieshout, Pascal H.H.M.
• Format: Article
• Language: English
• Subjects: Arousal; Content words; Emotion; Emotions; Emotions - physiology; Familiarity; Female; Fluency; Fluency disorders; Humans; Lexical characteristics; Linguistics - methods; Male; Oral reading; Phonetic properties; Readability; Reading; Reading passage; Research Design; Stuttering; Threat; Valence; Word familiarity
• ISSN: 0094-730X; 1873-801X
• DOI: 10.1016/j.jfludis.2016.06.003

•There is little information on properties of reading passages that affect reading (e.g., arousal and valence).•The three commonly used passages were found to contain emotionally valenced, high arousal, lower familiarity and polysyllabic content words.•The paper provides a new well-balanced (and ranked high on ease of readability) passage that minimizes the impact of these properties.•PWS error rates on a traditional passage and on the novel passage were correlated, yet many individuals showed a large difference between the two.•We suggest a combined procedure, using more than one passage. The details on passage characteristics can inform clinical practice. Fluency assessment in people who stutter (PWS) includes reading aloud passages. There is little information on properties of these passages that may affect reading performance: emotional valance, arousal, word familiarity and frequency and passage-readability. Our first goal was to present an extensive examination of these factors in three commonly used (“traditional”) passages. The second goal was to compare a traditional passage to a new passage, designed to minimize the impact of these properties. Content words were rated (129 participants) on arousal, valence and familiarity. Other linguistic features were analyzed based on available datasets. This information was used to assess traditional passages, and to construct a new well-balanced passage, made of neutral, low-arousal and highly-familiar words. Readability for all passages was tested using formula-based and CLOZE tests (31 participants). Finally, 26 PWS were evaluated on fluency comparing the commonly used “Rainbow” passage with the novel one. The three traditional passages contain a share of emotionally valenced (22-34%), high arousal (15-18%), lower familiarity (6-8%) and polysyllabic (5-9%) content words. Readability was highest for the novel passage (on formula-based scales). Average disfluencies percent for the Rainbow and our novel passage were not significantly different. Yet half of the individuals in this sample showed a large difference between the two passages. We provide detailed information on potential sources of variance using the traditional passages. Knowledge about these characteristics can inform clinical practice (and research). We suggest a combined procedure, using more than one passage to assess stuttering in individual cases.