Investigating the possible hierarchical order of reading skills using the many-facet Rasch Model

• Main Authors: Badrasawi, Kamal J I, Abu Kassim, Noor Lide
• Format: Conference or Workshop Item
• Language: English
• Published: Springer 2012
• Subjects: PE English
• Online Access: http://irep.iium.edu.my/31539/1/31539.pdf; http://irep.iium.edu.my/31539/

Reading ability in English as second or a foreign language is highly demanded as English has been extensively used in all fields of human knowledge. Thus, much research has been conducted to identify the nature of reading skill in L2. Two major views could be figured out: reading as a ‘unitary’ skill and reading as ‘multidivisible’ skill. Despite this, the literature shows a lack of consensus in determining the number of the skills/sub skills that reading includes, and whether they are hierarchically ordered. Some studies have found that item difficulty is influenced by a number of factors other than person ability such as item test features i.e., the interaction between item characteristics and item difficulty. These features include, among others, text type, context type, test length, item format etc. As such, it has become problematic to ascertain the hierarchical assumption of reading skills. To resolve this, the multifaceted approach has been recommended to examine item difficulty taking into consideration factors which affect item difficulty level. This study employs the Many-faceted Rasch Model to ascertain the hierarchical assumption of reading skills for support of a developmental of reading ability. A 42-MCQ item test was administered to 944 ESL secondary students in Malaysia. The test items were identified according to the skill areas associated with them (interpreting information, making inference, understanding figurative language, drawing conclusions, scanning for details, and finding out word meanings); context type (linear and non-linear); and text type (ads, notices, a chart, a story extract, short messages, a poem, a brochure, and a formal letter). As connectivity is established for skill areas and context type only, the results for these two aspects are reported as they can be compared directly and replicated. The results show that item context types are not equally difficult; linear context types tended to be more difficult than non-linear context types. And the skill categories do not have the same difficulty level. The most difficult skill category was interpreting information and the easiest one was finding out word meanings. This supports that notion that different reading skills exert differential cognitive demands and those that require higher order thinking skills such as analysis are more difficult than those requiring lower order skills such as finding out word meaning and scanning for information. To conclude,the results of the FACETS anlyses have provided the much needed evidence that there is a strong possibility of such reading hierarchy.