ACQ Vol 11 no 2 2009

Reading Scale (Hagley, 1987, cited in Nation & Snowling, 1997) tap different underlying skills. The Suffolk Reading Scale is a group-administered reading test, which uses a multiple choice sentence-completion (cloze) format (much like the WRMT-R), although there is a test time-limit of 20 minutes and the children are encouraged to spend the maximum time allowed. A total of 184 children, attending year 3 and year 4 of primary school, participated in the study. The results from Nation and Snowling’s study showed that the reading comprehension measure derived from the NARA was dependent on both word recognition and listening comprehension ability. In contrast, performance on the Suffolk Reading Scale was mostly dependent on word recognition ability as measured by single word reading. In practice this means that children who fit a specific comprehension deficit profile (i.e., reading comprehension difficulties as a result of listening comprehension deficits) may show adequate reading comprehension performance on tests using cloze formats, for example the WRMT–R or the Suffolk Reading Scale, but impaired performance on the NARA. Based on these results Nation and Snowling recommend the use of either the NARA or an individually administered test of listening comprehension to ensure adequate detection of specific comprehension deficits in children with reading difficulties. A more recent study into the NARA’s ability to differentiate between reading accuracy and reading comprehension problems, however, found that children with poor word recognition skills may be incorrectly identified as demonstrating a reading comprehension deficit (Spooner, Baddeley, & Gathercole, 2004). Because of the nature of the NARA, in which testing is discontinued once a child exceeds the number of permitted decoding errors on a passage, the correlation between reading accuracy and reading comprehension scores is inherently high for children with word recognition difficulties. In a response to this study Cain and Oakhill (2006) argued that the NARA is an effective tool in identifying children with specific comprehension deficits when used in conjunction with independent measures of listening comprehension skill and decoding ability. Keenan, Betjemann, and Olson (2008) attempted to confirm previous research regarding reading comprehension tests’ dependence on word recognition and listening comprehension skills. They also investigated if the

weaknesses in word recognition and/or listening comprehension skills, a reading comprehension deficit may go unnoticed (Bowyer-Crane & Snowling, 2005). To assess reading comprehension, most tests require children to read sentences or passages, either silently or out loud. Children’s comprehension of this written material is then tested by: 1) asking open-ended questions (with or without the text to refer back to), 2) asking multiple choice questions, or 3) using a cloze type procedure in which the children are asked to fill in the missing (key) words. Some frequently used reading comprehension assessments in Australia are listed in table 1. The main issues that need to be taken into consideration when choosing a specific test for reading comprehension include: • test format – the way in which the test measures reading comprehension; • passage dependency – do you need to read the passage to be able to answer the questions?; • a child’s test taking strategies – e.g., attention to detail, persistence. These factors will be discussed in more detail below. Test format The most commonly used test formats to measure reading comprehension ability are cloze tests and question-and- answer tests. An example of a test utilising a cloze task is the Passage Comprehension subtest from the Woodcock Reading Mastery Tests – Revised (WRMT–R: Woodcock, 1998), in which the child is required to silently read one or two sentences (the first 30 items), or a short passage (the remaining 38 items), and identify a missing key word. In contrast, the comprehension score derived from the Gray Oral Reading Tests – 4th edition (GORT—4; Wiederholt & Bryant, 2001) is based on the child’s ability to answer multiple choice questions which are read out by the examiner, after the child has read aloud a short passage and the text is removed. Other examples of question-and-answer tests are the Wechsler Individual Achievement Test – 2nd edition (WIAT-II; Wechsler, 2001) and the Neale Analysis of Reading Ability – 3rd edition (NARA; Neale, 1999). In both the WIAT and the NARA, children read out loud and are asked open-ended questions by the examiner while the text stays in view. Nation and Snowling (1997) investigated whether an older British version of the NARA (Neale, 1988) and the Suffolk

Table 1. Examples of frequently used reading comprehension assessments Reading comprehension Test procedure

Format

assessment NARA: Neale Analysis of Reading Ability – 3rd ed. (Neale, 1999)

Student reads passages out loud. Decoding mistakes are corrected by examiner.

The examiner asks

Testing is discontinued if the student makes more than 12 decoding mistakes in one open-ended questions passage. Open ended questions about the passage are asked immediately afterwards. The text stays in view.

GORT–4: Gray Oral Reading Tests – Student reads narrative or expository passages out loud, as quickly as possible.

The examiner asks

4th Ed. (Wiederholt & Bryant, 2001)

Examiner asks 5 multiple choice questions immediately following the reading. Text is multiple choice removed from view. Testing is discontinued when students answers 3 out of the 5 questions questions incorrectly. Student reads silently. There are 12 reading passages. Student reads a passage and Untimed. Modified then uses a cloze answer sheet to retell the passage, filling in the gaps in his/her own cloze procedure; cloze

TORCH: Test of Reading

Comprehension (Mossenson et al.,

1987)

words to demonstrate understanding.

style pencil and paper test

WRMT–R: Woodcock Reading Mastery Test – Revised : Passage Comprehension (Woodcock, 1998)

The student is required to silently read one or two sentences (the first 30 items), or a Cloze procedure

short passage (the remaining 38 items), and identify a missing key word.

82

ACQ Volume 11, Number 2 2009

ACQ uiring knowledge in speech, language and hearing