The effect of test method on reading comprehension: A study of three different test types used for measuring L2 reading comprehension.
Independent thesis Advanced level (degree of Master (One Year)), 20 credits / 30 HE creditsStudent thesis
The use of standardized test has grown rapidly over the last few decades. Most people will recognize the reading comprehension tests taken for educational use where the result is used for individual comparison and not seldom in setting ground for a course grade, college admission or academic-program placement (Koda, 2004). Due to the complexity of comprehension there has been, and still is, an ongoing discussion on how to best assess reading comprehension (Urquhart and Weir, 1998). There are mainly two categories of test types; selected-response test methods and constructed-response test methods (Cohen and Swerdlik, 2005). In the study at hand the test results from three tests were investigated, one free recall test (constructed-response method) and one matching test (selected-response test) and one multiple-choice test (selected response test).The subjects were 83 Swedish Biology students at Stockholm University. All subjects had Swedish as a native tongue and could be considered advanced L2 readers of English. The point of interest was to investigate if it is possible to see a tendency for one of the reading comprehension test methods being less of a challenge and to test the findings of Lee’s (1986) research i.e. a student writing a constructed-response test in his/her native tongue will show better understanding of the text than a student writing the same test in his/her L2. The results showed that reading comprehension performance is affected by different test methods. Out of the test types studied the selected-response tests were significantly less difficult than the constructed-response test.
The result from the recall task showed that the subjects writing in their native tongue, here Swedish, had a better score than subjects who chose to write in their L2, English, which in turn confirms Lee’s (1986) findings. T-test results showed that despite the fact that students performing the recall task in their native tongue showed a better performance on that test, they were inferior on the selected-response tests compared to the group of subjects who wrote the recall in their L2, English (p<0.05). This could indicate that subjects performing the free recall task in English were hindered by their language skills and not due to lack of comprehension.
A discussion of the use of reading comprehension tests and what the results can tell follows and it is argued, in accordance with Farr and Carey (1986), that it is necessary to recognize that test methods should not be the only way to gather information about individuals’ reading comprehension skills.
Place, publisher, year, edition, pages
2008. , 36 p.
IdentifiersURN: urn:nbn:se:su:diva-26755OAI: oai:DiVA.org:su-26755DiVA: diva2:211261