Effects Of Response Formats And Item Types On English Reading Comprehension Test Takers’Performance

Reading is not only a main channel to acquire information, but also one of the most basic skills of mastering a foreign language. In many large-scale English tests at home and abroad, reading comprehension usually accounts for a large proportion. To some extent, students’ performance in reading comprehension will greatly influence their total score of a test. Therefore, testers must design valid and reliable tests to assess test takers’ real reading ability.Bachman in1990proposed a framework of test method facets, exploring five major factors affecting testing, one of which is expected response format. Some researches in the past have explored this factor, yet different conclusions have been drawn. Some researchers (Samon,1983) believe response format exerts no influence over test takers’ performance in reading comprehension test while some researchers (Shohamy,1984; Kobayashi,2002) believe otherwise. Therefore, to investigate the effects of response format, enrich testing method and provide some feasible suggestions for test designers when constructing tests and for English teachers in teaching, the author proposed the following three hypotheses based on Bachman’s test method facets theory:1) There is significant difference among testing scores measured by the three response formats (multiple-choice question, short-answer question and true/false question);2) There is significant difference among testing scores measured by different item types (identifying main idea, deducing the meaning of words, making inferences, and skimming and scanning);3) Test-takers with low English proficiency are more susceptible to response formats and item types.In this research,50non-English major college students are divided into two groups according to their scores in National Matriculation English Test: students with high English proficiency and students with low English proficiency. Six almost identical passages are chosen to be the test paper. All the students were required to finish the same test paper at the same time. After the experiment, the author collected and organized all data, and then adopted one-way ANOVA and descriptive statistics to analyze data with the help of SPSS19.0and a computer.The results indicate:1) Response formats have effects over English reading comprehension performance, which is in accordance with Shohamy’s researches. Test takers perform best in true/false question and worst in short-answer questions. Based on this result, test designers can adopt multiple response formats to improve the validity and reliability of a test;2) Item types have influence over English reading comprehension performance. Test takers perform best in making inference and worst in grasping the main idea of the text. English teachers should encourage students to practice different item types in order to bring about the positive effects of reading comprehension tests;3) Test takers with low English proficiency are more likely to be affected by response formats and item types. To discriminate high-level students from low-level students in selective tests, test designers can adopt more short-answer questions.