Examination On The Fairness Of Practical English Test For Colleges Based On Its Test Contents

As a nationwide test, Practical English Test for College (abbr. Pretco) was designed in 1998 and has been administered since then. Along with the increasing influence upon its candidates and users, a question about its fairness is arising. Pretco is administered to the students of all majors from higher vocational and technical colleges. However, by examining the construction of the test, we can find that Proetco, with its stress on practical English, employs a great number of items related with business or trade, such as Contracts, Advertisements, Bids, etc. Naturally, the fairness of the test encounters the challenges from both English teachers and testees, who argue that candidates of business major may take the advantages of their background knowledge of the major while taking the test and they are likely to perform better than non-business candidates. As a result, this can be unfair to non -business majors. The research is meant to check whether test bias exists in Pretco and to examine its fairness based on its test contents and test tasks.Three research questions are raised in the research: 1) Do business candidates really perform better than non-business candidates on Pretco? If yes, is the difference between their performances significant or not? 2) Does candidates'performance on Pretco have relationship with their majors? 3) Are the differences in the performance of candidates on Pretco due to test bias in the test contents or due to their genuine language ability?In order to come up with the answers to the questions, an experiment is carried out. For the experiment, PETS (level 3) is employed as a comparative instrument with Pretco in the research, for there are no controversies with its content bias and it is widely considered as a general English test. Three groups of the subjects who took Pretco in January 2006 are invited to take PETS (level 3). During the experiment many measures are taken to guarantee the accuracy of data. Then the data drawn from the two tests are processed with SPSS for a comparative study. Two descriptive statistics are run respectively on the data of Pretco and that of PETS to see whether there are any differences in candidates'performance on the two tests. Then two one-way ANOVA tests are given to find out whether the differences are significant or not. And a Chi-square test is also carried out to see whether there is relationship between candidates'performance and their majors. Consequently, three paired-sample t-tests are run to see whether there is a significant difference between two test scores of each group and whether there is any correlation between them.The results of the data analysis indicate that 1) the difference in performance on Pretco between business candidates and non-business candidates is significant; 2) Candidates'performances on Pretco have relationship with their majors; and 3) the bias does exist in Protco.