| المستخلص: |
This thesis was conducted to investigate the effectiveness of two stage adaptive testing among a sample of second secondary scientific grade students , through answering the following questions : 1- What is the difficulty of each test item as estimated among the ( winsteps ) ? calibrated sample using 2- What is the ability of each student (θ1) as estimated by the ( Routing Test ) ? 3- What is the ability of each student (θ2) as estimated by the ( Second stage adaptive test ) and ( Routing Test ) ? 4- What is the ability of each student (θ3) as estimated by all test items ? 5- How close are the ability estimates (θ1 , θ2 ) to the ability estimates for the whole test ? 6- To what extent does the increase in the number of the items used to estimate the ability (θ) decrease the standard error of measurement in the three stages of testing ? 7- what is the correlation coefficient between ability (θ) estimates for each stage of testing and school grades ? 8- what is the correlation coefficient between ability (θ) estimates for each stage of testing and GCSE scores in biology ?
For the purpose of the study , an achievement test on genetics unit from biology text book for the second secondary scientific grade was constructed ; the test consisted of 55 multiple choice items , after calibrating the items using ( winsteps ) , designed by (Wright and Linacre , 1993 ) . six sub tests were derived : the first , a ( Routing Test ) which was used to direct students to one of the other five tests which formulated the second stage of the adaptive testing . three values for each student were estimated ,the first (θ1) on the ( Routing Test ) , the second (θ2) on the ( Routing Test ) and one of the tests for the second stage suitable to the ability test ,and the third (θ3) for the total items of the test . The standard error of estimation was calculated and Correlation coefficients between the ability for each student and the school grades and biology scores in GCSE . The study revealed the following finding : - The mean for the students ability estimation on the (Routing Test) was 1.29 logit ,the standard error of estimation was 0.87 logit . - The mean for the subjects ability estimation on ( Routing Test ) plus a version of second stage adaptive tests was 1.71 logit ,and the standard error was 0.64 logit . - The mean for the subjects ability estimation on total test items was 0.98 logit and standard error of estimation was 0.36 logit . - Using T-test ,the differences between (θ) values for the different testing stages were significant . - Correlation coefficient between (θ) values and the school grades were as follows: θ1 and school grades is 0.75 . θ2 and school grades is 0.70 . θ3 and school grades is 0.76 . - Correlation coefficient between (θ) values and biology scores in GCSE were as follows: θ1 and scores in GCSE is 0.46 . θ2 and scores in GCSE is 0.42 . θ3 and scores in GCSE is 0.52 . In general the use of the adaptive testing for ability estimation indicated higher values than those scored on the whole test items , also the standard error of estimation decreased when the number of items increased . The researcher recommends more studies on the implementation of adaptive testing especially the computerized adaptive testing .
|
