Item Arrangement as a Strategy for Enhancing Quality in Assessment in Science Education

ABSTRACT Performance in a series of real-life tasks tends to be maximized if such tasks are presented in some order from the least- to the most-demanding given the ability being tasked. Such arrangement enhances ease and objectivity at estimating the amount of ability expended to overcome a task, as well as the minimum level of ability demanded by each task before it could be overcomed without much extraneous influence from anxiety, demotivation, and waste of precious time by the person undertaking the task. In testing, it also makes it less difficult to determine the probability that an answer to a test item is guessed or copied during examination, or results from carelessness during examinee-by-item interaction. To test three hypotheses related to these views, performance and item data generated by administering 3 forms of a validated chemistry test, each with different item arrangements, to 461 randomly sampled senior secondary three (SS3) students from Akwa Ibom State in Nigeria were analysed using repeated measure and one-way ANOVA. It was found that item position in a test has significant influence on several person and item characteristics and hence influences the quality of assessment instruments. Based on these findings appropriate recommendations were made in an attempt to contribute solutions to this problem and to enhance adaptability of our testing procedures to practices supported by modern test theories.