گروه مهندسی کامپیوتر؛ دانشکده فنی و مهندسی؛ دانشگاه گلستان؛
عنوان مقاله [English]
Usability is an important attribute for software quality that can play a significant role in improving user experience. Various methods have been proposed for evaluating the usability of software systems. Selecting a proper method for collecting information and evaluating the usability of a software application is a challenge due to the variety of methods, and consequently the time and cost required for their implementation.
In this paper, ease of use is considered as a subset of usability and it is defined as the user’s effort to perform a task in the software application. We evaluate ease of use based on two methods; questionnaire and users’ performance in an empirical study and compare the results of these two methods. In this way, two different versions of an online store application with similar functionalities are produced and the users are asked to perform specific operations in the application and then answer questions regarding the ease of use of various software functionalities. During the experience, the interactions between user and the software are logged and consequently the user’s performance is measured based on the successful completion of the operation, the time of the task completion, the number of steps taken by the user, the number of errors, and the time taken to fix the error. These criteria have been selected based on observable criteria and ease of use measurement in relevant research.
The results of this study is based on the data of 48 applicants except one of the functionalities which has been expanded to 70 applicantsin order to achieve adequate level of confidence. Our results indicate that the evaluation of the questionnaire does not correlate with the performance of users in the measured criteria. In some functions, the performance of the users in two versions of the application do not have a significant difference, but their views are significantly different in the questionnaire. On the other hand, in some other functions, the performance of the users in two versions has a significant difference but the results of the questionnaire does not show a significant difference. Verifying the performance of each user with the results of his/her questionnaire does not indicate correlation between these two methods results. Also, according to the results of this study, the honesty of users in the questionnaire evaluation can be questioned, since 80% of the users who have not performed a functionality rated it higher than good in the questionnaire.