A WEB based decision support system developed for choosing hypothesis testing method

ABSTRACT One of the most important parts of scientific research is the statistical analysis of the data collected with experiments and observations. Unless appropriate statistical analysis method is not used, research results will not be reliable and valid. Most frequent problem the researchers face during the statistical analysis process is the uncertainty for choosing appropriate analysis method. They may choose inappropriate testing methods if they can not access professional help of statistics specialists. A decision support system was designed in order to prevent such errors. System is designed to be web-based in order to enable easy access. Another goal of this system is to become an education instrument not only for students of statistics and research methods, but also for people interested in these subjects. HTML was used for site design. Content of the site includes brief descriptions, schemes and pictures related with basic statistical concepts. Appropriate statistical analysis method is being determined with use of research related questions and the answers of the users. Visual materials about implementation of recommended method with one of the most common statistical software package SPSS (Statistical Package for Social Sciences) are supplied in the site. In addition, two imported JavaScript applications, one of which is designed for providing necessary descriptive statistics and the other is for checking the normality condition for data distribution, were added to the site. These applications are added to provide proper results from the site and to assist users with some calculation processes. As a conclusion, this web site is designed for choosing appropriate statistical analysis method and to assist researchers and students. Key Words: hypothesis testing, parametric test, nonparametric test, statistical analysis methods, decision support system, Internet.