SOCR Analyses - an Instructional Java Web-based Statistical Analysis Toolkit.

Area of Science:

• Statistics Education
• Computational Statistics
• Data Analysis

Background:

• The Statistical Online Computational Resource (SOCR) offers web-based educational tools.
• Existing SOCR resources have shown to improve student motivation and learning.
• A new component, SOCR Analyses, has been developed for data modeling and analysis.

Purpose of the Study:

• To introduce SOCR Analyses, a new component for statistical data modeling and analysis.
• To detail the pedagogical utilization and design framework of SOCR Analyses.
• To highlight the availability of open-source code and educational resources.

Main Methods:

• Implementation of parametric and non-parametric statistical techniques.
• Inclusion of graphical model diagnostics for analysis.
• Development of commonly used undergraduate statistical models like linear regression and ANOVA.
• Integration of hypothesis testing models such as contingency tables and Fisher's exact test.

Main Results:

• SOCR Analyses provides a suite of tools for data modeling and analysis.
• Implemented analyses cover linear models, sample comparison tests (t-test, Wilcoxon, Kruskal-Wallis, Friedman's), and hypothesis tests.
• The open-source code is available with general utilities and specific methods via an API.
• Free web-based access to HTML interfaces, tutorials, source code, activities, and data is provided.

Conclusions:

• SOCR Analyses serves as a valuable educational tool for undergraduate statistics.
• The open-source nature of the project encourages community contribution and development.
• Continuous improvement and addition of new functions are ongoing for SOCR resources.