WebStat provides a spreadsheet type environment in which statistical calculations can be performed. Data can be imported from and exported to files. Sample data files are provided. Documentation of the functions is provided.
WebStat is designed to allow students to perform a variety of statistical tests on data sets.
Target Student Population:
Anyone from students in introductory statistics classes to researchers. As soon as the basic ideas are clear, WebStat can be used.
Prerequisite Knowledge or Skills:
Knowledge of appropriate statistical tests
Type of Material:
Simulation, reference material
Designed to run on PC/MAC/UNIX. Tested on PCs with WIN 95, 98, 2000. The authors are available through email for technical support under firstname.lastname@example.org.
Evaluation and Observation
This is a versatile computational tool that almost has the ease of use of a spreadsheet and some functions that may not be in a spreadsheet. Software is superb with excellent results in all areas. 1) Data sets of various sizes can be created or imported and sample data sets are included. 2) A wide variety of statistical tests that include z-tests, t-tests, linear regression, control charts, hypothesis testing, confidence intervals, ANOVA and all topics covered in elementary statistics courses provide detailed results. 3) Graphing features allow use of colors and labeling for histograms, pie charts, box plots, scatter diagrams and others. 4) This software allows the student to perform statistical tests without knowledge of the language of a statistical package or a spreadsheet. Topics are organized as data, statistical tests, or graphics and results are presented in a format that students can easily interpret. 5) Instructors will be delighted with the ability to include a link on their webpage for students to access WebStat 2.0. Documentation is complete so that assignments using the software would be easy to create and data would could be easily formatted and imported.
Instructions distinguish between the z-test and t-test only by the availability of the population variance. A further note that a small sample size of less than 30 requires a population that is normally distributed, or neither test is appropriate, would reduce potential errors in utilization and conceptual misunderstanding.
Potential Effectiveness as a Teaching Tool
WebStat is the perfect support software for a basic statistics course if a spreadsheet is not available. Essentially any computation WebStat supports (and there are quite a few) can be automated. If you can incorporate a spreadsheet into the course,
then you can incorporate WebStat. This should be attractive to schools that don't want to commit to a commercial product. On the instructional side, beginning, non-mathematical students may find the learning curve for a spreadsheet intimidating and WebStat could give a smoother "link up". Excessive time spent in formula manipulation and numerical calculations are eliminated without sacrificing the learning of concepts, since the user must interpret the results provided. Financial constraints for students and schools would be another argument for WebStat (freely available spreadsheet products notwithstanding). WebStat is not intended to replace a textbook. Sample data sets are supplied with WebStat, which can be very convenient for effective class preparations. WebStat is highly effective as a teaching learning tool because it requires the user to develop problem-solving techniques by selecting the appropriate statistical test and interpreting the results. For example, the software does not warn the user that a t-test may not be appropriate and will actually perform the test on a sample size less than 30. Nor does it interpret the z-statistics and p values provided in hypothesis testing. Instead the software performs the calculations requested by the user and allows the user to interpret the results within the context of the problem
The lack of warnings if a test is appropriate or not can also be considered a concern (see above).
Ease of Use for Both Students and Faculty
Features are well organized in groups that include data, statistical tests and graphs. Navigation between the groups is excellent. Documentation is concise, readily available and easy to understand. It takes the user step-by-step through the computations. Statistical results are detailed and immediate, even when using slower processors and modems. Printing capabilities exist for data sets,
statistical results and graphs
As a learning tool it would be nice to have references to some background material. Given the assumption that the user has had or is attending an appropriate statistics class, this is not a large concern. In performing hypothesis testing using two-tailed and one-tailed tests, the statistical results did not seem to state the test used. It was only after several trials where I manually recorded the type of test requested, that I realized the slider bar would reveal information such as "two-tailed z test results" above the calculations presented. It would be helpful if this line could be included in the screen of current results.