Below are reviews of some statistics websites that my colleagues, Michael, Steve, Chris, Harry and I did. These were intended for our HRM-program students and instructors as well.
Reviewer: Michael Boyle
I looked at this website: http://www.kuleuven.ac.be/ucs/java/
It provides a visual introduction into basic statistics (distributions, correlations), allowing the user to vary assumptions and to see what happens visually. It would be helpful to identify this site to students who might want "to play".
Reviewer: Steve Hanna
I was assigned to review: http://ebook.stat.ucla.edu/calculators/
This is a collection of ten or so statistical and graphing utilities, posted by Jan de Leeuw at UCLA. These include some utilities for simple power/sample size calculations, graphing, look-up tables for sampling distributions, and some simple analyses (calculating a correlation, etc). It is not comprehensive. Most of these things would be available in a standard statistical package, so I expect these things were devised for specific teaching applications. There are no instructional resources, and very little documentation of the utilities, so you would mostly have to know what you are doing, or be following along in a text. The power calculations would be useful for very simple designs, like comparison of two means or proportions. The utilities for standard distributions are probably the only reason to visit.
Here is another site I strongly recommend and don't see on this list: http://www.math.uah.edu/stat/index.html
Virtual Laboratories in Probability and Statistics is an excellent introduction to mathematical statistics at the University of Alabama and is funded by the US National Science Foundation. It covers the basics of probability and statistical inference from a mathematical approach, so it would appeal primarily to learners with some experience in calculus and mathematical notation. It would have more limited appeal for students of applied statistics. It makes excellent use of examples and demonstrations to coincide with formal mathematical treatments of basic concepts. I highly recommend it.
I also highly recommend John Pezzulo's site that Chris has already reviewed: http://members.aol.com/johnp71/javastat.html This is the place to start if you are looking for any kind of calculator, particularly power.
Reviewer: Chris Woodward
The site I initially explored was not very interesting. It allows you to calculate sample size, does simple analysis, etc. However, following links, I ran across this site: http://www2.chass.ncsu.edu/garson/pa765/statnote.htm that I consider quite excellent, although the quality and level of the material varies somewhat by topic. You may wish to look at this site for potential materials related to topics you are teaching.
We would need to get permission of the faculty member who developed the site to use the material unless we simply use it as an electronic resource which clearly indicates that putting all this together is someone else's work.
Back to topReviewer: Lehana Thabane
Re: http://www.stat.vt.edu/~sundar/java/applets/
This is a great site! It is particularly good for introductory level biostatistics courses. It covers a wide range of basic statistics concepts from probability, the CLT (central limit theorem), hypothesis testing to applications in quality control, etc. For each statistics concept covered, it starts with a brief description of the theory and provides excellent visual display of simulation to illustrate the concept. For example, to illustrate the CLT, it first explains what it is and then allows one to generate data from different distributions to illustrate it. In addition, one can visualize how changes in the sample size and number of samples affect the distribution of the mean. A similar sensitivity analysis is provided for various problems such as hypothesis testing for means or proportions where the results may vary by (1) the amount of clinically important difference to be detected, (2) the level of significance, or (3) the sample size. Another impressive example is the illustration of power calculations and how power might be affected by the above quantities. All these are nicely demonstrated with interactive visual displays.
Several applications with live graphical displays of the results are provided. I was very impressed with the site and would recommend it to anyone teaching introductory to intermediate biostatistics courses. It is an excellent way of visualizing statistics.
Reviewer: Harry Shannon
My website to check was: http://www.ruf.rice.edu/~lane/stat_sim/index.html
There is an enormous amount of material accessible through this web site: a text "book", simulations which you can control to an extent, case studies and some analysis tools. The web address from Chris just got you to the simulations; you have to go back to the main menu for the other three components.
There is so much, it's hard to know where to start describing it all. Some of it is great, but there are flaws.
The text book has the contents page clearly outlined, and next to it is "Related Material" with links to other on-line texts, recommended hard copy books, on-line stats courses at various universities and colleges, and even some statistics jokes.
When you go into the text, there are hyperlinks to, for example, definitions and concepts, which is nice. One problem is that each new page has several new links so it feels that you can carry on almost forever going to new ones, getting lost in the maze.
Another thing I noticed was that there were some errors. For example, the following statement in one of the case studies: "Twenty one percent of the patients on the AHA diet had some sort of illness compared to only 10% of the patients on the Mediterranean diet. Thus, the chance of remaining healthy was twice as high for a patient on the Mediterranean diet." Also, sometimes I clicked on a link that wasn't there, and there were references to Greek symbols as population parameters, but they just ended up in Latin script (at least on my browser). This would cause students great confusion. Given how little I reviewed, this high error rate is a bit disturbing (unless I sampled badly).
I suspect that some of the texts you can link to may be better. For example, there are pieces written by Gerry Dallal--I got to know him a bit on sabbatical, and am sure he would be pretty much error-free in his concepts and content.
The case studies show you the concepts that will be covered. They sometimes include quite detailed how-to-do-it-on-the-computer sections for different packages, including SAS, JMP and SPSS, which would certainly be useful for our students.
Overall, this is a potentially very valuable resource, but one to be used with caution.
Other interesting sites:
http://acad.cgu.edu/wise -- This web interface is for statistics with links to more statistics resources, datasets, on-line journals and discussions.
http://www.statsoft.com/textbook/stathome.html -- This is an Electronic Statistics Textbook that provides general statistics training and applications.