Here is some feedback. Remember, I'm trying to see this from the point of view of a beginning Statistics student who will probably take things at face value.
Z DISTRIBUTION
I would prefer word descriptions to the image labels. The students will be tested verbally (either multiple choice or essay), not by picking pictures and reinforcing the terminology would be good. There is room at the bottom to display (for example) "two-tailed" followed by the appropriate image button. Yes, the SIP will cover this up but I can assume the used knows how to minimize the pop-up keyboard when she or he is ready to select the appropriate tail(s) for the test.
Instead of a checkbox for z, it would be clearer to the new student if there were two radio buttons where one could choose between z and P(z).
For any given z score, the program correctly computes the left (1-tailed) probability but displays the same value for the right (1-tailed) probability which should be 1 minus the left tailed probability. Curiously, the graph correctly displays and labels the right tailed probability.
The P(z) textbox permits an entry greater than 1. Although no computation is performed when a button is clicked, neither is the student warned of impossible input. When a negative value is input, the program automatically changes it to a positive value. I would prefer an error message. It is harder for students to learn when their mistakes or misunderstandings are automatically corrected for them.
The display uses P(z) whereas the other three distributions use alpha.
The default on this distribution is to compute P(z). For the other three distributions the default is to compute the test statistic. Better if the default is the same for all four distributions.
On the graph page, I enter a mean of 100 and an SD of 15 which produces a picture of an IQ distribution (very nice). But if I exit to the first page and then hit the graph button again, I get an error "An error occurred on sub main.areafill. Line number 746".
CHI-SQUARE DISTRIBUTION
A bug - I can't get the df box to get focus, so no input there is possible.
Again, I would prefer two radio buttons to the one checkbox for alpha.
No graphic (tails) image buttons, one has to push the +- button. After learning to push the buttons at the right on the Z page, one has to now figure out the correct button to push. Why not a graphical image button here, even if it is only one button? Yes, I know Chi-square distribution can vary a lot in shape, but even so a representative button would be more intuitive at this point (the button should also have a verbal description, such as mentioned above).
t DISTRIBUTION
Same comment on radio buttons and descriptions for image buttons.
A df > 1 not only does not generate an error message, it performs a calculation which the student (not understanding probability) might accept as correct. An error message should be generated.
Although I haven't done much in the way of checking results, I notice in passing that when calculating t for df=12 and alpha = .10, I get the correct result for the 2-tailed test, but the 1-tailed test returns 2.18 instead of the correct answer of 1.36.
F DISTRIBUTION
Again, an alpha > 1 does not generate a warning.
If a calculate an F for alpha = 0 and df = 2/22 I get a very large F. If I then correct alpha to a realistic value (say .10) and push the +- button an error is given "An error occurred on sub main.dchoice. Line number 466."
IN GENERAL
The exit door button directly below the standard windows "X" button is redundant. I found myself pushing it to see what it would do and, of course, had to restart the program.
Baloon tips would be a great learning aid.
It would be most useful for the program to detect when infinity is the correct answer, and to supply that answer.
Ideally, I would like the ability to graph all four distributions, not just the Z distribution. The t distribution in particular would be great because it is used just as the Z distribution is used, only for small samples. I know this is wishful thinking and I don't expect you to jump on it, but it sure would be nice . . .
I really hope I don't sound too nit-picky. I know you have put a lot of work into this program and I really appreciate it. Some of these changes I would make myself if I had the time (which I don't at the moment). My main priorities are (1) accuracy of results, (2) consistency in the user interface, (3) making the usage as intuitive as possible, and (4) giving student's a learning experience.
