Confidence Intervals

There are several choices to make:

1. How many draws per experiment? At the top, there is a place for you to type in how many tickets should be drawn for each experiment.
2. The approximate confidence percentage: Next to where you type in the number of draws per experiment, there is a place to choose what you want as the target percentage chance for the confidence interval. The choices are 99%, 95%, 90%, 85%, and 80%. (These percentages are approximate because the confidence intervals are calculated using the normal curve approximation.)
3. Are the draws with or without replacement? Choose "w rep" for drawing with replacement, and choose "w/o rep" for drawing without replacement.
4. How quickly should the tickets be drawn? The choices are
1. Fast: Which may not be very fast, depending on the environment
2. Slow: Which is useful for seeing each draw and its effect
3. Batch: This is the fastest. The draws are performed without updating the graphics after each draw, but just after all draws have been made.
5. How many experiments to perform? The choices are 1, 10, 25, 100, and 1000. Click on the button that you wish. You can click on the buttons more than once to request more experiments. For example, if you wish to have 250 experiments, click on "100" twice and "25" twice. (Don't click on them too quickly or they won't all register.) Right below the numbered buttons, there is a number showing how many experiments have been requested so far.
6. How many more experiments to perform? Even after the experiments have been completed, you can ask for more experiments by clicking on the numbered buttons again.
7. Start Over: This button sets everything back to 0. You can then make your choices again.
8. Pause: This button stops the drawing. You can restart the drawing from where it left off by pressing "Resume" (which is the same button), or you can hit "Start Over" to start over.

There are several results to watch:

• The draws for each experiment: At the bottom, a list of the values on the tickets drawn for the current experiment are given.
• The calculation of the confidence intervals: In the middle, these is an area that prints out the confidence interval for each experiment. It gives the Average +- z*(Standard Error of the Average) for each experiment, as well as whether the interval covered the box average (which is good) or missed the box average.
• The summary of results: Next to the printing out of the intervals, there is a record of how many intervals so far covered the box average, how many missed, and the percentage that covered.
• The graph of the intervals: Below the printing out of the intervals there is a plot that draws the interval for each experiment. The horizontal axis indicates experiments' numbers, and the vertical axis represents the values in the interval. The light blue line on the graph indicates the box average, so that intervals that cover the box average cover this line. Intervals that cover the line are blue, and those that miss are red.