This is a program to calculate the confidence limits around the proportion of successes in a series of Bernoulli trials (that is, measurements that can have only one of two outcomes such as yes/no, heads/tails, success/failure, and so on.)

This program calculates the limits using the "exact" method. (See the program help for an explanation and pointer to other ways to do the calculation.)


The program presents a dialog-like interface where the user must provide three pieces of information:

  • Trials: The number of measurements made. For example, in a coin flipping experiment, it would be the number of times the coin was flipped.
  • Events: Also called successes. For example, in a coin flipping experiment, this might be the number of "heads" observed.
  • Confidence: The certainty that the "true" proportion is included in the calculated confidence interval. Higher confidence settings produce wider confidence intervals.

Once the information above has been entered, clicking the "Calculate" button will start the calculation of the confidence limits. Depending on the data input, this could take a few moments to complete. After the calculation is complete, the upper and lower bounds of the confidence interval (that is, the confidence limits) will be displayed in the two result fields.

You can quit the program by clicking the "Quit" button. You can get some additional guidance by clicking the "Help" button.

Known Issues and Limitations

  • The maximum number of trials is about 1000, but can vary slightly depending on the input data.


Copyright © 2017 David D. Clark

Distributed under the Eclipse Public License either version 1.0. See the LICENSE file.