Planning an experiment

This is an Excel spreadsheet template for calculating how long you might need to run an experiment to see statistically significant differences in count rates.

You enter your expected baseline count rate, and your expected difference between counts under different conditions. The ss calculates total counts during increasing numbers of seconds, and the uncertainty in that total count number. As you'll see, the smaller the expected difference, the more counts you need to collect, and the longer the detector needs to run.

We used this for an experiment in a 15 story steel & concrete building. We used 2 counters set to count when a 2 way coincidence occurred. The counters were separated by about 60 cm vertical distance. We planned to take counts on the 15th floor, 8th floor, and the basement. We expected, based on just a hunch, that the building would have very little effect on counts. We decided to take enough counts to demonstrate a real difference at a 1% level. This spreadsheet would calculate about an hour of counting time to collect enough counts to prove a difference at a 1% level.

By the way, the actual difference was about a 2/3 reduction in counts as the detector moved from top to bottom of the building.

Here's the Excel spreadsheet for planning an experiment: Planning an Experiment