A common question: "My measurements always change, am I operating my equipment right?"

In short, yes. If you were getting the same exact number EVERY time, I would actually have some concerns.

Scientific facts are verified by repeatable careful observation.

If a Geiger counter is placed near a radioactive source, the number of counts (N) in a time interval (T) will exhibit random fluctuations with repeated trials of the same measurement. The most likely value of the counting rate would be the mean value N while the amount of "spread" about the mean is given by the standard deviation SD (or σ).

Radioactive decay and the emission of particles is a completely random process. Given 1,000 radioactive atoms we cannot predict which atom will decay at what point in time, the best we can say is x amount of 1,000 will decay within x time period.

Additionally radiation is emitted in all directions from the source, and the directions of the emitted radiation are also random - meaning the operator will never be seeking an absolute value as much as a range that is representative of the sample being analyzed.

Analyzing a sample is not a search for the highest possible measurement, rather it is a quest to establish the range of activity and reduce possible errors (random, operator, or otherwise) as much as possible.

What the operator should do is take repeated measurements and average out the results.

To demonstrate the random fluctuation of measurements I set up a detector with a scaler and took a series of measurements. The results were as follows.

The first thing that I did after powering the unit on was take 10 1-minute measurements.

The simplified average is just the sum of all ten measurements divided by 10.

But remember, as the operator I want to select a value that is most representative of the sample I'm looking at. I also know that there are fluctuations. By ignoring the highest and lowest values I can try to get a more definitive result.

Keeping in mind that the results should be repeatable, I took another 10 1-minute measurements. The results were in the same range as the first series.

I ran a third series of 1-minute measurements. The longer I operated the equipment the tighter the grouping of measurements became. (This can easily be seen by looking at the distance separating the simplified and adjusted averages of each series)

In short, yes. If you were getting the same exact number EVERY time, I would actually have some concerns.

Scientific facts are verified by repeatable careful observation.

If a Geiger counter is placed near a radioactive source, the number of counts (N) in a time interval (T) will exhibit random fluctuations with repeated trials of the same measurement. The most likely value of the counting rate would be the mean value N while the amount of "spread" about the mean is given by the standard deviation SD (or σ).

Radioactive decay and the emission of particles is a completely random process. Given 1,000 radioactive atoms we cannot predict which atom will decay at what point in time, the best we can say is x amount of 1,000 will decay within x time period.

Additionally radiation is emitted in all directions from the source, and the directions of the emitted radiation are also random - meaning the operator will never be seeking an absolute value as much as a range that is representative of the sample being analyzed.

Analyzing a sample is not a search for the highest possible measurement, rather it is a quest to establish the range of activity and reduce possible errors (random, operator, or otherwise) as much as possible.

What the operator should do is take repeated measurements and average out the results.

To demonstrate the random fluctuation of measurements I set up a detector with a scaler and took a series of measurements. The results were as follows.

The first thing that I did after powering the unit on was take 10 1-minute measurements.

The simplified average is just the sum of all ten measurements divided by 10.

But remember, as the operator I want to select a value that is most representative of the sample I'm looking at. I also know that there are fluctuations. By ignoring the highest and lowest values I can try to get a more definitive result.

Keeping in mind that the results should be repeatable, I took another 10 1-minute measurements. The results were in the same range as the first series.

I ran a third series of 1-minute measurements. The longer I operated the equipment the tighter the grouping of measurements became. (This can easily be seen by looking at the distance separating the simplified and adjusted averages of each series)

"All models are flawed, some are useful."

George E. P. Box

George E. P. Box