How do you calculate 2 sigma?

Answer 1

That's 2 Standard deviations from the mean.

Answer 2

There is a technique, applied to Normal or Bell distributions, that you can use without knowing much about statistics. It's called the Three Point Estimate (TPE). The TPE is usually accurate enough for business and similar applications. But I wouldn't use it for scientific research.

Here's the essence of the TPE:

Average = Expected Value = (Upper + 4XMost Likely + Lower)/6; where, Upper is the highest number in your data, Lower is the lowest number, and Most Likely is the mode (the value that shows up most in your data).

Standard Deviation = (Upper - Lower)/6; so, for 2 sigma, just multiply this answer by two.

EXAMPLE: Suppose you measure the resistance of 10 resisters and get: 2, 1, 3, 2, 1, 3, 3, 5, 4, 3 ohms each; in which case:

Upper = 5, Lower = 1, Most Likely = 3; so that:

Average = (5 + 4X3 + 1)/6 = 3.0; and
Standard Deviation = (5 - 1)/6 = 2/3 (2 sigma = 4/3 in this case)

If you calculate the Average the long way (get the sum and divide by ten), you'll find Average = 2.7, which you must admit is pretty close to the 3.0 we got by the TPE method.

Much of the error comes from the likelihood that the ten data points do not represent a perfect Normal or Bell curve. So using the TPE method is not advised if you are not sure your data distributions approximate a Normal or Bell curve.

Answer 3

Sigma Level Calculation Formula