Examine scatter in a repeated measurement
Use this tool when you have measured the same dimension from the same datum several times. Do not combine different door locations or different intended dimensions and interpret their variation as repeatability. It summarizes the entered readings; it does not infer the correct dimension.
Three distinct summaries
Mean = Σx / n
Observed range = maximum − minimum
Sample standard deviation s = √[Σ(x − mean)² / (n − 1)]
For hypothetical readings 10, 11, 12 in, mean is 11 in, range 2 in, and sample standard deviation 1 in. The squared deviations sum to 2 and the divisor is n−1 = 2. Two identical readings have zero observed scatter, which does not establish zero measurement error.
The standard-deviation definition is described in the NIST/SEMATECH variability handbook. Between 2 and 50 values are required; one reading cannot define this sample standard deviation. Blank or malformed readings are rejected. The mean/range retain exact input arithmetic; the square root is numerical. Displays round to three decimals.
Systematic bias, resolution, calibration, measurement force, temperature, and datum choice can affect accuracy without appearing in repeat scatter. No uncertainty budget, tolerance pass/fail, or compatibility conclusion is calculated. Use dimensional comparison for a measured value against an explicit target or range.