To provide an analysis method for statistical errors of measuring data with less restrictions, which provides resolution better than conventional methods, and measures wide ranges with a cheap device.

The method comprises a statistics analysis step in which a center position μ, a spreading width σ, and statistical errors σ(μ) at the center position μ are statistically calculated from a plurality of measuring data Xi (1=1 to N, N is an integer not less than 2) , and an output step in which the calculated center position μ, spreading width σ, and statistical errors σ(μ) at the center position μ are output.