Procedure for statistical analysis of one-parameter discrepant experimental data

Area of Science:

• Nuclear Physics
• Experimental Data Analysis
• Statistical Methods

Background:

• Experimental data often exhibit discrepancies due to various sources of error.
• Accurate estimation of experimental uncertainty is crucial for reliable scientific conclusions.
• Existing methods may not fully capture the impact of unrecognized experimental errors.

Purpose of the Study:

• To introduce a novel Mandel-Paule-type procedure for processing one-parameter discrepant experimental data.
• To develop a method for estimating and incorporating unrecognized experimental errors into uncertainty analysis.
• To define discrepant experimental data for an arbitrary number of measurements.

Main Methods:

• Developed a new statistical procedure based on the Mandel-Paule methodology.
• The procedure estimates the contribution of unrecognized experimental errors to total uncertainty.
• Applied the method to analyze experimental half-life data for 20 actinides.

Main Results:

• Calculated mean half-lives for 20 actinides, showing consistency with existing evaluations (ENSDF, DDEP).
• Demonstrated that the new procedure yields significantly larger uncertainties for discrepant data compared to ENSDF and DDEP.
• Attributed the increased uncertainties to the explicit inclusion of unrecognized experimental errors.

Conclusions:

• The new Mandel-Paule-type procedure effectively accounts for unrecognized experimental errors in discrepant data.
• This method provides more realistic uncertainty estimations, especially when data are inconsistent.
• The findings highlight the importance of considering all error sources for robust experimental analysis in nuclear science.