Statistical designs are developed to meet the scheme's objectives from the data type (quantitative, or qualitative whether ordinal or nominal), the statistical assumptions, the types of error and the expected number of results; where a customer or regulator mandates the design it can be taken from that specification, and a preliminary interlaboratory comparison can supply missing information. The provider documents the design and the analysis methods by which the assigned value is derived and results are evaluated, with the reasons for choosing them and the assumptions behind them, and can demonstrate that the assumptions are reasonable and the analyses follow prescribed procedures. The design gives careful consideration to the accuracy and uncertainty required of the assigned value for each property; the minimum number of participants needed, with documented alternative approaches provided to participants where numbers are insufficient; the relevance of significant figures and decimal places; the number of items and repeat measurements; the procedures behind the proficiency assessment standard deviation or any other criterion; the treatment of results from methods that are not technically equivalent; whether participant measurement uncertainty is reported and how it is used; the identification and handling of outliers; the evaluation of excluded values where relevant; and the design objectives and round frequency where appropriate.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.