Operator Variation in the Estimation of Percentage Heavy Minerals

Griffiths and Rosenfeld have demonstrated that all experimental procedures are characterised by two sources of error: (i) constant bias distinctive of each operator and (ii) inconsistency within one operator and among a group of operators. An experiment was organised by the writer to evaluate operator variation that may enter in the estimation of percentage heavy minerals by using a large sample of Ariyalur sandstone (Trichinopoly Cretaceous) as an example. This experiment revealed that heavy mineral work also involves, as many other experimental endeavours, sources of error such as (a) experimental (b) operator variation (c) constant bias within individual operators and (d) inconsistency of operators from sample to sample. It is pointed out that if all these sources of error are not properly evaluated, 'isolated and controlled by suitable analysis of variance techniques, the results of heavy mineral work will be far from any real value.