Statistical Considerations on the Results of Testing Methods of Film-Type Rust-Preventive Oils

Abstract

SummaryThis report deals with statistical studies on the results of several tests for evaluating the preventive ability of film-type rust-preventive oils, such as humidity cabinet, salt-spray, artificial weathering, settle gelatine drop, contact angle test, etc. The relative merits of these tests were observed on their sensitivities derived from the minor data obtained by factorial experiments by the inductive statistical method. The sensitivity is defined as variance ratio in analysis of variance of factorial experiments. As the results of factorial experiments carried out by the same L8(27)—orthogonal array to which three factors of two levels and one factor of four levels were laid out it was found that the difference in information on the effects of factors is caused by the difference in sensitivity of each test, and it is essential to inquire into the sensitivities of these tests or to improve their characteristics for the purpose of statistical treatment.