A method of computing the effectiveness of an insecticide

TLDR: In order to make experimental studies comparable and statistically meaningful, the article recommends the following formula: per cent control = 100(X - Y)/X, which eliminates errors due to deaths in the control sample which were not due to the insecticide.

Abstract: There are several statistical methods used in biology (entomology) for computing the effectiveness of an insecticide, based on relating the number of dead insects in the treated plat to the number of live ones in the untreated plat. In order to make experimental studies comparable and statistically meaningful, the article recommends the following formula: per cent control = 100(X - Y)/X, where X = % living in the untreated check sample and Y = % living in the treated sample. Calculation using this method eliminates errors due to deaths in the control sample which were not due to the insecticide. An example based on treatments of San Jose scale includes computation of probable errors for X and Y, and the significance of the difference between the two counts. Common biometric convention holds that when the difference between the results of two experiments is greater than three times its probable error, the results are significant and due to the treatment applied.