F -to-remove

n. in model-building procedures such as backward elimination or stepwise regression, the specific ratio of variances needed to justify keeping a variable as a predictor. Generally, an F-to-remove value around 4.0 is sufficient to allow a variable to be retained in an analysis.

Browse dictionary by letter

a b c d e f g h i j k l m n o p q r s t u v w x y z

Psychology term of the day

May 9th 2024

nonadditive

adj. describing values or measurements that cannot be meaningfully summarized through addition because the resulting total does not correctly reflect the underlying properties of and associations between the component values. For example, if two variables a and b interact to influence another variable y, the addition of the separate effects of a and b will not equal the total effect since the contribution of the interaction needs to be included. Compare additive.