Functional analysis of behavior depends upon accurate measurement of b
oth independent and dependent variables. Quantifiable and controllable
operations that demonstrate these functional relationships are necess
ary for a science of human behavior. Failure to implement independent
variables with integrity threatens the internal and external validity
of experiments. A review of all applied behavior analysis studies with
children as subjects that have been published in the journal of Appli
ed Behavior Analysis between 1980 and 1990 found that approximately 16
% of these studies measured the accuracy of independent variable imple
mentation. Two thirds of these studies did not operationally define th
e components of the independent variable. Specific recommendations for
improving the accuracy of independent variable implementation and for
defining independent variables are