Habit strength and distribution of learning

Habit strength is an abstract concept. How can it be an independent variable qualified as a satisfactory scientific construct? In Hull's systematic research of learning he discussed it and believed that habit strength was dependent upon a number of measurable antecedent conditions under which reinforcement occured, and that all of the different combinations of antecedent reinforcement variables would yield, other factors equal, in the action evocation exactly the same persistence of reaction, latency of reaction or others. He used the Simley's (1933) and William's (1938) experiments to verify his view point. According to Hull's concept the reaction latency and the resistance to experimental extinction may be regarded as indicators of habit strength. If it is true, then the habit strength represented by the reaction latency should correspond to that represented by the resistance to experimental extinction. But their imcompatibility has been suggested by Professor Cheng in his investigation of the 'function of punishment'.