The idea of informing power comes from information theory (or entropy). H. Kaiser (1978, reprinted from a 1974 publication) provided an introduction to the use of information theory for evaluating methods in analytical chemistry. Fitzgerald and Winefordner (1975) also described the concept and provided applications to molecular absorption, conventional phosphorimetry, and time-resolved phosphorimetry. Fetterolf and Yost (1984) determined the informing power of various tandem mass spectrometry configurations, incuding GC/QMS, QMS/QMS, GC/QMS/QMS (see Table 2-1).

The informing power, P_inf, of a measuring device is the number of bits required to encode the information potentially available from the device. Supposing, for example, that the device can report one of S possible values, the informing power of the device is log₂(S). If the device has a parameter x that can be varied over k possible values x₁, … ,x_k, and the device is capable of reporting S(x_i) measurement values at x_i, then the informing power is summed over the k settings, producing

Because of the log term, greater gains can typically be achieved by increasing the number of values for x than by increasing S(x).

If the parameter x can be varied continuously, then P_inf can be reformulated by introducing the concept of resolution for the parameter x, defined as R(x) = x/δx, where δx is the smallest distinguishable difference in x for practical purposes. As introduced by Kaiser, the informing power becomes

There might be a number of simplifications to this expression. It might be that S(x) is constant, S(x) ≡ S, for example when S is fixed by characteristics of the detector. Or, the resolution might be constant, R(x) ≡ R. Another common possibility is that δx is constant. If S(x) ≡ S and R(x)≡ R, then