Diverse psychological correlates have been ascribed to “P300,” the conspicuous P3b component of event‐related potentials (ERPs) recorded in many laboratory tasks. Traditionally, hypotheses on P3b have conceived of this component being independent from implementing the response to the present stimulus. This has changed in the recent decade when P3b has been related to aspects of the decision process. The present review first focusses on effects of the classic variables stimulus frequency and relevance on P3b amplitude. It turns out that already these effects are related to response requirements because effects of stimulus frequency actually are effects of frequency of response‐defined stimulus categories and effects of relevance may be defined as effects of graduating the response requirements. Then, constructs and hypotheses on psychological functions reflected by P3b are evaluated for their abilities in explaining those effects. The tested constructs are information, relevance, and capacity, and the hypotheses are priming, cognitive processing, memory storage, context updating, closure, response facilitation, decision, stimulus–response (S–R) link reactivation, and conscious representations. S–R link reactivation hypothesis performed best, closely followed by memory storage and closure hypotheses. To make further progress, more studies should conduct tests between competing hypotheses.

We had previously reported that the oddball effect on the P3b EEG potential evoked by infrequent vs. frequent S1 presented in a sequence of two stimuli, S1 and S2, gets reduced in a "combination task". In this task, responses were determined by the combinations of S1 and S2 rather than by S1 only. We had attributed this reduction of the oddball effect to increased task difficulty. The present study investigated possible reasons for this reduction of S1-evoked P3b in more detail, by making the combination task easier in several respects: allowing for forming associations from S1 to responses (Experiment 1), reducing the complexity of stimulus-response (S-R) mappings (Experiment 2), and decreasing S2 relevance in defining responses (Experiment 3). The results showed that only S2 relevance affected the oddball effect on S1-evoked P3b. Namely, when S2 attained some relevance by inducing a go/no-go decision for S1-defined responses, the oddball effect on S1-evoked P3b was intermediate between the large effect in the simple oddball task and the small effect in the combination task. The results may be explained in terms of the S-R link hypothesis of P3b which interprets P3b as reflecting reactivation of well-established S-R links.