Numerous theories have been proposed to explain how organisms choose between different alternatives. Two areas of research in particular have serious implications for the validity of these choice theories: one, is the investigation of whether difference or ratio controls choice, and another is the investigation into the role that temporal context exerts on conditioned reinforcement value. Recent developments in both areas of research, as well as the implications for various choice theories, are reviewed. Findings are generally consistent with delay reduction theory: however, further research is clearly needed. According to delay reduction theory (DRT), the conditioned reinforcing strength of a stimulus may be predicted by calculating the reduction in the duration of time to primary reinforcement signaled by that stimulus (Fantino, 1969; 1981; Squires & Fantino, 1971). The simplest form of DRT (from Fantino, 1981) states that: Reinforcement strength of stimulus A = f (T-[t.sub.A]/T] (1) where [t.sub.A] is the time between the appearance of stimulus A and primary reinforcement, and where T is the overall time between reinforcer presentations. Thus, a stimulus that signals a greater improvement in time to reinforcement will be a more effective conditioned reinforcer than a stimulus that signals less of an improvement. DRT has been extended, with some success, to self-control (Ito & Asaki, 1982; Navarick & Fantino, 1976), elicited responding (Fantino, 1982), three-alternative choice (Fantino & Dunn, 1983), observing (Case & Fantino, 1981; Fantino & Case, 1983), operant analogues to foraging (Abarca & Fantino, 1982; Fantino & Arbaca, 1985), percent reinforcement (Spetch & Dunn, 1987), and the serial-position effect in short-term memory (Wixted, 1989). Perhaps the most common procedure for assessing choice is the concurrent-chains procedure, as depicted in Figure 1, where responding to concurrent initial-link schedules (the choice phase) is reinforced by the presentation of two mutually exclusive terminal-link schedules (the outcome phase), completion of which leads to primary reinforcement. Following primary reinforcement, the initial links are reinstated and a new trial begins. Preference, and presumably conditioned reinforcement value, of the terminal links is measured by response allocation during the initial links. [FIGURE 1 OMITTED] Figure 1. Diagram of the concurrent-chains procedure. Panels A and B are the sequence of events when responses on the left and right are reinforced, respectively. Circles represent the response keys with the schedule requirement indicated within. The initial links are normally equal VI schedules with the same color, while the terminal links are on different schedules, each having a unique color. Responses to the terminal links are reinforced with food according to some schedule. Equation 1 represents the essence, or kernel, of DRT. Applying DRT to the concurrent-chains procedure however, requires that the delay reduction signaled by each alternative be weighted by the rate of primary reinforcement for that alternative. Thus, the Squires and Fantino (1971) form of DRT is written such that: [B.sub.L]/[B.sub.L] + [B.sub.R] = [r.sub.L] (T - [t.sub.L])/[r.sub.L] (T - [t.sub.L] + [r.sub.R] (T - [t.sub.R]) for [t.sub.L] and [t.sub.R] = 1, for [t.sub.L] , = 0, for [t.sub.L] where [B.sub.L] and [B.sub.R] are the choice responses on the left and right initial links, respectively; [r.sub.L] and [r.sub.R] are the overall rates of primary reinforcement on the left and right keys, respectively; T is the average overall time to primary reinforcement measured from the onset of the choice phase; and [t.sub.L] and [t.sub.R] are the average delays during the terminal links on the left and right keys, respectively. …