Choice in the repeated-gambles experiment

Humans chose 10 times between two roulette wheels projected on a monitor. During the first trial, the left wheel provided a hypothetical $100 with p = .94, and the right wheel provided $250 with p = .39. A titration procedure adjusted the probability of a $250 win across trials to permit estimation of an indifference point between alternatives. In Experiment 1, intertrial-interval duration (25 vs. 90 s) and whether sessions began with an intertrial interval or a trial were varied in a 2 × 2 design in this risky-choice procedure. Risk aversion (preference for the $100 wheel) increased with intertrial interval but was unaffected by whether sessions began with a trial or an intertrial interval. In Experiment 2, all sessions began with a trial, and subjects were informed that the experiment ended after 10 trials. Intertrial-interval duration had no effect on choice. In Experiment 3, intertrial-interval duration and whether subjects were given $10 or $10,000 before beginning were varied among four groups in a 2 × 2 design. In all other ways, the procedure was unchanged from Experiment 2. Intertrial interval had no effect on choice, but the $10,000 groups showed less risk aversion than the $10 groups. These results can be explained more readily in terms of Kahneman and Tversky's (1984) notion of “framing of the prospect” than in terms of Rachlin, Logue, Gibbon, and Frankel's (1986) behavioral account of risky choice.     

Pigeons' choices in situations of diminishing returns: fixed- versus progressive-ratio schedules.

In two different discrete-trial procedures, pigeons were faced with choices between fixed-ratio and progressive-ratio schedules. The latter schedules entail diminishing returns, a feature analogous to foraging situations in the wild. In the first condition (no reset), subjects chose between a progressive-ratio schedule that increased in increments of 20 throughout a session and a fixed-ratio schedule that was constant across blocks of sessions. The size of the fixed ratio was varied parametrically through an ascending and then a descending series. In the reset condition, the same fixed-ratio values were used, but each selection (and completion) of the fixed ratio reset the progressive-ratio schedule back to its minimal value. In the no-reset procedure, the pigeons tended to cease selecting the progressive ratio when it equaled or slightly exceeded the fixed-ratio value, whereas in reset, they chose the fixed ratio well in advance of that equality point. These results indicate sensitivity to molar as well as to molecular reinforcement rates, and those molar relationships are similar to predictions based on the marginal value theorem of optimal foraging theory (e.g., Charnov, 1976). However, although previous results with monkeys (Hineline & Sodetz, 1987) appeared to minimize responses per reinforcement, the present results corresponded more closely to predictions based on sums-of-reciprocals of distance from point of choice to each of the next four reinforcers. Results obtained by Hodos and Trumbule (1967) with chimpanzees in a similar procedure were intermediate between these two relationships. Variability of choices, as well as median choice points, differed between the reset and no-reset conditions.(ABSTRACT TRUNCATED AT 250 WORDS)     

Choice in transition: A comparison of melioration and the kinetic model

Transition-state choice behavior of pigeons was examined in two experiments designed to test predictions of melioration and the kinetic model. Both experiments began with an initial training condition during which subjects were maintained on concurrent variable-interval schedules. In Experiment 1, subjects were then exposed to concurrent variable-ratio schedules, whereas in Experiment 2, subjects were then exposed to concurrent extinction. Contrary to the predictions of melioration, but consistent with the kinetic model, acquisition of preference on concurrent variable-ratio schedules followed a negatively accelerated logistic trajectory, and preference remained stable in concurrent extinction. Predictions made by the kinetic model concerning rates of switching between alternatives were also supported.     

Analysis of a guided-response procedure in visual discriminations by rats.

A guided-response procedure was used to train a visual pattern discrimination by rats in a modified Sutherland box. The method consisted of guiding the animal to the correct choice by means of a retractable bridge that led to reinforcers, followed by gradually removing this prompt. This method was compared to a stimulus-fading procedure, in which the initial differences between discriminative stimuli were gradually faded until they differed only with respect to the critical dimension for discrimination, and to a trial-and-error procedure. Both gradual procedures resulted in fewer errors compared to the trial-and-error procedure. The higher efficiency of the fading procedures was attributed to less aversiveness derived from performance with few errors and to the use of step-by-step requirements relative to the criterion performance.     

A quantitative analysis of extreme choice

Six homing pigeons were trained on a variety of concurrent variable-interval schedules in a switching-key procedure. Unlike previous work, reinforcer ratios of up to 160 to 1 and concurrent extinction variable-interval schedules were arranged in order to investigate choice when reinforcer-frequency outcomes were extremely different. The data obtained over 11 conditions were initially analyzed according to the generalized matching law, which fitted the data well. The generalized matching law was then fitted only to conditions in which the reinforcer ratios were between 1 to 10 and 10 to 1. The deviations of choice measures from the other four more extreme reinforcer-ratio conditions were significantly more towards equal choice than predicted by this second generalized matching fit. A contingency-discriminability model, which predicts such deviations, described the data more effectively than did the generalized matching law, and also correctly predicted the maintenance of responding on both alternatives when one was associated with extinction.     

Independence of reinforcement delay and magnitude in concurrent chains

A three-component concurrent-chains procedure was used to investigate preference between terminal-link schedules that differed in delay and magnitude of reinforcement. Response and time allocation data were well described by a generalized matching model. Sensitivity to delay appeared to be lower when reinforcement magnitudes were unequal than when they were equal, but when obtained rather than programmed time spent responding in the initial links was used in the model, the difference vanished. The results support independence of delay and magnitude as separate dimensions of reinforcement value, as required by the matching law, and the assumption of the contextual choice model (Grace, 1994) that sensitivities to delay and magnitude are affected similarly by temporal context. Although there was statistical evidence for interaction between successive components, the effects were small and transient. The multiple-component concurrent-chains procedure should prove useful in future research on multidimensional preference, although it may be necessary to control obtained initial-link time more precisely.     

The applied importance of research on the matching law

In this essay, we evaluate the applied implications of two articles related to the matching law and published in the Journal of the Experimental Analysis of Behavior, May 1994. Building on Mace's (1994) criteria for increasing the applied relevance of basic research, we evaluate the applied implications of basic research studies. Research by Elsmore and McBride (1994) and Savastano and Fantino (1994) involve an extension of the behavioral model of choice. Elsmore and McBride used rats as subjects, but arranged a multioperant environment that resembles some of the complex contingencies of human behavior. Savastino and Fantino used human subjects and extended the matching law to ratio and interval contingencies. These experiments contribute to a growing body of knowledge on the matching law and its relevance for human behavior.     

The dimensional effects of missing information on choice processing

This paper demonstrates that choice processing may be different in missing information situations than in full information situations depending on whether inferences are used to fill in missing values and the overlap of the missing information itself. It is shown that when individuals do not form inferences to fill in missing values, fewer full attribute-based processes and more processes which accommodate for missing attribute values, alternative-based or given-dimension attribute-based, are used. It is also shown that when a processing shift due to missing information does occur, the overlap of the missing values will affect the type of shift that takes place. If overlap is high, a shift to given-dimension attribute-based processing is more likely, and when overlap is low, a shift to alternative-based processing is more likely. When individuals do form inferences to fill in missing values, processing is more similar to that in full information situations. Finally, it is shown that individuals will often partially fill in missing information, thus moderating the proposed effects.     

A behavioral economic analysis of concurrently available money and cigarettes.

In economic terms, consumption of a reinforcer is determined by its price and the availability and price of other reinforcers. This study examined the effects of response-requirement (i.e., price) manipulations on the self-administration of two concurrently available reinforcers. Six cigarette smokers participated in 4-hr sessions in which money and puffs on a cigarette were concurrently available according to fixed-ratio schedules of reinforcement. Once stable responding was obtained with both reinforcers available at Fixed Ratio 100, the response requirement for one reinforcer was systematically varied (Fixed Ratio 1,000 and 2,500), while the other reinforcer remained scheduled at Fixed Ratio 100. Increasing the fixed-ratio size for a reinforcer decreased its consumption, with a greater decrease occurring for monetary reinforcement. This finding was quantified in economic terms as own-price elasticity, with elasticity coefficients greater for money than cigarettes. The effects of fixed-ratio size on response output also differed across the two reinforcers. Although greater responding occurred for money at Fixed Ratio 100, increases in fixed-ratio size (for money) decreased responding for money, whereas the same increase in fixed-ratio size (for puffs) increased responding for puffs. Finally, increasing the fixed-ratio size for one reinforcer had little effect on consumption of the other concurrently available reinforcer. This finding was quantified as cross-price elasticity, with elasticity coefficients near 0.0 for most subjects, indicating little or no reinforcer interaction. The results indicate that the reinforcing effects of cigarettes and money in the setting studied here differed, and that the effects produced by changing the price of one reinforcer did not interact with the consumption of the other concurrently available reinforcer.     

Delay reduction and optimal foraging: variable-ratio search in a foraging analogue.

The present study investigated conditions under which the conditioned reinforcement principles of delay-reduction theory and views based on simple maximization of reinforcement rate make ordinally opposing predictions with respect to foraging-related choice behavior. The use of variable-ratio schedules in the choice phase also represents an extension of delay-reduction theory to schedules that may better mimic the effort involved in searching. Pigeons responded on modified concurrent-chains schedules in which equal variable-ratio schedules led to unequal variable-interval outcomes and unequal reinforcer amounts. All 4 subjects completed a minimum of two replications of conditions for which the predictions of delay-reduction theory and a simple rate-maximizing theory were opposed. Results were consistent with delay reduction's ordinal predictions in 11 of 11 replications of the divergent predictions favoring the smaller, more immediate alternative. The predictions of rate maximization were upheld only when they were consistent with those of delay reduction. Results are discussed in terms of conditioned reinforcement, sensitivity to reductions in delay to food, and possible rules of thumb that may be useful in characterizing foraging.