The living legacy of the Harvard Pigeon Lab: quantitative analysis in the wide world.

From the Harvard Pigeon Lab of the 1960s arose a behavior-analytic approach that was quantitative and rigorous, rooted in Herrnstein's matching law. Researchers modified the matching law to describe choice behavior in a variety of different settings and examined its relations with other quantitative models. Beginning in the early 1970s, researchers began using the Harvard Pigeon Lab's quantitative framework to study in the laboratory specific aspects of the world outside the laboratory. Much of this work concerned investigations of self-control-choice of a larger, more delayed reinforcer over a smaller, less delayed reinforcer. Experiments using a quantitative framework derived from the matching law have also been conducted outside the laboratory; however, these have been far less frequent. Current and future researchers will benefit the field by devising new, creative ways to investigate the matching law and related quantitative models outside the laboratory. Such research can help to demonstrate the validity of these models as basic principles of behavior, can enhance public opinion of and rewards for such research, and can stimulate further development of the Harvard Pigeon Lab's quantitative approach by using that approach with new variables.

Higher-education administrators: when the future does not make a difference.

The relationships between administrative actions and their long-term consequences were investigated in 14 people beginning training to become high-level college and university administrators, 19 people who had just finished that training, and 44 provosts (chief academic officers) at colleges and universities in the northeastern United States. The experienced administrators (those who had completed the training plus the provosts) were significantly more likely than the trainees to mention long-term consequences when describing their past and possible future administrative actions. However, in hypothetical choice situations, the experienced administrators were also significantly more likely than the trainees to choose smaller amounts offunds available immediately for their units versus larger amounts of promised future funds. With experience, administrators may both become more aware of their actions' long-term consequences and learn that they are unlikely to receive promised future funds. The contingencies in effect for higher-education administrators may lead them to make choices that do not result in their institutions meeting the highest standards.

Effects of deprivation level on humans' self-control for food reinforcers.

Deprivation level was manipulated in fourteen food- and water-deprived adult human females to examine its effects on self-control for food (choice of larger, more delayed access to apple juice over smaller, less delayed access to apple juice). Each subject was exposed to two treatments: (1) Consumption of a 500 g tomato soup preload just prior to self-control testing and (2) no soup preload. When subjects had consumed soup, they reported significantly less hunger and showed significantly more self-control as compared to when not having consumed soup. Additionally, when subjects had consumed soup, self-control decreased as a function of session time. Subjects who reported that they were currently dieting drank significantly less juice when they had previously consumed soup than when they do not previously consumed soup. Together, the results indicate that when subjects are more deprived they may be less able to wait for food reinforcers (i.e., show less self-control). Such behaviour may be adaptive in situations in which energy is needed to survive periods of food scarcity.

Effects of reinforcer type on rats' sensitivity to variation in reinforcer amount and reinforcer delay.

Twenty rats, divided into two groups, served as subjects in an experiment that examined choices between reinforcers that varied in amount or delay. One group received food reinforcers, and the other water reinforcers. No difference was found between the two groups in their values of sA (tendency of choice behavior to vary in accordance with variation in reinforcer amount). However, Group Food had significantly lower values of sD (tendency of choice behavior to vary in accordance with variation in reinforcer delay) than did Group Water, and Group Food's values of s As D (a measure related to self-control-choices of larger, more delayed over smaller, less delayed reinforcers) tended to be greater than for Group Water. Finally, over the session, sA and s As D showed no significant change, overall response rates decreased, and sD increased. Taken together, these results suggest that, independent of deprivation level, self-control for water is less than self-control for food.

Self-control and impulsiveness in children and adults: Effects of food preferences.

Experiment 1 used 6 preschool boys and Experiment 2 used 6 adult women to explore the effects of food preference on humans' choice in self-control paradigms. The boys showed a higher proportion of responses for more delayed, larger reinforcers (a measure of self-control) when those choices resulted in receipt of the most preferred food compared to when those choices resulted in the least preferred food. Further, the boys chose the less delayed, smaller reinforcers significantly more often when only those choices, as opposed to both choices, resulted in the most preferred food. Conversely, they chose the more delayed, larger reinforcers significantly more often when only those choices, as opposed to both choices, resulted in the most preferred food. Finally, the women demonstrated significantly less sensitivity to reinforcer amount relative to sensitivity to reinforcer delay (another measure of self-control) when they had a higher preference for the juice received as the less delayed, smaller reinforcer than for the juice received as the more delayed, larger reinforcer. Together, the results show that subjects' food preferences can influence self-control for food reinforcers.

The effect of variable delays on self-control.

Five pigeons served as subjects in an experiment that examined the effects of variable as opposed to fixed delays on preference in a self-control paradigm (choice between larger, more delayed and smaller, less delayed reinforcers). Nonindependent concurrent variable-interval schedules were used to measure choice. When delays to the larger, more delayed reinforcers were variable as opposed to fixed, the subjects showed an increased preference for that alternative (the self-control alternative). A series of regressions revealed that the hyperbolic decay model and incentive theory provided poor fits to the data, but a modified version of the generalized matching law provided an adequate fit. Together, consistent with a general prediction made by discounting models, the data supported the conclusion that variable delays can increase self-control. However, specific discounting models were not able to explain the present data well.

Self-control across species (Columba livia, Homo sapiens, and Rattus norvegicus).

Data from six previous studies of self-control behavior were compared against predictions made by the matching law and by molar maximization. The studies involved pigeons (Columba livia), rats (Rattus norvegicus), and 3-year-old, 5-year-old, and adult humans (Homo sapiens) who had received food as the reinforcer, and adult humans who had received points exchangeable for money as the reinforcer. Neither theory proved to be an accurate or better predictor for all groups. In contrast to the predictions of these theories, self-control was shown to vary according to species, human age group, and reinforcer quality. When the reinforcer was food, the self-control of different species was found to be negatively correlated with metabolic rate; that is, larger species showed greater self-control. These results suggest that allometric scaling may prove useful in describing and predicting species differences in self-control.

Predictors of adult humans' self-control and impulsiveness for food reinforcers.

Humans vary in the degree to which they demonstrate self-control--choice of larger, more delayed over smaller, less delayed reinforcers. When reinforcers consist of food, individual human subjects' choice behavior varies from virtually exclusive self-control to exclusive impulsiveness. The present experiment, using 26 men and 26 women subjects, explored some possible sources of this individual variation through assessing the correlation of behaviors exhibited in the self-control paradigm with various subject characteristics. The results suggest that self-control is negatively related to individuals' reported susceptibility to hunger (Factor 3 of the Three-Factor Eating Questionnaire) and positively related to overall rating of juice desirability, but not to gender or to various measures of caloric deprivation or personality. Relatively high self-perceived levels of food deprivation may decrease self-control, whereas assigning a relatively high value to the reinforcer may increase self-control. Both tendencies would have been adaptive for evolving humans.

Cocaine decreases self-control in rats: a preliminary report.

Cocaine abuse is often associated with behavior that takes into account short-term, but not long-term consequences. However, there has been no empirical research concerning the effects of cocaine on self-control (choice of a larger, more delayed reinforcer over a smaller, less delayed reinforcer). In the present research, when food-deprived rats repeatedly chose between a larger, more delayed food reinforcer and a smaller, less delayed food reinforcer, chronic intraperitoneal injections of 15 mg/kg cocaine (but not 10 mg/kg fluoxetine) decreased the rats' choices of the larger, more delayed reinforcer. Cocaine can decrease rats' self-control.

Choice in a self-control paradigm: effects of uncertainty.

Four experiments examined humans' self-control and impulsiveness as a function of reinforcer uncertainty. Uncertainty was manipulated by varying the probability and durations of interruptions, temporary periods of timeout from reinforcement. In Experiment 1 the probability of an interruption varied while its duration was constant. In Experiment 2 the duration of an interruption varied while its probability was constant. In Experiment 3, the subjects were interrupted only during the delay to the larger, more delayed reinforcer. In Experiment 4 the subjects experienced the same amounts and delays of reinforcement as in Experiments 1-3, but did not experience any interruptions. The subjects in Experiments 1 and 2 were not significantly more impulsive than the subjects in Experiment 4. In Experiment 3, the subjects exhibited significantly more impulsiveness than the subjects in Experiments 1, 2 and 4 when the reinforcement rates were equal, but not when they were unequal. Reinforcer uncertainty affects choice behavior to the extent that it influences the relative reinforcement rate.

Probability and delay of reinforcement: an examination of Mazur's equivalence rule.

It has recently been proposed that the probability of an event (e.g. delivery of a reinforcer) can be interpreted as functionally equivalent to the delay of an event. The present experiment examined the equivalence rule that states that reinforcer probability is functionally equivalent to a set of variable delays. A titration procedure was used in which pigeons chose between a standard, probabilistic alternative, and an adjusting, delayed alternative. Reinforcer probability was manipulated by varying the number of keypecks required to obtain a reinforcer. The duration of the adjusting alternative varied as a function of a pigeon's previous choices. There were six standard probabilities. The results indicated that the pigeons were sensitive to changes in the standard probabilities and that there were significant relations between: (1) reinforcer probability and the mean adjusted delay of reinforcement, and (2) the obtained and predicted adjusted delays. The present results support Mazur's conception of the relation between reinforcer probability and delay, and extend his conception to a different method of manipulating reinforcer probability.

Self-control and impulsiveness in adult humans when food is the reinforcer.

Moderately food- and water-deprived women repeatedly chose between longer access to juice after a long delay (the self-control alternative), and shorter access to juice after a short delay (the impulsive alternative). In contrast to previous experiments using points exchangeable for money as the reinforcer, which have found consistent self-control across subjects, behavior of the present subjects varied between complete self-control and complete impulsiveness. Individual subjects' degree of self-control was not related to their majors, to their past or present degree of reported actual food deprivation, to their ratings of desire for the juice, or to their latency to respond. However, subjects who reported that they were currently dieting were significantly more impulsive; they obtained juice earlier within individual trials but had smaller total juice access time. The operant conditioning laboratory paradigm developed for the present research can be useful in investigating general determinants of self-control and impulsiveness for food reinforcers, as well as in investigating the origins of individual differences in food choice.

Humans' sensitivity to variation in reinforcer amount: Effects of the method of reinforcer delivery.

Two experiments examined human subjects' sensitivity to variation in reinforcer amount under different methods of reinforcer delivery. Subjects chose between schedules varying in terms of amount and/or delay of reinforcement, the reinforcer being points exchangeable for money. In Experiment 1, reinforcer amount was manipulated by varying the monetary value of the points across conditions while the number of seconds of access to a consummatory response remained constant. Choice was strongly sensitive to reinforcer amount and indicative of self-control, as in previous experiments. In Experiment 2, reinforcer amount was manipulated by automatically delivering different numbers of points during the amount period, and the consummatory response was eliminated. Sensitivity to variation in reinforcer amount was significantly lower than in Experiment 1. Furthermore, the subjects in Experiment 2 exhibited significantly less self-control than did the subjects in Experiment 1. Humans' sensitivity to variation in reinforcer amount appears to be affected by factors that enhance the discrimi-nability of the consequences of responding.

Choice in a self-control paradigm: effects of reinforcer quality.

Pigeons were exposed to a self-control paradigm in which they chose between larger, more delayed and smaller, less delayed reinforcers. Reinforcer quality (i.e., the type of grain associated with each alternative) was varied across conditions. Choice behavior was influenced by grain quality; proportions of responses for the larger, more delayed alternative changed as a function of the grains associated with the response alternatives. Furthermore, the percentage deviation from mean baseline response proportions generally decreased as a function of the relative grain types associated with each response alternative. Manipulation of reinforcer quality can significantly influence the degree of self-control typically exhibited by pigeons.

Food preferences in families.

Preferences for a wide variety of foods were examined in families: 77 students, plus their siblings, mothers and fathers, for a total of 241 subjects. The food preferences of family members were more similar than would be expected by chance. However, this occurred entirely in the comparisons between spouses and between female family members. The results also indicated sex and age differences in family members' food preferences. For example, females tended to prefer low-calorie foods more when they were older, while males tended to prefer alcoholic beverages more when they were older. Both males and females showed a greater preference for coffee when they were older. Some of the variance in food preferences can be explained by an individual's family members' food preferences, as well as by the individual's sex and age.

Adjusting delay to reinforcement: comparing choice in pigeons and humans.

Indifference functions of amount and delay of reinforcement were compared for pigeon and human subjects by using Mazur's (1987) adjusting delay procedure. A model similar to the matching law (the simple reciprocal model), a hyperbolic model, three modified versions of the hyperbolic model, and a negative exponential model were evaluated. In Experiment 1 the subjects were pigeons, and in Experiments 2 and 3 the subjects were humans. In order to make the nonhuman and human situations more comparable, in Experiments 2 and 3 the reinforcer (points exchangeable for money) was discounted at a constant rate during the delay periods. The rate of this discounting varied between Experiments 2 and 3. The results of all three experiments demonstrated that a power function transformation of the hyperbolic model (in which 1 is added to the delays in the denominator of the simple reciprocal model) provided the best description of both nonhuman and human data.

Self-control in adult humans: variation in positive reinforcer amount and delay.

In five experiments, choice responding of female human adults was examined, as a function of variations in reinforcer amount and reinforcer delay. Experiment 1 used a discrete-trials procedure, and Experiments 2, 3, 4, and 5 used a concurrent variable-interval variable-interval schedule. Reinforcer amount and reinforcer delay were varied both separately and together. In contrast to results previously reported with pigeons, the subjects in the present experiments usually chose the larger reinforcers even when those reinforcers were delayed. Together, the results from all the experiments suggest that the subjects followed a maximization strategy in choosing reinforcers. Such behavior makes it easy to observe self-control and difficult to observe impulsiveness in traditional laboratory experiments that use adult human subjects.

Predictors of food preferences in adult humans.

Predictors of preferences for a wide variety of foods were examined in 303 male and female human subjects ranging from 14-68 years of age. The subjects completed questionnaires which requested information on the subject's sex, age, thinness, sensation seeking and ethnic background, as well as on the subjects' food preferences. Largely consistent with previous studies, female subjects reported higher preferences for low-calorie foods, candy and wine, and lower preferences for meat, beer, spicy foods and milk. Younger subjects reported higher preferences for sweet foods and lower preferences for foods such as chili pepper that are considered acquired tastes. Thinner subjects tended to rate both sweet foods and meat lower than did other subjects. Preferences for spicy foods or foods likely to cause illness were positively correlated with sensation seeking while preferences for sweet or bland foods or foods unlikely to cause illness were negatively correlated with sensation seeking. Subjects for whom the primary cuisine on which they were raised was Oriental cuisine preferred alcoholic beverages and non-Oriental foods less than did other subjects. A factor analysis of the food preferences yielded ten factors including those for meat and potatoes, alcohol, spices and junk food. Data on predictors of food preferences can assist research on the determinants of food preferences, however much of the variance in food preferences remains to be explained.