Managing the Deliverance of Food to Create Motivation
By Barbara Heidenreich
First presented at The Animal Behavior Management Alliance conference 2005
Abstract: In the animal training community, food is often used as a primary reinforcer for desired behavior. While some animals respond immediately to food reinforcers, some animals or situations may present challenges that reduce the value of food as a reinforcer. A common misconception is that animals must be severely deprived to create motivation for food. This paper will explore the many methods that can be used to create motivation for food without misuse of deprivation. These methods include examining the food items presentedhow and when food is provided, the utilization of training concepts to increase motivation and more. Case studies will be provided that demonstrate the application of various strategies.
Reinforcers
Reinforcers are used to increase behavior. Friedman defines positive reinforcement as the presentation of a stimulus following a behavior that serves to maintain or increase the frequency of the behavior. Positive reinforcers tend to be valued or pleasant stimuli.To get positive reinforcers, learners often enthusiastically exceed the minimum effort necessary to gain them.(Friedman, 2005) While negative reinforcement also serves to increase behavior for the purposes of this paper positive reinforcement and reinforcers will be the focus.
Food is a primary reinforcer. Primary reinforcers are defined as any reinforcer which is necessary for the survival of the species. It is also defined as a reinforcing stimulus that has acquired its properties as a function of species history. (Cambridge Center for Behavior Studies) Secondary reinforcers are reinforcers that are dependent on their association with other reinforcers. They are also called conditioned reinforcers. (Cambridge Center for Behavior Studies)
Both primary and secondary reinforcers are valuable training tools. Deciding which tool(s) to use requires some careful consideration. Trainers can perform a reinforcer assessment to see what activities, foods, locations, tactile interactions, etc. appear to be reinforcers. If it’s a slim list trainers can pair the existing reinforcers with a wide range of other stimuli to create a longer list.
In general the longer the list of reinforcers available, the more prepared a trainer can be for a variety of circumstances. In addition an arsenal of reinforcers allows the consideration of other strategies to increase performance of behavior, such as variable reinforcement.
Food as a Valid Training Tool
While it is understood that many types of reinforcers are beneficial, the focus of this paper is the use of food as a reinforcer. Under certain circumstances food has the potential to be a powerful reinforcer. Without diminishing the value of secondary reinforcers, it is important to recognize that using food is a valid training tool. Unfortunately in some training circles, using food to train has been viewed as failure on the part of the trainer. It has also been suggested that a “better” trainer would not have to use food to train or that a better trainer would only use secondary reinforcers.
In addition using food to reinforce behavior has sometimes been described as a “bribe” or an inappropriate “payoff” for behavior. While bribing, luring, baiting are well defined by behavior analysis and can serve a function in initiating a behavior, the implication here is more equated with social connotations as opposed to behavior analysis principles. Knowing that food is a naturally occurring reinforcer for behavior it becomes difficult to perceive the use of food in a training situation as failure.
In the wild and in captivity animals frequently perform behavior to earn food reinforcers. This behavior may be as simple as walking from point a to point b to feed from a bowl, to presenting complex foraging behaviors that require an animal to express its adaptations for feeding, to participating in a structured training session. Researchers have coined the term contrafreeloading to describe the phenomenon that animals choose to perform a learned response to obtain reinforcers even when the same reinforcers are freely available. For example, given a choice between working for food and obtaining food for free, animals tend to choose to work, often quite hard, with a bowl of free food placed right next to them. This phenomenon has been replicated with rats, mice, chickens, pigeons, crows, cats, gerbils, Siamese fighting fish, and humans (Osborne, 1977); starlings (Inglis & Ferguson, 1986); Abyssinian ground hornbills and bare-faced currasows (Gilbert-Norton, 2003); and captive parrots (Colton, et al., 1997). There are several interesting hypotheses explaining why this phenomenon occurs. For example, contrafreeloading behavior may be motivated by innate foraging behaviors that are otherwise frustrated in captivity; animals may be engaging in information seeking behaviors as they work to predict the location of optimal food sources; or they may be responding to the additional reinforcement provided by stimulus changes when one works for food such as the sound of a hopper. (Friedman, 2005)
Food also has the potential to reinforce behavior when conditions may not be supportive of the use of secondary reinforcers. For example an animal that has no positive reinforcement history with a trainer may not find attention or tactile stimuli reinforcing. This animal may also not demonstrate a level of comfort that would allow it to engage in play activities in the presence of the trainer. In these situations in particular, food can be an excellent reinforcer for any acceptable behavior. This can allow the animal to associate positive reinforcement with the presence of the trainer. This can change the dynamic of the relationship between the animal and the trainer often much quicker than not using food. This same concept can be applied to other objects or circumstances that may elicit a fear response in animals, such as items associated with restraint or the presence of unfamiliar items.
Once food is accepted as a positive reinforcer it also allows the trainer to pair food with other reinforcers to increase the list of reinforcers available for that animal. Food can be used to reinforce animals to interact with enrichment items, accept tactile stimuli, and even accept other food items as reinforcers.
When food is considered a training tool it can also open the door to different training strategies. For example some trainers rely on jesses to maintain the behavior of sitting on the hand with bird species such as toucans, corvids, and kookaburras. While using jesses is a practice with a long history for raptor species, it has the potential to present health risks to bird species with delicate legs. Jesses employ the use of negative reinforcement to create the behavior of sitting on the hand. Allowing the use of food to train the behavior of sitting on the hand for a period of time gives a trainer the opportunity to focus on using positive reinforcement techniques and avoid the potential injury that can be caused to non raptor species wearing jesses. (Edmonds, 2002)
Another advantage to using food to reinforce behavior is that a situation can be created in which the animal can learn very fast. For example during a twenty minute training session a trainer may have the opportunity to reinforce steps towards the desired behavior numerous times. Each reinforcement becomes a learning opportunity for that animal. For example a parrot can learn an entire behavior (such as step up on the hand, wave, turn around, kiss) in a single training session when food is offered as a positive reinforcer and the bird is receptive to eating.
Creating Motivation for Food without Misuse of Deprivation
Primary reinforcers are dependent on establishing operations. For example an animal that is satiated may find a food reinforcer of less value than a secondary reinforcer. However an animal that is not satiated may demonstrate intense vigor to present behavior in the presence of a primary reinforcer. Creating vigor to present a behavior for a primary reinforcer can be accomplished by considering a variety of strategies that do not rely on deprivation or compromising animal health.
The following are a list of strategies that can encourage a stronger response to a food reinforcer.
Free feed a base diet and reserve preferred food items for training sessions: This strategy allows an animal access to food at all times. However motivation for food is due to the desire to gain the preferred food items during the training sessions. Example: Parrots at a sanctuary had a pelleted diet in their bowl at all times. The birds received peanuts, sunflower seeds, fruit and vegetables during training sessions. In a collection of 50 psittacines, every bird responded strongly to the food items offered during training. A blue and gold macaw was trained to voluntarily participate in nebulization using this feeding strategy. (Heidenreich, 2004)
Manage the deliverance of food: The amount and type of food the animal receives is not altered. Instead of presenting the entire diet at one feeding time, the diet is offered in portions throughout the day in training sessions. By the end of the day the bulk or all of the diet should have been offered to the animal. Example: A skunk at a wildlife facility was receiving a measured diet presented in a bowl at 5:00PM each night. The empty bowl was removed in the morning. Motivation for food was created by taking that same diet and offering portions of it during training sessions during the day. Any food not used during the training session was offered to the skunk at the end of the last training session. The skunk received the entire diet daily. However the skunk’s behavior changed dramatically. Prior to managing the deliverance of the diet, the skunk was usually sleeping whenever keepers approached. After managing the diet and embarking on a training program the skunk was often waiting at the door in apparent anticipation of the training session. In addition the skunk readily consumed portions of the diet offered during the training session. The skunk was trained to present a chained behavior of walking from a kennel to a bowl of dirt, digging in dirt, walking to a bag of garbage, digging in the garbage, and walking into a kennel to be presented for outreach programs.
Feed until satiated each training session: This strategy involves using the animals diet during a training session. The training session continues as long as the animal is willing to eat. When the animal shows signs of satiation the session is ended and no more food is offered. Several more sessions are conducted wherein again the animal eats until satiated. No more food is offered at the end of the day. The animal’s weight and response to food in each session can be monitored to help identify the appropriate number of training sessions per day. Example: A mink at wildlife rehabilitation facility was offered pieces of mice during training sessions. Although the mink was eager to participate, eventually the mink would begin to cache mouse bits in the training area. The mink would continue to present behavior for food that it would then cache. However the caching behavior was taken as indication of satiation. The sessions were ended at the sign of repeated caching behavior. Two sessions per day allowed the mink to eat until satiated and maintain a relatively stable body weight and response to food during training sessions. The mink was trained to target, enter a kennel, get on a scale, allow touching, allow a harnessing, and present a short chained behavior of hopping on a stump and running across a log.
Use small reinforcers: A common practice in the pet community is the use of large reinforcers. Unfortunately using food reinforcers that are very large can quickly diminish motivation for food. While a large reinforcer may serve a useful purpose to provide a magnitude reinforcement for a behavior, in general using the smallest reinforcer an animal is willing to accept can allow a trainer many more opportunities to reinforce desired behavior. Example: A companion parrot owner with a sun conure wanted to train her parrot to learn to step up onto her hand without biting. The birds preferred food reinforcer was goldfish crackers. Early in the training the owner was offering an entire goldfish cracker for approximations towards the desired behavior. After two repetitions and two crackers the bird was satiated and no longer had an interest in the training session. For the next session the owner broke the cracker into tiny pieces. The bird progressed through all the approximations to the desired goal behavior in one session and consumed one cracker in the process.
Train just prior to a regularly scheduled feeding time: In this strategy animals are feed a diet that is typically more than they can eat in a day. However the diet may not remain in the enclosure the 24 hours. In the moments just prior to receiving a fresh supply of food often animals are motivated to eat and can participate in a training session. This can then be followed by the regular feeding. Example: A mixed species aviary that feeds from communal bowls normally received its daily feedings at 8:00AM. At 7:30 AM training sessions for entering travel cages and standing on a scale were successfully conducted for food items normally received in the diet. The birds continued to have access to free amounts of food throughout day. At 5:00 PM uneaten food was discarded.
Avoid overfeeding by weighing the animal and diet to determine a range that works for that animal: Monitoring the weight of an animal is helpful to monitoring general health. However by also measuring the amount of food an animal needs to maintain a healthy weight, a diet can be prepared that can allow an animal to maintain motivation to eat during training sessions, yet at the conclusion of training will satiate the animal completely. Often the goal is to work towards maintaining as high a weight as possible and the greatest amount of food possible, while still maintaining response to food in training sessions. Example: Often such deliberate measuring of food can be seen in birds such as hawks, eagles, and other birds flown in bird presentations. Typically the hawk is weighed in the morning. A diet is prepared that is estimated to maintain the current weight of the bird. During the training session the bird’s attention for the food item and willingness to present the behavior under normal, calm circumstances is observed. Depending on the response of the animal during the training session the diet may be adjusted. A weight range that usually corresponds to an acceptable response to food during the training session is identified. The diet is prepared to maintain the bird in that weight range and adjusted daily depending on response during training sessions. Other conditions may influence weight and behavior, such as weather, age, food items, etc. These are also taken into consideration when evaluating weights and diets.
Factors That Can Lower Motivation to Present a Behavior
While food can be a useful tool for training behavior, an animal’s motivation to eat can be effected by many factors. Increasing motivation can also be achieved by eliminating or reducing factors that lower motivation. The following list describes typical factors that can reduce an animal’s motivation to perform a behavior for food or other reinforcements.
New environment
New trainers/people
Other environmental distractions
Physical exhaustion
Illness
Excess heat or cold, rain
Wind: If training out of doors. Particularly influential to flying birds
Objects blocking the identified path animal is to take
Inappropriate paths or patterns (too angled, too far, etc)
Confusion, unclear cues, or too many people requesting behavior at the same time
Poor training strategies, reinforcing inappropriately
Well Trained or Deprived?
An interesting observation is that a well trained animal and a deprived animal may show some similarity in the vigor with which each responds to a trainers cues for behavior. This can present certain challenges. An inexperienced observer may conclude a well trained animal is deprived. A trainer that relies on excessive deprivation may not understand how to create behavior without the misuse of deprivation. Therefore it can be helpful to know some indicators to help one differentiate between well trained and deprived. Here are some behaviors to look for that can indicate an animal is receiving enough food. A “yes” answer indicates an animal is most likely not experiencing excessive deprivation.
Does the animal sit, stand or rest comfortably, or express other natural behaviors that indicate comfort for the species when it is not anticipating a training session?
Does the animal interact with enrichment when offered?
Does the animal solicit a bath or other normal grooming activities when the opportunity arises?
Does the animal express vocal behaviors that would be associated with natural behavior other than food seeking behavior?
Does the animal exhibit good physical condition performing active behaviors, such as running, flying? Does it have strong strides or wing flaps, good maneuverability, and the physical ability to express natural behavior easily?
Is the animal able to sustain aerobic behavior for a reasonable amount of time for the species?
Does the animal ever have a point in the day in which it appears to be satiated?
It is not recommended to misuse deprivation to create food motivation. The strategies suggested in this paper can help increase an animals desire to consume food without misuse of deprivation. Some signals that may indicate an animal should receive more food include the following:
Absence of the behaviors listed above
The animal eats food very quickly.
The animals physical condition is observably compromised (ex birds keel bone is prominent).
The adult animal continually presents baby like behaviors.
The animal has no interest in playing, vocalizing, bathing or moving much.
The animal is always focused on a trainer for an opportunity to gain food.
The animal may tire easily, no energy. Physically compromised animals do not have the energy to perform some behaviors.
The animal may express behaviors related to food competition (mantling in birds of prey, carrying food, etc.) In addition, aggressive behaviors may be presented in relation to food competition.
Some extremely anxious behavior about food can also be learned. It is important to look at many different factors that could be influencing behavior. An animal that appears too motivated by food, whether it is deprived or not, may also present training challenges. Too motivated for food is challenging in that some animals cannot focus on the task at hand and obsess over food. This can be seen in some dogs for example. In these situations less preferred food items or working after a meal or considering other reinforcers can lead to a successful session. In addition keeping the presence of food well hidden until the moment it is time to reinforce the behavior can also be helpful. Another strategy to better evaluate the strong motivation to eat is to allow the animal free access to food at all times for a long period of time. This may have to be applied gradually to avoid health issues that can arise from an animal rapidly consuming large quantity of food. In some cases this process can involve months. This strategy can allow food to lose its intense reinforcing value. Training can then be initiated again. Trainers can then apply the food management strategies carefully to avoid creating an overly anxious animal.
Conclusion
In general good training success relies on the use of all the tools in a trainer’s repertoire to create desired behavior. This can involve the use of different schedules of reinforcement, magnitude reinforcements, clear contingencies, appropriate application of bridges, selectively reinforcing strong response, effective use of time outs, etc. These and more, are all critical factors and should also be well understood and applied when training. The reinforcer is no small component of the training process. Combining a strong reinforcer with good application of training strategy allows trainers to achieve desired behavioral goals. By considering the role of food as a reinforcer trainers are given a potentially powerful tool to influence behavior for the better.
References
Cambridge Center for Behavioral Studies: Applied Behavior Analysis and Behavior Study: Behavior Analysis Glossary Terms. www.behavior.org
Coulton, L.E., Warren, N.K., Young, R. J. (1997). Effects of foraging enrichment on the behavior of parrots. Animal Welfare 6, 357-363.
Freidman, S.G. (2005). He said, she said, science says. Good Bird. Vol 1 issue 1.
Edmonds, M. (2002).Alternatives to the use of falconry equipment on non-raptors. Proceedings. International Association of Avian Trainers and Educators conference.
Gilbert-Norton, L. (2003). Captive birds and freeloading: The choice to work. Research News, 4 (1).
Heidenreich, B.E. (2004). Training birds for medical and husbandry behaviors. Proceedings Animal Behavior Management Alliance conference.
Inglis I .R., Ferguson, N. J. K. (1986). Starlings search for food rather than eat freely available food. Animal Behaviour, 34, 614-616.
Osborne, S. R. (1977). The free food (contrafreeloading) phenomenon: A review and analysis. Animal learning & Behavior, 5 (8), 221-235.