Psychologist and behaviorist, B. F. Skinner and his colleagues used rats and pigeons in most of their groundbreaking behavior-analytic research (Skinner, 1953). Skinner is also identified as the father of Operant Conditioning based on his works with Skinner boxes and the Law of Effect. This suggests that if a behavior is reinforced, it tends to be repeated. If a behavior is not reinforced, it tends to become extinguished or weakened. Skinner (1948) used positive reinforcement with hungry rats to strengthen the behavior of hitting a lever. After a while, the rats learned to go straight to the lever instead of exploring the box. Skinner (1948) used negative reinforcement by introducing an electrical current running throughout the box. Once the rat figured out the lever turned off the shock, it immediately went to it in later trials. The behavior of pushing the lever to turn off the electrical shock was maintained using negative reinforcement. Through these reinforcement procedures, the rats behavior was guided using positive or negative consequences. Nonhuman subjects continue to be used to gain insight in behavioral research, as is evident in articles published in the Journal of Experimental Analysis of Behavior (JEAB) and other basic-research outlets.
Not only do reinforcements maintain a target behavior, they can also change or shape a set of behaviors. Shaping consists of teaching new behavioral properties by differentially reinforcing successive approximations toward the behavioral objective (Mayer, Sulzer-Azaroff, & Wallace, 2014). The procedure used for shaping a new behavior begins with a response that is already in the individuals’ repertoire. After this, slight changes in the topography of the behavior are reinforced as they gradually approach or approximate the target behavior. These responses are labeled as successive approximation to describe a particular response that has shifted along some dimension, such as quality, rate, intensity as it more closely approaches it target behavior. An example of shaping was conducted by Skinner (1958) and his colleagues when they taught pigeons how to bowl. The initial goal was for the pigeon to send a wooden ball down a small alley toward a set of toy pins by hitting the ball with its beak. To condition the swiping response, a food pellet would be given after the initial swipe. However, after a few minutes of waiting, the pigeon did not engage in any behaviors. Through the process of shaping, researchers began to reinforce any behaviors that resembled swiping (i.e., looking at the ball) then progressed to select responses. After a few minutes of shaping, the pigeon engaged in the behaviors of swiping the wooden ball to hit the pins. This rapid rate of learning suggests that animals and humans, illustrate the law of effect by repeated behaviors that are reinforced and extinguishing other behaviors that are not.
In order to find the optimal reinforcers for animals and humans, researchers must examine what happened before and after the behavior. This investigation of operants to demonstrate a cause-and-effect relationship is called a functional analysis. Functional analysis can be described as “the examination of inappropriate behavior and its antecedents and consequences to determine one or more functions that the behavior might serve for the learner” (Mayer, Sulzer-Azaroff, & Wallace, 2014). These antecedents and consequences are manipulated to demonstrate their lawful effect on the interested behavior. For example, in order to prevent the behavior of flapping arms in class, a functional analysis was performed. First, the behavior of flapping arms was defined and the rate was recorded and graphed under a set of two distinct conditions. The first condition was to perform a set of activities he has difficulty doing (tying his shoe) and the second condition was him playing with his favorite toys. Both of the graphs will be examined in order to see what antecedent influenced the initial behavior the most. This functional analysis gives the researcher and idea of the type of antecedents that cause the problem behavior to occur. Based on these findings, preventative measures can be implemented to make sure these establishing operants do not happen.
Similar to human behaviors, animals’ behaviors respond to reinforcement and consequences. By using the basic principles of shaping and functional analysis, three interventions were conducted that effectively changed animals’ behavior.
Shaping Procedures with Horses
The process of loading horses into trailers can be very challenging for trainers because horses exhibit problem behaviors such as rearing, pulling back, head tossing, etc. This combination of a fighting horse and an owner who uses physical force can create a threatening situation. This ongoing conflict was the catalyst for researchers Ferguson & Rosales-Ruiz (2001) study implementing positive reinforcement as an effective method for loading horses safely into trailers. Five quarter horses were taught to load safely and quickly into a trailer on command. Shaping techniques were used such as food and clicks to produce the desired response of touching the nose to a target located outside the trailer. After this behavior occurred consistently outside the trailer, the horses were taught to touch the target as it was moved progressively deeper into the trailer. These reinforced responses demonstrate successive approximations that guide the entire shaping procedure. Throughout the study, the responses generalized to other trailers and the problem behaviors became extinct because there was no reinforcement contingent on those behaviors. Historically, physical and aversive stimuli have been used to treat these problem loading behaviors but this study suggests new evidence that shaping can be a simple yet effective alternative method.
Functional Analysis: SIB in Primates
As previously stated, a functional analysis examines the antecedents and consequences of a behavior in attempts to observe a cause and effect relationship. This analysis is important for problem behaviors such as self-injurious behavior (SIB) in captive animals to try and observe what factors are contributing to the cause of such problem behaviors.
Self-injurious behavior (SIB) is defined as “deliberate, repetitive, impulsive, non-lethal harming of oneself” (Mayer, Sulzer-Azaroff, & Wallace, 2014). The first published case of self-injurious behavior in primates was documented by Tinklepaugh (1928). In this study, he examined self-biting in a male Macacus rhesus monkey (Macaca mulatta) named Cupid. This behavior was observed after Cupid formed a significant monogamous, attachment for a female of a different species then had another female of his species introduced into his situation. His SIB repertoire included lacerations of his legs, abdominal, and tail. In recent times, SIB have been reposted in approximately 10% of captive, individually housed primates.
A more recent study using functional analysis in primates for SIB was conducted by Dorey, Rosales-Ruiz, Smith, & Lovelace (2009). Researchers performed a functional analysis on a 13-year-old female olive baboon (Papio hamadryas anubis) named Rafiki to identify reinforcers that maintained her SIB. Her frequent behaviors included biting her arms, legs, and pulling hair out of her head resulting in abundant bald spots. Researchers observed Rafiki from her enclosure and recorded the durations of hair pulling, hand biting, and foot biting. Researchers performed the functional analysis examining five conditions (ignore, attention, demand, play, and tangible) in efforts to find what served as a consequence for the SIB. Another factor taken into consideration was that each of the conditions could have been paired with a reinforcement contingency that maintained SIB.
Conditions
In the ignore condition, Rafiki was located in her original enclosure without any enrichment items such as toys, tools, etc. Researchers observed all of Rafiki’s behaviors in another room out of her sight. This condition was included to test if the SIB continued in an environment with no added stimulation.
In the attention condition, Rafiki was in her original enclosure presented with no external stimulations. Outside of her enclosure, the experimenter and observer sat next to each other and chatted. The experimenter had his back facing Rafiki but the observers was not. As soon SIB occurred, the observer would begin talking to Rafiki, signaling to the experimenter that a SIB occurred. Once a SIB behavior was identified, the experimenter turned around and gave attention in the form of statements to Rafiki (e.g., “don't do that,” “don't hurt yourself”) until the SIB stopped. This attention condition was necessary in the functional analysis because it suggests that SIB could be maintained by positive reinforcement in the form of contingent social attention.
In the tangible condition, Rafiki was given the opportunity to play with ball. The ball was attached to a chain that could be easily inserted or removed from her enclosure. The appearance of the ball in the enclosure was contingent on the occurrence of the SIB. After 30 seconds passed without SIB, the experimenter removed the ball and waited for the next sign of SIB. This condition was essential in the functional analysis because it suggests the presentation of play objects (i.e., ball) functions as positive reinforcement for Rafiki's SIB.
Lastly, the play condition involved presentation of the same ball stated in the tangible condition. This time, however, the ball was initially placed in the enclosure for 30 seconds and remained there if Rafiki was interacting with it for more than 30 seconds. If Rafiki did not interact with the ball after 30 seconds, the experimenter would place it closer to Rafiki’s face. If any SIB occurred during the presentation of the ball, it was removed immediately until the SIB stopped. This control condition suggests that minimal SIB occurred because Rafiki was obtaining non-contingent social interaction and had access to alternative activities (i.e., play time with the ball).
Overall, the results of the functional analysis indicated that Rafiki's SIB was fueled by operant contingencies, specifically positive reinforcement (i.e., attention from humans). The results of the functional analysis were supported by the effects of providing attention contingent on SIB during the treatment analysis (see Figure 1). Explain results. This is breakthrough study because it was the first to establish that a functional analysis could provide effective assessment and treatment of problem behavior in a nonhuman animal (i.e., an olive baboon). With this study as a foundation, future research was conducted examining undesirable behaviors in captive chimps (Martin & Bloomsmith, 2011).
Functional Analysis: Captive Chimps
Many animals in zoo captivity exhibit undesirable behaviors with trainers, researchers, and the public. For example, captive chimpanzees (Pan troglodytes) engage in problem behaviors such as spitting, throwing feces and other objects, screaming, etc., In a zoo setting, these aversive behaviors can harm relationships with other chimpanzees and caregivers, portray a destructive image to the public, and even present health risks to staff. Researchers Martin & Bloomsmith (2011) conducted a functional analysis in attempts to discover what the cause of these behaviors were as well as what was maintaining them. Through recorded observations and interviews with on-site chimpanzee caregivers, Martin & Bloomsmith (2011) developed two response patterns to these inappropriate behaviors. The first was the delivery of positive reinforcement in the form of attention (i.e., verbal statements) or food in an attempts to divert the chimps’ attention or calm them down. The second response pattern involved escaping the area as soon as the problem behavior occurs. This escape method may be functioning as negative reinforcement because of the removal of the staff but still maintaining the behavior. Based on these patterns and problem behaviors, the functional analysis focused on identifying the reinforcers maintaining feces throwing and spitting. The participant for this study was one 27-year-old male chimpanzee that was solitarily housed. He exhibited various problem behaviors such as throwing feces, spitting, screaming, and cage shaking. Each behavior was videotaped, observed, scored, and operationally defined. Throwing feces was scored when the chimp picked up fecal matter and threw it at least 30 cm from his body. Spitting was scored when liquid exited from the chimpanzee's mouth traveled at least 15 cm from his body. Screaming was defined as a general, high-pitched, loud vocalization. Cage shaking was defined as pounding or pushing on the cage with enough force to make an audible sound. In the treatment portion, an alternative response was added that included the addition of a plastic ring to the enclosure. The chimp needed to grab the ring with at least 2 fingers and hold on for at least 2 seconds for the alternative response to be scored. The data recorded in the functional analysis was measured on the frequency of each behavior (inappropriate and alternative behavior) and converted to responses per minute.
In the positive reinforcement condition, the experimenter began with their back facing the chimp. Once the chimp engaged in a problem behavior, the experimenter faced him and provided 20 seconds of attention in the form of verbal statements as well as access to fruit juice, a known desirable sugary reward for primates. In the negative reinforcement, the experimenter was stood in front of the chimp wearing a white hazmat suit and a prepped syringe ready for injection. If the chimp exhibited the desirable behavior of pressing his thigh up to the enclosure wall, the experimenter responded with praise and backed up a few steps from the enclosure. On the other hand, if the subject engaged in a target problem behavior, the experimenter exited the display and remained out of sight for 20 seconds.
Results from the functional analysis indicated that high rates of inappropriate behavior were observed during baseline (M= 1.4 responses per minute). This data is expected during the baseline measure because no other interventions are occurring on the behavior. Also, the rates of inappropriate behavior during extinction were relatively high (M= 0.8). Under DRA plus extinction, inappropriate behavior decreased by 90% (M= 0.2) and ring holding increased to an average of 1.8 responses per minute (see Figure 2). These results indicate that DRA plus extinction was successful in reducing inappropriate behavior and increasing the occurrence of the alternative behavior of ring grabbing. As seen in Figure 2, extinction alone may be a short-term solution to problem behaviors, but not a long-term. In order to be an effective, lasting, solution, trainers and caregivers should implement a DRA or alternative behavior for the animal to engage in.
Conclusion
Skinner’s initial behavioral experiments with rats and pigeons acted as a catalyst for the future research integrating ABA techniques with animals. With the foundations of operant conditioning and the Law of Effect published, present day research now investigates the origins and motivating forces driving targeted or problem behaviors through the use of a functional analysis. Now, problem behaviors in animals can be examined and treated with the same effective procedures used with humans. Studies conducted by Tinklepaugh (1928) successfully treated SIB in a Macacus rhesus monkey and Martin & Bloomsmith (2011) treated undesirable behaviors in a captive chimpanzee. Going even further with the properties of reinforcement, the use of shaping techniques guided frightened mares to safe and calm behaviors that allowed them to be loaded into trailers painlessly (Ferguson & Rosales-Ruiz, 2001). By using the basic principles of shaping and functional analysis, three interventions were conducted that effectively changed animals’ behavior. One potential pathway for future research involves animal rehabilitation; a career field close to my heart. By using various reinforcement and shaping techniques, a desired behavior could be obtained more rapidly and reliably. This would decrease the animals overall rehab time, potentially allowing it to be released into the wild faster. I believe these successful studies with ABA techniques with animals is just the beginning of a new realm of research.