Association for Behavior Analysis International

Animals are thought to acquire, apply, and exploit information in order to exhibit socially-learned behavior. In chacma baboons (Papio ursinus), Carter et al. (2016) suggested that it is possible to explain variation in a novel, socially-learned behavior by identifying phenotypic constraints (cognitive, social, behavioral, ecological, and demographic characteristics). We conducted a similar experiment using wild vervet monkeys (Chlorocebus pygerythrus) whose social structure varies significantly from chacma baboons. We presented a novel water patch to three troops of vervet monkeys (approximately 150 individuals) living on a private game reserve in South Africa during times of water scarcity and abundance. Using a Bayesian multi-level model, we tested whether the phenotypic constraints of age, sex, dominance rank, inter-individual behavioral differences, social network structure, and troop ecology influenced water patch exploitation. Here we discuss the sequential nature of information use and argue for the importance of considering phenotypic constraint in the ontogeny and evolution of social learning. We also provide a behavior analytic approach to the concepts of information use and social learning.

One approach to investigating the dog-human bond is to evaluate dogs’ preferences for their owner compared to other salient stimuli. Dogs typically preferred food to petting from their owners in familiar contexts, but were more likely to be sensitive to a thinning food schedule, or prefer their owner on all schedules of reinforcement in unfamiliar contexts. One issue with this prior research is that the dog was not equally deprived of food and access to the owner. Thus, in this study we tested dogs’ preference for their owners versus food when they had been deprived of both for at least four hours. After a dog was separated from their owner for the required time, we gave them a concurrent choice between a bowl of food or the owner. The dogs were presented with the food they were normally fed. The dog’s behavior was recorded for a minute-long session. We tested each dog in at least four sessions. Generally, dogs preferred their owner over food and also made the owner their first choice over the food. However, we detected individual differences in dogs. We are confident in our results as we conducted multiple sessions with each dog.

Reinforcer efficacy is an essential issue for training and maintaining desirable behavior in animals. Horses are often required to perform long sequences of behavior such that identifying reinforcers that can maintain long bouts of behavior or high effort behaviors would be useful. One way of testing reinforcer efficacy is using a progressive ratio schedule and measuring break points. Higher break points correspond to more effective reinforcers. We trained horses used in a riding instruction program to touch a target stick with their noses. After horses reached criterion levels of responding, they entered the testing phase in which we used three different grains with varying starch content as reinforcers to determine if starch content affected reinforcer value. Horses were tested on each grain each testing day and we tested each horse multiple days. We measured trials to criterion during training and break points during testing. We will discuss our results in terms of individual differences and sensitivity to different levels of starch using concepts of unit price, demand, and work.We are confident in our results given that we used a single-subject design and tested each horse multiple times on each grain type.

There is a common untested belief that domesticated dogs tend to sniff their owners more when they have been around unfamiliar dogs. In this study we aimed to determine whether this phenomenon is true in a controlled test. Six dogs interacted with a familiar caretaker while wearing sweatpants impregnated with one of four odors: odor of an unfamiliar dog, odor of self, almond extract, and no scent control. The Experimenter wore the sweatpants and stood in an enclosed space. The Experimenter called the dog’s name and pet the dog for ten seconds to encourage the dog to approach the experimenter, and then remained neutral for the duration of a 2 min session. The interaction was recorded and coded for the time the dog spent sniffing the Experimenter’s pants. Each dog went through each condition twice in a pseudorandomized order. Dogs spent significantly more time investigating sweatpants with an unfamiliar dog odor compared to self, control and almond extract. There was no difference in time sniffing between almond extract, self, and control. These results support that dogs do spontaneously recognize unfamiliar canine odors.

Non-native invasive plants (NNIP) have the potential to cause considerable economic and environmental harm. Ruminants, including goats (Capra hircus), are a NNIP management strategy that may be less environmentally harmful than herbicide use, less time consuming than physical weed removal, and more cost effective than either strategy. In this naturalistic observation study, the extent to which goats made facial contact with one NNIP--Rosa multiflora--was evaluated. The goat herd was led by the handler to a Rosa bush in a forest where they were free to roam or remain next to the Rosa bush with the handler. Momentary time samples of goat facial contact (any part of the goat’s head touching any part of a Rosa bush) with 9 goats (5 adult and 4 juvenile does) were collected during 49 video-recorded sessions. Although all goats were observed to contact Rosa, it did not occur frequently (M = 28.02%), was highly variable (SD = 26.75), and in only 2% of sessions was more than 50%. Individual differences in amount of Rosa contact occurred ranging from 29.17-7.83% per session. Operant conditioning methods are necessary to increase goats’ contact with NNIP and facilitate its destruction through stomping or eating.

Non-native invasive plants (NNIP) can cause considerable economic and environmental harm. Ruminants such as goats (Capra hircus) as a NNIP management strategy may be less environmentally harmful compared to traditional herbicide use, less time consuming than human weed removal, and more cost effective than herbicides and/or manual removal. An experimental demonstration of operant conditioning procedures to increase a goat’s consumption of one NNIP--Rosa multiflora--was performed. A multielement research design was used to compare a goat’s frequency of eating Rosa during training (prompting/fading, differential reinforcement) versus baseline conditions. Sessions occurred while the goat freely roamed with a herd in a forest containing native and non-native shrubs and trees. Data consists of observational coding of goat eating behavior during video recorded sessions. During all 15 sessions, a higher frequency of goat Rosa eating behavior occurred in the training condition compared to baseline (M = 45.53 versus 6.67, respectively). Conversely, the goat consumed vegetation other than Rosa less frequently in the training condition compared to baseline (M = 23.47 versus 47.87, respectively). Although these findings suggest that operant conditioning procedures may increase a goat’s consumption of Rosa, demonstration of maintenance and replication across other goats and NNIP are needed.

Separation-related issues, often characterized by excessive vocalizations, locomotory behaviors and/or property destruction are common canine behaviour problems. While other disciplines (e.g., veterinary behaviour and applied ethology) have extensively researched these phenomena, behaviour analysis has paid little attention to them. This data-supported case study investigated an operant approach as treatment for separation-related behaviors in a companion dog. Lucy, a 6-month old, mixture-breed dog was presented for excessive vocalization when left alone. Upon initial consultation and baseline taking, it was found that Lucy had no settle (i.e., lying-down on cue) response in her repertoire and displayed high instances of vocalizations (i.e., whining and barking) and pacing (i.e., walking around the apartment) when left alone. Treatment comprised (a) teaching a solid settle response; (b) shaping successively longer durations of separation; and (c) counterconditioning with a stuffed toy. A modification of ABC design was used for data analysis. Results showed overall improvements in settle responses, vocalizations, and pacing. A follow-up probe detected medium-term improvements across all behaviors. This case study suggests shaping procedures as potential treatments for canine separation-related issues. More research is needed to further investigate efficacy and generality of operant approaches for treating separation-related behaviors of varying magnitudes in companion dogs.

In a square experimental tank (ET), one experimentally naïve lake sturgeon (Acipenser fulvescens) was operantly conditioned to enter one of four target areas where it received an auditory feedback-response stimulus (an automated “click”) following its association with darkness reinforcement. The number of responses for each of the four targets was monitored by visual observation and computer software. The reliability of visual observation was confirmed by IOA among observers. The targets were in each corner of the ET and had to be entered by the subject to constitute a response. The study consisted of the following phases: (1) no-feedback baseline (NFB); (2) response feedback stimulus baseline (RFSB); (3) continuous reinforcement schedule (CRF); (4) extinction (EX); (5) test for spontaneous recovery; and (6) fixed-ratio schedule of reinforcement (FR1 ~ CRF, FR2, and FR4). It was found during phases 3 and 6 that the subject preferred the target area that resulted in reinforcement as it was indicated by an increase in target area responses for that corner; and decreased responding on the target during phases 4 and 5. Results during the other phases were less clear. This was to the best of our knowledge the first operant conditioning study conducted using this species.

Pet cats are often thought to be difficult to train, which may be due to a lack of knowledge of what motivates cats to respond (Vitale Shreve, Mehrkam, & Udell 2017). The present study compared two preference assessments with shelter cats – a free operant design (Vitale Shreve et al., 2017) and a paired stimulus design (Fisher et al., 1992). The purpose was to determine an ideal assessment for use in a shelter setting, including ease of use and reliability of results. Each cat completed both assessments, had their most-preferred stimulus used as a reinforcer in basic obedience training for five training sessions, and then was retested. On their initial assessments (n = 20), chicken-flavored wet food was selected most during the paired stimulus assessments (45%), while petting won the free operant assessments (50%). On their retests (n = 6), only two cats had consistent preferences. While data is still being collected, initial analysis demonstrate the paired stimulus assessment worked best overall (easier to set up, required less space, and resulted in less cat stress behavior) and that cat preferences are likely to change in as little as a week regardless of assessment used, indicating regular assessments are necessary.