15.2.1 Structured Behavioral Test

Under this framework, an animal is exposed to one or more stimuli under controlled conditions during a predetermined timeframe and their behavioral response is measured. Conventionally, these tests are administered only once, but repeat testing can be conducted as well. The stimuli presented are usually selected to reflect stimuli that may be encountered in a traditional home environment. For example, a hypothetical test to assess if a cat is prone to inter‐cat aggression could involve holding a second cat 1 ft. from the focal cat’s cage for one minute and measuring the behaviors associated with aggression or affiliation exhibited by the focal cat. The method of measuring these behaviors varies, but could include the presence or absence of a behavior (e.g., hissing), an ordinal rating of the severity of the behavior (e.g., 1 = pupils not dilated to 5 = pupils fully dilated), and the frequency, duration/proportion, or latency of a behavior (e.g., meowed four times, was in hiding box for 60% of test period, approached after 37 seconds). The conditions controlled within the testing environment usually include the characteristics of the room, the behavior of the tester, and the order in which the stimuli are presented, but controlling additional conditions may be requested as well. For example, a test may require a quiet room with no windows, the tester refrain from acknowledging the animal, and that animal be presented first with a novel object, then being pet with a stick, and finally introduced to an unfamiliar animal.

This framework has the advantage of producing a quantitative data set. If shelters set clear criteria for what results classify an “adoptable’ animal, this framework facilitates easy decision‐making, thereby reducing the time required to debate each case and reducing the burden of guilt often associated with rejecting an animal (UC Davis Koret Shelter Medicine Program 2019). These quantitative results can also be compared across individuals to understand how performance in these tests differs between variables such as intake type, sex, and reason for surrender. Findings could be used to generate a greater understanding of the different needs of these demographic categories, or within individuals to compare response to different stimuli (e.g., male or female handlers) to contribute to ideal placement recommendations.

Unfortunately, this framework has several disadvantages. Standardized behavior tests can use a lot of shelter resources (e.g., staff time, space for testing). It is also difficult (if not impossible) to present stimuli and conditions that adequately recreate real world situations, thus bringing into question the validity of results. Take for example, the hypothetical test to assess if a cat is prone to inter‐cat aggression given above. Introducing a new cat to a resident cat in a home setting is often stressful for one or both cats. However, a responsible owner will ensure this introduction is gradual and paired with positive stimuli. In the hypothetical test situation this introduction is instantaneous, and the shelter setting presents additional stressors (e.g., barking dogs, confinement) likely not present in a home. It is not difficult to imagine that this trigger‐stacking may result in a reaction that is not representative of how this cat would react to meeting a new cat in a home environment. Furthermore, while the controlled conditions limit the variability expected in the results (by reducing confounding factors), this may limit the ability to generalize the results outside of the confines of the test. For example, a test may be designed to assess tolerance of petting by measuring the presence/absence of swatting in response to touch, and the control conditions might stipulate that it is always conducted by the same handler to reduce variability in approach style. However, results of that test may not be generalizable to other handlers with different approach styles. Finally, the stimuli presented in these tests are often purposefully provoking, resulting in a high degree of false positives (Patronek and Bradley 2016). Patronek and Bradley (2016) demonstrate mathematically that behavioral tests designed to identify aggressive behaviors in dogs provide results that are “no better than flipping a coin.”

15.2.2 Scan Samples of Behaviors

This framework involves an observer approaching an animal’s enclosure and recording the presence or absence of specific behaviors or behavioral indicators. To be most impactful, this would be conducted at multiple timepoints, ideally on a regular schedule. Within a shelter this could be scheduled in association with other regular tasks, such as before cleaning/feeding. The regular schedule could therefore help in standardizing the conditions (e.g., before shelter is open to the public, before feeding/cleaning). Specified behaviors could include indicators of both positive (e.g., approach cage front, playing) and negative (e.g., hiding, hissing) emotions. Other indicators linked to stress could also be recorded, for example food intake (Ellis et al. 2014; Tanaka et al. 2012) or sickness behaviors (Stella et al. 2013; Tanaka et al. 2012). Certain behaviors could also be assessed on a continuum, such as pain (Epstein et al. 2015) or stress (e.g., the Cat‐Stress‐Score; Kessler and Turner 1997). These assessments require behavioral definitions and training of observers, but do not require previous experience with the individual animal.

This framework produces an ordinal data set and thus scores could be associated with specific criteria that can be used for decision‐making or to monitor progress in behavior modification. The system also produces information about the cat’s behavior daily, so information can be gathered quickly, problems can be identified early, and progress can be monitored easily. Unlike standardized behavioral tests, it does not use many resources (since assessments are conducted in the cat’s home environment and can be done alongside regular shelter activities), is collected at multiple time points by design (eliminating the risk of making interpretations based on a single data point that could be anomalous), and is not in response to purposefully provoking stimuli.

Unfortunately, behaviors observed under this framework would also be influenced by the inherently stressful conditions of the shelter. This means that behaviors observed may not reflect how the cat would behave in a home environment. Additionally, these scans are best suited to monitor chronic problems (such as fear) and not specific undesirable behaviors (such as petting‐induced aggression). Finally, training is required on the behavioral definitions and rating scales of the behaviors or behavioral indicators being measured, and ideally, interobserver reliability of raters should be assessed. This could be challenging for shelter environments where time is already stretched thin.

15.2.3 Ad Libitum Behavioral Observation

This framework involves observing and making inferences about an animal’s behavior outside of a structured context. Observations can be made at any time (e.g., intake, medical exam, volunteer interactions). Any notable behaviors (such as those associated with aggression or fear) should be reported and recorded, as should the context in which the behavior(s) took place. For example, a volunteer interacting with a cat may observe that the cat was pushing into her hand while she was petting him, but that he hissed and retreated to his hiding box after five minutes of petting. This observation could contribute to the inference that the cat is exhibiting petting‐induced aggression

A major advantage afforded by this framework is that it does not necessarily involve extra staff time. Notable behaviors are already being observed ad libitum by shelter staff and volunteers during their regular activities; all that is required is an organized system for reporting and recording them. Although the behavior exhibited in a shelter is always prone to the impact of stress, ad libitum observation of behavior in a free setting may avoid some of the behavioral abnormalities that may result from an artificial test situation. As a result, this framework is far less likely to result in false positives.

Unfortunately, this framework does not have clear criteria for interpretation, and thus does not lend itself to quick decision‐making. Since it is based on observations made opportunistically over time, it may take time for issues to emerge. It also suffers from a lack of standardization, which can make observations difficult to interpret. As these observations can be made by a range of different people, this introduces a lot of variability in terms of the quality of the observations and the context of the interactions. Dawson et al. (2019) found that people vary widely in their ability to interpret feline emotion from cats’ faces and that this ability was not related to owning a cat in the past. This study highlights the importance of a thorough training program in feline behavior for staff and volunteers if observations in a free setting can be trusted as a valid source of information. The context of the interaction needs to be considered as well. A cat exhibiting hissing and growling while a volunteer reads to them through their cage bars should be interpreted differently than the same behaviors exhibited while a technician is drawing blood. In shelters that are under resourced and cannot allocate much staff or volunteer time for positive enrichment‐based interaction, most of the interaction the cats experience may take place during unpleasant activities such as cleaning and intake exams. This may bias the data to include higher rates of negative behaviors. Finally, compliance with recording behaviors may not be high. Some people may elect not to record a behavior such as a bite or swat if they fear it may contribute to a euthanasia decision or if they are embarrassed they did not read warning signs and end the interaction earlier.

15.2.4 Trait Rating

Under this framework one or more individuals who have extensive experience with the animal rate them for various traits (e.g., agreeableness, neuroticism) using a Likert or visual analogue scale. Ratings on each trait can be compared between cats to identify differences, or within cats to identify common patterns of traits. This method is used broadly in the study of human individuality (e.g., the five‐factor model; McCrae and John 1992), quality of life in farm animals (e.g., Qualitative Behavior Assessment; Wemelsfelder and Lawrence 2001), and within cats specifically (e.g., Feaver et al. 1986).

This framework benefits from the ability to measure the subtle aspects of a cat’s behavior that do not fit well within the structures imposed by conventional methods of measuring behavior (Meagher 2009). Trainers familiar with each animal can rate them on traits in moments, and thus little staff time would need to be allocated to testing or data collection. Criteria could then be set for how to proceed based on an individual’s score (e.g., if an animal scores ≥4 for “aggressiveness”).

However, trait rating requires raters to have extensive experience with the individual animal, so even though the rating itself does not require significant staff time, ratings could not be made until the animal had been in care for enough time for the raters to have spent extensive time with them. Furthermore, the method is subjective and is thus subject to personal bias. Comparison between observer ratings is intended to help assess this problem (Meagher 2009), but this requires at least two observers, which could reduce the feasibility of using this method in shelters.

15.2.5 Qualitative Behavior History

This framework derives inferences about an animal’s behavior through free text descriptions provided by someone with extensive experience with the individual. Within animal sheltering, this type of information is often collected through conversations with surrendering parties or intake forms. This structure allows for the description of an animal’s temperament that may not be evident in the shelter situation, for anecdotal incidents that may be revealing about individual preferences or responses to specific stimuli, and for combinations of individual quirks.

However, it is easy to imagine how personal bias or emotion may play a role in the quality of information provided in owner reports, or how a lack of expertise in feline behavior may lead the owner to flawed conclusions. Details in intake profiles may be skewed to downplay behavioral issues if an owner thinks an honest description of the animal’s behavior may result in euthanasia. Similarly, these descriptions may convey an exaggerated representation of negative qualities, as owners are often very frustrated with their cats when filling out these forms. They also can describe behaviors without the relevant context required to understand the motivation for the behavior. Furthermore, the qualitative behavior histories provided by owners do not lend themselves to clear criteria for acceptability, resulting in less clear‐cut outcome decision‐making. This increases the time required to debate each case and increases the burden of guilt on shelter staff that is often associated with rejecting an animal.

15.3.1 Identify Propensity for Undesirable Behavior

A common goal of behavior assessments is to determine whether an individual animal is prone to specific undesirable behaviors. The propensity for such behaviors can influence whether a shelter feels equipped to accept an animal, whether an animal is deemed “adoptable,” the housing and training/management that might be required for that animal’s care while in shelter, and the home or environment deemed appropriate for that animal’s placement.

A standardized behavioral test administered shortly after arrival at the shelter can seem an ideal framework for determining a propensity for undesirable behaviors. Predetermined pathway plans associated with the results from such a test can eliminate the necessity to debate the outcomes of animals, and alternative endpoints (e.g., euthanasia, return to colony, working cat programs) can be pursued quickly in appropriate cases. This would reduce the unnecessary suffering associated with a prolonged length of stay resulting from indecision, reduce the burden of guilt often felt by shelter staff, and focus time‐consuming behavior modification efforts on individuals with the highest likelihood of being successfully rehomed. Furthermore, van der Borg et al. (1991) found that structured behavioral tests provided a better prediction of problem behavior than did the opinions of the staff based on ad libitum observations. However, there are several problems inherent in relying solely on the results of a standardized behavioral test including the stress of the shelter environment, the speed of assessment after intake, the reliance on the data from a single assessment, debate about the ability of the test to predict the exhibition of unwanted behavior in subsequent homes, and the relative lack of tests developed for use with shelter cats.

It can be difficult to get an honest representation of how an animal is likely to behave in a home environment based solely on their behavior in a shelter. Despite the best efforts of staff and volunteers, shelters are an inherently stressful place for animals due to exposure to novel stimuli, confinement, reduced enrichment, and an absence of stable positive relationships. As behavior is a key part of the stress response, it is logical that an animal’s behavior in a shelter environment would be impacted by the stress inherent in this environment. This is particularly true for cats as, unlike dogs, they are notoriously poorly socialized to novelty. Most cats are kept within the owner’s home for the majority of their lives (with the exception of negative experiences, such as veterinary visits). In times of stress, the behavioral response of most animals is fight, flight, or freeze (Skinner et al. 2003). In the traditional shelter cage environment, the option to flee is largely removed or eliminated. This greatly increases the risk of aggression—a behavior that places a cat at increased risk for euthanasia. It is important to note that aggressive behaviors exhibited under these conditions are a normal behavioral response and not indicative of a pathological behavioral condition.

While the shelter environment is likely to be stressful for animals for the duration of their stay, it is experienced most profoundly during the first few days or weeks. Kessler and Turner (1997) found a sequential significant decrease in Cat‐Stress‐Scores (CSSs) each day a cat was in a boarding facility until day five, and Ellis et al. (2014) found that CSS decreased from week one to week two, and fecal glucocorticoid metabolites decreased from week one to week five in a shelter‐like environment. These studies suggest that the average time for cats to habituate to a shelter environment ranges from five days to five weeks. While the ability to make quick decisions about an animal’s fate provided by structured behavioral tests is ideal from an efficiency standpoint, the profound stress experienced by a cat immediately following intake has the potential to greatly impact the validity of an assessment conducted during this period.

Many problem behaviors identified by behavior assessments depend on a range of conditions, such as the cat’s experience that day, previous experience with handler, hunger, or pain. For example, attempts to pet a cat who has spent all day napping could easily be met with a different outcome than a cat who has spent all day enduring painful medical procedures. Slater et al. (2013b) found that cats needed to be interacted with on multiple occasions in order to capture key behaviors that are indicators of socialization. Standardizing conditions for behavioral tests can be difficult even in a laboratory context, and in a busy shelter environment can be next to impossible. As decisions about animals are often based on the results of a single behavioral test, it is impossible to account for the impact of variables, such as affective state, on test results (Patronek and Bradley 2016).

There is a growing body of research questioning the ability of behavioral tests conducted in shelters to predict behavior in subsequent homes. Within the dog literature, Christensen et al. (2007) found that more than 40% of dogs that passed an in‐shelter behavioral test designed to identify a propensity for aggressive behavior exhibited aggressive behaviors in their subsequent home, and Mohan‐Gibbons et al. (2012) found that adopters of dogs identified as exhibiting food‐guarding behaviors in shelter via a standardized behavioral test reported this behavior as rare or absent in their home during follow‐up surveys. The standardized behavioral tests used in both of these studies had been published and subjected to the rigor of the peer review process. However, the incongruency of behavior in the home and shelter environment point to these tests having poor predictive value. See Chapter 9 for more information on assessing the behavior of shelter dogs.

Adding further difficulty to the use of standardized behavioral tests is the relative lack of such tests in the literature for cats as compared to dogs. The Feline Temperament Profile (FTP) may be the only published example of such a test, at least for use with socialized cats. The FTP was originally designed (Lee et al. 1983) to determine the suitability of individual cats for companion animals in nursing homes by identifying “acceptable” and “questionable” behaviors through 10 sequential mini‐tests of increasing contact (e.g., calling the cat, petting the cat, holding the cat) with an unfamiliar person. Due to the to the ease of administration, small time commitment, and absence of the need of observers with extensive knowledge of the animal, Siegford et al. (2003

Stay updated, free articles. Join our Telegram channel

Join