Author Names: Georgina Lewis (BSc (Hons) Animal Behaviour and Welfare) and Dr Carrie Ijichi
Very few studies have explored the differences between dog breed groups and their cognitive abilities. Exploring canine cognition is important for improving both dog welfare and the dog-owner bond. It is crucial that dog owners understand a dog’s cognitive abilities and the level of stimulation a dog needs to meet their mental requirements. In doing so this may prevent behavioural problems developing which frequently result in the breakdown of the canine-owner relationship and dog relinquishment. The aim of the current study was to investigate whether cognitive abilities differ between the working dog group and the toy dog group. To assess the cognitive performance of toy and working dogs three cognitive tests were administered (object permanence test, quantity discrimination test and discriminative cues test). Latencies to complete each test and the number of correct choices made were measured, allowing a comparison to be made between the two breed groups and cognitive performance. It was hypothesised that there would be a significant difference between the two canine breed groups and their cognitive abilities, due to the working group completing the cognitive tests more successfully. Overall, no significant difference was found between the toy dog group and the working dog group in terms of their cognitive abilities. This suggests that working and toy dogs have similar cognitive abilities. No significant difference was found between the two breed groups and latencies taken to complete the tests or the number of correct choices made during the discriminative cues and quantity discrimination tests. A better understanding of the cognitive abilities of dogs may help owners understand their dog’s needs more appropriately, which may lead to a reduction in the number of dogs relinquished into shelters due to behavioural problems.
In the UK the estimated dog population is 8.5 million, with around 24% of households owning a pet dog (PFMA, 2015). It is accepted that dogs play an important part in society providing companionship to their owners, yet it has been estimated that around 5% of the population of owned dogs are relinquished to shelters yearly (Case, 2013; Westgarth et al, 2007). Although shelters aim to rehome as many unwanted dogs as possible, approximately 56% of dogs surrendered to kennels are euthanised (Case, 2013). The predominant cause of relinquishment is due to the breakdown of bonds between dogs and their owners (Case, 2013). Many factors can interfere with the success and length of the human-canine bond. These include; owners being unaware of their dog’s needs, unrealistic expectations and behavioural problems such as hyperactivity and destructive behaviours developing due to a lack of and/or inappropriate training and potentially inadequate environmental enrichment (Case, 2013; Previde and Valsecchi, 2004). Environmental enrichment can be utilised to improve an animal’s physical and psychological wellbeing (Newberry, 1995). Providing appropriate stimuli can provide opportunities for an animal to use its cognitive skills and may allow animals to perform motivated behaviours (Meehan and Mench, 2007). Cognition can be defined as the mental process by which an animal acquires, stores and acts on information from the environment (Broom and Fraser, 2015; Broom, 2010). Although studies have revealed that some species such as dogs, pigs and parrots have more complex cognitive abilities than previously expected, cognitive abilities of individuals within a species may differ (Broom, 2010). A better understanding of canine cognition, may help owners meet their dog’s needs more appropriately, leading to a reduction in dogs being relinquished due to behavioural problems.
It is apparent that dog owners overestimate a dog’s cognitive abilities. Howell et al (2013) conducted a survey involving 565 dog owners where participants were asked to indicate their perception of dog’s cognitive abilities. Howell et al (2013) revealed that most participants overestimated the cognitive capabilities of dogs. This demonstrates that owner’s expectations and perceptions are unrealistic, which may be due to a lack of understanding or knowledge regarding canine cognition amongst dog owners. However, this study was overrepresented by female dog owners (90.1%), which may have biased the results as it has been shown that men and women differ in the level of attachment to their pets (Payne, Bennett and McGreevy, 2015). Misinterpreting a dog’s abilities such as their level of cognition can often lead to poor welfare and can be responsible for the breakdown of canine-owner bonds (Payne, Bennett and McGreevy, 2015). For instance, owners may provide enrichment that does not stimulate a dog’s needs sufficiently or present a challenge that a dog cannot effectively solve, potentially leading to frustration (Meehan and Mench, 2007). Frustration is defined as an emotional state that arises when an animal is highly motivated to perform a behaviour but is unable to do so because of physical barriers or other constraints (Webster, 2011). This can result in abnormalities of physiology and behaviours which are indicative of poor welfare (Broom, 1991). Frustration is associated with behaviours such as chewing, pacing, barking and play bouncing, which are generally considered as undesirable, and may increase the risk of relinquishment (Shore, Burdsal and Douglas, 2008; Stephen and Ledger, 2005).
Dog breeds are also perceived to differ in their level of trainability, intelligence and obedience (Ley, Bennett and Coleman, 2009). Working dogs are described as dogs which have been bred to perform specific roles such as guarding and hunting (Turcsán, Kubinyi and Miklósi, 2011). Toy dogs are described as small breeds, with the main purpose being to provide companionship (Turcsán, Kubinyi and Miklósi, 2011). Ley, Bennett and Coleman (2009) found that dog owners rate working dogs higher on the training focus subscale in comparison to toy dogs. These perceptions seem reasonable considering the roles working dogs have been bred to perform such as search and rescue, herding and hunting and their long history working alongside humans (Serpell and Duffy, 2014; Ley, Bennett and Coleman, 2009). The assumption that working dogs are more intelligent and trainable in comparison to toy dogs should be tested. This could be achieved by conducting a series of cognitive tests to determine whether cognitive abilities do differ between toy and working dogs.
Although, much research into canine cognition has been addressed, breed group differences in relation to their cognitive abilities has received little attention from researchers. Serpell and Hsu (2005) investigated breed differences in trainability revealing that dogs bred for working were significantly more trainable in comparison to show bred dogs or toy breeds, indicating that working dogs have higher cognitive abilities (Serpell and Hsu, 2005). However, these results were based on subjective data collected from a self-report questionnaire, and secondly this study was primarily interested in breed differences and therefore did not explore breed group differences (Serpell and Hsu, 2005). To prevent subjectivity influencing the results, future studies should aim to collect objective data by conducting experimental studies to test the cognitive performance of canine breeds.
Research on canine cognition has shown that cognitive tests are reliable for measuring cognitive abilities and that dogs are capable of using human gestural cues, solving detour tasks and successfully completing quantity discrimination and object permanence tests (Bensky, Gosling and Sinn, 2013). Arden and Adams (2016) conducted a study to assess individual differences in cognitive abilities of 68 border collies. Arden and Adams (2016) achieved this by administering multiple cognitive tests, these included a detour test, quantity discrimination and discriminative cues test. Discriminative cues tests are designed to test a dog’s ability to use human gestures to infer the location of a hidden item (Lakatos et al, 2011). Quantity discrimination tests are used to assess whether a dog can discriminate between two different quantities (Arden and Adams, 2016). Although, Arden and Adams (2016) primarily explored within breed differences, the study revealed that there was in-breed variability in performance and that dogs which performed highly on one test tended to do well in the following tests.
Gagnon and Dore (1992) conducted a series of object permanence tests to assess the level of object permanence in dogs and to discover whether cognitive performance was consistent amongst various breeds. Object permanence tests involve presenting an object to a subject and then removing it from the subject’s view (Gagnon and Dore, 1992). This will test whether a dog can represent and locate an object that is no longer available to its immediate perception (Gagnon and Dore, 1992). Gagnon and Dore (1992) demonstrated that dogs can successfully complete visible displacement tasks and that the performance of sporting, terrier and working dogs was similar during displacement tasks. Considering existing literature, it is evident that further research needs to be conducted to discover whether cognitive abilities differ between canine breed groups.
Cognitive abilities of dogs are generally misunderstood by dog owners. The majority of dog owners overestimate the cognitive abilities of dogs and it is likely that most owners will have preconceived ideas regarding which breeds are the most intelligent (Howell et al, 2013; Ley, Bennett and Coleman, 2009). Very little research has investigated the differences between two canine breed groups and cognitive abilities. The current study aims to investigate whether cognitive abilities differ between the working dog group and the toy dog group, two dog groups which have very different roles in society (Turcsán, Kubinyi and Miklósi, 2011). This was achieved by running three cognitive tests (object permanence test, quantity discrimination test and discriminative cues test) selected and adapted from previous research. Latencies to complete each test and the number of correct choices made during the quantity discrimination and discriminative cues test were measured. Performance was then compared between the two breed groups to identify whether there was a significant difference in their cognitive abilities. Many studies have demonstrated that there are differences between dog breeds and their cognitive abilities (Bensky, Gosling and Sinn, 2013). Working dogs have been bred specifically to complete complex working roles such as hunting, unlike toy dogs whose main function is to provide companionship (Turcsán, Kubinyi and Miklósi, 2011; Serpell and Hsu, 2005). Therefore, it was hypothesised that there would be a significant difference between the two dog breed groups and their cognitive abilities, with the working group completing the cognitive tests more successfully.
2.1 Subjects and Site
A total of eighteen dogs participated in the present study (Table 1). Eleven dogs were female and seven were male. Subjects were aged between one and ten years (mean=4.77 ± 2.51). Dogs over the age of 10 were excluded from the study due to the potential of age-related cognitive dysfunction (Nagasawa et al, 2012). Dogs recruited for this study were from private households and were familiar with food treats and either tennis balls or squeezy toys which ensured that dogs would respond well to the cognitive tests that were food and game orientated. Subjects were sourced through posting adverts on social media and handing out adverts in veterinary practices, pet shops, canine groomers and dog walking parks where canine owners were likely to visit.
Tests were administered within two sessions and were conducted in a controlled room at Hartpury College where all sessions were videotaped. One experimenter and one research assistant participated during this study as well as a camera operator.
Table 1: Details of the subjects involved in this study.
|Subject||Breed group||Breed||Sex||Age (years)|
|5||Working||Husky Cross Collie||F||8|
|6||Working||Sprocker (Springer Spaniel Cross Cocker Spaniel)||F||4|
|10||Working||Springador (Springer Spaniel Cross Labrador).||F||6|
|15||Toy||Cavalier King Charles Spaniel||F||4|
|16||Toy||Cavalier King Charles Spaniel||M||10|
|17||Toy||Cavalier King Charles Spaniel||F||4|
2.2 Study Design
2.2.1 Pilot Study
An initial pilot study was conducted prior to the main study to assess the suitability of the tests involved. The pilot study took place in the same room as that used for the main study, and the same experimenter, handler and camera operator were involved. Two dogs (Bedlington Whippet and Jack Russell Terrier) aged two and nine completed the same three cognitive tests used in the main study. Results from the pilot study revealed that both dogs during the quantity discrimination test showed a side-bias for the baited plate on the right. This test consisted of three trials. During each trial a different pairing of food was presented (15ml vs 2.5ml, 7.5ml vs 1.2ml and 5ml vs 2.5ml), and the dog could choose a plate to go to. This test was then repeated using different pairings of food which were more visually distinguishable (7.5ml vs 1.2ml, 15ml vs 1.2ml and 15ml vs 2.5ml). When these measures were used, it resulted in both dogs performing more accurately and removed side bias, therefore these measures were used in the main study. The pilot study also revealed that both the gravy bones and wet food were good motivators for both dogs.
2.2.2 Main Study
Prior to testing in the main study, dogs were introduced to the experimental room with their owner. Subjects could explore the room for five minutes to become familiar with the experimenters, environment and objects, preventing distractions arising during testing. The familiarisation period is important in ensuring that factors that could affect a subject’s performance are limited and do not confound the results. During testing all participants (owners) and other subjects remained outside of the experimental room. Three tests; object permanence, quantity discrimination and discriminative cues test were all carried out in the same room in a random order. The number of trials per test was limited to minimise opportunities for trial and error learning which could improve performance (Pongracz et al, 2003). Once the first test was completed the subject then went on to complete the second and third test. In between each test dogs were given between 1-2 minutes to rest if necessary and use the water bowls provided. All tests were recorded using a Sony HRD-CX33OE video camera fixed on a tripod and timed using a stopwatch to ensure each dog received the same time to complete a test.
2.3 Object Permanence Test.
Squeezable toys (green frog) and tennis balls were used as target objects to gain the subject’s attention and maintain the subject’s motivation during the experiment. During the five-minute introduction to the experimental room, each subject was given the chance to play with the squeezable toy and ball. This allowed the experimenter to assess which item motivated the subject most and would therefore be the target object for that individual dog. The target object was manipulated by a 1.25 metre of transparent nylon thread.
Gagnon and Dore (1992) used three screens that functioned to hide the target object, however in the current study a single plywood screen was used to hide the object. This is because the current study was only concerned with whether dogs could successfully locate a displaced object. This could be assessed using a single visible displacement task which involves hiding the object in a single location, meaning only a single screen was required. Gagnon and Dore (1992) however required three screens in their study as they were interested in exploring what level of object permanence dogs had. To assess this Gagnon and Dore (1992) utilised numerous tests, including single visible and invisible tests and successive displacement tests which involved moving the target object multiple times behind a series of screens before the object is finally hidden. The screen consisted of three sides with no bottom or back panel (20cm wide x 40cm high x 12cm deep).
Subjects were placed 1.5 metres away from the screen and were restrained by the research assistant. This involved the research assistant holding the subject by its front shoulders which prevented the subject from moving towards the screen or target object. The experimenter attracted the subject’s attention towards the target object by moving the object via the nylon thread into the subject’s immediate visual field. Once the subject’s attention was on the target object, the object was hidden behind the screen and the experimenter stood motionless, 0.5 metres behind the screen. The subject was then released by the research assistant and allowed to search for the hidden object. If the subject successfully located the object it was given 10- 20 seconds to play with the toy as Gagnon and Dore (1992) found that a subject’s motivation to search for an object was increased if they were given the opportunity to play with the object after locating it.
A trial was successful if the subject grasped the object with its mouth, touched the object with its paw or placed its muzzle on the object. A trial was unsuccessful if the subject did not display one of the three mentioned responses and therefore the subject was given the full-time score of 60 seconds. Three consecutive trials were carried out, for each trial the subject had 60 seconds from when the subject was released to locate and respond to the displaced object. A maximum of 60 seconds was chosen as dogs can complete this test within 1-minute (Pongracz et al, 2001). A trial was ended and resumed from the beginning if the subject did not follow each step during the manipulation process of the object displacement.
Figure 1: Experimental set up for the object permanence test.
2.4 Quantity Discrimination Task
Two plates (20cm in diameter) were used during each trial, wet dog food was placed onto the centre of each plate. Four different sized spoons (1.2ml, 2.5ml, 7.5ml, 15ml) were used to measure out the wet dog food.
Quantity discrimination was tested by offering two quantities of food presented on two separate plates, then counting how many times the subject chose to go to the plate that held the larger amount of food. Subjects were brought to a marker (1.5 metres from the baited plates) and restrained by the research assistant who held the lead attached to the subject. This prevented the subject from moving towards either of the baited plates. The protocol of Arden and Adams (2016) was adapted such that instead of drawing various sized circles on to the plates and then filling the circle marked with wet dog food, the current study measured the wet dog food out using 4 different sized spoons (1.2ml, 2.5ml, 7.5ml, 15ml) which was then placed onto the centre of each plate. Using spoons was found to be a more precise method of measuring out the dog food as it ensured the accurate amount was presented on the plate each time. Also, using smaller volumes of food ensured that subjects still had an appetite and therefore motivation to complete the task.
Once the subject’s attention was directed towards the experimenter, the experimenter held two plates both containing a different amount of food. Both plates were shown to the subject and then placed on the floor 60cm apart and 0.5 metres in front of the experimenter. Once the plates were on the floor and the subject had seen the food, the subject which had been previously held on a lead was released by the research assistant allowing the subject to choose a plate to go to and eat. To prevent the research assistant subconsciously influencing the subject’s performance during the release, avoiding the “Clever Hans Effect”, the lead used previously to restrain the subject was simply dropped (Schmidjell et al, 2012). Once the subject had gone to a plate the other plate was removed immediately. This was repeated for three trials, during each trial a different pairing of measures were used (7.5ml vs 1.2ml, 15ml vs 1.2ml and 15ml vs 2.5ml). On average, it takes 2.4 seconds for a dog to complete the quantity discrimination task (Arden and Adam, 2016). Dogs in this study were therefore given 5 seconds to complete the task before the trial was ended. The number of times each dog correctly chose the plate with the larger quantity was measured as well as latency in seconds from the dog’s release to choice.
Figure 2: Experimental set up for the quantity discrimination task
2.5 Discriminative Cues Test
During training and testing the same experimenter and assistant were present. Two bowls (plastic tubs), both the same colour (white): 23.5cm in diameter and 9cm in height/depth functioned to hide the bait (piece of gravy bone).
The protocol described by Lakatos et al (2011) was followed and adapted to fit this study. During training the two bowls were placed in front of the experimenter, 1 metre apart. Subjects were positioned 2 metres away from the bowls where the subject could witness the hiding of the bait, which involved placing the bait into one of the two bowls. The subject was then released by the research assistant which allowed the subject to retrieve the treat from the bowl. This was repeated twice for each bowl to ensure that the subject knew that either bowl could contain food (Lakatos et al, 2011).
The position of subjects and bowls was the same as mentioned above, the main difference being that during testing the subject was prevented from observing the hiding of the bait. Subjects were positioned on a marker 2 metres from the two bowls set a metre apart. During training and testing subjects were restrained by the research assistant instead of the subject’s owner as described in Lakatos et al (2011). Owners of subjects were not present during testing to prevent them from influencing the subject’s behaviour which could lead to invalid results (Schmidjell et al, 2012). The research assistant restrained the subject gently by holding the free end of the lead that was attached to the subject. The experimenter picked up both bowls turned away from the subject and placed the food into one of the bowls. Both bowls were then returned to the allocated points at the same time and the experimenter stood 0.5 metres back from the middle line of both bowls to face the subject. It was essential that the experimenter stood equidistant from both bowls as Lakatos et al (2011) found that dogs showed a slight preference to choose the bowl that was closer to the experimenter. Therefore, if the experimenter stood closer to one bowl it could influence the subject’s decision and influence the results.
Once the subject’s attention was on the experimenter (sounds such as calling the subject’s name or clapping could be used to draw the dog’s attention) the visual cue was given. The pointing gesture (cue) used was momentary distal pointing: the experimenter pointed with an extended arm and index finger towards the correct location for 1-2 seconds before the arm was lowered beside the experimenter’s body. Once the arm was lowered the research assistant released the subject. To avoid the “Clever Hans Effect” the research assistant released the subject by letting go of the lead, which prevented unintentionally cueing the subject for instance through subliminally guiding the subject towards the correct direction (Schmidjell et al, 2012).
Subjects had up to 30 seconds to respond to the visual cue and make a choice (approach a bowl). If a subject did not react to the pointing gesture, it was repeated to a maximum of three times before the test was ended, resulting in the subject failing the test and receiving the full-time score of 30 seconds. Subjects could choose only one bowl, the bowl not chosen was removed as soon as the subject approached the other bowl. If the baited bowl was chosen, the subject could eat the contents. Test sessions consisted of 5 trials. The pointed side was presented in a pre-determined random order, to prevent bias towards one side a random letter generator app was utilised which offered 5 randomly allocated sides for each subject (Random Letter Sequence Generator, 2017). Food was not hidden in the same bowl more than twice in a row to avoid side bias (Szetei et al, 2003). Latency was measured in seconds from the dog’s release to choice and the number of times the dog went to the indicated side was measured.
Figure 3: Experimental set up for the discriminative cues test.
2.6 Ethical Note
Prior to commencement of the study, all participants were provided with a consent form that outlined details of the tests involved. Participants were asked to read and sign the consent form if they were happy to give permission to involve their dog in the study. All data provided by participants was held in accordance with the Data Protection Act (1998). All participants were able to remove their dog from the study at any point, up until the point of data analysis.
To minimise subjects becoming stressed and their welfare being impacted negatively, the following procedures were put in place. Subjects were familiarised with the experimental room prior to testing, the same handler and experimenter were involved, and the equipment used was safe and appropriate for small and large canine breeds. Stress was also monitored and upon any indication that a subject was becoming stressed (yawning, whining, trembling, constant head turning) the subject was either removed immediately from the experiment or put on a break before any further testing (Mariti et al, 2012). During two of the three tests food was involved and could potentially be consumed by the dogs if they completed a task successfully. As nutritional requirements vary from dog to dog, owners were provided with information regarding the food that would be involved. This ensured that the welfare of dogs with specific health or dietary requirements was not impacted. Water bowls were also provided within the experimental room to prevent subjects becoming dehydrated which could lead to dogs becoming anxious, potentially affecting their performance and welfare (Hatch, 2007).
2.7 Data Analysis
Data was statistically analysed using the SPSS statistical software package (IBM SPSS statistics for Windows, version 24.0. Armonk, NY: IBM Corp). To discover whether the data was parametric or nonparametric the Shapiro-Wilk test was run, revealing that all data was nonparametric. Therefore, the Mann-Whitney U test was employed. This test is suitable for running data on small sample sizes and compares the differences between two independent groups (Nachar, 2008). Mann-Whitney U was used to determine whether there was a significant difference between the two canine breed groups and the variables “time” taken to complete each of the three cognitive tests and “correct choices”, the number of correct choices made during the quantity discrimination and discriminative cues tests.
The Mann-Whitney U test revealed that there was a slight difference found between the two breed groups and the mean ranks for the object permanence test. To investigate this further, dogs were placed into one of two groups; those dogs which did not complete the test at all and those which completed the test at least once successfully. The Fishers exact test was employed to determine whether there was a significant difference between the two breed groups and whether they completed the object permanence test. This test was chosen due to the small sample size (N=18).
Mann-Whitney U test revealed that there was no significant difference between the two dog breed groups and time taken to complete the object permanence test (U=22, N1=12, N2=6, P=0.213; Figure 4). However, the mean rank was slightly higher for the toy dog group (M=11.83, IQR=43.20) in comparison to the working dog group (M=8.33, IQR=174.16). Fishers exact test revealed that there was no significant difference found between working and toy dog breeds and whether they completed the object permanence test (P=0.316; Table 2).
Figure 4: Comparison of the time taken to complete the object permanence test, (U=22, N1=12, N2=6, P=0.213).
Table 2: Results from the Fishers Exact Test: Comparison of successfully completed tests between working and toy dogs (P=0.316).
Mann-Whitney U test revealed that there was no significant difference found between the two dog breed groups and time taken to complete the discriminative cues test (U=32, N1=12, N2=6, P=0.75; Figure 5). There was no significant difference found between working dogs and toy dogs and the number of correct choices during the discriminative cues test (U=33.5, N1=12, N2=6, P=0.820; Figure 6).
Figure 5: Comparison of time taken to complete the discriminative cues test, (U=32, N1=12, N2=6, P=0.75).
Figure 6: Comparison between the two breed groups and the number of correct choices during the discriminative cues test, (U=33.5, N1=12, N2=6, P=0.820).
Mann-Whitney U test revealed that there was no significant difference between the two dog breed groups and time taken to complete the quantity discrimination test (U=31, N1=12, N2=6, P=0.682; Figure 7). No significant difference was found between the two canine breed groups and the number of correct choices during the quantity discrimination test (U=30.5, N1=12, N2=6, P=0.616; Figure 8).
Figure 7: Comparison of time taken to complete the quantity discrimination test, (U=31, N1=12, N2=6, P=0.682).
Figure 8: Comparison between the two breed groups and the number of correct choices during the quantity discrimination test, (U=30.5, N1=12, N2=6, P=0.616).
Behavioural problems are often the main cause for the breakdown of canine-owner relationships and have also been found to increase the risk of dog relinquishment (Case, 2013; Shore, Burdsal and Douglas, 2008). Providing adequate stimulation can help prevent or reduce undesirable behaviours, whilst increasing desirable behaviours and improving canine welfare (Herron, Kirby-Madden and Lord, 2014; Hubrecht, 1993). The aim of the current study was to discover whether cognitive abilities differ between the toy dog group and the working dog group. To investigate this, subjects completed three cognitive tests and the number of correct choices made and latencies to complete each test were measured. The results from the current study suggests that there was no significant difference between the two canine breed groups and cognitive performance.
From visual inspection of the graphs, it initially appears that the toy group took longer to complete the object permanence test and the discriminative cues test in comparison to the working dog group. However, this was not significantly different. The current study did not find a significant difference between the two dog breed groups and time taken to complete the three cognitive tests. Additionally, no significant difference was found between the two dog breed groups and the number of correct choices made during the quantity discrimination and discriminative cues tests. These results suggest that the working group and toy group have similar cognitive abilities. Therefore, these results do not support the hypothesis that there would be a significant difference between the working group and toy group in terms of their cognitive abilities. Furthermore, these results indicate that both breed groups can discriminate between two quantities, use human gestural cues to locate hidden items and understand object permanence to a similar level.
A potential explanation as to why the toy and working group completed all three cognitive tests within similar latencies and with a similar rate of correct choices is artificial selection. Artificial selection has been ongoing for approximately 14,000 years, which has given rise to a great variety of canine breeds (Turcsan, Kubinyi and Miklosi, 2011). Dogs have been selected for their behaviour, physical characteristics and perhaps even intelligence or trainability (Turcsán, Kubinyi and Miklósi, 2011). Toy breeds such as the Cavalier King Charles Spaniel have been recognised since the 16th century and are valued for their companionship (Serpell and Duffy, 2014; Coren, 2006). The results from the current study indicate that working and toy dogs have retained similar cognitive abilities. Intelligence may therefore be a trait which is valued by humans across all canine breeds, not only those selected for working roles. This might be a potential explanation as to why the toy group performed similarly to the working group in the current study. If intelligence was not appreciated by humans in toy breeds, it is expected that this trait would have been bred out.
King, Marston and Bennett (2009) conducted a questionnaire amongst Australian citizens to discover which characteristics were considered important in a dog. King, Marston and Bennett (2009) found that obedience was considered the most important characteristic. These findings support the theory that intelligence is valued in dogs regardless of breed group and supports our findings that working, and toy dogs have a similar level of intelligence. An obedient dog is one which is likely to be more trainable (King, Marston and Bennett, 2009). Therefore, an obedient dog may display fewer undesirable behaviours as owners may be able to teach their dog to behave appropriately (King, Marston and Bennett, 2009). This is likely to improve owner satisfaction and improve the human-canine relationship (King, Marston and Bennett, 2009). However, behavioural problems may still arise if a dog’s physical and mental requirements are not met (Landsberg, Hunthausen and Ackerman, 2012). According to the Kennel Club, the toy breeds involved within the current study, except for the Bichon Frise are described as intelligent and alert (The Kennel Club, 2009). These breed standards are clearly reflected in the current studies results, which show that toy dogs have similar cognitive abilities to working dogs. This is also consistent with breed rankings on perceived intelligence. Coren (2006) ranked the Papillion, a toy breed within the top ten breeds recognised as being the most intelligent and obedient. While Coren’s breed ranking is widely accepted, it is also criticised. Coren did not assess breed differences in trainability or performance but instead based rankings on the ratings of expert judges of trainability (Helton, 2010).
As the performance of the toy group and the working group was similar on all three cognitive tests, we would expect both breed groups to have similar levels of trainability. Serpell and Hsu (2005) investigated trainability between eleven common breeds. Data was collected from 1,563 canine owners who completed a canine behavioural assessment and research questionnaire (Serpell and Hsu, 2005). This study detected differences between breeds and trainability, revealing that toy breeds such as the Yorkshire Terrier ranked lowest in relation to trainability (Serpell and Hsu, 2005). Whereas working breeds such as the Labrador were ranked amongst the highest for trainability and therefore learning (Serpell and Hsu, 2005). The results from this study indicate that cognitive abilities do differ between working and toy breeds which is not in accordance with the current study’s findings. It must be noted that Serpell and Hsu (2005) relied upon self-rated measures and the opinions of dog owners to assess breed trainability which may have resulted in subjectivity, confounding the results. In comparison the results from the current study may be more reliable as an experimental design was employed. Each subject’s performance was assessed by measuring latency to complete each test and measuring the number of correct choices made during the quantity discrimination and discriminative cues tests. Therefore, the current studies results were not influenced by subjectivity as per self-reported measures.
However, a limitation of the current study that needs to be considered is the sample size utilised (N=18). This is a small sample size in comparison to existing studies that have investigated canine cognition (Arden and Adams, 2016; Ward and Smuts, 2007). There was also an overrepresentation of working dogs (N=12) in comparison to toy dogs (N=6) in the current study. Utilising a larger sample size with an equal representation of working and toy breeds may have resulted in a significant difference being found and would also allow the findings to be generalised. Furthermore, participants in the current study self-selected which may have resulted in a biased sample of only those participants interested in taking part or those who expected their dog to perform well. Consequently, the results may have been confounded and are unlikely to be representative of the wider population (Binfet and Passmore, 2016).
Though the current study suggests there to be no significant difference between the two breed groups and cognitive performance, few studies have found significant differences between breed groups and performance. Wobber et al (2009) conducted two experiments, the latter experiment investigated dogs’ (N=41) cue following abilities. Wobber et al (2009) found that working dogs (Shepherds and Huskies) historically bred for working purposes outperformed dogs typically bred for companionship such as the Toy Poodle and Basenji. This suggests that breeds traditionally bred to complete roles involving interactions with humans use human social cues more skilfully in comparison to non-working breeds bred for companionship. Although this study concluded that working breeds were more successful at using human gestural cues, the non-working group only consisted of two companion breeds, the Toy Poodle and Basenji, these two breeds alone cannot be used to represent all non-working breeds. Future studies should consider using multiple breeds per breed group alongside a larger sample size to strengthen their findings.
In comparison, Pongratz et al (2005) did not find any differences between dog breed groups and performance in a detour task. Although this study focussed on the breed groups; working, utility and hunting and therefore this is not completely relevant to the current study’s findings. The findings from this study indicate that cognitive abilities are independent of breed or breed group, suggesting that the findings from the current study are likely to be true (Pongratz et al, 2005). The differing results between these two studies might be due to the differences in their methodologies. Pongratz et al (2005) investigated cognitive abilities using a detour task conducted outside, while Wobber et al (2009) used a cue following task conducted within a novel indoor room. Whilst a vast amount of studies have explored canine cognition, many of these have only involved a single cognitive test to measure cognition unlike the current study. The current study involved three cognitive tests to assess cognitive abilities, increasing the reliability of the results. As the current study did not find a significant difference between the two breed groups and cognitive abilities across all three-cognitive tests, this indicates that the results found are likely to be valid.
Differently from toy breeds, working breeds have been selectively bred to optimise their performance in their working role (Turcsan, Kubinyi and Miklosi, 2011). This may have involved selecting for attentiveness or behavioural skills such as stalking in herding dogs (Turcsan, Kubinyi and Miklosi, 2011; Gacsi et al, 2009). Therefore, it would be expected that working breeds would outperform toy breeds in their cognitive abilities. However, the current study indicates otherwise, suggesting that toy breeds have similar cognitive abilities to working breeds as both group completed the cognitive tests within similar latencies and correct choices. These findings might be explained by selection pressures. Selection pressures have resulted in phenotypically different breeds, such as brachycephalic breeds (breeds with short noses) and dolichocephalic breeds (breeds with long noses) (Gacsi et al, 2009). Brachycephalic breeds such as the Cavalier King Charles tend to have more frontally positioned eyes, whereas dolichocephalic breeds such as sheepdogs tend to have more laterally located eyes (Gacsi et al, 2009). Gacsi et al (2009) recently demonstrated that brachycephalic breeds were more successful at using human gestural cues to locate hidden food items in comparison to dolichocephalic breeds. These results highlight the possibility that through selective breeding which has resulted in forward facing eyes, this may have also resulted in improved sensitivity or attentiveness (Helton and Helton, 2010). This is likely to improve or aid a dog’s performance during tasks, including those utilised in the current study. This theory has also been proposed by a few studies, suggesting that attention span can affect learning abilities (Gacsi et al, 2009; Range et al, 2008).
Factors such as training may have affected the results of the current study as this was not controlled for. Therefore, the effects of training cannot be ruled out. Marshall-Pescini et al (2008) discovered that training experiences affects the performance of dogs during problem solving. In comparison to untrained dogs (those with basic or no training), dogs which have received higher levels of training were more successful in problem solving tasks (Marshall-Pescini et al, 2008). This suggests that cognitive abilities are likely to be influenced by training experiences. However, Marshall-Pescini et al (2008) failed to put measures in place to minimise the effects of owner presence on a dog’s performance. Many studies have shown that owners can influence a dog’s decision (Schmidjell et al, 2012; Prato-Previde, Marshall-Pescini and Valsecchi, 2008). Although training experience was not controlled for in the current study, age was controlled. Nagasawa et al (2012) demonstrated that dogs over the age of ten have poorer cognitive performance in comparison to dogs younger than ten years. Therefore, dogs over the age of ten years were excluded from the current study, this ensured that a subject’s cognitive performance was not affected by age.
A further limitation of the current study is that hunger states of subjects were not considered during data collection. Assuming subjects differed in appetitive motivations due to being tested at various times throughout the day, our results which suggest that the two breed groups performed similarly may have been confounded. Future studies should consider asking owners not to feed their dogs for four hours prior testing. This might ensure that all dogs are highly motivated to complete the food orientated tests (Prato-Previde, Marshall-Pescini and Valsecchi, 2008).
Although the current study did not find a significant difference between the two canine breed groups and cognitive abilities, it would be beneficial to repeat this study with minor alterations. Future studies should involve a larger sample size to identify if the sample size used in the current study was too small to discover any significant differences. Arden and Adams (2016) conducted a study involving a sample size of 68 dogs. Future studies should aim to utilise larger sample sizes similar to that used by Arden and Adams (2016), this would enhance the reliability of the study’s findings. The current study only involved a selection of breeds to investigate breed group differences. Future studies should include all toy and working breeds to allow the findings to be generalised whilst strengthening the findings. Finally, since the current study found that the toy and working group could complete the three cognitive tests with similar latencies and rate of correct choices, future studies could implement cognitive tests that are more challenging to discover if there is a significant difference between the two breed groups and cognitive performance when using more advanced tasks. For instance, the current study tested for object permanence using single visible displacement, perhaps using invisible displacement or three successive displacements would have revealed a significant difference in cognitive abilities between the toy and working group (Gagnon and Dore, 1992).
The current study investigated whether there were differences between the toy dog group and the working dog group regarding their cognitive abilities. It was hypothesised that there would be a significant difference between the two dog breed groups and their cognitive abilities, due to the working group completing the cognitive tests more successfully. However, results from this study revealed no significant differences between the two dog breed groups and latency to complete the three cognitive tests or the number of correct choices made. Therefore, the current study suggests that dogs from the working dog group and the toy dog group have similar cognitive abilities. These findings may be useful in improving knowledge amongst dog owners. A better understanding of a dog’s cognitive abilities may allow owners to meet their dog’s mental needs more appropriately. This may lead to a reduced number of dogs surrendered into shelters because of broken or undeveloped canine-owner bonds due to behavioural problems developing.
The findings from the present study are in accordance with existing research which did not discover a difference between dog groups and cognitive performance. However, some studies have found differences between working and non-working breeds and their cognitive abilities. Due to this discordance between studies and their findings, more research is required to establish whether cognitive abilities are affected by breed group. Although the current study suggests there to be no difference between the two dog breed groups and their cognitive abilities, this study needs to be repeated with a larger sample size to confirm these findings, considering factors potentially associated with influencing cognitive performance such as training experience. Considering the current studies small sample size and the lack of supporting literature, the current study’s hypothesis cannot be accepted or rejected.
Arden, R. and Adams. M.J. (2016). A General Intelligence Factor in Dogs. Intelligence. [Online]. 55, pp. 79-85. [Accessed 4th September 2017].
Bensky, M.K., Gosling, S.D. and Sinn, D.L. (2013). The world from a dog’s point of view: a review and synthesis of dog cognition research. Advances in the Study of Behaviour. [Online]. 45, pp. 209-406. [Accessed 3rd February 2018].
Binfet, J.T. and Passmore, H.A. (2016). Hounds and homesickness: The effects of an animal-assisted therapeutic intervention for first-year university students. Anthrozoös. [Online]. 29(3), pp. 441-454. [Accessed 30th January 2018].
Broom, D.M. (1991). Animal welfare: concepts and measurement. Journal of Animal Science. [Online]. 69(10), pp. 4170. [Accessed 17th October 2017].
Broom, D.M. (2010). Cognitive ability and awareness in domestic animals and decisions about obligations to animals. Applied Animal Behaviour Science. [Online]. 126, pp. 1-29. [Accessed 16th October 2017].
Broom, D.M. and Fraser, A.F. (2015). Domestic Animal Behaviour and Welfare. 5th Edition. United Kingdom: CABI.
Case, L.P. (2013) The dog: Its behaviour, nutrition, and health. 2nd Edition. United States: Wiley: John & Sons.
Coren, S. (2006). The intelligence of Dogs: A Guide to the Thoughts, Emotions and Inner Lives of our Canine Companions. New York: Free Press.
Data Protection Act 1998. [Online]. Chapter 29. (1998). Data Protection Act 1998. Available from: https://www.legislation.gov.uk/ukpga/1998/29/contents. [Accessed 24th January 2018].
Gácsi, M., McGreevy, P., Kara, E. and Miklósi, Á. (2009). Effects of selection for cooperation and attention in dogs. Behavioural and Brain Functions. [Online]. 5(1), pp. 31. [Accessed 26th January 2018].
Gagnon, S. and Dore. F.Y. (1992). Search behaviour in various breeds of adult dogs: object permanence and olfactory cues. Journal of Comparative Psychology. [Online]. 106(1), pp. 58-68. [Accessed 4th September 2017].
Hatch, A. (2007). The view from all fours: a look at an animal-assisted activity program from the animals’ perspective. Anthrozoös. [Online]. 20(1), pp. 37-50. [Accessed 24th January 2018].
Helton, W.S. (2010). Does perceived trainability of dog (Canis lupus familiaris) breeds reflect differences in learning or differences in physical ability? Behavioural processes. [Online]. 83(3), pp. 315-323. [Accessed 21st January 2018].
Helton, W.S. and Helton, N.D. (2010). Physical size matters in the domestic dog’s (Canis lupus familiaris) ability to use human pointing cues. Behavioural processes. [Online]. 85(1), pp. 77-79. [Accessed 30th January 2018].
Herron, M.E., Kirby-Madden, T.M. and Lord, L.K. (2014). Effects of environmental enrichment on the behaviour of shelter dogs. Journal of the American Veterinary Medical Association. [Online]. 244(6), pp. 687-692. [Accessed 8th February 2018].
Howell, T.J., Toukhsati, S., Conduit, R. and Bennett, P. (2013). The perceptions of dog intelligence and cognitive skills (PoDIaCS) survey. Journal of Veterinary Behaviour: Clinical Applications and Research. [Online]. 8(6), pp. 418-424. [Accessed 4th February 2018].
Hubrecht, R.C. (1993). A comparison of social and environmental enrichment methods for laboratory housed dogs. Applied Animal Behaviour Science. [Online]. 37(4), pp. 345-361. [Accessed 9th February 2018].
King, T., Marston, L.C. and Bennett, P.C. (2009). Describing the ideal Australian companion dog. Applied Animal Behaviour Science. [Online]. 120(1), pp. 84-93. [Accessed 21st January 2018].
Lakatos, G., Gacsi, M., Topal, J. and Miklosi, A. (2011). Comprehension and utilisation of pointing gestures and gazing in dog–human communication in relatively complex situations. Animal Cognition. [Online]. 15(2), pp. 201-213. [Accessed 19th July 2017].
Landsberg, G.M., Hunthausen, W.L. and Ackerman, L.J. (2012). Handbook of Behaviour Problems of the Dog and Cat. 3rd Edition. Edinburgh: Saunders.
Ley, J.M., Bennett, P.C. and Coleman, G.J. (2009). A refinement and validation of the Monash Canine Personality Questionnaire (MCPQ). Applied Animal Behaviour Science. [Online]. 116(2), pp. 220-227. [Accessed 21st January 2018].
Mariti, C., Gazzano, A., Moore, J.L., Baragli, P., Chelli, L. and Sighieri, C. (2012). Perception of Dog’s Stress by their owners. Journal of Veterinary Behaviour: Clinical Applications and Research. [Online]. 7(4), pp. 213-219. [Accessed 20th September 2017].
Marshall-Pescini, S., Valsecchi, P., Petak, I., Accorsi, P.A. and Previde, E.P. (2008). Does training make you smarter? The effects of training on dogs’ performance (Canis familiaris) in a problem-solving task. Behavioural processes. [Online]. 78(3), pp. 449-454. [Accessed 23rd January 2018].
Meehan, C.L. and Mench, J.A. (2007). The challenge of challenge: can problem solving opportunities enhance animal welfare. Applied Animal Behaviour Science. [Online]. 102(3), pp. 246-261. [Accessed 16th October 2017].
Nachar, N. (2008). The Mann-Whitney U: A test for assessing whether two independent samples come from the same distribution. Tutorials in Quantitative Methods for Psychology. [Online]. 4(1), pp. 13-20. [Accessed 13th January 2018].
Nagasawa, M., Yatsuzuka, A., Mogi, K. and Kikusui, T. (2012) A new behavioural test for detecting decline of age-related cognitive ability in dogs. Journal of Veterinary Behaviour: Clinical Applications and Research. [Online]. 7(4), pp. 220-224. [Accessed 26th March 2017].
Newberry, R.C. (1995). Environmental enrichment: increasing the biological relevance of captive environments. Applied Animal Behaviour Science. [Online]. 44(2), pp. 229-243. [Accessed 10th February 2018].
Payne, E., Bennett, P.C. and McGreevy, P.D. (2015). Current perspectives on attachment and bonding in the dog–human dyad. Psychology Research and Behavior Management. [Online]. 8, pp. 71-79. [Accessed 17th October 2017].
PFMA. (2015) Pet Population 2017. Available from:
http://www.pfma.org.uk/pet-population-2017. [Accessed 10th October 2017].
Pongracz, P., Miklosi, A., Kubinyi, E., Gurobi, K., Topál, J. and Csányi, V. (2001). Social learning in dogs: the effect of a human demonstrator on the performance of dogs in a detour task. Animal Behaviour. [Online]. 62(6), pp. 1109-1117. [Accessed 17th January 2018].
Pongracz, P., Miklosi, A., Kubinyi, E., Topal, J. and Csanyi, V. (2003). Interaction between individual experience and social learning in dogs. Animal Behaviour. [Online]. 65(3), pp. 595-603. [Accessed 17th February 2018].
Pongrácz, P., Miklósi, Á., Vida, V. and Csányi, V. (2005). The pet dog’s ability for learning from a human demonstrator in a detour task is independent from the breed and age. Applied Animal Behaviour Science. [Online]. 90(3), pp. 309-323. [Accessed 29th January 2018].
Prato-Previde, E., Marshall-Pescini, S. and Valsecchi, P. (2008). Is your choice my choice? The owners’ effect on pet dogs’ (Canis lupus familiaris) performance in a food choice task. Animal Cognition. [Online]. 11(1), pp. 167-174. [Accessed 23rd January 2018].
Previde, E.P. and Valsecchi, P. (2004). The Bond that Never Developed: Adoption and Relinquishment of Dogs in a Rescue Shelter. Journal of Applied Animal Welfare Science. [Online]. 7(4), pp. 253-266. [Accessed 10th October 2017].
Random Letter Sequence Generator. (2017). Random Letter Sequence Generator. Available from: http://www.dave-reed.com/Nifty/randSeq.html. [Accessed 17th August 2017].
Range, F., Aust, U., Steurer, M. and Huber, L. (2008). Visual categorization of natural stimuli by domestic dogs. Animal Cognition. [Online]. 11(2), pp. 339-347. [Accessed 30th January 2018].
Schmidjell, T., Range, F., Huber, L. and Viranyi, Z. (2012). Do Owners Have a Clever Hans Effect on Dogs? Results of a Pointing Study. Frontiers in Psychology. [Online]. 3, pp. 558. [Accessed 4th September 2017].
Serpell, J.A. and Duffy, D.L. (2014). Dog breeds and their behaviour. Domestic dog cognition and behaviour. [Online]. pp. 31-57. [Accessed 19th January 2018].
Serpell, J.A. and Hsu, Y.A. (2005). Effects of breed, sex, and neuter status on trainability in dogs. Anthrozoös. [Online]. 18(3), pp. 196-207. [Accessed 27th January 2018].
Shore, E.R., Burdsal, C. and Douglas, D.K. (2008). Pet owners’ views of pet behaviour problems and willingness to consult experts for assistance. Journal of applied animal welfare science. [Online]. 11(1), pp. 63-73. [Accessed 4th February 2018].
Stephen, J. and Ledger, R. (2005). An Audit of Behavioural indicators or poor welfare in Kennelled Dogs in the United Kingdom. Journal of Applied Animal Welfare Science. [Online]. 8(2), pp. 79-95. [Accessed 17th October 2017].
Szetei, V., Miklósi, Á., Topál, J. and Csányi, V. (2003). When dogs seem to lose their nose: an investigation on the use of visual and olfactory cues in communicative context between dog and owner. Applied Animal Behaviour Science. [Online]. 83(2), pp. 141-152. [Accessed 24th January 2018].
The Kennel Club. (2009). Breed Information Centre. Available from: https://www.thekennelclub.org.uk/services/public/breed/standard.aspx?id=6157. [Accessed 19th January 2018].
Turcsán, B., Kubinyi, E. and Miklósi, Á. (2011). Trainability and boldness traits differ between dog breed clusters based on conventional breed categories and genetic relatedness. Applied Animal Behaviour Science. [Online]. 132(1), pp. 61-70. [Accessed 3rd February 2018].
Ward, C. and Smuts, B.B. (2007). Quantity-based judgments in the domestic dog (Canis lupus Familiaris). Animal cognition. [Online]. 10(1), pp. 71-80. [Accessed 26th January 2018].
Webster, J. (2011). Management and Welfare of Farm Animals: The UFAW Farm Handbook. 5th Edition. Library of Congress in cataloging-in-publication-data: Wiley Blackwell.
Westgarth, C., Pinchbeck, G.L., Bradshaw, J.W., Dawson, S., Gaskell, R.M. and Christley, R.M. (2007). Factors associated with dog ownership and contact with dogs in a UK community. BMC Veterinary Research. [Online]. 3(1), pp. 5. [Accessed 4th February 2018].
Wobber, V., Hare, B., Koler-Matznick, J., Wrangham, R. and Tomasello, M. (2009). Breed differences in domestic dogs (Canis familiaris) comprehension of human communicative signals. Interaction Studies. [Online]. 10(2), pp. 206-224. [Accessed 29th January 2018].