This study provides much novel and confirmatory information on sociodemographic factors associated with childhood ownership of different pet types, including those that may influence the nature of the relationship children have with their pets. Such data are likely to be useful to researchers in the fields of public health, social science and veterinary science, as well as those studying the field of human-animal interactions. We have demonstrated that children are keen to tell us all about their pets, regardless of pet type, and we suggest that in order to maximise compliance and understanding of the study, researchers should be present during data collection with the children where possible.
A strength of this study is that the sample of children was relatively large, not convenience-based, and had high response rates, due to the specific context of data collection where all children were captured for a time period set aside purely for this purpose. We also used multivariable regression modelling, to adjust for the confounding effects that demographic variables can have on each other, which are methods lacking from much previous HAI literature.
However, the data are limited by the nature of self-report, as we did not see the actual pets or the dog types ourselves for verification, nor did we ask the parents. The nature of what constitutes pet ‘ownership’ may also be questioned, and may differ between adults and children, but for the purposes of this study, both the children and investigators inferred it to mean living with a pet in the household in which they spent most of their time, or in the case of horses, the child feeling that the horse belonged to their household. We also did not ask whether the children were from dual or single-parent families, which may have been interesting in terms of attachment to pets. The data were collected from a specific population, 9–10 year-old children attending primary schools in a region of Liverpool that contains areas of considerable deprivation and, thus, may not be generalisable to other UK cities or countrywide, or other age groups. In the 2001 Census the Liverpool Local Authority Area was noted to have lower employment, fewer people with qualifications and lower rates of home ownership, when compared with the whole of England . The study was also cross-sectional and, therefore, causation cannot necessarily be implied.
Dogs are often thought to be the most common pet type owned by households in the UK, closely followed by cats [15, 16, 25]. However, in this study, dogs were much more commonly reported than cats, which may be a regional difference as cats are reported to be less popular in the north . It may also be due to the nature of this sample being households with children, rather than all household types. However, in other childhood pet ownership data from a UK sample in the 1990s, cats were the most common pet type . It is also noted that the age of the pet dogs reported was skewed towards younger age. This may be due to recall bias or guesswork in reporting of age of the dog by the children, older dogs in this population being more likely to be relinquished or abandoned, or most likely due to evidence that families with young children commonly do not acquire a dog or puppy until the child is older , meaning that by a child age 9–10 years the dog is still relatively young.
Girls were more likely to report owning pets than boys, for all pet types except rabbits. Previous research in the UK and Ireland suggests that females are more likely to own cats than males [15, 17, 18], and this relationship has also been suggested to apply to children in respect to cats, rabbits and rodents specifically but not other pets , or pets in general . In contrast, others found no difference in pet ownership by gender [11, 28]. Thus, there is good evidence that, during childhood, girls are more likely to own pets than boys, and it seems likely that this is a true phenomenon rather than girls just being more likely to tell us that they own a pet. However, there was no evidence that girls were more or less attached to their favourite pet than were the boys, which is similar to previous findings of no gender differences in frequency of play or care-giving related to pets owned by children , but contrasts with other studies, where females scored higher on pet attachment than did males [13, 29]. These contrasting findings could reflect differences in the populations studied or the tools used to measure the relationship/attachment.
Ethnicity may act as an individual factor, although it frequently forms complex relationships with the religion, history and culture of specific ethnic groups; a simple measure of ‘ethnicity’ may not be fully representative of the beliefs and behaviour within that group. Due to the sample size we were limited in the analysis of ethnicity that could be performed. However, here we report that children of white ethnicity were more likely than non-white children to own dogs, rodents and ‘other’ pets. In data from the USA, whites have previously been observed to be more likely to own pets in adolescents aged 12–17 , in university students  and in the 21-64yrs age group . We also found that white children were no more or less attached than non-white children to their favourite pet owned, disagreeing with previous findings which suggest higher white attachment [13, 30] or rating of importance of the pet .
These differences may be due to limited sample sizes, combining results from ethnic minorities into a category of ‘non-white’ or, alternatively, may represent children and their families who have been domiciled in the UK for a prolonged period, even generations, and thus have incorporated 'western' influences. The use of different measures of relationship/attachment in each study may also contribute to differences in findings. It has been suggested that attachment scales may bias towards whites by considering western attitudes as a 'baseline,' rather than considering that western attitudes may in fact be more 'positive' and other ethnic attitudes 'baseline' . More research into the potential cultural-bound phenomenon of the role of pets in the family is required , particularly in countries other than the USA. As some diseases or risk factors for disease are more common within specific ethnic groups, for example overweight and diabetes mellitus [31, 32], pet ownership has the potential to counteract risk factors in these groups, via health promoting behaviours such as increased physical activity  and/or emotional support .
Previous research suggests that children with younger siblings have fewer pets than those with no younger siblings or singletons ; the presence of an older sibling increases the likelihood of ownership of dogs, rodents, birds and fish being reported ; or pet ownership does not vary with sibling status [11, 28]. In our study, single children with no brothers and sisters were no more (or less) likely to own pets than those with younger brothers or sisters. We do, however, report novel and strong evidence that single children were more attached to their favourite pet, which concurs with the observation of Siegel (1995) that adolescents with no siblings in the household rated their pet as more important to them than those living with siblings. It has also been suggested previously, from parental report, that the youngest sibling plays more with a pet . We also provide some evidence to suggest that children who were the youngest of their siblings were more attached to their favourite pet, although no more or less likely to report owning a pet in the first place.
Thus, research is contradictory as to whether presence of siblings (and their comparative ages) affects frequency of pet ownership, but single children and possibly youngest siblings may have stronger attachment to, and interact more with, their pets, than children with younger brothers and sisters. This is contrary to the findings of McConnell et al. (2011) who suggested that the support from pets provided (to adults) complements rather than competes with other human resources of support; however, they also noted that this required further investigation in people who are more socially isolated, as may be the case here of children with no brothers or sisters to play with.
There was good evidence to suggest that children who owned dogs lived in more deprived areas than those without dogs, strengthened by the dose–response relationship observed: the more deprived the area, the more dogs were owned. To our knowledge, this has not been previously reported, and may be unique to this particular area, although the findings are compatible with the general observation that dog ownership decreases as education level or social class of the owners increases [15, 18, 19, 34]. Our data also suggested that ‘Pit Bull or cross’ or ‘Bull Breed’ dogs were more commonly found in more deprived areas, implying that ownership of these types of dogs has an inherent regional or cultural component linked to social deprivation. There was no evidence in our study that children who reported that their favourite dog was a Bull Breed were any more or less attached to the dog than those with a non-Bull-Breed dog. This fits with the observations of Maher and Pierpoint (2011) that so-called ‘status’ or ‘weapon’ dogs play a role of companionship, socialisation and protection in youth gangs in deprived inner city areas. There was also some suggestion from our data that non-white children were more likely to report owning a Pit Bull type (but not broader Bull Breeds) than white children. This may be a reflection of actual ownership of preferred dog types by different ethnic groups, or it could be due to non-whites being less inhibited in reporting that their dog is a Pit Bull, due to social and cultural reasons.
Two percent of all children sampled reported living with a ‘Pit Bull or cross’, and ten percent were living with similar ‘Bull Breed’ dogs including Staffordshire Bull Terriers, the most popular breed reported. Reported Pit Bull type dogs were surprisingly common, considering this is an illegal breed in the UK (Dangerous Dogs Act 1991), although not so surprising considering anecdotal knowledge of this region. Dog bites in this region have been reported in recent media, including fatalities in children, and are often attributed to Pit Bull and ‘status’ type breeds [35–37]. No data were collected on dog bites, and there was no specific mention of this information during conversation. Our data suggest that, if a considerable proportion of children are living with types of dogs that are often deemed to be ‘dangerous’ , the question can be raised as to why even more aggressive incidents are not reported.
Dogs were the pet that children most frequently reported owning, contacting outside their home, and were also the most common ‘favourite pet’ owned and scored the strongest attachment. This is likely to be due at least partly to real relationship differences between children and dogs compared to other pet types; Siegel (1995) also observed that fish owners felt that their pet was less important to them than dog or cat owners. However, it is also likely due to the nature of the questions asked in the CENSHARE Pet Attachment Scale, it was impossible to score highly when answering questions about a pet fish, as they are very unlikely to sleep on the bed or require daily grooming, and this was something that a number of children were observed to be frustrated by when completing this part of the questionnaire. This is an issue that should be addressed in future studies of this nature.