A framework for understanding how activities associated with dog ownership relate to human well-being

There is notorious inconsistency regarding mental health benefits of dog ownership, partially due to repeated cross-sectional studies comparing dog owners and non-owners, without taking into account the heterogeneity of dog-owner dyads, especially the activities with which the owners are involved. This study aimed to develop a comprehensive framework of the most important dog human related activities and their impact on owner well-being. Six focus groups with 35 dog owners were conducted, and their audio transcripts thematically analysed. Dog human related activities and themes of activities were linked to their reported changes in well-being through matrix coding. A framework of 58 dog human related activities linked with their specific hedonic well-being, life satisfaction and eudaimonic well-being outcomes was generated. Most activities were reported to improve owner’s well-being, (e.g. human–dog tactile interaction increases owner’s self-esteem), and a minority was mainly associated with negative outcomes. The richness of the framework presented in this study reinforces the importance of assessing dog ownership well-being outcomes based on specific dog human related activities with which dog owners are involved. This new and systematic investigative approach should decrease inconsistencies in the field and facilitate mental health interventions and study designs of a higher level of evidence.

Eighteen well-being outcome themes also emerged for the framework (Fig. 2). Aspects of hedonia most frequently associated with dog human related activities were an increase in 'positive valence-high arousal' states (195 mentions, 22.9%), such as excitement, happiness, and an increase in 'positive-valence-low arousal' states Being followed by the dog Being greeted by dog Dog waiƟng for owner to do something Dog's presence in the house 10. Shared acƟvity outside the house Dog's company at work Dog's presence while outside the house Pet therapy with own dog Trip with dog 11. Social interacƟons   , kissed by dog  Cuddle, snuggle with dog  Pet the dog  Sleep, lie on bed with dog  TacƟle interacƟons (nonspecific)    www.nature.com/scientificreports/ (136, 16.0%), e.g. calmness, relaxation. For eudaimonia, 'positive relations with others' (93, 10.9%), such as social interactions, and 'purpose in life' (57, 6.7%) prevailed. Details of all well-being outcomes and the dog human related activities reported to lead to them are given in the heat map in Fig. 2. The darker the cell in the map, the higher the percentage of mentions (up to ≥ 30%) of the relevant activity within the well-being outcome column.

Teaching or learning
'negative valence-high arousal' and dog human related activities. In this aspect of hedonic wellbeing negative feelings such as 'annoyed' , 'angry' , 'stressed' , 'worried' , and 'frustrated' emerged. Increase in feelings of this aspect were mainly occasioned by unwanted dog behaviours, such as barking; but also more benign social interactions with dog/people, especially negative encounters while out with their own dog. For example, participant 15: It is quite annoying when people don't ask if a dog is friendly first, you shouldn't just go straight and stroke the dog, you shouldn't be touching a dog if he has an owner.
Also, teaching or learning dog-related things, particularly training the dog; and non-specific ownership routines, such as looking after a puppy, had a negative impact on this aspect of well-being. Some activities, however, were reported as protective against these negative feelings, especially shared activities with the dog outside and inside the house, simply due to the dog's presence/company in these contexts. Tactile interactions with the dog, such as cuddling and sleeping together, and exercises with the dog, notably walking, also played a key role in decreasing feelings of negative valence-high arousal. For example, participant 4: It is very hot in [name of the country] so I jog at night, like at midnight, so I take her with me, it makes me feel protected, it is quite nice. She gives me company, like someone is with me, so I don't have to be afraid, it reduces my fear.
'negative valence-low arousal' and dog human related activities. This aspect of hedonic wellbeing included states such as 'sad' , 'tired' , 'unhappy' , 'lonely' , 'depressed' . Elements of this aspect increased in dog owners mostly when they felt they failed to meet their dog's expectations/needs, especially for leaving the dog alone/behind, and when their dogs performed an unwanted behaviour, notably the destruction of objects. For example, participant 25 said: If it [destruction of object] is in the middle of the day it is very tiring, it is not even anger, you are exhausted.
Additionally, having an old dog and coping with the end of life of the animal, such as euthanasia, was associated with negative feelings within this aspect. Negative emotions of low arousal were nevertheless improved by tactile interactions with the dog, such as cuddling, and by sharing activities with the dog both inside the house, especially by being greeted by the animal, and outside the house, in particular, the dog's presence/company. Also, exercising together, mainly walking, was frequently reported to decrease these negative feelings. For example participant 16, explained: It was difficult at times to walk him, but I was determined, me and him. For me, walking is on the higher top of my list, it helped in my recovery [from depression] 'positive valence-high arousal' and dog human related activities. This includes states such as 'happiness' , 'joy' , 'fun' , 'excitement' , 'activation' . Owners reported improvements in feelings of this aspect from exercising with their dogs, especially walking; from watching their dog, mainly when the dog is happy in a situation; and from teaching the dog, such as training exercises. For example, participant 7: It makes me happy to see how happy she [the dog] is, how excited she is.
Positive feelings of high arousal also arose from tactile interactions with the dog, such as cuddling and lying next to each other, playing with the dog, e.g. tug and ball games; and shared activities inside the house, notably being greeted by the dog. No activity was identified to directly decrease this element of well-being.
'positive valence-low arousal' and dog human related activities. In this aspect of well-being, feelings of calmness, relaxation, peace, and love emerged. They reported growth in feelings of 'positive valence-low arousal' when they were involved in tactile interactions with their dogs, such as cuddling, sleeping together and lying next to each other. For example, participant 1: Snuggling with the dog, on the sofa or on the bed, they just come up and flop on you, you can just relax and forget how busy it is and things like that, it is really calming.
Additionally, exercising with the dog, mainly walking, and sharing activities in the house, notably being greeted by the dog, were also associated with increases in feelings of this aspect. No activity was reported to directly decrease this element of well-being.
The results relating to dog human related activities effects on the valence-arousal (hedonic well-being) are summarised in Fig 'Autonomy' and dog human related activities. In this element of eudaimonic well-being, 'non-specific ownership routines' , such as being a dog owner and being able to look after an animal were frequently mentioned by owners as activities that make them feel more autonomous. For example, participant 8: As a pet parent I do feel protective of him and I have that sense that I can go out and take care of him, so I feel independent.
Increases in this element were also associated with shared activities with the dog (outside and inside the house), especially having the dog's presence/company; and with exercises with the dog, such as walking, running, hiking. In contrast, owners' autonomy decreased when they struggled to leave their dog behind/alone. For example, participant 5: Sometimes dogs restrict that [autonomy, independence], you don't always have the opportunity to take dogs somewhere, we have a family in [another country far away] and to go and travel there means you have to leave the dog somewhere, and sometimes there is no place to leave the dog. www.nature.com/scientificreports/ 'environmental mastery' and dog human related activities. In this element, owners reported their ability to fit or cope with their surrounding environment, including the performance of tasks relating to daily living. Improvements in environmental mastery were associated with dog ownership and looking after the animal on a regular basis, which is part of the theme 'non-specific ownership routines' . Also, performing exercises with the dog, mainly walking, and providing for the animal, particularly feeding it, were reported to increase the sense of environmental mastery of owners. For example, participant 10: It [feeding the dog] gives me something to do, otherwise I would skip a few meals. They have their own dinner, they have their own routines so I need to keep my routines as well. They have routine in the morning, it is good, it reminds me of doing things. One of the dogs has tablets, so it reminds me of my tablets as well.
However, owners also mentioned that having an old dog and giving support to this old animal hindered their own environmental mastery. For example, participant 31: The negative side of dog ownership is when they get old, you need to adjust your life to help your dog, […] you need to adjust loads of your routine.
'personal growth' and dog human related activities. Personal growth was reported as a sense of selfachievement in particular aspects of the owner's life (or life as a whole). The activities that mainly contributed to this element of well-being were being a dog owner and being capable of looking after a dog. Also, teaching and learning were associated with personal growth, especially learning dog-related things, such as dog behaviour and dog travel regulations, and training the dog successfully, e.g. to perform a new skill. For example, participant 35: He is the first dog I have had as an adult so I had to learn something about dog behaviour, how to look after them, about visas, how to get a [nationality of dog] dog into this country, that helped my personal growth.
No activity was reported to decrease this element of well-being.
'purpose in life' and dog human related activities. Having a dog, looking after the animal and having a routine because of the dogs' needs were associated with improvements in their purpose in life. Additionally, exercising with the dog, primarily through walks, increased this element as well, as did some non-specific activities; for example participant 23: Because I got a dog I got quite a good routine now. I need to make sure she walks, she is fed, all the needs are met, so that gives me purpose in life.
Some activities, however, decreased or hindered owners' sense of purpose in life, such as the responsibilities linked to dog ownership and not being able to leave the dog behind/alone. For example, participant 33: If I didn't have my dogs I would travel more and be more free, I would do more things, it is more negative to purpose in life, not autonomy, because I think in a big picture.
'positive relation with others' and dog human related activities. Dog owners reported changes in their social relations and empathy for others. Increases were attributed to having more interactions with other people, especially when out with the dog, and explicitly meeting other dog lovers on purpose. Exercising with the dog, particularly walking, and being accompanied by the dog while outside the house improved owners' relations with others as well. For example, participant 20: When you go out for a walk you meet different people, it seems okay to talk to them because you have a dog, the dog is an introducer. If you are just walking by yourself it is different, the dog is the connector.
Several owners also said that simply having a dog was beneficial to their social relations. In contrast, contact with others was also detrimental to their social relations, especially when having disruptive encounters, such as inappropriate interactions from others towards their dogs. Unwanted behaviours, particularly those related to aggression, such as when their dogs bark, bite or lunge towards others, were also detrimental to their social relations. For example, participant 25: Sometimes it is annoying meeting other people. Yesterday, for example, there was a guy sitting with three dogs, and then a woman approached and tried to pet them, then all dogs started barking, and the guy was trying to calm the three dogs. Sometimes I think 'please don't interact with my dogs, just leave me' 'Self-acceptance' and dog human related activities. Self-acceptance included owners' self-esteem and acceptance of their good and bad qualities. This element improved through the simple fact of having a dog and being able to look after the animal successfully, as well as due to activities shared with the dog in the house, particularly being greeted by the dog. For example, participant 2: Being licked, being cuddled, greeted [by dog] helps with self-esteem and self-acceptance because you can feel that you are loved, like when you feel sad, they pick up that and they try to cheer you up, it helps you accept who you are. www.nature.com/scientificreports/ Tactile interactions with the dog, such as cuddling and being licked/kissed by the dog, and teaching or learning dog related things, such as dog training and learning about dog behaviour, had a beneficial impact on this element as well. By contrast, owners' self-acceptance was hindered by dog's unwanted behaviours, especially due to the feeling of not being able to control their animal's behaviour, such as when the dog was acting aggressively or getting a recall when it was needed. For example, participant 27: On training she is great but on public she won't come back to me when I call her, then suddenly she starts to bark to someone with no reason, and I can't control her, it makes me feel completely useless.
Training the dog and looking after a puppy also decreased owner's self-esteem, particularly due to the frustration they felt for not being able to train a specific skill or manage their puppy.
These results relating to the themes of dog human related activities associated with eudaimonic well-being are summarised in Fig. 4.

Discussion
The recurrent inconsistencies observed across studies investigating dog ownership and human well-being, without taking into consideration the individuality of each dog-human relationship, drove the use of a new investigative approach in this study. A framework based on the full spectrum of activities/events dog owners could be involved with, due to the existence of a dog in their lives (defined by us as dog human related activities -see Introduction) was created. The use of the activities (or themes of activities) of our framework in future investigative studies is likely to increase consistency across studies, as patterns of well-being outcomes are expected to be observed for each particular identified activity. We believe this is the first systematic attempt to identify the most important specific dog human related activities that have an impact on the perceived well-being of owners. The framework is comprised of 58 dog human related activities and 15 themes of activities, which emerged as the most important dog human related events to dog owners' hedonic well-being, life satisfaction and eudaimonic well-being. Importantly we considered not only positive but also negative impacts, since the balance between these may be critical in the final well-being outcome.
As might be expected, the majority of dog human related activities were associated with positive well-being outcomes but three important groups of activities had a predominantly negative impact on owners' well-being: (1) aging and end of dog's life, such as euthanasia, (2) failure to meet dog's needs/expectations, such as leaving the dog alone at home, and (3) unwanted behaviours, such as problematic tendencies and poor obedience. The richness of the framework revealed in this study reinforces the problem with oversimplified dichotic comparisons www.nature.com/scientificreports/ based purely on ownership or not frequently observed in studies of the "pet effect" 21,37 . Our data support the need for better mapping of activity to outcome: for example, being greeted by a dog, physical interactions with the animal and dog training, were repeatedly referred to as boosters for owner's self-acceptance, by contrast walking the dog had a stronger impact on owner's social interactions, autonomy and purpose in life. It should also be noted that having to leave the dog alone at home, which may be related to owner work status, or looking after an old dog was related to important negative feelings of low arousal, such as sadness, and to decrease in autonomy or environmental mastery. In contrast, the occurrence of unwanted behaviours and obedience issues were predominantly linked with negative feelings of high arousal, such as anger, and worsening on positive relation with others and self-acceptance. It is not surprising that the most frequently reported activities to impact on well-being in this study, have also been reported to be important in previous investigations. Physical interactions with dogs, for example, have been shown to decrease stress [38][39][40][41] , anxiety [42][43][44] and fear 45 , and increase feelings of positive valence 15 . Being greeted by the dog is suggested to provide comfort and positive feelings for owners 46 . Dog walking often catalyses owners' social interactions 15,47 , improves owner's emotional state 24 and gives them purpose in life 15 . In contrast, unwanted behaviours may hinder owners' social interactions 46 and generate feelings of negative valence 48 . The loss of a pet 46 and the care of an old animal 49 have also been linked to negative feelings of low arousal, such as sadness and depression, as has a sense of failure to fulfil the dog's expectations, e.g. when owners fail to walk them 26 or leave them alone at home. However, our study extends this list and systematically integrates this more comprehensive inventory of activities into a coherent framework for the first time (Figs. 2, 3 and 4).
Changes in affect (hedonia) were mainly reported to be caused by specific events, such as playing, training, being greeted by the dog, while changes in eudaimonic well-being (eudaimonia) were mainly linked to more general events, such as being a dog owner and looking after the dog. This difference is not surprising since hedonia (represented by affect) amounts to the pursuit of pleasure through the experience of discrete life events, responsible for changes in feelings 32,50 ; by contrast, eudaimonia, represents a lifestyle, ("eudaimonic living") 51 , in which intrinsic values and long-enduring outputs are aimed for 36 , making eudaimonia more stable over time than hedonic well-being 33,52 . Thus, it seems likely that either a set of activities or a high intensity/frequency/duration of activities is required to yield transformations in eudaimonia, rather than the experience of a single event. For example, walking a dog once a year may suffice to increase one's momentary happiness and be identified as an activity that improves this affect. However, this same event may not be enough to change the meaning or quality of one's life (aspects of eudaimonia).
Changes in life satisfaction, one of the components of subjective well-being, was not linked to dog human related activities by the majority of participants. This might be a product of the study design, which required participants to focus on the four most important dog human related activities for their affect and/or life satisfaction in the same exercise. This might have encouraged them to focus more on specific events and thus subjective feelings. Another possible limitation is the lack of males in the sample, even though an effort was made to increase their representation in the selection phase. Women are believed to have higher psychological openness than men 53,54 , thus, future studies may need to minimise the disclosure of information about the nature of the research to recruit more men. Nonetheless, the activities and changes in well-being reported by male participants in the focus group sessions were not noticeably different to those reported by their female counterparts, and sessions were continued until redundancy of the information provided by all participants was achieved, but this does not negate the possibility of a selection bias limiting the information provided.
To increase the consistency across future studies, we recommend the selection of individual activities provided in our framework or combinations of them to test associations with the potential well-being outcomes as identified here. For example, testing if a higher frequency (or duration) of tactile interactions with dogs is positively associated with self-esteem. However, if our list of activities does not mirror a specific targeted population (e.g. dog owners with physical disability; dog owners who live on farms), new specific activities (or themes) could be added to the existing framework to make it even more comprehensive or a new framework specific for this group could be created based on the specific dog human related activities of interest. By assessing individual activities, rather than dog ownership per se (or even dog-assisted intervention per se), future investigations are likely to identify patterns of increase or decrease in particular well-being outcomes in relation to specific activities, especially if representative sampling of their targeted population and control for potential mediators (e.g. demographics, closeness to pet) are applied. Thus, it might be evidenced that dog ownership per se is not the key to change elements of well-being but dog-related activities.
Future studies by the authors will quantitatively assess the relationship between dog human related activities and changes in well-being identified here, through a survey, as a prelude to more comprehensive investigations of causality. It should also be noted that there may be cultural differences, particularly between countries, and so these results should not be thought to be globally comprehensive. Likewise, the relationship with other pets could be explored, since the activities reported are likely to differ from one species to another. Even though our framework has a good representability of dog owners (different ages, genders, dog-related expertise, dog sizes, dog ages and dog ownership duration) and saturation of themes has been reached in our thematic analysis, the framework cannot encompass all possible activities existent in all dog-owner dyads. More activities are expected to be found in studies targeting specific types of dog owners, i.e. those who have a particular goal associated with their relationship, such as a form of work or a specific hobby or interest. Any new activities might be rare at a general population level and/or not likely to change the well-being of dog owners at a population level. For example, even though some owners might be happy while dressing their dogs with different clothes and that might be a very important activity in their lives, this activity is not likely to have a strong effect on the well-being at a population level but could still be added into our framework to make it even more complete.
The creation of a definition for dog human related activities and the framework of activities and their wellbeing outcomes provided in this qualitative study provides a more robust basis for future research in the field Scientific RepoRtS | (2020) 10:11363 | https://doi.org/10.1038/s41598-020-68446-9 www.nature.com/scientificreports/ examining the mental health of dog owners (or even in the field of dog-assisted interventions), which should help to resolve current inconsistencies. This new, more specific approach opens the door to the investigation of changes in human well-being which acknowledges the individuality of each dog-owner dyad. This framework also provides the foundation for the development of more robust study designs and treatment plans for patients with mental health issues, which might be helped by dog-assisted interventions or dog owners simply hoping to improve their mental health.

participants.
Initially, convenience and voluntary samplings were used to recruit self-identified dog owners via the University of Lincoln's Petscando database (volunteer owners), social media, in person leaflet distribution and leaflets placed at pet shops, food markets and buildings of the university. 91 dog owners filled a recruitment form to volunteer to the study. Some of them were selected through purpose sampling to generate greater sample diversity with regards to owners' age, gender, dog-related expertise or not (e.g. dog trainer, veterinarian of small animals, volunteer in dog-assisted interventions), size of their dogs, age of their dogs and duration of the dog ownership. Diagnosable mental health problems were not directly investigated in the study (e.g. as a selection criterium), owners with and without diagnosable mental health conditions were welcome to participate and, during the focus group, they were free to disclose as much or as little of their mental health status as they wanted.
The population was thus mental health stigma-free, and accordingly honest and unbiased by such issues. Based on the 91 registrations, a total of 45 participants were selected. Ten of them were not included as they either did not attend or arrived too late for the focus group meeting. The 35 dog owners who comprised the final sample, were aged from 18-24 years to 65-74 years old (median 25-34 years), 26 were female (74.3%), 15 owners had some dog-related expertise (42.9%), eight owners had small dogs (23%), 14 medium dogs (40%), 13 large dogs (37%), their dogs' age ranged from a few months old to up to more than 15 years old (median 1-3 years old) and dog ownership lasted from a few months to more than 15 years (median 1-3 years). Finally, allocation to one focus group or another was performed according to participants' time availability.
pilot study and focus group sessions. Focus groups consist of an organised discussion with a selected group of people to gather information on a topic 55 , and are recommended for exploratory research 56 , such as to the development of theoretical frameworks 57 , justifying their use in this work. This study was approved by the ethical review committee at the University of Lincoln (reference 2019-Jul-0503), and all methods were carried out in accordance with the university Research Ethics Policy and with the BPS Code of Ethics and Conduct. Written informed consent was obtained from all participants. A pilot session, with four volunteers, was conducted to assess the quality of the moderator guide and time management before the main focus groups. All focus groups took place at the University of Lincoln, UK and were focused on dog human related activities and their impact on well-being. They were moderated by the first author, who had been trained in focus group moderation and supported by an assistant. Two audio devices were used simultaneously for recording: Homder Digital Audio Recorder and iPhone 8 (Apple). In line with Guest et al. 58 , which reports that 90% of themes in focus groups are usually discovered within three to six sessions, a total of six sessions (besides the pilot) were performed in this study, and saturation of themes was achieved. Each of the 35 participants could take part in one session only. Group sizes ranged from four to eight dog owners, and session duration varied from 80-103 min (mean 91 min).
Based on a pre-defined semi-structured moderator guide, dog owners were first introduced to the concepts of dog human related activities, hedonic well-being, life satisfaction and eudaimonic well-being in a similar way they have been defined in the Introduction of this paper but with simpler words (e.g. instead of using the word affect in hedonic well-being, the moderator referred to it as emotions and moods) and through a PowerPoint presentation aimed to facilitate the comprehension of the concepts. Examples of dog human related activities during this explanation were minimal so as not to influence owners' replies in the next step.
Second, owners were asked about the four most important activities for their own subjective well-being (positive and negative affect plus life satisfaction), and the four most crucial activities linked to their eudaimonic well-being. In order to prevent inter-participant bias and to generate a greater diversity of activities, dog owners were initially instructed to individually write these activities on a piece of paper provided. Participants were free to write more than one well-being outcome for each activity and it was made clear that they could also write less than four dog human related activities for each aspect of well-being if they wished. When asked about activities important to their life satisfaction and hedonic well-being, life satisfaction itself or any aspect of affect (e.g. happiness, sadness, calm) was a potential outcome. No predetermined fixed-options of affect was used to help participants characterise freely, with any word, the nature of the affect they experience from each activity reported, as recommended by Scherer 59 . In contrast, when participants wrote the activities essential for their eudaimonic well-being, they were asked to use the theory of the six elements of this well-being (e.g. autonomy, self-acceptance) as a guide for their reply, as these elements are part of the core concept of eudaimonic well-being.
Third, once all owners had finished writing, they shared out loud their list of dog human related activities and provided some background to justify their connection with the well-being outcome mentioned. At this point, dog owners had the opportunity to interact with each other, for example, discussing and elaborating on each other's points. Several new dog human related activities emerged from that discussion and participants had the opportunity to agree or disagree with each other's perceptions, which was important to evidence both positive and negative well-being outcomes of the same activity. In that discussion, probes and prompts were used by the moderator to clarify and gather further information from participants. transcription and data analysis. Word-for-word transcription of the audio records and data analysis were performed on Microsoft Word and NVivo 11, respectively. Thematic analysis of the transcripts was con-