Attitudes towards free-roaming dogs and dog ownership practices in Bulgaria, Italy, and Ukraine.

Free-roaming dog population management is conducted to mitigate risks to public health, livestock losses, wildlife conservation, and dog health and welfare. This study aimed to determine attitudes towards free-roaming dogs and their management and describe dog ownership practices in three European countries. We distributed an online questionnaire comprising questions relating to dog ownership practices and attitudes towards free-roaming dogs using social media. We used logistic regression and ordinal probit models to determine associations between demographic and other factors with ownership practices and attitudes towards free-roaming dogs. This study found that most surveyed respondents wanted to see a reduction in free-roaming dog numbers, and felt that this should be achieved through sheltering, catch-neuter-release, and by controlling owned dog breeding. We identified significant associations between both attitudes and ownership practices with gender, religious beliefs, age, education level, reason for dog ownership, previous experience with free-roaming dogs, and country of residence. Respondents who identified as: (i) male, (ii) holding religious beliefs, (iii) owning dogs for practical reasons, (iv) being young, or (v) having no schooling or primary education had a lower probability of neutering and a higher probability of allowing dogs to roam. Respondents who identified as: (i) female, (ii) feeling threatened by free-roaming dogs, (iii) older, or (iv) having more education had a higher probability of answering that increases in free-roaming dog numbers should be prevented. These findings can help to inform future dog population management interventions in these countries. We emphasise the importance of considering local attitudes and dog ownership practices in the development of effective dog population management approaches.

Introduction

Domestic dogs (Canis familiaris) are one of the most abundant species of canids in the world, with total population estimates between 700 million to 1 billion [1, 2]. Around 75% of this population are classified as “free-roaming”, indicating their ability to roam and reproduce freely [1]. Where free-roaming dogs exist in high densities, there are important implications for public health [3-5], livestock losses [6-10] and wildlife conservation [11-13], in addition to issues for the welfare of the dogs themselves [14-17]. The management methods applied to control population sizes involve culling, reproductive control (e.g. through catch-neuter-release; CNR), and the use of shelters to house unowned or unwanted dogs [18].

Human behaviour can shape the success of a population management programme. This includes actions of local communities, the teams involved in dog population management and the governments imposing management interventions. Indeed, the World Organisation for Animal Health (OIE) has identified that understanding public attitudes is important for developing effective dog population control [19]. In order for interventions to be successful, there must be public support for both the management method and its aims (e.g. reducing or stabilising the number of free-roaming dogs). Different communities may have different attitudes towards free-roaming dogs and management methods due to culture, religion, and the specific risks to humans, wildlife, livestock, and other companion animals in the area. Organisations involved in dog population management should consider these cultural, religious, and risk factors to ensure interventions are effective. For example, free-roaming dog populations can be an important part of a community, providing protection to people and livestock [17]. Where management methods aim to reduce free-roaming dog numbers, there may still be demand for dogs in a community. Unless this demand reduces, new dogs may be acquired to replace those removed by population management (either bought, adopted, or by uncontrolled immigration of free-roaming dogs). Prior to implementing management interventions, the level of acceptance of free-roaming dogs in the area should be gauged (i.e. determine whether the public prefer to have fewer free-roaming dogs in the community) so that those involved in dog population management can work towards a goal that benefits the community.

Dog ownership practices can also influence the success of population management. Encouraging responsible ownership practices is included as an objective in Chapter 7.7 of the OIE Terrestrial Animal Health Code on free-roaming dog population control [19]. The OIE describes the requirements of responsible ownership as: “When a person takes on the ownership of a dog, there should be an immediate acceptance of responsibility for the dog, and for any offspring it may produce, for the duration of its life or until a subsequent owner is found” [19]. Unrestricted owned dogs (free-roaming owned) and abandoned dogs are sources of free-roaming dog population increase [19]. Dog ownership practices that allow owned dogs to roam and do not prevent reproduction can hinder efforts to control free-roaming dog populations. Those involved in dog population management must determine the extent to which owned dogs contribute to the free-roaming dog population so that management interventions can be tailored appropriately (e.g. by enforcing responsible ownership through legislation and education programmes).

Questionnaire surveys are frequently used to gain insight into attitudes, opinions, behaviours, and the demographic and sociological factors associated with these. In terms of dog population management, different attitudes, opinions and behaviours about and towards dogs have been associated with responder gender [20-24], age [20, 22, 25], education [21, 22, 26], and previous life experiences (e.g. experience of keeping dogs in childhood) [27]. Questionnaires aiming to describe dog ownership, attitudes, and knowledge have been conducted in many countries around the world, but few published studies have been carried out in European countries.

This study determines attitudes towards the presence of free-roaming dogs and of dog ownership practices in three European countries–Bulgaria, Italy, and Ukraine. We selected these focal countries due to the networks established with collaborating organisations (VIER PFOTEN International and Istituto Zooprofilattico Sperimentale dell’Abruzzo e del Molise “Giuseppe Caporale”; IZSAM) that provided local knowledge to facilitate data collection. The focal countries are culturally and environmentally distinct, allowing comparison of the collected data between different countries within Europe.

In this study, we define attitudes as the thoughts, feelings, and opinions of respondents, as reported in the questionnaire. We define ownership practices as the actions taken to acquire, provide care, and relinquish ownership of dogs, as reported in the questionnaire. The objectives of this study are to: (i) determine attitudes towards the presence of free-roaming dogs; (ii) determine local ownership practices and attitudes, including whether owned dogs were free-roaming or neutered, the level of dog abandonment, and the reasons for dog abandonment; and (iii) investigate whether demographic and other factors (including age, gender, education level, religious beliefs, and previous experience with dogs) influence ownership practices and attitudes towards free-roaming dogs. This information can inform interventions so that education campaigns can target groups who are at-risk of irresponsible dog ownership behaviours [20, 26], as well as provide a baseline for evaluating the impact of interventions on human behaviour and attitudes [26].

Materials and methods

Study design

This was a cross-sectional study, with target populations of Bulgaria, Italy, and Ukraine. The study populations were residents who used social media. Participants were recruited through social media using an online questionnaire (Online Surveys [28]) that was open between the 8th of March 2019 and the 21st of December 2019, available in four languages: Bulgarian, Italian, Ukrainian, and Russian. The social media outlets used to distribute the questionnaire included Facebook [29] and Twitter [30]. Facebook advertising was used to increase the visibility of the questionnaire to the study population and increase the number of respondents. Facebook advertising targeted Facebook users who: (i) were recorded in their online profile as living in Bulgaria, Italy, or Ukraine; and (ii) were over the age of 18. The Facebook adverts invited respondents to provide their opinion on free-roaming dogs and dog ownership practices (see S1 File for English translation of adverts). All adverts were used in all countries to an equal extent, though it is not possible to know which advert respondents had seen. Minimum sample sizes were calculated for the three study areas, using Eq 1. A sample size of 385 respondents per study country was necessary to provide estimates with a 95% confidence interval that was within a 5% margin of error.

Eq 1. Sample size calculation

Where N = population size, e = margin of error, z = z-score, p = population proportion.

Ethical approval

Prior to completing the questionnaire, all participants gave consent within the online survey through a digital signature by selecting yes in a tick box to confirm that they (i) had read and understood the information sheet explaining the project (S2 File), (ii) understood that they had the opportunity to ask questions about the project, (iii) knew they could withdraw from completing the questionnaire, prior to submitting, (iv) agreed for their responses to be collected, stored, and analysed in an anonymised form for the purpose of reports and publication (see S2 File for information). Those who did not consent were not able to complete the questionnaire and were therefore not included in the study. No directly identifiable information was collected; all data obtained remains anonymous. Participants were able to withdraw from the questionnaire prior to completion, but as the data was collected anonymously, participants could not withdraw after the questionnaire was submitted. No minors were included in this study, as respondents who reported being under the age of 18 were not able to complete the online questionnaire. The study was approved by the University of Leeds Ethical Committee (reference BIOSCI 17–003).

Questionnaire design

The questionnaire was developed in English and translated into Bulgarian, Italian, Ukrainian, and Russian. The questionnaire comprised closed questions regarding the respondents’ attitudes towards free-roaming dogs and their management. Respondents that reported owning a dog were asked to also complete questions relating to dog ownership practices. Likert-type scales were used to estimate the level of agreement with specific questions. The questionnaire consisted of three sections: (i) socio-demographic information of the respondent (all respondents); (ii) ownership practices (only dog owners); and (iii) attitudes towards the presence of free-roaming dogs and the management of the free-roaming dog population (all respondents). A copy of the questionnaire in English can be found in S2 File.

Statistical analyses

All predictor and response variables are described in Table 1. Bernoulli logistic regression models were used to test associations between demographic parameters and respondent experience on the binary response variables: (i) Neutering status of owned dogs (binary response: neutered/not neutered); and (ii) respondents’ answers to the question “Do you think an increase in dogs on the street should be prevented?” (binary response: Yes/no). Ordinal probit models [31] were used to test associations between demographic parameters and respondent experience on: (i) Roaming status of owned dogs; (ii) I do not like the presence of free-roaming dogs around my home or work; and (iii) respondents’ answers to the question “Would you prefer to see: no free-roaming dogs, fewer free-roaming dogs, do not mind free-roaming dogs, more free-roaming dogs”. Ordinal variables are categorical variables with a natural order, for example Likert-type scales [32]. Ordinal variables are assumed to have an underlying continuous latent variable that cannot be measured directly (e.g. the attitude of a respondent). This underlying latent variable is therefore split into discrete options that can be measured (e.g. Strongly agree or Agree). The intervals between these discrete options may not be equal (i.e. not equidistant), an assumption required by metric models [33], and responses to ordinal questions may have non-normal distributions. Ordinal predictor variables can be problematic if analysed metrically, leading to Type I (false positive) and Type II (false negative) errors [31]. Ordinal models deal with issues in potential non-equidistant responses and non-normal distributions.

Table 1

Response and predictor variables (self-reported responses to questions) included in the statistical analyses and their levels.

Variables	Levels
Age	* 18–24, 25–34, 35–44, 45–54, 55–64, 65–74 and 75 and above
Children in household	Children in household, no children in household
Country	Bulgaria, Italy, Ukraine
Dog ownership	Dog owner, non-dog owner
Education status	* No education, primary, secondary, tertiary
Gender	Male, Female, NA (including option Other)
Neutering status of owned dogs	Neutered, not neutered
Owning a dog for practical reasons	Practical, not practical
Religious belief	Religious, non-religious
Feeling physically threatened by dogs on the street	* Strongly disagree, disagree, neither agree nor disagree, agree, strongly agree
Been attacked by dogs on the street	Been attacked, not been attacked
Respondent or family members have been bitten by dogs on the street in last 12 months	Been bitten, not been bitten
Roaming status of owned dogs	Never, Sometimes, Always
I do not like the presence of free-roaming dogs around my home or work	Strongly disagree, disagree, neither agree nor disagree, agree, strongly agree
Should an increase in dogs on the street be prevented?	Yes, No
Would you prefer to see dogs on the street?	No free-roaming dogs, fewer free-roaming dogs, do not mind free-roaming dogs, more free-roaming dogs

* Ordinal predictor variables analysed as continuous variables in statistical models.

Dog ownership practices

Model 1 tested associations between demographic parameters and respondent experience with neutering of owned dogs using a Bayesian Bernoulli logistic regression model. The response variable was neutering status of owned dogs with fixed effects of gender, age, education status, religious belief, owning a dog for practical reasons and country (Table 1). Model 2 tested associations between demographic parameters and respondent experience with the roaming status of owned dogs using a Bayesian ordinal probit model. The response variable was roaming status of owned dogs and fixed effects were the same as for Model 1.

Attitudes towards free-roaming dogs

Model 3 tested associations between demographic parameters and respondent experience with agreement to the statement I do not like the presence of free-roaming dogs around my home or work using a Bayesian ordinal probit model. The response variable was I do not like the presence of free-roaming dogs around my home or work and fixed effects were dog ownership, gender, age, education status, children in household, feeling physically threatened by dogs on the street, been attacked by dogs on street, respondent or family members have been bitten by dogs on the street in last 12 months, and country.

Model 4 tested associations between demographic parameters and respondent experience with the question Do you think an increase in dogs on the street should be prevented? using a Bayesian Bernoulli logistic regression model. The response variable was should an increase in dogs on the street be prevented, with fixed effects the same as in Model 3.

Model 5 tested associations between demographic parameters and respondent experience with response to the question Would you prefer to see: no free-roaming dogs, fewer free-roaming dogs, do not mind free-roaming dogs, more free-roaming dogs using a Bayesian ordinal probit model. The response variable was Would you prefer to see dogs on the street, with fixed effects the same as in Model 3.

To fit the statistical models using a Bayesian analysis framework, the package “brms” version 2.12.0 [34] was used in R version 3.6.1 [35]. All models were run with four chains, each with 2000 iterations (1000 used for warmup and 1000 for sampling). Thinning was set to one. The total number of post-warmup samples was 4000. Where a response was missing (i.e. a respondent did not answer a question), the response was omitted from the statistical analysis (see S1 Table for number of no responses per variable).

Collinearity in the predictor variables was checked using the “vif” function in R package “car” [36] and values lower than three were considered not collinear. Model parameters were summarised by the mean and 95% credible intervals of the posterior distribution (CI; 95% most probable values). A significant association was determined if the 95% credible intervals of the posterior distribution did not contain zero on the log odds or probit scale. Probabilities were converted from the logit scale to the probability scale by , and are converted to odds using exp (x), where x is the posterior value on the logit scale.

Results

Descriptive analyses

Demographics

The numbers of respondents were 5,434 in Bulgaria, 3,468 in Italy, and 19,323 in Ukraine. All demographic information is provided in S2 Table. Respondents were from multiple regions within Bulgaria, Italy, and Ukraine (see S3 to S5 Tables). A broad range of ages between 18 and 64 were represented in all three study countries. Most respondents were female in all three study countries (87.5%, n = 4,754 of 5,434, in Bulgaria, 83.1%, n = 2,882 of 3,468 in Italy, and 87.1%, n = 16,832 of 19,323 in Ukraine). In Bulgaria 68.9% (n = 3,743 of 5,434), in Italy 42.0% (n = 1,457 of 3,468), and in Ukraine 67.3% (n = 13,011 of 19,323) of the respondents considered themselves to be religious. In Bulgaria 36% (n = 1,970 of 5,434), Italy 43% (n = 1,474 of 3,468) and Ukraine 57% (n = 10,928 of 19,323) of respondents lived in households with children.

Ownership practices

Sixty-five percent of respondents in Bulgaria (n = 3,528 of 5,434), 75% in Italy (n = 2,581 of 3,468) and 56% in Ukraine (n = 10,797 of 19,323) reported owning a dog. The main reason for dog ownership in all three study countries was for pleasure and company (Bulgaria 85.5%, n = 3,017 of 3528, Italy 87.7%, n = 2,263 of 2,581, Ukraine 70.7%, n = 7,631 of 10,797; see S6 Table for detailed responses on ownership practices). In Italy, a higher percentage of respondents acquired their dog from a dog shelter (38.1%, n = 986 of 2,581), compared to in Bulgaria (9.7%, n = 341 of 3,528) and Ukraine (9.9%, n = 1,070 of 10,797) (Fig 1). In Bulgaria and Ukraine, more respondents found their dog on the street (Bulgaria 35.5%, n = 1,252 of 3,528, and Ukraine 34.6%, n = 3,734 of 10,797) or received their dog from friends/family (Bulgaria 32.6%, n = 1,149 of 3,528 and Ukraine 27.9%, n = 3,015 of 10,797). More respondents in Italy answered that they prevent their dog from breeding through neutering (65.4%, n = 1,689 of 2,581), compared to 40.4% (n = 1,424 of 3,528) in Bulgaria and 35.4% (n = 3,828 of 10,797) in Ukraine. When asked the reason why respondents did not prevent breeding, 37.6% (n = 82 of 218) of respondents in Bulgaria, 34.6% (n = 75 of 217) in Italy, and 13.7% (n = 242 of 1770) in Ukraine answered: “A dog should reproduce at least once” (Fig 1). When respondents were asked if they allowed their dog to roam outside unsupervised, 59.0% (n = 2,082 of 3,528) in Bulgaria, 92.1% (n = 2,377 of 2,581) in Italy and 79.4% (n = 8,571 of 10,797) in Ukraine responded Never, and 29.5% (n = 1,039 of 3,528) in Bulgaria, 6.3% (n = 162 of 2,581) in Italy and 16.3% (n = 1764 of 10,797) in Ukraine responded Sometimes.

Fig 1

Ownership practices of respondents in Bulgaria, Italy, and Ukraine.

The percentage of respondents who answered each of the answer options regarding (A) acquisition of dog, (B) reasons for not preventing breeding, (C) the outcome of the dog, and (D) reason for relinquishment. * Multi answer question: Percentage of respondents who selected each answer option (i.e. 100% would indicate that all respondents chose this option).

Most respondents in all study countries responded that they had never given up a dog (Bulgaria 98.5%, n = 3,474 of 3,528, Italy 92.4%, n = 2,386 of 2,581, and Ukraine 92.2%, n = 9950 of 10,797). Those respondents who had given up a dog mostly answered that this was because of an Animal behavioural problem (Bulgaria 27.3%, n = 9 of 33, Italy 36.5%, n = 66 of 179, and Ukraine 23.8%, n = 161 of 676), or Other reason (Bulgaria 39.4%, n = 13 of 33, Italy 57.5%, n = 104 of 179, and Ukraine 45.3%, n = 306 of 676) (Fig 1), such as family illness; a change in circumstances (e.g. birth of new child in home); moving home; owners going on a long trip away; the dog having puppies; or the dog not getting along with other dogs in the household.

Attitudes

In Bulgaria and Ukraine, high percentages of respondents had seen a free-roaming dog on the day they filled in the questionnaire (73.3%, n = 3,983 of 5,434, and 77.3%, n = 14,934 of 19,323 respectively), compared to only 15.4% (n = 534 of 3,468) of respondents in Italy (Fig 2; see S7 Table for detailed results). A higher percentage of respondents in Bulgaria (21.6%, n = 1,174 of 5,434) and Ukraine (26.5%, n = 5,129 of 19,323) had been attacked by dogs on the street ever in their lifetime, compared to few (4.2%, n = 147 of 3,468) in Italy. Higher percentages of respondents in Bulgaria answered that they provided care to free-roaming dogs by giving food (90.6%, n = 4,911 of 5,434), water (71.0%, n = 3,847 of 5,434), and shelter (34.8%, n = 1,886 of 5,434), compared to Italy (53.7%, n = 1831 food, 44.2%, n = 1508 water, 19.0%, n = 647 of 3468 shelter) and Ukraine (67.5%, n = 13,045 food, 29.6%, n = 5,721 water and 9.7%, n = 1,882 of 19,323 shelter) (S6 Table).

Fig 2

Attitudes of respondents towards free-roaming dogs in Bulgaria, Italy and Ukraine.

The percentage of respondents who answered each of the answer options regarding (A) observation of free-roaming dogs and (B) agreement with statement “I do not like free-roaming dogs being present around my home or work”.

When respondents were asked their level of agreement with the statement “I do not like free-roaming dogs being present in the streets around my home or work”, responses were varied across the full range of options between strongly disagree and strongly agree in Bulgaria and Ukraine (varying between 14 and 25% for all answer options) (Fig 2). Most respondents in Italy strongly disagreed with this statement (35.8%, n = 1,242 of 3,468). In all three study countries, most respondents disagreed (Bulgaria 20.8%, n = 1,130 of 5,434, Italy 19.3%, n = 671 of 3,468, and Ukraine 25.3%, n = 4,890 of 19,323) and strongly disagreed (Bulgaria 42.2%, n = 2,295 of 5,434, Italy 56.4%, n = 1,956 of 3,468 and Ukraine 31.6%, n = 6,097 of 19,323) with the statement “I feel physically threatened by free-roaming dogs”.

Respondents answered most often that the municipality government and volunteer organisations should be responsible for managing the free-roaming dog population (Fig 3; S6 Table). Respondents most often answered that they would like to see no (Bulgaria 52.4%, n = 2,848 of 5,434, Italy 70.2%, n = 2,435 of 3,468, and Ukraine 45.2%, n = 8,740 of 19,323) and fewer (Bulgaria 32.8%, n = 1,780 of 5,434, Italy 24.3%, n = 841 of 3,468, and Ukraine 40.6%, n = 7,846 of 19,323) free-roaming dogs. Respondents who answered that they would like to see no or fewer free-roaming dogs answered that this should be achieved through sheltering, CNR, and controlling the breeding of owned dogs (Fig 3). Few answered that the free-roaming dog population should be reduced through culling (Bulgaria 1.7%, n = 92 of 4628, Italy 1.6%, n = 56 of 3276, and Ukraine 7.3%, n = 1,216 of 16586).

Fig 3

Attitudes of respondents towards dog population management in Bulgaria, Italy, and Ukraine.

The percentage of respondents who answered each of the answer options for: (A) who should be responsible for dog population management? and (B) how should free-roaming dogs be reduced? These were multi answer questions: Percentage of respondents who selected each answer option (i.e. 100% would indicate that all respondents chose this option).

All models converged (for all parameters Rhat = 1.00 and effective sample size >1000, see Supplementary information). There was no collinearity in the predictor variables (all values less than three). All raw model results (including the posterior mean values, standard deviations and 95% credible intervals, the 2.5% and 97.5% percentiles of the posterior distribution) are presented in S8 to S12 Tables. Estimates for mean and 95% CIs for probabilities are reported for each model and presented in Table 2. Odds ratios (OR) are reported for predictor variables in the Bernoulli logistic regression models (Models 1 and 4).

Table 2

Associations between predictor and outcome variables on the probability scale.

Predictor variable		Model 1	Model 2	Model 3	Model 4	Model 5
		Probability of neutering (95% CI)	Probability of answering Never allow dog to roam (95% CI)	Probability of answering Strongly agree to question “I do not like the presence of free-roaming dogs around home or work” (95% CI)	Probability of answering Yes to the question “Should an increase in free-roaming dogs be prevented” (95% CI)	Probability of answering No free-roaming dogs to question “Would you prefer to see”(95% CI)
Gender	Male	0.38 (0.35–0.41)	0.74 (0.71–0.76)	0.17 (0.16–0.18)	0.97 (0.96–0.97)	0.53 (0.51–0.55)
	Female	0.48 (0.47–0.49)	0.81 (0.80–0.82)	0.15 (0.14–0.15)	0.98 (0.98–0.99)	0.60 (0.59–0.61)
Religious belief	Religious	0.44 (0.43–0.45)	0.79 (0.78–0.80)	NI	NI	NI
	Non-religious	0.54 (0.52–0.56)	0.83 (0.82–0.85)	NI	NI	NI
Reason for dog ownership	Practical	0.31 (0.28–0.33)	0.71 (0.69–0.74)	NI	NI	NI
	Non-practical	0.48 (0.47–0.49)	0.81 (0.80–0.82)	NI	NI	NI
Been attacked by dogs on the street	Attacked	NI	NI	0.16 (0.15–0.17)	0.98 (0.98–0.99)	0.61 (0.59–0.62)
	Not attacked	NI	NI	0.15 (0.14–0.15)	0.98 (0.98–0.99)	0.59 (0.58–0.59)
Respondent or family members have been bitten by dogs on the street in last 12 months	Bitten	NI	NI	0.19 (0.18–0.21)	0.98 (0.98–0.99)	0.65 (0.63–0.67)
	Not bitten	NI	NI	0.15 (0.14–0.15)	0.98 (0.98–0.99)	0.58 (0.57–0.59)
Dog ownership	Yes	NI	NI	0.15 (0.15–0.16)	0.98 (0.98–0.99)	0.59 (0.59–0.60)
	No	NI	NI	0.15 (0.14–0.16)	0.98 0.98–0.99)	0.56 (0.55–0.57)
Children in household	Yes	NI	NI	0.15 (0.15–0.16)	0.98 (0.98–0.99)	0.59 (0.58–0.60)
	No	NI	NI	0.15 (0.14–0.15)	0.98 (0.98–0.99)	0.59 (0.57–0.59)
Age	18–24	0.40 (0.39–0.42)	0.78 (0.76–0.79)	0.14 (0.14–0.15)	0.98 (0.98–0.98)	0.57 (0.56–0.58)
	25–34	0.44 (0.43–0.45)	0.79 (0.78–0.80)	0.15 (0.14–0.15)	0.98 (0.98–0.99)	0.58 (0.57–0.59)
	35–44	0.47 (0.46–0.48)	0.80 (0.80–0.81)	0.15 (0.15–0.16)	0.98 (0.98–0.99)	0.59 (0.58–0.60)
	45–54	0.51 (0.49–0.52)	0.82 (0.81–0.83)	0.16 (0.15–0.16)	0.99 (0.98–0.99)	0.60 (0.59–0.61)
	55–64	0.54 (0.52–0.56)	0.83 (0.82–0.84)	0.16 (0.15–0.17)	0.99 (0.98–0.99)	0.61 (0.60–0.63)
	65–74	0.57 (0.55–0.59)	0.84 (0.83–0.86)	0.16 (0.15–0.17)	0.99 (0.99–0.99)	0.62 (0.61–0.64)
	75+	0.61 (0.58–0.63)	0.85 (0.84–0.87)	0.17 (0.15–0.18)	0.99 (0.99–0.99)	0.63 (0.61–0.65)
Education level	None	0.30 (0.24–0.35)	0.73 (0.67–0.78)	0.14 (0.12–0.16)	0.95 (0.92–0.97)	0.55 (0.51–0.60)
	Primary	0.35 (0.32–0.39)	0.75 (0.72–0.79)	0.14 (0.13–0.16)	0.96 (0.95–0.98)	0.57 (0.54–0.60)
	Secondary	0.41 (0.39–0.43)	0.78 (0.77–0.80)	0.15 (0.14–0.16)	0.98 (0.97–0.98)	0.58 (0.56–0.60)
	Tertiary	0.48 (0.47–0.49)	0.81 (0.80–0.82)	0.15 (0.15–0.16)	0.98 (0.98–0.99)	0.60 (0.58–0.60)
Threatened by dogs on the street	Strongly disagree	NI	NI	0.05 (0.047–0.054)	0.97 (0.97–0.98)	0.49 (0.48–0.50)
	Disagree	NI	NI	0.13 (0.12–0.13)	0.98 (0.98–0.98)	0.57 (0.56–0.58)
	Neutral	NI	NI	0.26 (0.25–0.27)	0.99 (0.99–0.99)	0.65 (0.64–0.66)
	Agree	NI	NI	0.44 (0.43–0.45)	0.99 (0.99–0.99)	0.73 (0.72–0.74)
	Strongly agree	NI	NI	0.64 (0.62–0.65)	1.00 (0.99–1.00)	0.79 (0.78–0.81)
Country	Bulgaria	0.41 (0.39–0.43)	0.61 (0.59–0.62)	0.20 (0.18–0.21)	0.99 (0.98–0.99)	0.55 (0.53–0.56)
	Italy	0.62 (0.60–0.64)	0.92 (0.91–0.93	0.10 (0.10–0.11)	0.98 (0.98–0.99)	0.76 (0.74–0.77)
	Ukraine	0.37 (0.36–0.38)	0.81 (0.81–0.82)	0.17 (0.16–0.17)	0.98 (0.98–0.98)	(0.44–0.46)

Significant results are highlighted in bold. NI = predictor variable not included in the model.

Summary of statistical associations

Respondents were less likely to answer that they neutered their dog(s) and more likely to answer that they allow their dog(s) to roam if they identified as (i) male, (ii) religious, (iii) owning dogs for practical reasons, (iv) young, and (v) having no schooling or primary education. Respondents were more likely answer that an increase in free-roaming dogs should be prevented if they identified as (i) female, (ii) feeling threatened by free-roaming dogs, (iii) older, and (iv) having more education. Below we report the detailed statistical findings.

Respondent demographics and owned dog neutering status

Gender, age, education level, owning a dog for practical reasons, religious beliefs and country had significant associations with the neutering status of owned dogs (S8 Table). Probabilities of neutering are presented in Table 2. Female respondents had a higher probability of neutering, compared to males (OR 1.47; 95% CI 1.28–1.64). Holding religious beliefs (OR 0.66, 95% CI 0.61 to 0.72) and owning dogs for practical reasons (i.e. guarding or hunting, compared to for pleasure and company; OR 0.49, 95% CI 0.43 to 0.54) were both negatively associated with neutering. Respondent age (OR 1.15; 95% CI 1.12–1.18) and education level (OR 1.29; 95% CI 1.17–1.41) were both positively associated with neutering (i.e. the older and more educated a respondent, the more likely they were to neuter). Respondents from Italy had a higher probability of neutering compared to Bulgaria (OR 2.32; 95% CI 2.05–2.62) and Ukraine (OR 2.73; 95% CI 2.44–3.01). Respondents from Ukraine had a lower probability of neutering compared to Bulgaria (OR 0.37; 95% CI 0.33–0.40).

Respondent demographics and owned dog roaming status

Gender, age, education level, owning a dog for practical reasons, religious beliefs, and country had significant associations with the roaming status of owned dogs (S9 Table). Probabilities of answering Never allow dog to roam for predictor variables are presented in Table 2. Females had a higher probability of answering that they Never allowed their dog to roam. Respondents who held religious beliefs, and respondents who owned dogs for practical reasons were less likely to answer Never. Age of respondent was positively correlated with answering Never (i.e. older respondents were less likely to allow their dog to roam). Increased education level of the owner was positively associated with answering Never (i.e. respondents with higher levels of education were less likely to allow their dog to roam). Respondents from Italy had the highest probability of answering Never, respondents in Bulgaria had the lowest probability of answering Never.

Respondent demographics and experience with free-roaming dogs and respondent answers to “I do not like the presence of free-roaming dogs around home or work”

Predictor variables gender, age, owning a dog for practical reasons, feeling threatened by dogs on the street, having been attacked by dogs on the street, respondent or family members have been bitten by dogs on the street in last 12 months, and country were significantly associated with the statement I do not like the presence of free-roaming dogs around my home or work (S10 Table). Probabilities for answering Strongly agree for predictor variables are presented in Table 2. Female respondents had a lower probability of agreeing with the statement. Respondents who answered Yes to the question Have you ever been attacked by dogs on the street? had a higher probability of agreeing with the statement. Respondents who answered Yes to the question Have you or your family members been bitten in the last 12 months? had a higher probability of agreeing with the statement. Respondent age was positively associated with agreement to the statement (i.e. older respondents were more likely to agree). Agreement with the statement I feel physically threatened by dogs on the street was positively associated with agreement with the statement I do not like the presence of free-roaming dogs around my home or work (i.e. respondents who felt threatened were more likely to agree with the statement that they did not like the presence of dogs around their home or work). Respondents from Italy had the lowest probability of answering Strongly agree, and respondents from Bulgaria had the highest probability of answering Strongly agree.

There was no evidence of significant associations between: dog ownership, education level, and children in household and agreement with the statement I do not like free-roaming dogs present around my home or work (S10 Table).

Respondent demographics and experience with free-roaming dogs and respondent answers to “Should an increase in free-roaming dogs should be prevented?”

There were significant associations between gender, age, education level, and feeling threatened by dogs on the street and answering Yes to the question Do you think an increase in dogs on the street should be prevented? (S11 Table). Female respondents had a higher probability of answering Yes (OR 2.14; 95% CI 1.67–2.65). There was a positive association between answering Yes and respondents’ agreement with the statement “I feel physically threatened by dogs on the street” (OR 1.53; 95% CI 1.37–1.68); their age (OR 1.11; 95% CI 1.03–1.19); and education (OR 1.54; 95% CI 1.24–1.83).

There was no evidence of significant associations between dog ownership, children in household, having been attacked by dogs on the street, respondent or family members have been bitten by dogs on the street in last 12 months or country and answering Yes to the question Do you think an increase in dogs on the street should be prevented? (S11 Table).

Associations between demographic parameters and respondent experience with the question “Would you prefer to see: No free-roaming dogs, fewer free-roaming dogs, do not mind free-roaming dogs, more free-roaming dogs”

Dog ownership, gender, age, feeling threatened by dogs on the street, having been attacked by dogs on the street, respondent or family members have been bitten by dogs on the street in last 12 months, and country had significant associations with response to this question regarding preference for observing free-roaming dogs (S12 Table). Probabilities for answering No free-roaming dogs for predictor variables are presented in Table 2. Male respondents had a lower probability of answering No free-roaming dogs (Table 2). Dog owners had a higher probability of answering No free-roaming dogs. Respondents who answered Yes to the question “Have you ever been attacked by dogs on the street?”, or Yes to the question “Have you or your family members been bitten in the last 12 months” or had children in their household had a higher probability of answering No free-roaming dogs. Agreement with the statement “I feel physically threatened by dogs on the street” was positively correlated with answering No free-roaming dogs (i.e. respondents who feel threatened by dogs on the street are more likely to answer No free-roaming dogs). Respondent age was positively correlated with answering No free-roaming dogs (i.e. older respondents had a higher probability of preferring to see No free-roaming dogs). Respondents in Italy had the highest probability of answering No free-roaming dogs, and respondents in Ukraine had the lowest probability of answering No free-roaming dogs.

There was no evidence of significant associations with children in household and education level on the probability of preference of observing free-roaming dogs (S12 Table).

Discussion

This study provides a summary of attitudes towards the presence and management of free-roaming dogs, and of dog ownership practices in Bulgaria, Italy, and Ukraine. We have also identified associations between responder characteristics and ownership practices or attitudes towards free-roaming dogs and their management in Bulgaria, Italy, and Ukraine. Respondents who identified as male, religious, owning dogs for practical reasons, young, or uneducated had a lower probability of neutering and a higher probability of allowing dogs to roam. Respondents who identified as: female, feeling threatened by free-roaming dogs, older, or having more education had a higher probability of answering that increases in free-roaming dog numbers should be prevented. This information could be used to target interventions towards groups who are at-risk of irresponsible dog ownership behaviours [20, 26].

Ownership practices

Responsible ownership is an important component of dog population management [19]. In order to effectively target dog population management interventions, it is important to understand the actions taken by dog owners to acquire, provide care, and relinquish ownership. Most respondents in Italy reported that they acquired their dogs from a shelter, whereas in Bulgaria and Ukraine most respondents reported acquiring their dogs from friends or by finding a dog in the street. The differences in dog acquiring behaviour could be due to a lack of public awareness of local shelters, or perceived differences in shelter quality between the study countries. However, there is currently little research to substantiate these explanations and more work on public awareness is needed. In all study countries, many respondents had adopted a dog directly from the street, potentially reflecting the prevalence of free-roaming dogs in these countries. This may also provide an explanation for the lower uptake from shelters. Where free-roaming dogs are prevalent, people may easily adopt dogs from streets near their homes, rather than travelling to a shelter to adopt a dog. Fewer respondents in Italy paid for their dog. Previous studies have suggested that dogs who are received for little cost are at higher risk of relinquishment [37]. However, the number of respondents who answered that they had given up a dog was low across the study countries. These numbers are likely to be an underestimate, given the taboo around relinquishing dogs. A study by Hsu, Severinghaus and Serpell (2003) [27] found similar estimates, where only 5.3% of respondents answered that they had given up a dog, though far more respondents answered that they knew someone who had given up a dog (31.9%). This indicates that respondents may underreport relinquishment of owned dogs. Additionally, the self-selection process of recruitment for this questionnaire may result in respondents who are more highly engaged with their dog and dog ownership and less likely to relinquish their dogs.

Responsible dog ownership requires that an owner provides care for a dog until it dies or is transferred to another owner [19]. Most respondents who had relinquished a dog in Italy and Ukraine reported they had given their dog to a friend (Fig 1), complying with responsible ownership [19]. In Bulgaria, a higher percentage of respondents answered they had “Let free” their dog (Fig 1). Letting a dog free to the street directly increases the free-roaming dog population. Previous studies have found that respondents prefer to let a dog free to the street as it offers the dog an opportunity to live, unrestricted, outside of a shelter and offers the possibility to find another owner through adoption from the street [27]. This suggests that some dog owners may perceive letting a dog free to the street as responsible ownership. Further research is required to understand why respondents in Bulgaria chose to let a dog free, instead of giving to a shelter or to another owner.

Preventing the production of unwanted puppies is an important part of responsible ownership [19]. Most respondents answered that they prevented their dogs from reproducing; 50.8% of respondents in Bulgaria, 65.3% in Italy, and 35.3% in Ukraine answered that they did so through neutering. This compares to study sample neutering percentages of 54% in the United Kingdom [38], and up to 80% in Australia [39, 40]. The results of the present study should be interpreted with caution, as the self-selection process of recruiting for questionnaires can result in biased samples of the populations. It is possible that respondents who were more likely to neuter their dogs (such as those with higher levels of education) were more likely to complete the questionnaire. The true proportion of neutered owned dogs in the study countries may be lower. Neutering of owned dogs can prevent unwanted offspring and, if owned dogs are free-roaming, can help to prevent unowned dogs from reproducing. When respondents were asked why they did not neuter their dog, the most common answer (if one was provided) across all countries was that a dog should reproduce at least once (Fig 1). Few respondents answered that it was for cost reasons. This contrasts with previous findings in Taiwan [27] and Brazil [41], where respondents cite cost and “too much trouble” as primary reasons for not neutering. As cost, in this study, was not found to be a primary reason that owners did not neuter their dogs, this suggests that in Bulgaria, Italy, and Ukraine, whilst low-cost or free neutering interventions may be important [42], interventions should also address owner knowledge, attitudes, and practices towards reproduction, in order for interventions to have a greater impact.

Our findings of significant associations between country, gender, religious belief, reason for ownership, age, and education level and the probability of neutering (S9 Table) reflect those reported in other studies [20–24, 26]. For example, a study by Fielding (2007) in New Providence, The Bahamas [21] also found that respondents with higher levels of education were more likely to have neutered their dog. Similarly, Costa et al., (2017) [26] found that respondents with higher levels of education were more likely to answer that neutering was the best way to control the overabundance of free-roaming animals in Brazil. Respondents with higher levels of education may have a higher level of awareness of responsible ownership and the benefits of neutering, in addition to potentially having a higher income and ability to pay for neutering. Lower education levels may mean less knowledge of possible effects (and lack thereof) of reproducing on a dog’s health and behaviour. Fielding, Samuels & Mather (2002) [20] also found significant associations between owner age and neutering probability, suggesting that younger owners may have a greater desire to breed from their dog, compared to older owners. These findings suggest interventions could be targeted towards younger owners and those with a lower level of education to increase knowledge of the possible effects of neutering and awareness of responsible ownership practices.

Owned dogs that are free-roaming directly increase the size of the free-roaming population. Owned free-roaming dogs therefore contribute to the issues, such as the risks to public health [43, 44] and wildlife [11, 12, 45–49]. Efforts encouraging responsible ownership may help reduce the number of dogs roaming, and may therefore help to reduce the impacts of the free-roaming dogs on public health and wildlife [19]. This study found evidence for significant associations between gender, religious beliefs, reason for dog ownership, age, education, and country, and the probability of allowing owned dogs to roam (S9 Table). It is clear from these results that interventions should be targeted using these demographic risk factors to prevent roaming behaviour, particularly in countries where higher percentages of owned dogs are free-roaming, such as Bulgaria and Ukraine.

Most respondents (59–92%) across all three countries answered that they never allowed their dogs to roam (S5 Table). Again, these results might be biased, as more highly engaged dog owners may have been more likely to participate in this study and may also be less likely to allow their dogs to roam. The results are higher than those reported in studies using similar sampling approaches (i.e. relying on voluntary participation in questionnaires/interviews) in the Bahamas 57% [50], Bhutan 50% [51], Cameroon 37.7% [52], Guatemala 25.7% [53], urban households in Haiti 54% [54], Kenya 19% [55], Mexico 44.9% [56], Ethiopia 15.7% [57], Tanzania 22% [58], and Uganda 21.7% [59], but lower than those reported in semi-urban households in Haiti 62% [54] and Taiwan 79% [27]. There was a significant association between study country and roaming probability (S9 Table), with respondents in Bulgaria more likely to allow their dogs to roam, compared to Italy and Ukraine. The significant association with country may reflect differences in dog ownership practices and attitudes between the study countries.

Attitudes towards free-roaming dogs

In Bulgaria and Ukraine, almost all respondents answered that they had seen free-roaming dogs on the street, whereas in Italy 18.7% of respondents had never seen a free-roaming dog (Fig 2). These results may indicate that the populations of free-roaming dogs are larger in Bulgaria and Ukraine. Within Italy, there are differences in dog population management: some regions permit CNR and the presence of “community dogs” (free-roaming dogs owned by the municipality), whilst other regions only permit dog population management through sheltering. Respondents living in regions that do not permit community dogs, or in regions with smaller free-roaming dog populations, may be expected to observe fewer free-roaming dogs. Higher percentages of respondents in Bulgaria and Ukraine answered that they felt threatened by free-roaming dogs, and that they or a member of their family had been bitten in the last 12 months. These results may also indicate a greater free-roaming dog population size and related problems in Bulgaria and Ukraine.

A large proportion of respondents across all countries answered that they provided care for free-roaming dogs (S6 Table). For example, 90.6% in Bulgaria, 53.7% in Italy, and 67.5% of respondents in Ukraine answered that they provided food for free-roaming dogs. For Bulgaria and Ukraine, these numbers are similar to those reported by Costa et al., (2015) in Brazil, where 61.9% of respondents reported that they or their neighbours fed free-roaming animals, and Massei et al., (2017) [60] in Nepal, where 47% of respondents provided food and care for free-roaming dogs. In a previous study by Slater et al., (2008) [61] in central Italy, only 5% of respondents reported that they provided care for free-roaming dogs. This is much lower than the numbers reported in this study, where 71.5% of Italian respondents answered that they provided care for free-roaming dogs. This may be explained by the potential bias in the recruitment process of this study, respondents who provide care for free-roaming dogs may also have been more motivated to complete the questionnaire. Data was collected by Slater et al., (2008) using an anonymous telephone survey and had a high response rate (74%). Providing care for free-roaming dogs is controversial. Providing food may alleviate welfare issues associated with lack of nutrition in the free-roaming dog population [62-64], but also increases the carrying capacity for the free-roaming dog population.

Most respondents across all study countries felt that the municipal government and volunteer organisations should be responsible for managing free roaming dog populations, and mostly by methods such as sheltering, CNR, and by controlling the breeding of owned dogs (Fig 3). These results are similar to those found in previous studies [61, 65, 66]. For example, a study by Ortega-Pacheco et al., (2007) [66] in Yucatan, Mexico found that 52.8% of interviewed households supported the neutering of dogs for dog population management, and felt that the government and society were responsible for dog population management. The results in this study suggest there is support for dog population management through sheltering, CNR, and restricted breeding of owned dogs. Few respondents answered that culling should be used to control the free-roaming dog population (Fig 3). These results are similar to those found by Beckman et al., (2014) [65], but are much lower than results by Costa et al., (2017) [26], where culling was supported by 26.8% of respondents.

As attitudes can play an important role in determining the success of dog population management, it is important that organisations involved in dog population management gauge the level of support for reducing free-roaming dogs in the area. Across all three countries, most respondents answered that they would prefer to see fewer or no free-roaming dogs, and that an increase in free-roaming dogs should be prevented. With regards to Italy, these responses correspond with previously reported attitudes in the Teramo province in the Abruzzo region of Italy [61].

Implications for future interventions

The results of this study suggest that there is a preference in all three study countries for a reduction in free-roaming dog numbers, and for this to be achieved through sheltering, CNR and responsible ownership, rather than culling. There is therefore support for the management interventions that are taking place in these study countries. Targeted interventions that can influence the behaviour of those less likely to practice responsible ownership may help to improve responsible ownership and reduce free-roaming dog numbers. For example, as there was evidence for significant associations between gender and age on response variables roaming and neutering, interventions could be adapted to target men and younger people on responsible ownership practices. For example, interventions could target these groups to increase knowledge of the necessity to neuter dogs and possible effects of neutering on dog health and behaviour.

Questionnaires are important tools for evaluating the impact of interventions on human attitudes and dog ownership practices. This includes monitoring attitudes and behaviour (such as responsible ownership) to determine whether education campaigns are having a significant impact. There have been numerous studies on attitudes towards free-roaming dogs and dog ownership practices, but few repeated surveys to assess the effectiveness of dog population control on human attitudes and dog ownership practices [26, 67]. The results from this present study can be used to target interventions to those who are less likely to practice responsible ownership and the results can also be used as a baseline for monitoring the effect of dog population management interventions on dog ownership behaviours and attitudes in Bulgaria, Italy, and Ukraine.

Limitations of questionnaire research methods

As discussed throughout, there are limitations in using questionnaires to determine attitudes and behaviours. The self-selection process involved in the recruitment for questionnaires can result in a biased sample of the target population, as certain members of the public may be more motivated to complete the questionnaire, for example dog owners, or those with strong views about the subject. In this questionnaire, as with other similarly themed questionnaire [61], a high percentage of the respondents were female. As there were fewer responses from men, the questionnaire results may not necessarily reflect the views of the wider population. A high percentage of respondents reported to have or be in tertiary education, which is not representative of the wider populations. The questionnaire was also primarily advertised through social media; therefore, members of the public who do not have access to social media are likely to have been missed. Although this is a limitation, social media provides opportunities to recruit a large and diverse range of respondents (see [68, 69] for review). We used four different adverts with slightly different wording in order to attract as many respondents as possible. The different adverts may have attracted different subsets of people, which could lead to biases in results. Though we were unable to determine which adverts respondents saw, all adverts were distributed equally across the three countries and are unlikely to lead to differences between countries.

In questionnaire surveys, missing data can occur at two levels: (i) missing data of the complete questionnaire (as described above), and (ii) missing data when respondents do not complete specific questions or sections. In this study, higher percentages of missing responses were observed (S1 Table) for questions relating to the respondents’ religion (13.3%), whether they did not like the presence of free-roaming dogs (10.2%), and if they felt threatened by free-roaming dogs around their home or work (8.2%). This missing data can result in biased estimates, and as such, these results should be interpreted with caution. Despite the clear biases in questionnaire surveys, given the range of respondents in this study (for example, in terms of ages and regions), the results provide an indication of ownership practices and attitudes, and the statistical models still give us information about the risk factors for behaviours and attitudes.

Conclusions

When planning dog population management interventions, it is important to understand how human behaviour may impact the success of an intervention. This involves understanding how dog ownership practices may influence intervention success, and gauging the level of public support for management interventions. This study found evidence for significant associations between demographic factors and ownership practices and respondent attitudes. These results can be used to inform future dog population management interventions in these countries. Interventions should consider also carrying out periodic questionnaire surveys to evaluate changes in respondent attitudes towards responsible ownership and the free-roaming dog population.

Facebook adverts.

(DOCX)

Click here for additional data file.

English copy of questionnaire.

(DOCX)

Click here for additional data file.

Answer option to question “Are your dog(s) registered and identified” in Bulgarian questionnaire.

(DOCX)

Click here for additional data file.

Number of "No responses" to outcome and predictor variables in statistical analysis.

(DOCX)

Click here for additional data file.

Demographic information about respondents in Bulgaria, Italy and Ukraine.

(DOCX)

Click here for additional data file.

Number of respondents in Bulgaria, split by oblasts in Bulgaria.

(DOCX)

Click here for additional data file.

Number of respondents in Italy, split by regions in Italy.

(DOCX)

Click here for additional data file.

Number of respondents in Ukraine, split by oblasts in Ukraine.

(DOCX)

Click here for additional data file.

Respondents answers to questions about ownership practices in Bulgaria, Italy and Ukraine.

(DOCX)

Click here for additional data file.

Respondents answers to questions about attitudes to free-roaming dogs in Bulgaria, Italy and Ukraine.

(DOCX)

Click here for additional data file.

The posterior mean values, error estimates, the 2.5 and 97.5 percentiles of the posterior distribution (CI), Rhat values and bulk and tail effective sample sizes (ESS) for Model 1.

(DOCX)

Click here for additional data file.

The posterior mean values, error estimates, the 2.5 and 97.5 percentiles of the posterior distribution (CI), Rhat values and bulk and tail effective sample sizes (ESS) for Model 2.

(DOCX)

Click here for additional data file.

The posterior mean values, error estimates, the 2.5 and 97.5 percentiles of the posterior distribution (CI), Rhat values and bulk and tail effective sample sizes (ESS) for Model 3.

(DOCX)

Click here for additional data file.

The posterior mean values, error estimates, the 2.5 and 97.5 percentiles of the posterior distribution (CI), Rhat values and bulk and tail effective sample sizes (ESS) for Model 4.

(DOCX)

Click here for additional data file.

The posterior mean values, error estimates, the 2.5 and 97.5 percentiles of the posterior distribution (CI), Rhat values and bulk and tail effective sample sizes (ESS) for Model 5.

(DOCX)

Click here for additional data file.

8 Sep 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

I Anna S Olsson, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf.

2. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

3. Thank you for stating the following in the Competing Interests section: "The authors declare that: A.M.M. and S.H. are employed by VIER PFOTEN International, a global animal welfare organisation; L.M.C has received a research grant from VIER PFOTEN International; and L.M.S.’s research has been funded by VIER PFOTEN International."

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests).  If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: With much interest I have read and reviewed the manuscript on the study of ‘Public attitudes towards free-roaming dogs and dog ownership practices in Bulgaria, Italy, and Ukraine’. The study discusses 1) feelings, thoughts and preferences regarding stray/free roaming dogs, which are studied as attitudes towards these dogs, 2) the practice of neutering owned dogs and allowing owned dogs to roam, which are studied as dog ownership practices. The study also determines 3) dog acquisition and abandonment practices. Data was collected in three countries, descriptive data is presented and the three topics are studied for associations with demographic variables, including country of origin of the respondent and for associations with for instance feelings of fear for stray/free roaming dogs.

I would recommend the study to be accepted for submission by PloSOne and feel the study has value as it covers a study area that is of much relevance to dog welfare and the animal-human bond. A strong point of the study is that it was done in three countries, allowing for comparison between these countries’ respondents. Yet, the manuscript would benefit from some improvements, for which I would suggest:

Abstract

• Would you have enough words left to add information on the study method? The abstract now moves from aim (line 21) directly to results (line 23).

• Line 27 indicates associations between ‘public attitudes and dog ownership practices’. However, is it not 1) associations between attitudes and gender (…) and 2) associations between dog ownership practices and gender (…) that you studied?

• Line 30 and 33: are the studied variables jointly leading to higher probabilities or separately? You now use the word ‘and’. Would ‘or’ be a better choice?

• Please consider a different final sentence for the abstract. The current choice suggests something that was not studied and working cohesively towards a shared goal may require another approach than stated here.

Introduction

• The final sentence of the first paragraph (line 48-50) doesn’t seem to connect this paragraph to the next. It is not the assessment of management programs that is key to your study, right? The next paragraph is on public attitudes and their role in dog population (management). Therefore, it may help the flow of your text to stay with your topics and to choose a connecting sentence with content on these topics. (As to avoid that the reader expects the next paragraph to provide details on assessment of population management). A similar disruption of text flow is in line 64-65. Here ‘Reduction in numbers (…) disease control (20-22)’, hints upon the (out of scope) topic of carrying capacity. However, you wish to discuss the role of public attitudes. Perhaps rephrase to something like: ‘Without a change in demand for dogs in the community, new dogs may be bought, adopted or not prevented from moving into a community. The latter shift in dog populations may be consequential to the community’s habitat offering food, shelter, and/or social requirements’. (And check if the references still apply.)

• In line with the suggestions on terminology below, your readers will be helped by indicating somewhere in the introduction how you define/study ‘attitude’ and ‘ownership practices’ in this study, as both concepts theoretically cover a broad range of factors.

Consistency in terminology and choice thereof

• Please check the manuscript for consistency in terminology and choice thereof. For example, both ‘free-roaming dogs’ and ‘stray’ are used.

• Another example is the use of the word ‘social factors’ in line 37-38 of the abstract, however, is this a correct reflection of the factors that you studied?

• In the introduction: ‘strategies’ or ‘intervention (programme)s’, or both? (line 54 and 56)

• Is the OIE quote formulated as a definition? ‘there should be’ indicates that it is a list of requirements, not a definition?

• Are ‘dog abandonment level and reasons’ dog ownership practices? Are ‘acquisition reasons and manners’ dog ownership practices?

• Line 484: new terminology of ‘culture’: do you need to introduce this new terminology here? Of can you stay closer to one of the terminologies/definitions/concepts previously used/addressed in your study?

Study sample

• Would it be possible to add early on a clarification on the study sample: was it one study sample that was questioned on both topics (attitudes versus ownership practices)? From S2 Table 2 I would conclude that you used one sample and for some questions only the responses of the dog-owners were used. The clarification of variables (eg Table 1) provides the answer in the levels of dog owner versus non-dog owner. However, an earlier clarification would help the reader and please do consider to include the ratio. P.s. You may even reconsider using the word ‘public’ in ‘public attitudes’ in the title. The nature of the study may have attracted only a certain part of ‘the general public’. The study approach of social media targeting may also have affected who is in the sample. Like many studies (on dogs) the female-male ratio is disbalanced in this study also. These matters were discussed nicely in lines 558 onwards, and you may want to opt to adjust the title accordingly.

Data and Methods

• Table 1 indicates percentages of >7.5% missing values for ‘religion’, ‘do not like stray presence’, ‘feel threatened’, would you want to comment on this in the discussion, particularly on the latter two?

Results

• The descriptive analyses are given in percentages with the numbers of overall respondents in line 206. This facilitates readability of the text, but the downside is less accuracy of the info, which would increase by adding (N=xx of xxx).

• In the statistical analyses section subsections are titled with ‘effect of…’ this is correct as it refers to the statistical method/model used, however it may unintentionally suggest a cause-effect relationship to less experienced readers. As PloSOne is open access, you may want to consider to change the titles of the subsections. A minor thing in these titles is that there is some inconsistency in ‘on answer to’, ‘on answering’, ‘on the question’. Shorter, more aligned formatting will help the reader. Please also look at the title on page 18 again: is ‘respondent experience’ precise enough for readers to know what the section is about (and note the dot at the end of the title).

• In Table 2 the description under Model 3 would benefit from adding the topic (of stray dog presence), as was done in the other cells of this row. Some of the lower listed probabilities in this Table have three decimal places, whereas above these all have two (eg final column for children in household ‘0.593’ and see in particular Model 4 – age and below There also seems to be an indication ‘t0’ in column Model 3, 6th row from below).

Discussion

• The first paragraph of the Discussion section on Ownership practices seems to address elements not studied. For instance ‘feeding and vaccination’. Also, was ‘quantifying the prevalence of ownership practices’ key of what you did? In addition: would you not want to mention the attitudes here also?

• The discussion could be more strongly structured. You could opt for a structure of summing up key results and then discussing the findings, after which you discuss limitation. Alternatively, you could present a key result and discuss it. You could also look at stronger starts of each paragraph. For instance, line 402 now starts with ‘There were (…)’ and line 450 with ‘This study found evidence’: such sentence parts make your writing less concise. Consequentially, it is harder for the reader to find the flow or logic between the paragraphs.

• Regarding the content of the discussion:

- Line 404 onwards indicates ‘a lack of public awareness of local shelters’, but a few sentences onwards you indicate how availability of dogs on the streets may affect ‘acquisition’. Yet, without indicating that this could also be an explanation for lesser uptake from shelters, that is if dogs are available more so ‘close by on the street’, why go to a shelter?

- Line 411 re. ‘low number of giving up’: well done that you cite the literature and possible underreporting. An addition could be to reflect a little on how your approach to gather participants may have impacted on your results? (Could these participants have been more highly engaged with their dog or with dog ownership?) You do touch upon this in lines 431 onwards and there is a nice section on it from line 500 onwards, so perhaps by restructuring the Discussion this comment becomes superfluous.

- With regard to line 411 onwards, there are some references on neutering percentages in study samples. It may put your data in perspective? (See for instance Diesel G, Brodbelt D, Laurence C. Survey of veterinary practice policies and opinions on neutering dogs. Vet Rec Open. 2010;166(15):455-458. & Bennett PC, Rohlf VI. Owner-companion dog interactions: Relationships between demographic variables, potentially problematic behaviours, training engagement and shared activities. Appl Anim Behav Sci. 2007;102(1-2):65-84. And this ref may also be of interest: Rohlf VI, Bennett, PC, Toukhsati S, Coleman G. Why do even committed dog owners fail to comply with some responsible ownership practices? Anthrozoös. 2010;23(2):143-155.)

- Lines 450-461: and how about your study’s finding of owners thinking ‘a dog should reproduce at least once’: could lower education levels also mean lesser knowledge of possible effects (and lack thereof) of reproducing on a dog’s health and behaviour? You may have a nice opportunity here to write about attitude, knowledge, skills and behaviour and argument the statement you make later how attitudes need to be addressed as well when aiming for effective dog population management.

- Line 463 onwards: would you not want to link with mental health (next to health)? You can discuss your findings on respondents that were bitten/had family bitten and not liking the presence of stray dogs, which may indicate mental health risks for these respondents?

- Line 476: yes, and why would you anticipate this lower likeliness? Some brief words on more highly engaged dog owners being more likely to partake in studies on dogs (as discussed well in the limitation section)? Well done that you indicate percentages from other studies for comparison: an addition could be to indicate whether the study approaches differed from yours?

- Lines 538-545: You end this interesting paragraph with age as a targeting option. Would you not want to add how the attitude aspect that associated neutering in your study could be addressed specifically?

- Line 548: Mind that questionnaires do not measure on behaviour itself.

Further suggestions

- Abstract (line 20): ‘behaviour and outlook’, but your aim (one sentence lower) regards attitudes of dog owners and non-dog owners & practices of dog owners. Is ‘behaviour and outlook of local communities’ the right step up to your study aim? (Also in line 94 of the Introduction).

- Abstract (line 23 & 26): This study – this questionnaire: you may want to reword it (esthetical reasons only).

- Abstract (line 35): Can your findings help to ‘inform’ or ‘shape’ management interventions?

- Introduction (line 61): ‘these factors’: is it clear to the reader which factors?

- Introduction (line 83): ‘encouraging’ – but note that legislation may have an effect other than or next to an encouraging effect.

- If available, indications of ‘other reasons’ (eg Results line 239) would be informative when percentages are high.

Reviewer #2: Thank you for the opportunity to review this manuscript that details dog ownership behaviors and attitudes towards free-roaming dogs in Bulgaria, Italy, and Ukraine. The study employed an online questionnaire and recruited participants via social media. The data showed differences in self-reported behaviors and attitudes between countries, and found statistically significant associations between demographic factors such as age, gender, education, and the measured outcomes.

The submitted article satisfies PLOS ONE’s 7 publication criteria. It is very well written and technically sound, and the limitations are adequately addressed in the discussion. There are a few instances in which the authors draw conclusions or make statements that go beyond the reach of their data, but this can be easily fixed with minor editing.

Going beyond the data: At several points in the manuscript the authors discuss the efficacy of free-roaming dog population control (e.g., abstract, lines 37, 95), but efficacy was not assessed in their study. They surveyed attitudes and behaviors, then hypothesized how those behaviors might affect efficacy, but they did not assess it as they claim.

Causality is sometimes implied, but causal relationships cannot be determined from the questionnaire data. This is true when the authors describe the relationship between demographic data and attitudes and behavior: I would caution against saying there was a “significant effect of gender/age/education/feeling threatened” on response variables because cause and effect were not assessed. Instead, I suggest saying there was a statistically significant association between gender/age/etc. and the response variable.

I have concerns about the advertisements used for recruitment. The four advertisements are very different, and depending on which advertisement a participant saw, their responses may have been biased. For example, Advert 3 includes the language “stray dog overpopulation is a global problem which is of public health, animal welfare and environmental concern,” whereas Advert 1 simply states that they want a wide range of opinions on stray dogs. This is problematic and needs to be addressed in the methods and/or discussion. How were these advertisements distributed? Was Advert 1 used in one country or in one language while Advert 3 was used in another? If so, significant differences cannot necessarily be attributed to region as they could instead be attributed to the condition of the advertisement. If they were randomly distributed across country/language then it may not be a confound, but there could still be an effect of advertisement and this should be discussed.

Line 119 – do you mean Power analysis (instead of sample size)?

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: I.R. van Herwijnen

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: Plos review July 2021.docx

Click here for additional data file.

15 Nov 2021

Dear Editor,

We would like to thank both you and the reviewers for your helpful feedback on our manuscript. We have addressed each of the comments, as outlined below.

All line numbers below refer to the revised manuscript with track changes (full markup).

Reviewer 1

With much interest I have read and reviewed the manuscript on the study of ‘Public attitudes towards free-roaming dogs and dog ownership practices in Bulgaria, Italy, and Ukraine’. The study discusses 1) feelings, thoughts and preferences regarding stray/free roaming dogs, which are studied as attitudes towards these dogs, 2) the practice of neutering owned dogs and allowing owned dogs to roam, which are studied as dog ownership practices. The study also determines 3) dog acquisition and abandonment practices. Data was collected in three countries, descriptive data is presented and the three topics are studied for associations with demographic variables, including country of origin of the respondent and for associations with for instance feelings of fear for stray/free roaming dogs.

I would recommend the study to be accepted for submission by PloSOne and feel the study has value as it covers a study area that is of much relevance to dog welfare and the animal-human bond. A strong point of the study is that it was done in three countries, allowing for comparison between these countries’ respondents. Yet, the manuscript would benefit from some improvements, for which I would suggest:

Abstract

• Would you have enough words left to add information on the study method? The abstract now moves from aim (line 21) directly to results (line 23).

Thank you for this suggestion – we’ve included details on the questionnaire and statistical analysis in the abstract now (lines 25-29).

• Line 27 indicates associations between ‘public attitudes and dog ownership practices’. However, is it not 1) associations between attitudes and gender (…) and 2) associations between dog ownership practices and gender (…) that you studied?

That’s correct – we’ve amended to “We identified significant associations between both attitudes and ownership practices with…”

• Line 30 and 33: are the studied variables jointly leading to higher probabilities or separately? You now use the word ‘and’. Would ‘or’ be a better choice?

The variables each separately lead to higher probabilities, so we’ve changed “and” to “or” as suggested.

• Please consider a different final sentence for the abstract. The current choice suggests something that was not studied and working cohesively towards a shared goal may require another approach than stated here.

We’ve omitted this statement.

Introduction

• The final sentence of the first paragraph (line 48-50) doesn’t seem to connect this paragraph to the next. It is not the assessment of management programs that is key to your study, right? The next paragraph is on public attitudes and their role in dog population (management). Therefore, it may help the flow of your text to stay with your topics and to choose a connecting sentence with content on these topics. (As to avoid that the reader expects the next paragraph to provide details on assessment of population management). A similar disruption of text flow is in line 64-65. Here ‘Reduction in numbers (…) disease control (20-22)’, hints upon the (out of scope) topic of carrying capacity. However, you wish to discuss the role of public attitudes. Perhaps rephrase to something like: ‘Without a change in demand for dogs in the community, new dogs may be bought, adopted or not prevented from moving into a community. The latter shift in dog populations may be consequential to the community’s habitat offering food, shelter, and/or social requirements’. (And check if the references still apply.)

Thank you for these suggestions, we have omitted the final sentence in the first paragraph and included a sentence, based on your suggestion:

“Unless this demand reduces, new dogs may be acquired to replace those removed by population management (either bought, adopted, or by uncontrolled immigration of free-roaming dogs)” Lines 72-74.

• In line with the suggestions on terminology below, your readers will be helped by indicating somewhere in the introduction how you define/study ‘attitude’ and ‘ownership practices’ in this study, as both concepts theoretically cover a broad range of factors.

We have included the definitions:

“In this study, we define attitudes as the thoughts, feelings, and opinions of respondents, as reported in the questionnaire. We define ownership practices as the actions taken to acquire, provide care, and relinquish ownership of dogs, as reported in the questionnaire.” (lines 108-111).

Consistency in terminology and choice thereof

• Please check the manuscript for consistency in terminology and choice thereof. For example, both ‘free-roaming dogs’ and ‘stray’ are used.

We have changed all to “free-roaming dogs” for accuracy and consistency.

• Another example is the use of the word ‘social factors’ in line 37-38 of the abstract, however, is this a correct reflection of the factors that you studied?

Now omitted from the abstract.

• In the introduction: ‘strategies’ or ‘intervention (programme)s’, or both? (line 54 and 56)

We have changed all occurrences of “strategies” to “interventions”.

• Is the OIE quote formulated as a definition? ‘there should be’ indicates that it is a list of requirements, not a definition?

Thank you for this – we have changed to “The OIE describes the requirements of responsible ownership as” (line 84).

• Are ‘dog abandonment level and reasons’ dog ownership practices? Are ‘acquisition reasons and manners’ dog ownership practices?

We have amended to “determine local ownership practices and attitudes” (line 113).

• Line 484: new terminology of ‘culture’: do you need to introduce this new terminology here? Of can you stay closer to one of the terminologies/definitions/concepts previously used/addressed in your study?

We have amended to “The significant effect of country may reflect differences in dog ownership practices and attitudes between the study countries.”.

Study sample

• Would it be possible to add early on a clarification on the study sample: was it one study sample that was questioned on both topics (attitudes versus ownership practices)? From S2 Table 2 I would conclude that you used one sample and for some questions only the responses of the dog-owners were used. The clarification of variables (eg Table 1) provides the answer in the levels of dog owner versus non-dog owner. However, an earlier clarification would help the reader and please do consider to include the ratio. P.s. You may even reconsider using the word ‘public’ in ‘public attitudes’ in the title. The nature of the study may have attracted only a certain part of ‘the general public’. The study approach of social media targeting may also have affected who is in the sample. Like many studies (on dogs) the female-male ratio is disbalanced in this study also. These matters were discussed nicely in lines 558 onwards, and you may want to opt to adjust the title accordingly.

Thank you for this suggestion. We agree and have omitted “public” in the title as suggested.

We have amended the section “Questionnaire design” to include:

“The questionnaire comprised closed questions regarding the respondents’ attitudes towards free-roaming dogs and their management. Respondents that reported owning a dog were asked to also complete questions relating to dog ownership practices.”(lines 158-161)

And

“ The questionnaire consisted of three sections: (i) socio-demographic information of the respondent (all respondents); (ii) ownership practices (only dog owners); and (iii) attitudes towards the presence of free-roaming dogs and the management of the free-roaming dog population (all respondents)” (lines 162-166)

We have included the number of respondents for this section of the questionnaire in lines 221-222 “Sixty-five percent of respondents in Bulgaria (n=3528), 75% in Italy (n=2581) and 56% in Ukraine (n=10797) reported owning a dog.”

Data and Methods

• Table 1 indicates percentages of >7.5% missing values for ‘religion’, ‘do not like stray presence’, ‘feel threatened’, would you want to comment on this in the discussion, particularly on the latter two?

We have included in the “Limitations of questionnaire research methods” section:

Lines 680-683: “In this study, higher percentages of missing responses were observed (S1 table) for questions relating to the respondents’ religion (13.3%), whether they did not like the presence of free-roaming dogs (10.2%), and if they felt threatened by free-roaming dogs around their home or work (8.2%)”

Results

• The descriptive analyses are given in percentages with the numbers of overall respondents in line 206. This facilitates readability of the text, but the downside is less accuracy of the info, which would increase by adding (N=xx of xxx).

Added throughout results.

• In the statistical analyses section subsections are titled with ‘effect of…’ this is correct as it refers to the statistical method/model used, however it may unintentionally suggest a cause-effect relationship to less experienced readers. As PloSOne is open access, you may want to consider to change the titles of the subsections. A minor thing in these titles is that there is some inconsistency in ‘on answer to’, ‘on answering’, ‘on the question’. Shorter, more aligned formatting will help the reader. Please also look at the title on page 18 again: is ‘respondent experience’ precise enough for readers to know what the section is about (and note the dot at the end of the title).

Thank you for this suggestion – we’ve changed the subsections titles and the use of “effect” throughout to prevent this occurring.

• In Table 2 the description under Model 3 would benefit from adding the topic (of stray dog presence), as was done in the other cells of this row. Some of the lower listed probabilities in this Table have three decimal places, whereas above these all have two (eg final column for children in household ‘0.593’ and see in particular Model 4 – age and below There also seems to be an indication ‘t0’ in column Model 3, 6th row from below).

We’ve amended the table to include the suggested column titles and amend the rounding and typo errors.

Discussion

• The first paragraph of the Discussion section on Ownership practices seems to address elements not studied. For instance ‘feeding and vaccination’. Also, was ‘quantifying the prevalence of ownership practices’ key of what you did? In addition: would you not want to mention the attitudes here also?

We have amended the first paragraph of the discussion to read:

“This study investigated dog ownership practices and attitudes towards the management of free-roaming dogs in Bulgaria, Italy, and Ukraine.”

And, in the second paragraph:

“In order to effectively target dog population management interventions, it is important to understand the actions taken by dog owners to acquire, provide care, and relinquish ownership.”

• The discussion could be more strongly structured. You could opt for a structure of summing up key results and then discussing the findings, after which you discuss limitation. Alternatively, you could present a key result and discuss it. You could also look at stronger starts of each paragraph. For instance, line 402 now starts with ‘There were (…)’ and line 450 with ‘This study found evidence’: such sentence parts make your writing less concise. Consequentially, it is harder for the reader to find the flow or logic between the paragraphs.

We have amended the discussion structure to improve flow and clarity of logic.

• Regarding the content of the discussion:

- Line 404 onwards indicates ‘a lack of public awareness of local shelters’, but a few sentences onwards you indicate how availability of dogs on the streets may affect ‘acquisition’. Yet, without indicating that this could also be an explanation for lesser uptake from shelters, that is if dogs are available more so ‘close by on the street’, why go to a shelter?

This is a good point – we’ve included in this paragraph:

“In all study countries, many respondents had adopted a dog directly from the street, potentially reflecting the prevalence of free-roaming dogs in the study countries. This may also provide an explanation for a lesser uptake from shelters. Where free-roaming dogs are prevalent, people may be easily adopt dogs from streets near their homes, rather than having to travel to a shelter to adopt a dog.” (Lines 482-486)

- Line 411 re. ‘low number of giving up’: well done that you cite the literature and possible underreporting. An addition could be to reflect a little on how your approach to gather participants may have impacted on your results? (Could these participants have been more highly engaged with their dog or with dog ownership?) You do touch upon this in lines 431 onwards and there is a nice section on it from line 500 onwards, so perhaps by restructuring the Discussion this comment becomes superfluous.

We have amended to include:

Lines 494-4956 “Additionally, the self-selection process of recruitment for this questionnaire may result in respondents who are more highly engaged with their dog and dog ownership and less likely to relinquish their dogs.”

- With regard to line 411 onwards, there are some references on neutering percentages in study samples. It may put your data in perspective? (See for instance Diesel G, Brodbelt D, Laurence C. Survey of veterinary practice policies and opinions on neutering dogs. Vet Rec Open. 2010;166(15):455-458. & Bennett PC, Rohlf VI. Owner-companion dog interactions: Relationships between demographic variables, potentially problematic behaviours, training engagement and shared activities. Appl Anim Behav Sci. 2007;102(1-2):65-84. And this ref may also be of interest: Rohlf VI, Bennett, PC, Toukhsati S, Coleman G. Why do even committed dog owners fail to comply with some responsible ownership practices? Anthrozoös. 2010;23(2):143-155.)

Thank you for suggesting these articles. We’ve now included their results for comparison:

Lines 514-515“Most respondents answered that they prevented their dogs from reproducing; 50.8% respondents in Bulgaria, 65.3% in Italy, and 35.3% in Ukraine answered that they did so through neutering. This compares to study sample neutering percentages of 54% in the United Kingdom (36), and up to 80% in Australia (37,38).”

- Lines 450-461: and how about your study’s finding of owners thinking ‘a dog should reproduce at least once’: could lower education levels also mean lesser knowledge of possible effects (and lack thereof) of reproducing on a dog’s health and behaviour? You may have a nice opportunity here to write about attitude, knowledge, skills and behaviour and argument the statement you make later how attitudes need to be addressed as well when aiming for effective dog population management.

Thank you for this suggestion, we have amended this paragraph to include your suggestion and included:

Lines 545-548: “These findings suggest interventions could be targeted towards younger owners and those with a lower level of education to increase knowledge of the possible effects of neutering and awareness of responsible ownership practices.”

- Line 463 onwards: would you not want to link with mental health (next to health)? You can discuss your findings on respondents that were bitten/had family bitten and not liking the presence of stray dogs, which may indicate mental health risks for these respondents?

This is an interesting idea and could make for a useful follow-on study, however any comments made on this topic here would be conjecture as our data does not relate specifically to this.

- Line 476: yes, and why would you anticipate this lower likeliness? Some brief words on more highly engaged dog owners being more likely to partake in studies on dogs (as discussed well in the limitation section)? Well done that you indicate percentages from other studies for comparison: an addition could be to indicate whether the study approaches differed from yours?

We have amended to now read:

Lines 563 onwards: “Again, these results might be biased, as more highly engaged dog owners may have been more likely to participate in this study and may also be less likely to allow their dogs to roam. The results are higher than those reported in studies using similar sampling approaches (i.e. relying on voluntary participation in questionnaires/interviews) in …”

- Lines 538-545: You end this interesting paragraph with age as a targeting option. Would you not want to add how the attitude aspect that associated neutering in your study could be addressed specifically?

Thank you for this suggestion, we have now included: Lines 638-640 “For example, interventions could target these groups to increase knowledge of the necessity to neuter dogs and possible effects of neutering on dog health and behaviour.”

- Line 548: Mind that questionnaires do not measure on behaviour itself.

Thank you – we’ve taken this out.

Further suggestions

- Abstract (line 20): ‘behaviour and outlook’, but your aim (one sentence lower) regards attitudes of dog owners and non-dog owners & practices of dog owners. Is ‘behaviour and outlook of local communities’ the right step up to your study aim? (Also in line 94 of the Introduction).

Amended

- Abstract (line 23 & 26): This study – this questionnaire: you may want to reword it (esthetical reasons only).

We have replaced survey for questionnaire throughout for consistency. However we would like to keep questionnaire and study, as the questionnaire (the questions and responses) was part of the study (which also includes description and statistical analyses of responses).

- Abstract (line 35): Can your findings help to ‘inform’ or ‘shape’ management interventions?

Amended

- Introduction (line 61): ‘these factors’: is it clear to the reader which factors?

We have amended to “Organisations involved in dog population management should consider these cultural, religious, and risk factors to ensure interventions are effective” (lines 68-69)

- Introduction (line 83): ‘encouraging’ – but note that legislation may have an effect other than or next to an encouraging effect.

We have amended to “enforcing”.

- If available, indications of ‘other reasons’ (eg Results line 239) would be informative when percentages are high.

We’ve included some other reasons “such as family illness; a change in circumstances (e.g. birth of new child in home); moving home; owners going on a long trip away; the dog having puppies; or the dog not getting along with other dogs in the household.” (Lines 279-282)

Review 2:

Thank you for the opportunity to review this manuscript that details dog ownership behaviors and attitudes towards free-roaming dogs in Bulgaria, Italy, and Ukraine. The study employed an online questionnaire and recruited participants via social media. The data showed differences in self-reported behaviors and attitudes between countries, and found statistically significant associations between demographic factors such as age, gender, education, and the measured outcomes.

The submitted article satisfies PLOS ONE’s 7 publication criteria. It is very well written and technically sound, and the limitations are adequately addressed in the discussion. There are a few instances in which the authors draw conclusions or make statements that go beyond the reach of their data, but this can be easily fixed with minor editing.

Going beyond the data: At several points in the manuscript the authors discuss the efficacy of free-roaming dog population control (e.g., abstract, lines 37, 95), but efficacy was not assessed in their study. They surveyed attitudes and behaviors, then hypothesized how those behaviors might affect efficacy, but they did not assess it as they claim.

We have amended the text throughout.

Causality is sometimes implied, but causal relationships cannot be determined from the questionnaire data. This is true when the authors describe the relationship between demographic data and attitudes and behavior: I would caution against saying there was a “significant effect of gender/age/education/feeling threatened” on response variables because cause and effect were not assessed. Instead, I suggest saying there was a statistically significant association between gender/age/etc. and the response variable.

We have amended the terminology throughout to avoid any reader confusion.

I have concerns about the advertisements used for recruitment. The four advertisements are very different, and depending on which advertisement a participant saw, their responses may have been biased. For example, Advert 3 includes the language “stray dog overpopulation is a global problem which is of public health, animal welfare and environmental concern,” whereas Advert 1 simply states that they want a wide range of opinions on stray dogs. This is problematic and needs to be addressed in the methods and/or discussion. How were these advertisements distributed? Was Advert 1 used in one country or in one language while Advert 3 was used in another? If so, significant differences cannot necessarily be attributed to region as they could instead be attributed to the condition of the advertisement. If they were randomly distributed across country/language then it may not be a confound, but there could still be an effect of advertisement and this should be discussed.

All adverts were used in all countries using the translated versions of the adverts (i.e. they were randomly distributed across country). We have included in the methods:

“ All adverts were used in all countries to an equal extent, though it is not possible to know which advert respondents had seen.” Lines 134-136.

And

“We used four different adverts with slightly different wording in order to attract as many respondents as possible. The different adverts may have attracted different subsets of people, which could lead to biases in results. Though we were unable to determine which adverts respondents saw, all adverts were distributed equally across the three countries and as such are unlikely to lead to differences between countries” Lines 667-671.

Line 119 – do you mean Power analysis (instead of sample size)?

We prefer to use sample size calculation, as power wasn’t used in the calculation.

Additional requirements:

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf.

We have ensured our manuscript meets PLOS ONE’s style requirements.

2. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

We have amended our ethics statement: “Prior to completing the questionnaire, all participants were asked to consent (by selecting yes in a tick box) to their responses being collected, stored, and analysed in an anonymised form for the purpose of reports and publication.” Lines 147-148.

3. Thank you for stating the following in the Competing Interests section: "The authors declare that: A.M.M. and S.H. are employed by VIER PFOTEN International, a global animal welfare organisation; L.M.C has received a research grant from VIER PFOTEN International; and L.M.S.’s research has been funded by VIER PFOTEN International."

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

We confirm that this does not alter our adherence to PLOS ONE policies and have included this in our statement.

14 Feb 2022

Submitted filename: R1_Review_Free Roaming Dogs.docx

Click here for additional data file.

15 Feb 2022

Dear Editor,

We are very grateful to both you and the reviewers for your feedback on the manuscript. We have addressed each of the comments, as outlined below:

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

We have reviewed the reference list to ensure it is complete and correct. We have added some references with newly included text, outlined below.

Additional Editor Comments (if provided):

This paper presents a questionnaire study carried out in three countries: Ukraine, Italy and Bulgaria. However, the author list represents research institutions in Italy and the UK and international organisations based in Austria and Belgium. Please comment on the absence of Bulgarian and Ukrainian researchers as authors of the paper. Also, please justify the choice of the three countries for the questionnaire study.

VIER PFOTEN International is an international organization with country specific offices within Bulgaria and Ukraine. We consulted with VIER PFOTEN International employees in Bulgaria and Ukraine to understand the challenges in dog population management specific to these countries, and to ensure the questionnaire was appropriate. Whilst we acknowledge these individuals in the acknowledgement section, their contribution did not merit authorship.

We have now included justification of the choice of the countries in the introduction: “We selected these focal countries due to the networks established with collaborating organisations (VIER PFOTEN International and Istituto Zooprofilattico Sperimentale dell’Abruzzo e del Molise “Giuseppe Caporale”; IZSAM) that provided local knowledge to facilitate data collection. The focal countries are culturally and environmentally distinct, allowing comparison of the collected data between different countries within Europe.”

Statistical analysis: I’m not familiar with the methods you have used for the statistical analysis, but it seems to me that a justification should be presented for the use of different approaches (Bernoulli logistic regression versus ordinal probit models) depending on which outcome variables you analysed.

We have included further detail in the statistical analysis section (i.e. that Bernoulli logistic regression are for binary outcome variables, whereas ordinal regression are used for ordered categorical outcome variables).

Results section overall: I appreciate your detailed and comprehensive presentation of the outcome of the statistical analysis, but for the paper to be accessible to the wider readership that the study topic merits, ideally this presentation should be complemented with a more straightforward summary of the results. Please consider summarizing in words the most important associations that you found, potentially as a first paragraph of the Results section. Your paper is likely to be of interest to practitioners in dog population management (e.g. shelter directors, veterinary officials), and it is important that they are able to read and understand the main findings!

We have now included a subsection “Summary of statistical associations” which reads:

“Respondents were less likely to answer that they neutered their dog(s) and more likely to answer that they allow their dog(s) to roam if they identified as (i) male, (ii) religious, (iii) owning dogs for practical reasons, (iv) young, and (v) having no schooling or primary education. Respondents were more likely answer that an increase in free-roaming dogs should be prevented if they identified as (i) female, (ii) feeling threatened by free-roaming dogs, (iii) older, and (iv) having more education. Below we report the detailed statistical findings.”

219-313 For readability, consider only including the percentage of responses, and not the n=x of y, in the written description of the results. You are also presenting these results in figures and supplementary material, and in particular the supplementary material is the best option for the additional detail.

We initially had only the percentages in the text when we first submitted the paper, but, as we included the numbers (n=x of y) as requested by reviewer 1 in the first round of feedback from our submission to this journal, we would like to keep this addition to increase the accuracy of information.

Lines 307 Is the last sentence of the figure legend not redundant?

Respondents could choose multiple answers to some of the questions in the questionnaire, whilst in others they could only choose one. We would like to indicate which results were multi answer questions, so would like to keep this sentence for clarity.

Line 303 Have you used the term CNR earlier and defined it? If it is – as a I think – used here for the first time it should be written out rather than using the abbreviation.

Thank you for spotting this – we’ve now included the full term in the introduction when we first mention reproductive control.

The authors have addressed most of my comments, concerns, and questions. However, there are still a few issues with wording that should be corrected for clarity.

Abstract: “Assessing dog population management interventions is important to determine their long-term impact. It is essential to also determine how the attitudes and dog ownership practices within local communities may influence the efficacy of dog population management” (lines 18-21)

The phrasing and positioning of these statements still suggest to me that population management was directly assessed in this study. The statements should either a) be removed entirely or b) be edited and moved to the end of the abstract and discussed as potential future research.

We have now removed these sentences.

Line 95: The red text below needs to be deleted from any methods or sections that talk about what the study did/aimed to do, because it is misleading. You absolutely can talk about future research or how your data could be used to hypothesize on the efficacy of dog population management programs, but that should be in the discussion section.

This study assesses how the attitudes and dog ownership practices within local communities may influence the efficacy of dog population management, by gauging attitudes towards the presence of free-roaming dogs and of dog ownership practices in three European countries – Bulgaria, Italy, and Ukraine.

We have now removed this and checked throughout to ensure readers are not misled.

Submitted filename: Response_to_reviewers.docx

Click here for additional data file.

17 Feb 2022

Public attitudes towards free-roaming dogs and dog ownership practices in Bulgaria, Italy, and Ukraine

PONE-D-21-15770R2

Dear Dr. Smith,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

I Anna S Olsson, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

21 Feb 2022

PONE-D-21-15770R2

Attitudes towards free-roaming dogs and dog ownership practices in Bulgaria, Italy, and Ukraine

Dear Dr. Smith:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. I Anna S Olsson

Academic Editor

PLOS ONE