Knowledge, attitudes, and practices regarding animal welfare among pet owners in Vietnam Authors: Bui Hai Phong, Y Pham Van To Nhu, Nguyen Thi Huong Giang, Tran Xuan Minh Published: 2026 ABSTRACT Background:Animal welfare has become an increasingly important social and veterinary issue regarding companion animals. However, the existing literature on the knowledge, attitudes, and practices of pet owners in Vietnam related to animal welfare is limited.Aim:This study aimed to assess the knowledge, attitudes, and practices of pet owners regarding animal welfare and how these factors relate to one another within the Knowledge, Attitudes, and Practices framework.Methods:A cross-sectional survey was conducted among 297 pet owners using a structured questionnaire with a five-point Likert scale.Results:The mean scores for knowledge, attitudes, and practices related to animal welfare were moderate (3.11–3.19). Knowledge was moderately correlated with attitudes (r = 0.509, p < 0.01), whereas attitudes were strongly correlated with practices (r = 0.645, p < 0.01). Regression analysis identified knowledge as a positive predictor of attitudes (β = 0.509, p < 0.001) and practices (β = 0.407, p < 0.001). When attitudes were included in the model, they emerged as the strongest predictor of practices (β = 0.590, p < 0.001), whereas the direct effect of knowledge on practices decreased but remained statistically significant (β = 0.107, p = 0.039). Mediation analysis indicated that attitudes partially mediated the relationship between knowledge and practices (indirect effect = 0.301; 95% CI: 0.226–0.378).Conclusion:The largest share of the work in animal welfare improvement rests on the education of their owners in a veterinary practice environment. There is a need to improve knowledge and attitudes to encourage sustainable and responsible animal care. Introduction Animal welfare (AW) has been gaining growing international concern over the past years. Animal welfare is currently viewed as both an ethical issue and an important factor in sustainable livestock production and social development (Keeling et al., 2019). The Five Freedoms are widely regarded as a global standard for assessing animal welfare. These include freedom from hunger and thirst; discomfort; pain, injury, and disease; freedom to express normal behavior; and freedom from fear and distress (McCulloch, 2013). Along with economic growth and rapid urbanization, the number of pets in Vietnam has increased significantly in recent years. According to survey data, only about 56% of Vietnamese citizens regard dogs and cats as family members (TGM Research, 2025). The young generation is especially susceptible to this trend, which was exacerbated by the COVID-19 pandemic, when the necessity of having people around and feeling their emotional support was more tangible (TGM StatBox, 2024). An estimated 12,000,000 dogs and cats were found in Vietnam, comprising 6,480,000 dogs and 5,580,000 cats. The number of dogs and cats in Vietnam is projected to reach 16,000,000 by 2027 (Nguyet-Minh, 2024). Public interest has also increased online, with searches related to dogs and cats increasing from 1,000,000 in 2019 to 1,900,000 in 2021 (Tung, 2021). The sudden rise in the number of pets and their widespread popularity in society raise new issues concerning the physical health and psychological well-being of animals, as well as the conditions of their living environments. Understanding pet owners and their knowledge, attitudes, and practices (KAP) is essential for creating educational initiatives, welfare interventions, and animal protection policies based on evidence (Grigg and Kogan, 2019; Philpotts et al., 2019; Forrest et al., 2023; Siano et al., 2024). Previous research on animal welfare has shown that knowledge of animal welfare is a mandatory requirement for developing positive attitudes and proper care practices (Balzani and Hanlon, 2020; Alemayehu et al., 2022). However, increased knowledge and positive attitudes do not always translate into improved welfare practices (Suharto et al., 2025). Demographic variables, such as gender, age, level of education, and pet experience, may also influence animal welfare behavior (Wahyuwardani et al., 2025). In recent years, Nghe An province has experienced noticeable socio-economic changes, including increased urbanization and a gradual shift from traditional rural lifestyles to more urban lifestyles. These transformations may influence patterns of companion animal ownership and people’s perceptions of animal welfare, as pets are increasingly kept for companionship rather than functional purposes. Although several studies in Vietnam have examined companion animals in relation to rabies and other zoonotic diseases (Mai et al., 2011; Dang and Nguyen, 2025; Phu et al., 2025; Truong et al., 2025), systematic research focusing on pet owners’ knowledge, attitudes, and practices regarding animal welfare remains limited, particularly in central regions such as Nghe An province. This study addresses this gap by providing empirical evidence on welfare-related KAP beyond disease-oriented perspectives. Furthermore, previous studies have usually not applied an integrated analytical framework that combines the KAP approach with behavioral theories. This study analyzes the KAP framework along with the Theory of Planned Behavior (TPB), facilitating a more thorough comprehension of the behavioral mechanisms that inform animal welfare practices. In addition, mediation analysis is employed to explore the potential indirect pathways through which knowledge may influence practices through attitudes, particularly in the context of how urbanization and socio-economic changes in Nghe An affect the understanding and implementation of animal welfare practices by pet owners. Thus, the aims of the research were as follows: (1) To evaluate the knowledge, attitudes, and practices of animal welfare among companion animal owners of Nghe An province; (2) To determine correlations between companion animal care and welfare knowledge, attitudes, and practices. Materials and Methods Study area The research was conducted between May and November 2025 in six wards located in Nghe An Province, Vietnam (18.675°N, 105.66°E) (Fig. 1). The selected wards cover approximately 166.25 km² and have a population of more than 568,700. These wards were selected for their relatively high pet ownership rates and representation of urban, peri-urban, and semi-urban areas. This diversity in socioeconomic conditions, education levels, and access to veterinary services allowed for a broader assessment of the knowledge, attitudes, and practices of pet owners regarding animal welfare. Location of the study areas. Study design The study employed a cross-sectional mixed-methods design to evaluate the KAP of pet owners regarding animal welfare in Nghe An Province and to examine the relationships among these variables (Fig. 2). The quantitative component consisted of a structured survey administered to 297 participants, whereas the qualitative component involved semi-structured in-depth interviews with 15 purposively selected participants to contextualize and enrich the interpretation of the survey findings. Eligible participants were individuals aged 15 years who currently owned at least one companion animal (including dogs, cats, ornamental birds, or ornamental fish) and were directly involved in daily care activities. Individuals who were not primary caretakers or who submitted incomplete questionnaires were excluded from the final analysis. All participants were provided with written information describing the study objectives and confidentiality procedures. Participation was voluntary, and withdrawal was permitted at any time without consequences. Conceptual framework of KAP regarding animal welfare among pet owners in Nghe An Province. The arrows indicate the hypothesized relationships among the variables. H1 examines the influence of demographic factors on knowledge, attitudes, and practices. H2 (+) indicates a positive relationship between knowledge and attitude. H3 (+) represents knowledge’s positive effect on practice. H4 (+) indicates a positive effect of attitude on practice. The questionnaire was developed based on the KAP framework, integrated with the TPB and the Five Freedoms of Animal Welfare, as defined by the World Organization for Animal Health (Table 1). The questionnaire items were developed based on previous studies on animal welfare and KAP research involving animal owners (Forrest et al., 2023; Liang et al., 2024). The draft questionnaire was reviewed by three experts in veterinary medicine and animal welfare research to ensure content validity. Minor revisions were made to improve clarity and contextual relevance. Before the main survey, a pilot test was conducted with 20 pet owners to assess the questions’ comprehensibility and relevance. Feedback from the pilot test was used to refine the wording and structure of the questionnaire. It was divided into 4 sections as follows: Table 1.Description of items used to assess pet owners’ knowledge, attitudes, and practices regarding animal welfare.Item codeItem descriptionScaleKKnowledge of animal welfare1 = Not knowledgeable at all; 2 = Very limited knowledge; 3 = Moderate knowledge; 4 = Good knowledge; 5 = Extreme knowledgeK1Understanding the “Five Freedoms” as basic animal welfare standardsK2Recognizing that pets can experience a range of emotional statesK3Identifying abnormal behaviors in pets indicating stress or illnessK4The importance of vaccination and regular veterinary check-upsK5Understanding Vietnamese regulations related to animal welfareK6Knowing that neutering/spaying improves pet welfare and healthK7Creating a species-appropriate living environmentK8Knowing the behavioral indicators used to assess animal welfareK9Understanding the negative effects of long-term confinement on mental healthAAttitudes toward animal welfare1 = Strongly disagree; 2 = Disagree; 3 = Neutral; 4 = Agree; 5 = Strongly agreeA1Considering pets as family membersA2Feeling responsible for the physical and mental health of petsA3Pets have the right to live in an appropriate environmentA4Willingness to change living conditions to meet pet needsA5Supporting the Enactment of Animal Welfare Legislation in VietnamA6Willingness to spend additional money to improve pet welfareA7Supporting strict penalties for animal abuseA8Feeling concerned when pets show signs of distressA9Not abandoning pets even under financial difficultiesA10Animal welfare should be taught in schools and communitiesPPet care and welfare practices1 = Never; 2 = Rarely; 3 = Sometimes; 4 = often; 5 = AlwaysP1Providing adequate nutritional food at appropriate timesP2Providing sufficient clean drinking water dailyP3Regular cleaning of pets’ living areasP4Spending time with and interacting with petsP5Observing behavior to detect stress or diseaseP6Regular veterinary visits or vaccinationsP7Allowing free movement or walking dailyP8Providing toys, sleeping areas, and an appropriate living environmentP9Allowing pets to express their natural behaviorsP10Not confining pets in small spaces for more than 4 consecutive hoursNote: K1–K9 denote knowledge items, A1–A10 denote attitude items, and P1–P10 denote animal welfare practice items. All items were assessed using a 0 Likert scale, with separate response options for the knowledge, attitude, and practice components. Description of items used to assess pet owners’ knowledge, attitudes, and practices regarding animal welfare. Note: K1–K9 denote knowledge items, A1–A10 denote attitude items, and P1–P10 denote animal welfare practice items. All items were assessed using a 0 Likert scale, with separate response options for the knowledge, attitude, and practice components. Section A Demographic characteristics and pet ownership profiles, such as gender, age, educational level, occupation, income, type of pets kept, total number of pets kept, length of time of pet ownership, and the main reason for keeping the pets. Occupational categories were classified based on the respondents’ primary employment type (e.g., government officer/civil servant, office employee, and self-employed). Section B Knowledge of animal welfare (9 items). Knowledge was measured using a five-point Likert scale ranging from 1 = not at all knowledgeable to 5 = extremely knowledgeable. Topics to be discussed include understanding the Five Freedoms, recognizing pets’ emotions and abnormal behaviors, health care knowledge, legal considerations, and understanding pets’ environmental needs. Section C Attitudes toward animal welfare (10 items), measured on a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree. This section included affective attachment to pets, a sense of responsibility in taking care of the pets, and support of the legislation and community education on animal welfare. Section D Practices concerning animal welfare (10 items), measured on a five-point Likert scale from 1 = never to 5 = always. Topics covered included nutrition and water supply, hygiene and housing, interaction and exercise, enrichment, health, and clinical sign monitoring. The mean scores were grouped into intervals of classes to classify the levels of knowledge, attitudes, and practices, assuming that the distances between the Likert-scale points were equal, as implied by Alkharusi (2022). Since the five-point Likert scale was between 1 and 5, the width of the classes was as follows: Class width = (5–1)/5 = 0.80(1) The KAP levels were coded as follows: L1 = very low (1.00–1.80), L2 = low (1.81–2.60), L3 = moderate (2.61–3.40), L4 = high (3.41–4.20), and L5 = very high (4.21–5.00). This approach was consistently applied across all three KAP components for comparison, enabling easy results interpretation. Determination of sample size and sampling procedure The sample size was calculated using the formula proposed by Thrusfield et al. (2018). In the absence of prior studies in the study area, the expected proportion was assumed to be P exp = 0.50. We applied a 95% confidence level (Z = 1.96) and an absolute precision of d = 6%. The following formula was used to calculate the required sample size: where n is the required sample size, P exp is the expected proportion, and d is the desired absolute precision. Substituting the values into the following formula: The sample size was adjusted using Cochran (1977) correction formula to account for a potential non-response rate of 10%: The final sample size for the study was set at 297 pet owners. A two-stage sampling procedure was used. First, six representative wards in Nghe An Province were selected by stratified random sampling by geographic area. Second, convenience sampling was used to recruit participants from veterinary clinics, pet shops, public parks, and online pet-owner communities on social media within these wards to ensure a sufficient sample size (n = 297). Although this approach allowed access to a diverse group of respondents, sampling bias may be introduced, which should be considered when interpreting the results. Data collection Data were concurrently collected using quantitative and qualitative approaches. For the survey component, 297 valid questionnaires were obtained. Of these, 178 responses were collected through face-to-face administration conducted by interviewers, whereas 119 responses were obtained through an online survey platform distributed via social media groups dedicated to companion animal caretakers. To assess potential mode effect bias, differences in overall knowledge, attitude, and practice scores between face-to-face and online respondents were statistically examined at the 0.05 significance level to evaluate the comparability of the two data collection modes. The qualitative component involved 15 semi-structured interviews purposively selected from survey respondents who agreed to participate in the follow-up. Selection ensured variation in age, occupation, type of companion animal, and overall knowledge level, categorized as low, moderate, or high based on survey scores. The interviews explored pet ownership motivations, animal emotional and physical needs perceptions, vaccination and disease treatment attitudes, feeding and housing practices, and perceived barriers to proper animal welfare. The interviews were audio-recorded with consent, transcribed verbatim, and analyzed using deductive content analysis guided by the KAP framework. Two researchers independently coded the transcripts, discussed discrepancies, and achieved substantial inter-coder agreement. Qualitative findings were used to contextualize and interpret quantitative results, particularly the mediating role of attitudes in the relationship between knowledge and welfare practices. Data analysis Data were coded, cleaned, and filtered using Microsoft Excel and then imported into SPSS version 26 for statistical analysis. The characteristics of the participants and their KAP levels regarding animal welfare were summarized using descriptive statistics, including frequencies, percentages, and mean values. Cronbach’s alpha and corrected item-total correlations were used to determine the internal consistency of the KAP scales. Construct validity was evaluated using an exploratory factor analysis with the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy, Bartlett’s test of sphericity, and eigenvalues exceeding 1. The correlations among knowledge, attitudes, and practices were examined using Pearson correlation coefficients and linear regression. The mediating effect of attitudes between knowledge and practices was also evaluated using PROCESS Macro version 4.2 (Model 4) with 5,000 bootstrap samples and a 95% confidence interval. Ethical approval This study was reviewed and approved by the Animal Research Ethics Committee of the Institute of Agriculture and Natural Resources, Vinh University, Vietnam (Approval No. 05/QD-NNTN-DHV/2025, dated 15 May 2025). Results Demographic characteristics and ownership profile of the pet Table 2 shows the demographic characteristics and pet ownership profiles of the 297 participants. Among the participants, 68.0% were female, and 32.0% were male. The majority of respondents were aged 18–25 years (30.0%) and 26–35 years (24.9%), with the rest aged 36–50 years (20.2%) and above 50 years (19.9%). Regarding education, most respondents (69.7%) had a university education, while 14.1% had a lower level of education. Students had the largest percentage of occupation (37.7%), followed by civil servants/public employees and self-employed workers (22.2%). Monthly income was concentrated mostly in the < 114.20 USD/month (33.7%) and 266.60–571.40 USD/month (32.7%) groups. Turning to the experience of keeping pets, almost half of all respondents (47.1%) kept pets for 6 months to 2 years, and 64.0% of respondents cited companionship or leisure as the primary reason for keeping pets. Table 2.Demographic characteristics and pet ownership profiles of the study participants (n = 297).CharacteristicCategoryFrequency (n)Percentage (%)GenderMale9532.0Female20268.0Age group< 18 years155.118–25 years8930.026–35 years7424.936–50 years6020.2> 50 years5919.9Educational levelBelow the high school4214.1High school graduate51.7Vocational/College82.7University degree20769.7Postgraduate3511.8Occupation*Student11237.7Government officer/civil servant6622.2Office employee289.4Self-employed6622.2Business owner196.4Homemaker31.0Architect31.0Monthly income**< 114.2 USD/month10033.7114.2–266.6 USD/month268.8266.6–571.4 USD/month9732.7> 571.4 USD/month7424.9Pet ownership experience< 6 months4816.26 months to 2 years14047.12–5 years4816.2> 5 years6120.5Purpose of Pet OwnershipCompanionship/Recreation19064.0Guarding/Security5618.9Commercial purposes155.1Other3612.1Type of the pet***Dogs17244.9Cats16041.8Ornamental birds184.7Ornamental fish277.0Note: *Occupational categories such as “government officer/civil servant” and “office employee” were defined based on the employment type of the respondents. **Monthly income was converted from VND to USD using an exchange rate of 1 USD = 26,265 VND. Percentages and monetary values were rounded to one decimal place. ***Participants could own more than one type of pet; therefore, the total number of pets exceeded 297. Demographic characteristics and pet ownership profiles of the study participants (n = 297). Note: *Occupational categories such as “government officer/civil servant” and “office employee” were defined based on the employment type of the respondents. **Monthly income was converted from VND to USD using an exchange rate of 1 USD = 26,265 VND. Percentages and monetary values were rounded to one decimal place. ***Participants could own more than one type of pet; therefore, the total number of pets exceeded 297. Knowledge of animal welfare The reliability analysis revealed that the animal welfare knowledge scale demonstrated very high internal consistency, with a Cronbach’s alpha of 0.889 (Table 3). The corrected item-total correlation coefficient of all items was above 0.5, which is a sufficient contribution of single items to the overall scale. In addition, the Cronbach’s alpha values remained low when any item was deleted, indicating that all items were retained in the scale. Table 3.Reliability and descriptive statistics of the knowledge scale on animal welfare (n = 297).ItemItem descriptionMeanSDCITCα if item deletedCronbach’s αK1Understanding the “Five Freedoms” as basic animal welfare standards3.131.2430.5630.8830.889K2Recognizing that pets can experience a range of emotional states3.131.1840.6070.879K3Identifying abnormal behaviors in pets indicating stress or illness3.161.1820.6140.879K4The importance of vaccination and regular veterinary check-ups3.161.2180.7000.872K5Understanding Vietnamese regulations related to animal welfare3.181.2180.6190.878K6Knowing that neutering/spaying improves pet welfare and health3.111.2300.6370.877K7Creating a species-appropriate living environment3.151.2240.6940.872K8Knowing the behavioral indicators used to assess animal welfare3.141.1910.6480.876K9Understanding the negative effects of long-term confinement on mental health3.151.1970.7010.872Note: CITC = corrected item–total correlation; SD = standard deviation. Mean scores were calculated using a 0 Likert scale, ranging from 1 (not at all knowledgeable) to 5 (extremely knowledgeable). Mean values around 3 indicate a moderate level of knowledge, whereas values above 4 indicate a high level of knowledge regarding animal welfare. Cronbach’s α indicates the knowledge scale’s internal consistency reliability. Reliability and descriptive statistics of the knowledge scale on animal welfare (n = 297). Note: CITC = corrected item–total correlation; SD = standard deviation. Mean scores were calculated using a 0 Likert scale, ranging from 1 (not at all knowledgeable) to 5 (extremely knowledgeable). Mean values around 3 indicate a moderate level of knowledge, whereas values above 4 indicate a high level of knowledge regarding animal welfare. Cronbach’s α indicates the knowledge scale’s internal consistency reliability. Based on the predefined classification scale, the level of knowledge among pet owners about animal welfare yielded mean scores of 3.11–3.18, indicating a moderate level of knowledge. The questions on legal provisions concerning animal welfare, the significance of vaccination and regular veterinary visits, and the implications of long-term confinement on the psychological well-being of pets demonstrated a higher mean score than the other questions. Conversely, the mean scores for knowledge about the benefits of neutering for animal welfare and health, as well as general knowledge of the Five Freedoms, were lower, indicating gaps in owners’ knowledge in these domains. The relatively broad range of item scores suggests substantial variation in the knowledge levels of participants. Animal welfare attitude The reliability tests revealed that the animal welfare attitude scale had very high internal consistency, with a Cronbach’s alpha of 0.897 (Table 4). The item-correlation coefficients were also corrected ≥0.5, indicating that all items were strongly related to the scale. These results confirm the reliability of the attitude scale. Table 4.Reliability and descriptive statistics of the animal welfare attitude scale (n = 297).ItemItem descriptionMeanSDCITCα if item deletedCronbach’s αA1Considering pets as family members3.151.1730.6310.8870.897A2Feeling responsible for the physical and mental health of pets3.161.1780.6640.885A3Pets have the right to live in an appropriate environment3.151.2380.7040.882A4Willingness to change living conditions to meet pet needs3.181.2080.6040.889A5Supporting the Enactment of Animal Welfare Legislation in Vietnam3.161.1970.6250.888A6Willingness to spend additional money to improve pet welfare3.121.1960.6410.887A7Supporting strict penalties for animal abuse3.151.2000.6600.885A8Feeling concerned when pets show signs of distress3.151.1990.5900.890A9Not abandoning pets even under financial difficulties3.161.1940.6690.885A10Animal welfare should be taught in schools and communities3.131.2150.6300.887Note: CITC = corrected item–total correlation; SD = standard deviation. Mean scores were calculated using a 0 Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Mean values of approximately 3 indicate a moderate level of agreement with statements related to animal welfare attitudes. Cronbach’s α indicates the internal consistency of the Attitude scale. Reliability and descriptive statistics of the animal welfare attitude scale (n = 297). Note: CITC = corrected item–total correlation; SD = standard deviation. Mean scores were calculated using a 0 Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Mean values of approximately 3 indicate a moderate level of agreement with statements related to animal welfare attitudes. Cronbach’s α indicates the internal consistency of the Attitude scale. The mean score of the attitude items ranged from 3.12 to 3.18, indicating a moderate attitudinal profile according to the predetermined classification scale. Statements about responsibility for the physical and psychological well-being of pets, readiness to change the living conditions to suit the needs of pets, and opinions about the right of pets to live in the proper environment received a higher mean score than other statements. On the contrary, there was lower agreement with statements about willingness to spend more money to improve pet and animal welfare and support animal welfare education in schools and communities. Altogether, the findings suggest that pet owners in Nghe An province generally favor animal welfare, although not strongly. Animal welfare practices The internal consistency of the animal welfare practice scale was high, with a Cronbach’s alpha of 0.897 (Table 5). The item-total correlation coefficients for all items were above 0.5, indicating that the observed variables were adequate for the overall scale. These findings indicate the satisfactory reliability of the practice scale. Table 5.Reliability and descriptive statistics of the practice scale on animal welfare (n = 297).ItemItem descriptionMeanSDCITCα if item deletedCronbach’s αP1Providing adequate nutritional food at appropriate times3.141.1830.5960.8900.897P2Providing sufficient clean drinking water daily3.171.2160.6710.885P3Regular cleaning of pets’ living areas3.111.2180.6980.883P4Spending time with and interacting with pets3.151.2060.6510.887P5Observing behavior to detect stress or disease3.191.2150.6710.885P6Regular veterinary visits or vaccinations3.121.1860.5640.892P7Allowing free movement or walking daily3.151.2180.6280.888P8Providing toys, sleeping areas, and an appropriate living environment3.171.2190.6840.884P9Allowing pets to express their natural behaviors3.151.1900.6560.886P10Not confining pets in small spaces for more than 4 consecutive hours3.131.2080.6090.889Note: CITC = corrected item–total correlation; SD = standard deviation. The mean scores were calculated using a 0 Likert scale ranging from 1 (never) to 5 (always). Mean values of approximately 3 indicate a moderate frequency of recommended animal welfare practices. Cronbach’s α indicates the practice scale’s internal consistency reliability. Reliability and descriptive statistics of the practice scale on animal welfare (n = 297). Note: CITC = corrected item–total correlation; SD = standard deviation. The mean scores were calculated using a 0 Likert scale ranging from 1 (never) to 5 (always). Mean values of approximately 3 indicate a moderate frequency of recommended animal welfare practices. Cronbach’s α indicates the practice scale’s internal consistency reliability. The average scores of the practice items ranged from 3.11 to 3.19, indicating a moderate level of animal welfare practices. Behavioral observation practices, such as identifying stressors or illnesses, providing daily adequate and clean drinking water, and establishing proper toys, sleeping, and living conditions, received comparatively high average scores. Conversely, lower mean scores were recorded when the practices required more time, money, or adjustments to current habits, such as regular veterinary visits and vaccinations, frequent cleaning of living spaces, and reduced access to long-term confinement in small areas. Exploratory factor analysis of the main study scales was performed The KAP scales were analyzed using Exploratory Factor Analysis (EFA) to independently evaluate the factor structure and item convergence (Table 6). All three scales were above the KMO measures, and the sphericity test was statistically significant (p < 0.001), indicating that the data were highly suitable for factor analysis. Table 6.Exploratory factor analysis of the knowledge, attitude, and practice scales (n = 297).ScaleKMOBartlett’s testExtracted factorsEigenvalueVariance explained (%)Factor loading rangeKnowledge0.941< 0.00114.61451.270.673–0.780Attitude0.949< 0.00115.25752.570.692–0.770Practice0.941< 0.00115.00850.080.664–0.774Note: To evaluate the unidimensionality of the Knowledge, Attitude, and Practice scales, exploratory factor analysis was conducted using principal component analysis (PCA). Factor retention was determined based on eigenvalues >1 and factor loadings ≥ 0.60. Exploratory factor analysis of the knowledge, attitude, and practice scales (n = 297). Note: To evaluate the unidimensionality of the Knowledge, Attitude, and Practice scales, exploratory factor analysis was conducted using principal component analysis (PCA). Factor retention was determined based on eigenvalues >1 and factor loadings ≥ 0.60. The EFA identified one factor with an eigenvalue of 4.614 for the knowledge scale, accounting for 51.267% of the total variance. The factor loadings ranged from 0.673 to 0.780, indicating good convergence between the observed variables. The attitude scale yielded a single factor with an eigenvalue of 5.257, accounting for 52.569% of the total variance. The factor loadings for all items were >0.69, indicating a clear and consistent scale structure. In the practice scale, the EFA identified a single-factor solution with an eigenvalue of 5.008, accounting for 50.083% of the total variance. The factor loadings ranged from 0.664 to 0.774, which is within the acceptable range for retaining items. In general, the EFA outcomes indicate that the KAP scales exhibit acceptable levels of convergent validity and unidimensionality, making them suitable for further analyses, such as Pearson correlation, linear regression, and hypothesis testing. Correlations among knowledge, attitudes, and practices related to animal welfare The correlation analysis, conducted using Pearson’s method, revealed positive, statistically significant linear relationships among the three components of the KAP model associated with animal welfare (Table 7). The correlations between knowledge and attitudes (r = 0.509, p < 0.01) and between knowledge and practices (r = 0.407, p < 0.01) were moderate. Practices were also highly associated with attitudes (r = 0.645, p < 0.01). Because all the correlation coefficients were positive and significant, omitting any variables in further regression analyses was not possible. These correlations provide empirical support for further examination of the relationships among knowledge, attitudes, and practices in the proposed research model. Table 7.Pearson correlation matrix between knowledge, attitude, and practice regarding animal welfare (n = 297).VariableKnowledge (K)Attitude (A)Practice (P)Knowledge (K)10.509**0.407**Attitude (A)0.509**10.645**Practice (P)0.407**0.645**1Note: All correlations are positive and statistically significant. **Indicates significance at p < 0.01 (two-tailed), suggesting moderate to strong relationships among knowledge, attitudes, and practices regarding animal welfare. Pearson correlation matrix between knowledge, attitude, and practice regarding animal welfare (n = 297). Note: All correlations are positive and statistically significant. **Indicates significance at p < 0.01 (two-tailed), suggesting moderate to strong relationships among knowledge, attitudes, and practices regarding animal welfare. Regression analysis and attitudes’ mediating role The direct linkages between KAP and animal welfare were tested using linear regression (Table 8). The findings revealed that animal welfare knowledge had a positive, statistically significant influence on the attitudes of pet owners (Model 1: β = 0.509, p < 0.001), explaining 25.9% of the variance in attitudes. Knowledge also directly affected animal welfare practices (Model 2: β = 0.407, p < 0.001). Attitudes had a strong, positive impact on practices (Model 3: β = 0.645, p < 0.001). When knowledge and attitudes were input into the model at once, attitudes were the key predictor of practices (Model 4: β= 0.590, p < 0.001), and the direct impact of knowledge on practices was reduced, yet still significant (β = 0.107; p = 0.039), supporting the possible mediating role of attitudes. Table 8.Regression and mediation analysis of the relationships among KAP (n = 297).ModelDependent variableIndependent variable(s)Bβt p -valueR²95 percent CI(1)Attitude (A)Knowledge (K)0.5120.50910.162< 0.0010.259[0.413; 0.611](2)Practice (P)Knowledge (K)0.4070.4077.658< 0.0010.166[0.302; 0.512](3)Practice (P)Attitude (A)0.6410.64514.483< 0.0010.416[0.554; 0.728](4)Practice (P)Attitude (A)0.5870.59011.479< 0.0010.424[0.486; 0.688]Knowledge (K)0.1070.1072.0720.039[0.005; 0.208]Indirect effect (a × b)Practice (P)K → A → P0.3010.038–––[0.226; 0.378]Note: B = unstandardized regression coefficient; β = standardized regression coefficient; CI = 95% confidence interval. The mediation analysis was conducted using PROCESS Macro v4.2 (Model 4) with 5,000 bootstrap samples. Model 4 indicates that attitudes partially mediate the relationship between knowledge and animal welfare practices. The indirect effect was considered statistically significant when the 95% bootstrap confidence interval excluded 0. Regression and mediation analysis of the relationships among KAP (n = 297). Note: B = unstandardized regression coefficient; β = standardized regression coefficient; CI = 95% confidence interval. The mediation analysis was conducted using PROCESS Macro v4.2 (Model 4) with 5,000 bootstrap samples. Model 4 indicates that attitudes partially mediate the relationship between knowledge and animal welfare practices. The indirect effect was considered statistically significant when the 95% bootstrap confidence interval excluded 0. The indirect effect of knowledge on practices via attitudes was statistically significant, as determined through mediation analysis using the PROCESS Macro with 5,000 bootstrap samples (indirect effect = 0.301; 95% CI: 0.226–0.378). These results indicate that attitudes partially mediate the relationship between knowledge and animal welfare practices among pet owners. Discussion This study demonstrates that the knowledge and attitudes of pet owners regarding animal welfare in Nghe An province are moderate to moderately high, indicating a growing awareness of the physical, psychological, and behavioral needs of companion animals. However, these levels were not fully reflected in animal welfare practices. Such findings suggest a gap between knowledge and actual practices, particularly for behaviors that require additional time, financial resources, or changes in daily routines, such as routine veterinary check-ups, complete vaccination, and improvements in pets’ living conditions. Previous KAP studies have reported similar patterns, indicating that positive knowledge and attitudes do not always translate into adequate animal welfare practices (Wambui et al., 2018; Rian Hari et al., 2025; Wahyuwardani et al., 2025). The sociodemographic characteristics of the respondents may also help explain the observed patterns in KAP. In this study, a large proportion of the respondents were young adults and students, and most had university-level education. Younger pet owners may be more exposed to information about animal welfare through digital media, educational institutions, and online communities, which may contribute to the moderate levels of knowledge and awareness identified in this study. However, younger individuals and students often have limited financial resources, which may restrict their ability to fully implement the recommended welfare practices, particularly those that require regular veterinary care or additional expenditures. Previous studies have also shown that age, education level, and socioeconomic conditions can influence attitudes toward animal welfare and the quality of animal care practices (Balzani and Hanlon, 2020; Alemayehu et al., 2022). Contextual factors in the study area may also contribute to the knowledge–practice discrepancy. Veterinary services remain limited in many locations, particularly in the peri-urban and semi-urban areas of the Nghe An province. The uneven distribution of veterinary clinics, behavioral counseling services, and preventive health care programs may limit the ability of pet owners to translate knowledge into practical actions. Similar findings have been reported in other developing countries, where access to veterinary services and infrastructure strongly influences animal care practices (Lanas et al., 2018; Tariku et al., 2025). Financial constraints may represent another important barrier to the implementation of recommended animal welfare practices. Although many pet owners expressed positive attitudes toward animal welfare, the costs of veterinary care, vaccinations, sterilization, and appropriate nutrition may limit the extent to which these practices. Economic factors have been identified as determinants of animal care behavior in developing regions, including Southeast Asia. Limited financial resources and the cost of veterinary services may prevent pet owners from accessing necessary care, contributing to neglect or abandonment in some circumstances (Rah and Choi, 2023; Dogbey et al., 2024; Kitole, 2025). In addition, broader socioeconomic challenges, such as unemployment, unstable income, or inadequate housing conditions, may further influence the quality of pet care (Dogbey et al., 2024; McDowall et al., 2024). These conditions, combined with limited policy support and affordable veterinary services, may widen the gap between awareness and practical care behaviors (Holden, 1999). Therefore, multidimensional policy approaches, including financial support mechanisms, community education programs, and accessible veterinary services, have been suggested to reduce economic barriers and improve animal welfare outcomes (Benson and Tincher, 2024; Cvetkovic, 2025). In addition, animal welfare may sometimes be interpreted in a limited sense, focusing primarily on basic care, such as food and shelter, rather than a comprehensive approach that includes behavioral and psychological needs, as reported in previous KAP studies in Vietnam (Mai et al., 2011; Sinclair and Phillips, 2019). The results also highlight the potential importance of community veterinary services in addressing the KAP gap. Community-based veterinary initiatives, particularly those offering low-cost or free services, can improve access to veterinary care for socioeconomically disadvantaged households while simultaneously promoting responsible pet care practices (Mueller et al., 2018; Kogan et al., 2021). These services may serve as an important bridge between policymakers, veterinary professionals, and pet owners, facilitating sustainable improvements in companion animal welfare, especially in resource-limited settings (Lavallee et al., 2017; Niemiec et al., 2024). When combined with vaccination campaigns and mobile veterinary clinics, education programs focusing on nutrition, disease prevention, and routine healthcare have been shown to improve companion animal health (Mueller et al., 2018). Furthermore, community veterinary services can strengthen trust between local communities and veterinary professionals, encouraging pet owners to seek health care services more proactively (Aguirre Siliezar et al., 2025). Therefore, these approaches may help address the KAP gap identified among pet owners in Nghe An province. Community veterinary programs must be flexible, responsive to cultural circumstances, and incorporate human behavior change models to increase animal welfare and pet abandonment success (Park et al., 2021; Powdrill-Wells et al., 2025). Finally, the mediation analysis confirms that attitudes play a central role in the relationship between knowledge and animal welfare practices. Although knowledge directly influences practices, this effect weakens when attitudes are included in the model, indicating that attitudes partially mediate the relationship between knowledge and practices. In other words, knowledge alone may not be sufficient to translate into behavioral change unless it is accompanied by positive attitudes toward animal welfare and a sense of responsibility toward companion animals. This finding is consistent with the TPB, which suggests that behavioral change is not driven solely by knowledge but also by individuals’ attitudes toward the behavior. Similar observations have been reported in other KAP studies on animal welfare, where attitudes are identified as a key factor in translating awareness into practice (Alemayehu et al., 2022; Suharto et al., 2025). In the Vietnamese context, where socioeconomic constraints also influence pet care practices, the partial mediating effect of attitudes suggests that interventions that address both knowledge and attitudes while also supporting practical enabling conditions are required to improve animal welfare. Conclusion The results indicate that pet owners in Nghe An Province demonstrate relatively favorable levels of knowledge and attitudes toward animal welfare; however, these do not always translate into corresponding welfare practices. This gap highlights the need for practical interventions tailored to the local context to improve companion animal welfare. Strengthening community education, expanding accessible community veterinary services, and improving supportive local policies may promote responsible pet ownership and enhance companion animal welfare.