Introduction

Over the last three decades, wolves (Canis lupus) have recolonized several European regions now heavily transformed by humans (Cimatti et al., 2021). Reintegrating wolves in human-dominated landscapes is a challenging endeavour (Pettersson et al., 2021), because fear for personal safety, concerns about losing traditional ways of life, and economic precarity can give rise to negative attitudes and intolerance towards the species. However, negative perceptions extend beyond the local level. Following an incident involving a wolf attack on her family’s pony, European Commission’s president von der Leyen commented on the perceived risk wolves pose stating that “the concentration of wolf packs in some European regions has become a real danger for livestock and potentially also for humans” (European Commission, 2023). These remarks show how policy decisions can be greatly swayed by prevailing, but often biased, human perceptions. Notably, this incident has prompted discussions on potentially reviewing the wolf’s conservation status in the EU rather than on addressing livestock protection issues.

The Portuguese wolf population (Canis lupus signatus) is part of the Northwest Iberian population, which constitutes the southwestern edge of European wolf range. During the twentieth century, its abundance drastically declined, mostly due to human persecution, lack of prey, and habitat loss (Blanco & Cortés, 2009; Torres & Fonseca, 2016). Now occurring at only 20% of its original range, the Portuguese wolf population is divided into two subpopulations: a larger and more stable one north of the Douro River, ranging across the northwestern region of the Peninsula and connected with the northern Spanish population; and a smaller one inhabiting the region south of the Douro, which shows great instability and no connectivity with the rest of the Iberian population (Torres & Fonseca, 2016). Although protected by law since 1988 and listed as “Endangered” in the Portuguese Red Data List of Mammals (Pimenta et al., 2023), wolves are still heavily persecuted by humans. Between 1999 and 2019, the wolf mortality monitoring system implemented by the official conservation agency (Instituto da Conservação da Natureza e das Florestas) recorded a total of 22 dead wolves in the southern subpopulation, whose size ranged from 18 to 49 wolves at the time of the last published census (Pimenta et al., 2005). Most wolves (73%) died due to human-related causes, including car accidents, shootings, snares, and poison (Barroso & Pimenta, 2008). Wolf persecution is an outcome of the conflict caused mostly by livestock depredation (Pimenta et al., 2018), but ancestral beliefs and narratives portraying wolves negatively also persist, fuelling fear of the species (Álvares et al., 2011). The belief that today’s wolves are different from those of the past due to captive wolves being released into the wild, and the assumption that wolves appearing in areas where they have been absent until recently constitute evidence of their reintroduction, are modern-day myths often perpetuated by local people and sometimes even amplified by the media (Espírito-Santo, 2007; Rainha et al., 2018). Contrary to these misconceptions, wolf reintroduction has never been carried out in Europe (Blanco & Sundseth, 2023). Nevertheless, these myths, fears, and beliefs influence peoples’ attitudes towards wolves and jeopardize conservation efforts (Torres et al., 2020).

Socio-demographic factors such as age, gender, and education level are well-known correlators of public attitudes towards carnivores (Williams et al., 2002). But factors such as knowledge, fear, and perceived costs and benefits of wolf presence, while also modulated by socio-demographic drivers, may have a greater influence over attitudes (Kansky & Knight, 2014). Attitudes mainly consist of affective (i.e., emotive response) and cognitive (i.e., beliefs and opinions) components (Frank et al., 2016). The affective component is based on past experience with the stimulus (i.e., wolves) and perceptions of its benefits and costs, which can be tangible (e.g., economic benefits from ecotourism, economic costs from livestock attacks) or intangible (e.g., role in ecosystem, aesthetic/symbolic value, fear) (Bruskotter & Wilson, 2014; Kansky & Knight, 2014). Positive affective responses relate to higher perceived benefits and lower perceived costs (Bruskotter & Wilson, 2014). Also, the extent to which people are exposed to and informed about a species influences both the cognitive and affective components, which can affect attitudes and thus species conservation (Karlsson & Sjöström, 2007; Arbieu et al., 2019).

There is limited availability of published data on the human dimension in wolf-related studies in Iberia. Most findings are confined to grey literature, such as theses, project reports, and conference communications not readily accessible to the scientific community. In this context, Rosalino et al. (2023) identified human dimensions as one of the research priorities for Iberian wolf conservation. Twenty years have passed since the last published human dimension study focusing on our subject population (Espírito-Santo, 2007). During this time, the socio-demographic context of the region has changed. Most municipalities have suffered population loss and report an increased ageing index, although the average schooling level has increased in the last decades. Supported by the history of Human-Wolf coexistence and social dynamics, we contribute new information about attitudes towards wolves inhabiting the area south of the Douro River. We developed a semi-structured survey to assess local people’s attitudes, fears, and knowledge about wolves, as well as their perceptions of costs and benefits of wolf presence and conservation. Specifically, we test the following hypotheses: (i) attitudes (H1), fear (H2), and knowledge (H4) vary across interest groups, peoples’ socio-demographic characteristics and exposure to wolves; (ii) fear is associated with knowledge about wolves and knowledge of recent wolf attacks on livestock (H3); (iii) attitudes are primarily driven by fears and perceptions of costs and benefits (H5) (Table 1).

Table 1 Hypotheses, rationale, and predictions regarding attitudes, fear, and knowledge levels toward the Iberian wolf population inhabiting the region south of the Douro River (Portugal)
Full size table

Methodology

Study Area

The study area comprises the wolf range south of the Douro River, central-north Portugal. We first selected municipalities inside and bordering wolf range (N = 37), and conducted a cluster analysis using k-means clustering algorithm to assign them to one of three clusters based on the: (i) proportion of the resident population older than 66 years (INE, 2022); (ii) proportion of the resident population with only primary level of education (INE, 2022); (iii) regular presence of wolves (Pimenta et al., 2005). These clusters were defined to encompass different socio-economic contexts (See Suppl. Material). We then randomly selected 17 municipalities for our survey (Fig. 1).

Fig. 1
figure 1

Study area depicting the 17 municipalities randomly selected for survey implementation, and number of surveys collected in each parish

Full size image

Survey Design and Data Collection

The survey consisted of 25 items divided into five constructs: attitudes, fear, knowledge, benefits, and costs (Suppl. Material, Table S1). We did not differentiate between tangible and intangible benefits, but as fear is a strong determinant of attitudes, we analysed it separately from tangible (i.e., economic) costs. We calculated a mean score for each of these constructs. We included some of the same attitudinal, beliefs, and knowledge items used in previous human dimension studies in Portugal (Espírito-Santo, 2007; Torres et al., 2020). Responses to the first 23 items were organized on 5-point Likert scale ranging from ‘‘strongly disagree’’ to “strongly agree,” with a “no opinion” option in the middle. The last two items pertained to changes in the respondents’ attitude in the past years (24), and receptiveness to a larger future wolf population in their municipality (25). These allowed us to understand respondents’ self-perception of their attitude and to identify possible discrepancies between it and their attitudinal score. We were also able to compare receptiveness to a larger wolf population across attitudes and between areas of regular and irregular/absent wolf presence. We included a section related to socio-demographic characteristics, in which data collection was restricted to gender, age, school education, and whether the respondent is a hunter and/or owns livestock. We also inquired if the respondent knew someone who suffered from an attack on livestock in recent years (up to five years) and, in the case of livestock owners if they had suffered a recent attack. We did not previously identify or specifically target any interest group. Interviewees were guaranteed anonymity in the handling and storing of their data and were assured that their answers could not be traced back to them. Only those who agreed to participate were interviewed. We targeted people over 20 years of age who had their primary residence in the 17 selected municipalities. Interviews were mostly face-to-face to ensure a higher response rate and allow for clarifications if necessary (Espírito-Santo, 2007), and were conducted using physical mediums (paper and pen) by the same interviewer who wrote down the answers. Some surveys were handed out in local coffee shops for collection when completed. This study was approved by the Ethics Committee for the Collection and Protection of Scientific Data of the University of Lisbon, Portugal (ref. CEC/38/2023).

Socio-Demographic Data

Gender was coded as binary (M/F), and age of respondent was recoded into four discrete classes (Age_class: 20–29, 30–44, 45–65, 66 + years old). Education level (School) was recoded into six classes: NoSchool, i.e., no school education; School1, School2, School3 relating to the 1st, 2nd, and 3rd cycle of primary education, respectively; Highsch high school; and SUP university degree. Residence was categorised at the parish level, based on Espírito-Santo’s (2007) classification and the latest available census (INE, 2022) (i.e., rural: < 50 residents/km2 & < 500 residents; semi-rural: ≥ 50 residents/km2 & ≥ 500 residents; semi-urban: ≥ 100 residents/km2 & ≥ 2000 residents; urban: ≥ 500 residents/km2 & > 5000 residents). When population density and size did not match, we prioritised population size to characterise residence type. We categorised wolf presence (Wolf_regpres) at the parish level as a binary variable (Y – regular presence of wolves; N – irregular presence or absence of wolves). Knowledge of a wolf attack in recent years (Know_attack) was categorised as binary (Y/N). For livestock owners, having suffered from a wolf attack in recent years (Had_attack) was categorised as a factor with three levels (Y/N/DN; DN – does not know if wolf or dog).

Constructs

We conducted a reliability analysis using the Cronbach’s alpha coefficient (Field, 2009), which ranges from 0 to 1, and measures the degree to which items in our survey determine their respective psychological construct. Alpha values ranging from 0.7 to 0.80 indicate that the items are acceptable for measuring their respective constructs (Field, 2009), but values below 0.7 are also to be expected when analysing psychological constructs (Kline, 1999). So, we considered a minimum threshold of 0.65. Additionally, we examined the “item-total correlation” to evaluate the strength of the relationship between each item and the total score of the construct, and the “alpha if-item-deleted” to investigate the impact of removing each item on the overall reliability of the construct. Items with low correlation (r < 0.3) and those that when removed significantly increased the overall alpha coefficient were flagged to be dropped from the analysis (Field, 2009). We used the alpha function in the “psych” R package (Revelle, 2023). Then, we derived scores of each construct using different aggregation methods. For attitudes, we computed the mean score across items which rendered an average sentiment directly interpretable on the original 5-point Likert scale. For other constructs, a cumulative approach was chosen to capture the overall intensity of the construct and to align with previous studies, ensuring that our findings are directly comparable (see Supplementary Material for greater detail).

Data Analysis

We explored our hypotheses 1, 2 and 4 of attitudes, fear and knowledge varying across groups and socio-demographic characteristics by comparing their respective scores using Kruskal-Wallis and Wilcoxon signed-rank tests for pairwise comparisons. We looked for a possible “Not In My Backyard” (NIMBY) effect by comparing answers to items 1 and 2 (Suppl. Material Table S1) across interest groups, using a paired sample t-test to assess if respondents scored differently for the two questions. Also, to understand how attitudes relate to receptiveness to a future wolf population increase we employed an ordinal logistic regression using Att_score and Wolf_regpres as predictors, and answers to item 25 as response variable (five ordered categories from “disappear” to “increase and expand”, see Suppl. Material Table S1). We investigated the interaction between predictors to determine if the relationship between attitudes and views on population increase differed based on the regularity of wolf presence in the area. Analyses were performed using the polr function, with cumulative logit link, from the MASS package in R (Venables & Ripley, 2002). To test our hypothesis 3 of an association between fear and knowledge we used Kendall Rank Correlation Coefficient, and for association with knowledge of an attack we used Wilcoxon signed-rank test. Finally, we explored which factors drive attitudes (hypothesis 5) of the general public and livestock owners, by means of a multiple regression analysis (normal distribution and identity link function) using backward selection (Zuur et al., 2007). The hunters group was not analysed due to the limited sample size which compromised a reliable assessment of normality. We first standardised all continuous variables (e.g., Att_score_std). Model ability to represent data variability was assessed by the Akaike Information Criteria (AIC) and AICc corrected for small samples (livestock owners), and by the adjusted R2 (Zuur et al., 2007). Model selection was performed based on the ΔAIC/AICc score (i.e., the best model is the one with the lowest ΔAIC/AICc score; Burnham & Anderson, 2002) and predictors influence through CI95% of its model coefficient (Suppl. Material Tables S4 and S5). We used p < 0.05 as the significance level criterion for all statistical tests.

Results

We retrieved a total of 220 answered surveys: 191 from personal interviews and 29 from self-completion. Of the total, 175 were from the public, 13 from hunters, and 32 from livestock owners (see Supplementary Material).

Attitudes towards Wolves

The strong positive correlation (r > 0.73) between respondents’ self-perception of their attitudes and the derived attitudinal scores indicates a high level of consistency, suggesting the robustness of our scoring method.

Attitudes Across Groups (H1)

Results showed significant differences in attitudinal scores across interest groups (χ² = 26.60, df = 2, p < 0.01), corroborating our first hypothesis. Livestock owners presented significantly lower attitudinal scores than the public and hunters’ groups (p < 0.01), which showed mostly positive attitudes towards wolves (Public \(\:\stackrel{-}{x}\) = 3.7 ± 0.9; Hunters \(\:\stackrel{-}{x}\) = 3.8 ± 0.6) while livestock owners showed negative attitudes (\(\:\stackrel{-}{x}\) = 2.7 ± 0.9) (Fig. 2). Livestock owners with positive attitudes towards wolves (attitudinal score greater than 3.3) were generally younger (average age of 44 years, range 25–57), with higher levels of education, and perceived more benefits than costs of wolf presence (Suppl. Material Fig. S1).

Fig. 2
figure 2

Mean, median, and interquartile range of attitudinal, fear, and knowledge scores, per interest group

Full size image

Attitudes Across Socio-Demographic Characteristics (H1)

We did not find significant differences in attitudes between genders (χ² = 1.91, df = 1, p = 0.16), but attitudes among age classes differed significantly between older and younger respondents (χ² = 34.85, df = 3, p < 0.01). Mean attitudinal scores were all on positive, yet the oldest respondents exhibited lower scores (Age 66 + years \(\:\stackrel{-}{x}\) = 3.2 ± 1.0; Age 45–65 years \(\:\stackrel{-}{x}\) = 3.4 ± 0.9), whilst the youngest (20–29 and 30–44 years) presented scores of 4.1 ± 0.6. Attitudes between the lower and higher levels of education (i.e., No school – Highschool/University; 1st cycle – Highschool/University; 2nd cycle - Highschool/University; and 3rd cycle - Highschool/University) were found to be significantly different (χ² = 40.39, df = 5, p < 0.01). People with no school education presented the lowest attitudinal scores (\(\:\stackrel{-}{x}\) = 2.6 ± 0.5), those with only primary school showed neutral to positive attitudes (1st cycle \(\:\stackrel{-}{x}\) = 3.2 ± 0.9; 2nd cycle \(\:\stackrel{-}{x}\) = 3.1 ± 0.8; 3rd cycle \(\:\stackrel{-}{x}\) = 3.5 ± 0.9), and respondents who completed high school or had a university degree showed the most positive attitudes (Highschool \(\:\stackrel{-}{x}\) = 4.0 ± 0.7; University \(\:\stackrel{-}{x}\) = 4.1 ± 0.6). We found that attitudes differed significantly between respondents living in rural areas and those living and semi-rural and semi-urban settings (χ² = 21.27, df = 3, p < 0.01). Mean attitudinal scores were all on the positive side of the scale: people from rural areas presented mean scores of 3.3 ± 0.9, whilst those living in semi-rural and semi-urban areas presented scores of 3.8 ± 0.9. We found a significant difference in attitudinal scores between areas of regular and irregular presence/absence (χ² = 12.73, df = 1, p < 0.01): people from areas of irregular/absent wolf presence presented lower attitudinal scores (\(\:\stackrel{-}{x}\) = 3.3 ± 0.8) than those from areas of regular wolf presence (\(\:\stackrel{-}{x}\) = 3.7 ± 0.9).

Not in My Backyard Effect

We found statistically significant differences in means between answers to items 1 and 2, among the public and livestock owners. Hunters’ sample size was too limited to draw any conclusion. There is a general acceptance of wolves inhabiting the country, expressed by the mean neutral to positive answers provided to item 1 (Suppl. Material Table S2). In contrast, acceptance decreases when the scenario focuses on respondents’ own municipalities, as seen in item 2’s means. Livestock owners exhibited the largest difference in mean scores between the two questions (mean difference = 1.16), indicating a more pronounced difference in acceptance of wolves living freely in the country versus in their municipality.

Relationship between Attitudes and Receptiveness to a Larger wolf Population

We found a significant positive association between attitudes and receptiveness to a larger population (Estimate = 1.89, t = 7.49, p < 0.01; Suppl. Material Table S3). Living in areas of irregular/absent wolf presence (Wolf_regpres N) was associated with a significant decrease in the likelihood of preferring a larger wolf population (Estimate = -4.59, t = -3.71, p < 0.01). However, the positive interaction term between attitudes and areas of irregular/absent wolf presence suggests that positive attitudes play a major and significant role in increasing people’s receptiveness (Suppl. Material Table S3). In these areas, as attitudes become more positive the odds of being in a higher category of receptiveness increase at a rate that is 1.59 units higher than in areas of regular wolf presence (Fig. 3, and Suppl. Material Table S3).

Fig. 3
figure 3

Views on future wolf numbers in the respondent’s municipality and regarding current wolf presence

Full size image

Fear of Wolves (H2, H3)

Results showed significant differences in fear scores across interest groups (χ² = 6.47, df = 2, p = 0.03), corroborating our second hypothesis. Hunters showed significantly lower fear scores (\(\:\stackrel{-}{x}\) = 9.1 ± 3.8) than the general public and the livestock owners (p < 0.05), which presented a moderate fear of wolves (Public \(\:\stackrel{-}{x}\) = 11.5 ± 4.3; Livestock owners \(\:\stackrel{-}{x}\) = 12.8 ± 5.0) (Fig. 2). As for gender differences, results confirm our prediction that women show more fear of wolves than men (χ² = 8.15, df = 1, p < 0.01) (Women \(\:\stackrel{-}{x}\) = 12.8 ± 4.6; Men \(\:\stackrel{-}{x}\) = 10.8 ± 4.1). Contrary to our prediction, we did not find significant differences between education levels (χ² = 5.14, df = 5, p = 0.39), nor any significant association with knowledge about wolves (Kendall’s tau = -0.09, p > 0.10) or with knowledge of wolf attacks (χ² = 0.41, df = 1, p > 0.50). No other socio-demographic characteristic presented significant differences regarding fear levels.

Knowledge about Wolves (H4)

In accordance with our prediction, all groups presented very low knowledge scores. We found significant differences only between the public and livestock owners (χ² = 3.97, df = 1, p < 0.05). Public and hunter respondents scored the lowest (Public \(\:\stackrel{-}{x}\) = 1.7 ± 1.0; Hunters \(\:\stackrel{-}{x}\) = 1.8 ± 1.1), whilst livestock owners scored slightly higher (\(\:\stackrel{-}{x}\) = 2.0 ± 1.0) (Fig. 2). In accordance with our prediction, we found significantly different knowledge levels between genders (χ² = 13.34, df = 1, p < 0.01), with men presenting higher scores (\(\:\stackrel{-}{x}\) = 1.9 ± 0.9) than women (\(\:\stackrel{-}{x}\) = 1.4 ± 1.1). No other socio-demographic characteristic presented significant differences regarding knowledge levels. Of all knowledge-related items, the one which got most incorrect answers was item 15 (related to the myth of wolf reintroduction), with over 96% of respondents failing to answer it correctly (Suppl. Material Fig. S4).

Factors Driving Attitudes towards Wolves (H5)

In the stepwise model, dropping the Knowledge variable but retaining fear, benefits, and costs, resulted in a simpler model, as good as the Full in terms of ΔAIC and Adjusted R2 (Suppl. Material Table S4) and thus the best for explaining the variation of the general public’s attitudes towards wolves (Table 2). We could determine the direction of the influence of these factors as their 95% confidence interval of the coefficients did not contain 0. Fear and costs scores were negatively associated with attitudes, whilst benefits score was positively associated. These results are in accordance with our fifth hypothesis, in which we expected attitudes to be better explained by fear of wolves and perception of benefits and costs, than by knowledge (H5, Table 1).

Table 2 Predictors that best explain the variability in attitudinal scores of the general public. Coefficients (coef), standard errors (SE), significance value [Pr (>|z|)], and 95% confidence intervals (CI 95%). Predictors whose CI95% do not include 0 are in bold. Adjusted R2 = 0.69
Full size table

Similarly, the best model for livestock owners was also the stepwise (Suppl. Material Table S5) and included benefits, costs, and wolf presence as significantly influential variables (Table 3). Costs and irregular presence/absence of wolves (expressed in the intercept) were negatively associated with attitudes, whilst benefits and regular wolf presence were positively associated.

Table 3 Predictors that best explain the variability in attitudinal scores of livestock owners. Coefficients (coef), standard errors (SE), significance value [Pr (>|z|)], and 95% confidence intervals (CI 95%). Adjusted R2 = 0.85
Full size table

Discussion

Our study highlights the importance of addressing the human dimension for the effective conservation of an endangered large carnivore species in a coexistence scenario, as different groups perceive unique challenges that require tailored solutions (Bath, 2009). Our findings confirm that the sampled population generally shows neutral to positive attitudes towards wolves - a pattern commonly observed in Portugal (Espírito-Santo, 2007; Milheiras & Hodge, 2011; Torres et al., 2020; Valente et al., 2024) and other European regions (Majić & Bath, 2010; Gosling et al., 2019; Žunna et al., 2020). However, variation exists within this general pattern, supporting the notion that attitudes towards wolves are context-specific and not governed by a one-size-fits-all solution.

Attitudes towards Wolves

Despite the general neutral trend, we found significant differences in attitudes between interest groups. Hunters were more positive, a pattern previously detected in northern Portugal (Milheiras & Hodge, 2011). This is likely because, in our study area, wolves are not currently seen as competitors for game species, such as wild boar (Sus scrofa). The general public also showed average positive attitudes, which are probably related to the lack of direct contact with wolves and wolf-related issues (Torres et al., 2020; Valente et al., 2024). As expected, livestock owners displayed the most negative attitudes, with mean values similar to those found in other regions of the country (Milheiras & Hodge, 2011; Espírito-Santo, 2014). Other authors attributed this negative stance to wolf depredation on livestock (Torres et al., 2020). However, according to our analysis, having suffered from an attack or having knowledge of its recent occurrence had no significant effect over attitudes. We agree with Espírito-Santo (2007) that the perceived loss rather than actual loss is probably more influential in shaping attitudes. The continuous negative perception of this interest group might come from deep-rooted beliefs and narratives, or from the fear of threats to their way of life (Pettersson et al., 2021). When exploring a possible “Not In My Backyard (NIMBY) effect in our sampled population, there were significant differences in attitudes towards wolves living freely in the country versus wolves living freely in the respondent’s municipality for the general public and livestock owners. The attitudes of the general public, were positive regardless of the scenario. In contrast, livestock owners exhibited a more pronounced difference in mean scores between the two scenarios. They demonstrated a negative mean attitude towards the scenario of wolves living freely in their municipalities, indicating a possible NIMBY sentiment.

Fear of Wolves

Fear levels ranged from low to moderate across groups, but were especially low in the hunters group, which is in accordance with past research (Røskaft et al., 2003; Espírito-Santo, 2014). Although the average fear score was highest in the livestock owners group (a pattern in accordance with Torres et al. (2020); and value similar to Espírito-Santo (2014), it was still moderate. Overall, women experienced higher levels of fear than men, although still moderate in magnitude, which is in accordance with previous findings (Zimmermann et al., 2001; Røskaft et al., 2003; Espírito-Santo, 2007; Torres et al., 2020). Men generally show a preference for predators, whilst women tend to fear them (Røskaft et al., 2007). Nevertheless, the difference found might be artificial as it is often more socially permissible for women to admit fear (Røskaft et al., 2003). Contrary to our prediction, we did not find a significant correlation between fear and knowledge about wolves, which contrasts with a previous study conducted in our study area (Espírito-Santo, 2007), but is consistent with recent research from Valente et al. (2024) and another study conducted north of the Douro (Espírito-Santo, 2014). We believe this pattern might be linked to the overall low knowledge levels or to the type of knowledge items surveyed.

Knowledge about Wolves

Our prediction of low knowledge levels across groups was confirmed. There were significant differences between the public and livestock owners, with the latter presenting higher levels. This contradicts previous results where knowledge levels were highest in hunters (Espírito-Santo, 2014; Torres et al., 2020), but it is likely due to the limited size of our sample of hunters. Nevertheless, these results align with our prediction that groups with frequent contact with nature (livestock owners and hunters) score higher than those without, a pattern previously described in Portugal (Espírito-Santo, 2007; Torres et al., 2020; Valente et al., 2024). Our findings support the idea that knowledge about wolves, and likely biodiversity in general, is not gained through formal education at any level. As predicted, men exhibited greater knowledge about wolves than women. Most respondents answered all questions incorrectly except those pertaining to solitary habits of wolves animal and their protection status under Portuguese law (Suppl. Material Figs. S2 and S6). In fact, only one respondent (0.45%) correctly answered all items. Only 3% of respondents correctly dismissed the myth of wolf reintroduction in the country: This pervasive belief continues to undermine trust and exacerbate conflicts between livestock owners, wolf scientists, and conservationists.

Drivers of Attitudes towards Wolves

We found that attitudinal differences among interest groups relate to distinct drivers, despite some being shared. We confirmed our prediction of knowledge not being an important driver of attitudes of the general public and livestock owners, which is in line with previous findings (Torres et al., 2020; Valente et al., 2024). We argue that knowledge of ecological facts does not always translate into positive attitudes (Heberlein, 2012). Nonetheless, the observed below-average knowledge levels in our study might have further contributed to its lack of influence. Living in areas of regular wolf presence was found to be an important factor in shaping livestock owners’ attitudes, suggesting that continued coexistence with the species influences their views (Espírito-Santo & Petrucci-Fonseca, 2017). Apart from this, attitudes of livestock owners appear to be largely influenced by their perception of benefits and costs associated with sharing the landscape with wolves. Naturally, as they depend on animal farming, livestock owners are economically impacted by attacks on their livestock.

The public respondents recognized costs and benefits of the presence of wolves, but also expressed a negative attitude of fear (see also Espírito-Santo, 2007; Torres et al., 2020; Valente et al., 2024). We also found that individuals who perceive higher tangible costs related to wolf presence (e.g., negative economic impacts on livestock owners or financial costs associated with wolf conservation) tend to hold significantly lower attitudinal scores. Conversely, those who recognise the benefits of wolf presence and conservation, such as their role in regulating the ecosystem, or aesthetic and symbolic values, exhibited higher attitudinal scores. Slagle et al. (2012) highlighted that support for wolf recovery in the United States was closely tied to the perceived benefits. In our case, benefits seem to drive more favourable attitudes, which in turn are significantly associated with a higher receptiveness to increasing wolf numbers. Notably, this association was particularly pronounced in areas where wolves are either irregularly present or absent. Despite average attitudes being close to neutral in these areas, the most positive respondents are more receptive to an increase in wolf numbers than the most positive respondents from areas of regular wolf presence. This probably occurs because people from the former areas have not experienced or heard of conflicts with wolves for a long time (Dressel et al., 2015). Respondents from regular wolf presence areas showed preference for wolf numbers to remain stable (more than 50% of respondents stated they would prefer wolf numbers to “stay as they are”). In a similar line of reasoning, these people probably have a better understanding that conflicts could arise from a larger wolf population. Similar to Bisi et al. (2007), the strongest negative views, suggesting a preference for wolf disappearance, came predominantly from areas where wolf presence is irregular or absent, denoting very polarized views. In conclusion, while the baseline receptiveness of wolf populations in areas of irregular/absent presence is lower than in areas of regular presence, the impact of positive attitudes in raising this receptiveness is much stronger in the former.

Conclusions and Management Considerations

Our results suggest that the way knowledge about wolves is conveyed to local populations should be reevaluated, as evidenced by not just the persistence of the decades-old myth of wolf reintroduction but also the low knowledge level among the respondents. Management actions should continue to target the factors contributing to negative attitudes by actively listening to the concerns and complaints of all interest groups, including conservationists. Understanding the key issues and proposed solutions from each group is crucial for finding common ground and ultimately achieving an agreement that gathers the support of all interested parties. This collaborative approach is essential for promoting sustained coexistence (Bath, 2009; Rainha et al., 2018). Of course, efforts should not be limited to these actions alone. It could also be beneficial to focus on imparting more practical insights about wolf ecology (e.g., activity patterns) to help people minimise encounters and livestock damages. Promoting initiatives among all interest groups, but especially livestock owners, to address the role of free-ranging dogs in attacks on livestock is also pivotal (Lino et al., 2023), and it encourages collaboration and works towards minimizing conflicts. Although the neutral attitudinal pattern can be seen with some optimism, it can still be improved. The seemingly indifferent Portuguese population might be effectively reached through untargeted interventions, such as TV documentaries and fact-based coverage in news outlets. The media holds a responsibility in shaping public perceptions and can be a valuable ally for providing accurate information and revealing misconceptions. Fostering intergenerational learning through children can also be an important and effective strategy often leading to intentional and behavioural changes (Damerell et al., 2013).

As we have shown, attitudes are not solely driven by fear and perception of costs, but are also positively influenced by the perception of benefits. Espírito-Santo and Petrucci-Fonseca (2017) reported that most livestock owners surveyed in their study informally questioned about the benefits of having wolves in the wild, denoting a possible link between a better perception of benefits and increased acceptance. We share their idea of a communication strategy targeting common myths and addressing the ecological and economic benefits of wolf presence. Leveraging the positive aspects associated with wolf presence in the territory is an opportunity that both generates economic revenue for local communities, through nature-based economies such as tourism, and increases support of wolf conservation. This is pivotal in areas where wolves are still not regularly seen, such as the eastern and southeastern municipalities of our study area. As we have shown, fostering positive attitudes in these areas could significantly increase people’s receptiveness to wolf presence in the future. Finally, we suggest that these awareness and communication campaigns should be mindful of the schedule constraints tied to traditional gender roles, such as household responsibilities, to ensure increased participation from women.