Public Perception of Drinking Water Quality in an Arsenic-Affected Region: Implications for Sustainable Water Management

Abstract

This work explores the relationship between water quality and public trust in their water supply, in an arsenic-affected region of Serbia. The results from an online public survey are presented and subjected to Pearson’s correlation, cluster analysis, and principal component analysis. In general, survey respondents in settlements with known arsenic issues had a poor opinion on the quality of their tap water. This poor opinion was underlined by their consumption of bottled water, with more than 43% of responders purchasing at least 5 L of bottled water a week. In addition to the great economic cost, the relatively low plastic recycling rate in Serbia means that this also has a very negative effect on the environment, as most of the plastic bottles are sent to landfill, to degrade slowly into microplastics, whilst leaching a variety of chemical contaminants into the surroundings. In the area studied, the poor public opinion of the water quality is a realistic reflection of water at the tap. Although significant, the additional environmental pollution from bottled water consumption should nonetheless be of secondary consideration in comparison to the health risks associated with chemical contaminants in the study area, with local waterworks requiring significant financial assistance if they are to meaningfully improve tap water quality.

1. Introduction

The sustainable management of drinking water resources, both environmentally and economically, is becoming increasingly complex with each passing year. As living standards improve, water consumption correspondingly rises, while climate change is anticipated to have a profound impact on drinking water supply by altering precipitation patterns, intensifying the frequency and severity of droughts, and diminishing the availability of freshwater [1]. Understanding consumer perceptions of drinking water quality is essential for effective water management because public trust and behaviour directly influence water usage, conservation efforts, and the use of less sustainable alternative water sources such as bottled waters [2].

The relationship between public perceptions and environmental quality is built from a complex mixture of past experience, social and cultural factors, demographics, and environmental influences. For drinking water, the reliability of water supply, accessibility, affordability, and the visibility of infrastructure all contribute to the overall perceptions of water quality [3,4]. The public perception of water quality is influenced by the extent of community engagement, institutional transparency, and trust. Communities that are knowledgeable about water quality and have a clear understanding of the management of their water sources are more likely to develop a sense of ownership and trust in their water supply. Conversely, if communities are denied access to the decision-making process and information regarding water quality, trust may diminish, resulting in negative perceptions, even in the absence of actual water quality issues [5]. It is therefore essential to understand the distinct characteristics of different communities and stakeholders when addressing the complexities of water resource management [3]. Scientific communication is facilitated by a deeper understanding of individual environmental concerns. Incorporating these concerns into policy development enhances the likelihood of securing broader public support during implementation [6].

Improving consumer trust in public water supplies is particularly challenging in regions negatively impacted by the presence of arsenic in groundwater sources. In Vojvodina, the northern province of Serbia, many of the groundwater sources are affected by geogenic arsenic contamination [7]. Chronic arsenic exposure is associated with several serious health risks, including cardiovascular diseases and cancers of the skin, bladder, and lungs [8]. Even low levels of long-term exposure have been linked to neurotoxicity, diabetes, and adverse developmental outcomes, underscoring the urgent need for effective risk mitigation strategies [9]. Regional arsenic concentrations can be as high as 300 µg/L in some wells [7], whereas globally, levels above 1 mg/L are not unheard of [10]. Most countries, including Serbia, have adopted the World Health Organization (WHO) recommended maximum concentration of 10 µg/L as their legal maximum [11], and ensuring compliance with this water quality standard in arsenic-affected regions is a public health imperative.

In addition to its organoleptic (taste, odour, and appearance) properties, other factors which influence public perceptions of drinking water quality include prior experiences, education, trust in the water provider, the flow of information within the community, and demographic variables. While aesthetic concerns such as taste, colour, and odour often serve as immediate triggers for consumer dissatisfaction, they may mask more serious but less perceptible risks, such as chronic arsenic exposure. In regions where arsenic contamination exists, limited public awareness of its health effects may lead to the underestimation of chemical risks, while consumer behaviour remains shaped primarily by visible and sensory cues. Some individuals are more likely than others to utilise alternatives such as bottled water as problems persist [12]. It is important to be aware of this divergence during the design of effective risk communication and public health strategies.

Public concerns about tap water contamination, particularly regarding its quality and safety, are among the major drivers behind the global rise in bottled water consumption, a market that has been growing at an annual rate of approximately 5% [13]. In the United States, usage increased consistently each year from 1977 through 2016, while by 2013, China had become the world’s largest bottled water market, reporting a rapid annual growth rate of 14.3% [14]. This upward trend is largely fuelled by consumers’ perceptions that bottled water is safer and more convenient than tap water—a belief reinforced by intensive marketing strategies [15,16]. In many high-income countries, bottled water has become a primary drinking source despite the availability of safe municipal supplies. For instance, 31% of Canadians and 38% of Americans primarily consume bottled water [17,18], while in France it represents 64% of the soft drinks market, despite being significantly more expensive than tap water [19]. This widespread consumption is often attributed to perceived health benefits, improved taste, and aggressive marketing that promotes the purity of bottled water, especially among target groups such as children, athletes, and pregnant women. These perceptions, although sometimes overstated, continue to influence consumer behaviour in the Global North [13]. In contrast, bottled water consumption in low- and middle-income countries is often a necessity driven by inadequate infrastructure, unreliable public water supply systems, and persistent contamination concerns. For example, in Mexico, public trust in municipal water systems plummeted following cholera outbreaks, leading to more than 80% of the population relying on bottled water as their primary source. Similar trends have been reported across Asia, Africa, and Latin America, including in Indonesia, Ghana, Nigeria, and the Philippines, where bottled water usage remains high due to widespread concerns over water quality and safety [13]. In these contexts, bottled water use is not a lifestyle choice but a health-protective response. The role of arsenic as a driver of bottled water consumption is particularly notable. Jakus et al. [20] demonstrated that in several U.S. communities with known arsenic contamination in drinking water, perceived mortality risk from arsenic exposure significantly influenced household spending on bottled water, with average monthly expenditures reaching USD 27. The study found that as perceived arsenic risk increased, so did bottled water purchases, providing quantitative evidence of health-based decision-making. Similarly, in central Florida counties where private wells exceeded the arsenic limit of 10 μg/L, the Florida Safe Water Restoration Program provided bottled water to affected households. Nearly 60% of these households reported using bottled water as their primary source of drinking water, affirming its role as an essential protective measure [21]. These cases highlight that in arsenic-affected areas, bottled water serves not only as a perceived safer option, but also as a practical public health intervention. Although bottled waters are generally safe, they are typically hard mineral waters that may not offer additional health benefits unless an individual is deficient in calcium, magnesium, or similar minerals. Public satisfaction with bottled water remains high, though the financial burden is frequently cited as a limitation [22]. Ultimately, while bottled water provides immediate risk reduction in contaminated areas, overreliance on it may divert public attention and funding away from investments in sustainable, equitable, and resilient public water systems.

In this work, we explore the relationship between public trust in drinking water supply, the quality of available water resources, and the environmental impacts of public water consumption habits, in order to draw out some guidance for future developments in the public water supply sector. By investigating the local perspectives on water quality, we build on the knowledge and experience of the general public, in order to identify key water quality issues and bridge the gap between the water sector, consumers, and academia. These findings are intended to inform context-sensitive policy development and to contribute to broader discussions on sustainable water management, particularly in areas where public trust and environmental health are closely intertwined.

2. The Study Area

This study was conducted in the Autonomous Province of Vojvodina (see Figure 1), an area of 21,614 km² which lies entirely within the geographical confines of the Pannonian basin, site of the Pannonian Sea during the Miocene and Pliocene epochs. Some researchers have reported that sedimentary minerals in this region may average as much as 185 mg/kg As [23]. Arsenic leaches into aquifers through natural processes such as desorption, dissolution, and reduction, particularly in the anoxic conditions common in deeper aquifers. Arsenic concentrations in Vojvodina’s groundwater frequently exceed the World Health Organization (WHO) guideline of 10 μg/L for drinking water [24]. Recorded concentrations range from 50 to 300 μg/L, making Vojvodina one of the most arsenic-affected regions in Europe, with an estimated 600,000 residents (more than one third of the total population of the province) affected [7,25,26]. A recent study identified geogenic sources, combined with agricultural activities and sewage, as the primary contributors to arsenic pollution in parts of Vojvodina [27]. Iron and manganese are also present in this groundwater, which together with the prevalence of arsenic, indicates a reducing environment within the aquifers. This geochemical condition promotes the release of arsenic through the reductive dissolution of Fe/Mn oxyhydroxides. This results in source waters with Fe concentrations ranging up to 21 mg/L (average concentration: 1.1 mg/L; n = 155) and Mn concentrations as high as 0.09 mg/L. The groundwaters in this region also contain significant amounts of natural organic matter and sodium attributed to the plain’s geological history [28]. While the arsenic contamination poses a more severe health risk, iron and manganese notably affect organoleptic water properties. Therefore, this study aims to identify how the different water quality parameters influence public perception.

In addition to the poor groundwater quality, a number of other issues make water supply strategy in Vojvodina challenging, such as the low population density and the widespread agriculture (approximately 1,600,000/2,000,000 ha of arable land). As of the 2022 census [29], the total population of Vojvodina was 1,740,000, of which more than 664,000 (38.2%) live in widely distributed small settlements. Based on data published by the Institute of Public Health of Serbia, a total of 43 public water supply systems in urban areas and 227 in rural areas were monitored in Vojvodina, making a total of 270 public water supply systems [30]. The smaller communities lack both the financial and the human resources necessary to significantly improve tap water quality beyond disinfection, which is usually carried out by sodium hypochlorite. According to the Health Statistical Yearbook of the Republic of Serbia [31] and the previous report by the Institute of Public Health of Serbia [30], 16.9% of analysed samples did not meet the prescribed physicochemical water quality standards. The most common causes included elevated concentrations of iron, manganese, ammonia, nitrates, and KMnO₄ consumption (an indirect measure of the organic carbon content) as well as turbidity and the colour of the water. Perhaps more alarmingly, 5.5% of microbiological water quality samples were non-compliant, mainly due to the presence of aerobic mesophilic bacteria, Pseudomonas aeruginosa, total coliforms, and faecal coliform bacteria [30]. Microbiological contamination is widespread in rural areas where water treatment infrastructure is insufficient or outdated—many of the water supply systems have seen little development since they were originally constructed back in the 1970s. This issue further heightens the health risks associated with drinking untreated or minimally treated groundwater, especially for vulnerable rural populations. It should be noted that in Serbia, public water utilities are administered at a municipal level. Figure 2 therefore presents a summary of the water quality in each municipality in Vojvodina.

The availability, usability, and utilisation of surface water for drinking water supply also come with significant challenges. Although Vojvodina may seem rich in surface waters, the vast majority of the available surface water is transit water, and the very flat topography of Vojvodina yields limited potential for water storage. Furthermore, as already pointed out, the economy of Vojvodina is heavily reliant on agriculture. The Danube and its main tributaries, the Sava and Tisa, play an important role for irrigation. Intensive agricultural production, followed by extensive fertiliser application, contributes to elevated levels of nitrates and phosphates in surface water bodies [32]. Meanwhile, industrial activities and the discharge of untreated wastewater introduce heavy metals and organic contaminants into water sources [33]. One study from 2020 suggests that over 65% of industrial facilities in Vojvodina lack adequate wastewater treatment systems, exacerbating pollution levels [34]. Emerging pollutants, including pharmaceutical residues and contaminants of emerging concern (CECs), have also been detected, posing serious risks to ecosystems and human health. This leaves the rural areas of Vojvodina heavily dependent on the available groundwaters, and the basic water-bearing complex (BWC) in particular, an aquifer at a depth from 50 to 250 m (sometimes even deeper).

It should be noted that shallow groundwater aquifers, indirectly connected to the river through the riverbank filtration (RBF) system, are used for water supply in Novi Sad, the capital of Vojvodina. This is due to favourable hydrogeological conditions in the Danube alluvium, such as the adequate thickness and hydraulic conductivity of the aquifer, the hydraulic connection between the Danube and the aquifer, and the excellent filtering properties of both the riverbed and the aquifer. Unfortunately, suitable RBF systems are only available in certain locations, and significantly increasing their use would require substantial infrastructure investments, in addition to the improved maintenance of existing systems.

3. Materials and Methods

3.1. Methodology of the Public Survey

Consumer perception of tap water quality in the province of Vojvodina was investigated using an online questionnaire. The questionnaire was entirely voluntary, anonymous, and no personal data were requested or stored during the survey. The water consumer questionnaire consisted of 24 questions, including the place of residence of the responders, and the other questions summarised below in

Table 1. Statistical summary of responses to the questionnaire, separated into responses from Novi Sad and outside Novi Sad. Yes/No answers were coded as +1/−1, so that an exact 50:50 split of responses would result in a mean of 0.0.

	Question	Novi Sad (n = 135)				Rest of Vojvodina (n = 144)
		Mean	SD	Min	Max	Mean	SD	Min	Max
Q1	How old are you?	43.26	11.05	26.00	86.00	43.80	12.50	10.00	75.00
Q2	How many members are there in your household?	3.03	2.80	1.00	33.00	3.43	1.10	1.00	8.00
Q3	Do you drink the tap water? (+1: Y, −1: N)	0.53	0.85	−1.00	1.00	−0.06	1.00	−1.00	1.00
Q4	How would you rate the tap water quality (from 1: very bad, to 5: excellent)	3.22	0.95	1.00	5.00	2.21	1.24	1.00	5.00
	I use tap water for (+1: Y, −1: N)
Q5	drinking	0.44	0.90	−1.00	1.00	−0.14	0.99	−1.00	1.00
Q6	cooking	0.97	0.23	−1.00	1.00	0.48	0.88	−1.00	1.00
Q7	cleaning and household maintenance	0.99	0.17	−1.00	1.00	0.94	0.34	−1.00	1.00
Q8	personal hygiene	0.99	0.17	−1.00	1.00	0.91	0.42	−1.00	1.00
Q9	recreational purposes (filling pools)	−0.81	0.59	−1.00	1.00	−0.52	0.86	−1.00	1.00
Q10	watering the garden (irrigation)	−0.62	0.79	−1.00	1.00	−0.11	1.00	−1.00	1.00
Q11	other (e.g., washing cars)	−0.63	0.78	−1.00	1.00	−0.23	0.98	−1.00	1.00
	Are these the reasons you do not drink the tap water (+1: Y, −1: N)
Q12	water colour	−0.96	0.28	−1.00	1.00	−0.44	0.90	−1.00	1.00
Q13	water taste	−0.60	0.80	−1.00	1.00	−0.27	0.97	−1.00	1.00
Q14	water odour	−0.74	0.68	−1.00	1.00	−0.36	0.94	−1.00	1.00
Q15	microbiological contamination	−0.90	0.43	−1.00	1.00	−0.52	0.86	−1.00	1.00
Q16	presence of chemical substances	−0.78	0.63	−1.00	1.00	−0.17	0.99	−1.00	1.00
Q17	other	−0.88	0.48	−1.00	1.00	−0.67	0.75	−1.00	1.00
Q18	Are you familiar with the issue of arsenic in drinking water in your region? (+1: Y, −1: N)	0.03	0.68	−1.00	1.00	0.25	0.72	−1.00	1.00
Q19	Is there an issue with arsenic in the drinking water where you live? (+1: Y, −1: N)	−0.46	0.55	−1.00	1.00	0.11	0.94	−1.00	1.00
Q20	Are you aware of the effects of arsenic on human health? (+1: Y, −1: N)	0.14	0.63	−1.00	1.00	0.21	0.74	−1.00	1.00
Q21	Is the presence of arsenic in drinking water one of the reasons why you do not drink tap water? (+1: Y, −1: N)	−0.22	0.57	−1.00	1.00	0.32	0.71	−1.00	1.00
Q22	What quantity of bottled water do you purchase weekly? (1: <5 L, 2: 5–15 L, 3: >15 L)	1.33	0.80	0.00	3.00	1.78	0.88	0.00	3.00
Q23	Do you have a water purification device? (+1: Y, −1: N)	−0.77	0.64	−1.00	1.00	−0.82	0.58	−1.00	1.00

. It should be noted that the questions were posed in the Serbian language. The questionnaire was carried out as part of a wider market analysis for our NanoCompAs project, which developed an Fe-Mn binary oxide nanocomposite filter medium for arsenic removal. The survey aimed to assess whether the public is aware of arsenic in their drinking water sources, whether they have general concerns about drinking water quality, and whether they have installed or would consider installing a point-of-use water filter to improve their water quality. One of our chief concerns in structuring the survey was to avoid leading our respondents to making claims about having concerns about arsenic before giving their overall assessment of their water quality. For this reason, arsenic was deliberately not even mentioned until the second half of the survey.

The questionnaire was launched for the general public in March 2024 using the LimeSurvey platform (LimeSurvey GmbH, Hamburg, Germany), with the intent of capturing the opinions of people living in arsenic-contaminated areas and in Serbia as a whole. The links to the survey opened an initial landing page containing a disclaimer informing participants that the survey would be used for research purposes only. Given the entirely anonymous and voluntary nature of the survey and the contents of the disclaimer, no prior ethical approval was deemed necessary to carry out this activity.

Below the disclaimer was the button to begin the survey, which created a unique time-stamped entry in the database. LimeSurvey recorded 453 unique entries in this database, leading to a total of 324 completed surveys submitted. Responses were excluded if the respondents (a) were from outside the region of Vojvodina (44 exclusions, mainly from Belgrade); (b) gave no location (1 exclusion); and (c) were less than 16 years old (1 exclusion). This left 135 valid responses from Novi Sad and 144 from the rest of Vojvodina. In order to perform the statistical analysis on the results, the responses to the survey were numerically codified. Where Yes/No questions were asked, the answers were coded as +1 (Y) and −1 (N). Table 1 summarises the responses, sorted into two groups: from Novi Sad, and from the rest of Vojvodina. It should be noted that the age distributions of the responses from Novi Sad and the rest of Vojvodina were statistically very similar.

In Figure 3, the age demographics of the 17 municipalities which were covered by the participants in our survey are presented. This figure highlights several structural challenges for future water supply in Vojvodina: (1) 7 of the municipalities are very small, with fewer than 25,000 inhabitants in total; (2) in 6 of the municipalities, the number of inhabitants in the surrounding villages is greater than the population of the municipal centre, demonstrating a low population density in those municipalities; and (3) in 14 of the municipalities, the “65 and older” demographic is larger than the “less than 19” demographic. According to Eurostat data, Serbia has the 5th highest median age in a generally ageing Europe. Ageing populations put increasing financial strain on the workforce, and smaller economies like Serbia already have less gross domestic product (GDP) to maintain critical systems such a drinking water supply infrastructure [35].

It should be noted that the gender demographics of the population in the municipalities covered by responses to the survey are essentially equal to Serbia as a whole, with the census reporting 48.3% of the population as male and 51.7% as female, within the survey area.

The age distribution of our survey responders is given in the insert of Figure 3 and reflects a certain sample bias which is to be expected in online surveys, which tend to exclude individuals without internet access or digital literacy. As a result, they often overrepresent younger, more educated, or more technologically savvy respondents, leading to non-representative samples. Only 5% of our responders were of retirement age, in comparison to 21% of residents in the municipalities surveyed. In total, 41% of responders were 20–39, and 53% were 40–64. This is a slightly older distribution than the working population in the study area, which breaks down as 24% and 35% of the overall population for those same age groups.

3.2. Statistical Analysis

In order to gain better insight into the factors which drive public perceptions of water quality in a region with known arsenic contamination issues, the responses to our survey from outside Novi Sad were analysed using Version 5.2 of the Past data analysis software (Natural History Museum, University of Oslo, Norway). Novi Sad was excluded as an outlier from our analyses: it is much more populous than the other municipalities in Vojvodina and therefore represents 135 out of a total 279 responses from Vojvodina in our survey. Furthermore, Novi Sad is the only settlement to utilise river bank filtration and does not have any problems with arsenic in the source water. Pearson’s correlation, cluster analysis, and principal component analysis (PCA) were performed. Prior to the cluster analysis and PCA, the data were standardised. These analyses were employed to reduce the complexity of the survey data, identify underlying patterns and enhance interpretability. PCA reduces dimensionality by transforming correlated variables into a smaller set of uncorrelated components, preserving as much variance as possible. Cluster analysis then groups respondents based on similar response patterns, revealing latent segments within the data. Together, these methods enable more efficient data analysis. The cluster analysis was carried out with the hierarchical Unweighted Pair Group Method with Arithmetic Mean (UPGMA) and the Gower similarity index, which is specifically designed for the comparison of datasets consisting of numerical, categorical, and binary questions. Cluster validity was assessed using the cophenetic correlation coefficient, which yielded a value of 0.8474, indicating a good fit between the dendrogram and the original data structure.

4. Results and Discussion

The questionnaire was completed by 324 people over the 12-month period included in this analysis. Respondents from 42 different locations in Serbia participated in the questionnaire, including Novi Sad, and settlements from 16 other municipalities in Vojvodina. For this work, we chose to analyse only those responses from the study area, i.e., those from Vojvodina. The average age of our survey participants was 43.3 in Novi Sad, and 43.8 in the rest of Vojvodina. These numbers are in good agreement with the 2022 Census data, which state that the average age in Vojvodina was 43.6 [29]. As discussed above, Novi Sad is by far the largest settlement in Vojvodina, and the only one to utilise river bank filtration for its water supply. Given the good quality of the water at the tap, the responses from Novi Sad represent an interesting baseline for assessing public perceptions of water quality. More than 75% of participants from Novi Sad drink the tap water (average perception of water quality: 3.22, good to very good), which is in significant contrast to the rest of Vojvodina, where less than 50% of participants drink the tap water (average rating: 2.21, bad to good). The entirety of Novi Sad is supplied from the same waterworks, and although the distribution network is unlikely to be perfect, only slight localised deviations in tap water quality can be expected. The normal distribution of responses to how participants rate the tap water provides a good demonstration of how opinions vary in a population, and contrasts strongly with the rest of Vojvodina, where once again public perceptions of water quality are considerably worse (see Figure 4).

As a first point of comparison, Figure 5 presents the water quality data given in Figure 2, together with the water quality perceptions of our respondents, for each of the municipalities where we received responses. With a few interesting exceptions, public perceptions of water quality seem to quite accurately reflect the quality of the tap water they receive. In Zrenjanin, Kikinda, and Temerin, where the water quality is poor, public perceptions of water quality are low. In contrast, in Novi Sad and Sombor, where water quality is good, public perception largely agrees. Vrbas and Beočin stand out as municipalities where public perceptions are considerably worse than the water quality data would suggest. In the case of Vrbas, this may be due to a heightened sense of environmental pollution brought about by the proximity of the DTD canal, much publicised as one of the most contaminated surface water bodies in Serbia [36]. Although the problems with the DTD canal do not influence the quality of their groundwater, opinions on surface water may well negatively impact drinking water quality perceptions. Kanjiža is also interesting, as a municipality where public perception of water quality is considerably better than the quality reported. Kanjiža is a famous spa town in Vojvodina, and the thermal waters are widely regarded in Serbia as having medicinal properties [37]. It is therefore at least possible that the local perceptions of the drinking water quality are positively inflated despite the poor chemical water quality.

Statistical Analysis of the Responses from the Rest of Vojvodina, Outside of Novi Sad

The qualitative relationships between the responses to our questionnaire were analysed using the Pearson correlation test (Figure 6). The first interesting finding of note is the lack of correlation between the age of our responders and any of the survey questions. In a similar study, Sajjadi et al. [22] reported a correlation between age and satisfaction with water quality, suggesting that as residents live longer, their observations of improvements in society increase their satisfaction in general. However, in our survey, age had little impact on participant response.

Questions Q3 to Q11 (see Table 1) deal with how respondents utilise their tap water, questions Q12–Q14 deal with their perceptions of organoleptic quality (colour, taste, and odour), and Q15 and Q16 asked about microbiological and chemical contamination. Q18–Q21 cover the presence of arsenic in the tap water, and Q22 and Q23 ask whether survey participants use bottled water or point-of-use filtration systems.

Before analysing the various correlations observed, it is worthwhile noting that correlation does not imply causation. This well-known statistical principle states that even when strong correlations are identified, it is erroneous to assume that one causes the other. As expected, whether or not people drink their tap water (Q3) was strongly correlated with their perceptions of tap water quality (Q4). Questions Q5–Q11 dealt with different types of water usage. In addition to direct consumption, whether respondents utilise the water for cooking (Q6) was also correlated with quality perceptions, although the effect was smaller. Water usage for cleaning, personal hygiene, recreation, and irrigation was not correlated with any of the other responses.

The strongest negative correlation with Q3 (“Do you drink the tap water?”) was with the responses to Q16 (chemical quality concerns), which was around twice as strong as the correlation with Q15 (microbiological concerns). This suggests that public perceptions of water quality are in fair agreement with the water quality problems in Vojvodina shown in Figure 5. Of the organoleptic properties, the largest negative correlation with water consumption was taste (Q13), followed by odour (Q14) and then appearance (Q12). The perceived presence of chemical substances (Q16) suggests a suspicion of contamination among the public, further discouraging tap water usage. These observations are in line with other studies showing how sensory perception can strongly influence the public’s trust in the safety of drinking water [38].

There was a large difference observed in whether respondents utilise their tap water for drinking (Q5; average value: −0.14) and cooking (Q6; average value: +0.48). Boiling water during cooking will significantly improve microbiological safety, but for most types of chemical contamination, it does not have a significant positive impact on water safety. Clearly, many of the participants believe that they have a problem with tap water quality but are less well informed about the nature of that problem. Although the low acceptability of tap water may unintentionally reduce arsenic exposure by encouraging the use of alternative sources such as bottled water, this should not be viewed as a viable or recommended mitigation strategy. Instead, it underscores the urgent need for improved public awareness and targeted risk communication in arsenic-affected regions. In areas with elevated concentrations of arsenic and other naturally occurring minerals, better public communication is needed to address widespread misconceptions about water treatment methods, particularly the belief that boiling water can remove chemical contaminants. While boiling eliminates microbial pathogens, it does not remove inorganic pollutants such as arsenic, nitrates, or heavy metals; in fact, prolonged boiling may increase their concentration due to evaporation. Therefore, it is essential to implement targeted risk communication strategies that educate the public about these limitations and promote safer alternatives such as certified filtration systems or access to treated water. Effective strategies may include community outreach workshops, public service announcements via local media, printed guidance in water utility bills, school-based education programmes, visual risk zone maps, and two-way communication platforms such as hotlines. These efforts are critical for protecting public health but also for enhancing the policy relevance and effectiveness of water safety interventions.

The responses to questions Q18–Q21 (relating specifically to arsenic) show that outside of Novi Sad, the respondents to our survey are generally aware of the problems with arsenic in Vojvodina, and that this awareness strongly impacts their water usage.

To improve the cluster and principal component analyses, the results of Pearson’s correlation analysis were used to exclude non-correlated survey questions (with a significance of p > 0.05). These analyses were therefore performed on Q3–Q5, Q12–Q16, Q19, Q21, and Q22. It should be noted that Q18 was also eliminated—although it weakly correlates with some questions, it relates to the entire country and not the respondents’ own situation, and was therefore not deemed necessary for this analysis.

The cluster analysis in Figure 7 illustrates the similarities between responses, which are colour-coded by municipality of residence. Five main clusters are observed, and the dependence of the clusters on place of residence is immediately apparent. The first two clusters are dominated by responses from Zrenjanin and Temerin, two municipalities with chemical water quality issues according to the Health Statistical Yearbook of the Republic of Serbia [30]. Zrenjanin is the largest municipality in Vojvodina after Novi Sad, so the uniform clustering of public opinion responses is notable. In contrast, the other large municipalities (with multiple responses), although clearly clustered together, show more variation. Cluster 3 is almost entirely made up of responses from Kikinda, which has problems with organoleptic water quality, but is also found in clusters 1, 2, and 4. This variation amongst members of the public is to be expected, as demonstrated by the broad distribution of responses from Novi Sad, where consumers are all supplied with the same good quality water, but opinions range from very bad to excellent (see Figure 4). Clusters 4 and 5 are dominated by municipalities with better water quality, locations such as Sombor, Subotica, and Bačka Palanka. (It should be noted that the respondent from Titel is isolated; they were the only respondent analysed who did not answer Q4, “How do you rate your water quality?”).

The identification of geographically distinct clusters indicates that public perceptions of drinking water quality in Vojvodina are potentially reflecting localised differences in water infrastructure and environmental conditions. These findings have practical implications for water management, as they support the development of geographically targeted strategies to address specific concerns, allocate resources more efficiently, and improve public confidence in water services. In this context, the municipalities of Kikinda, Temerin, and Zrenjanin stand out as requiring significant public efforts to improve perceptions of drinking water quality.

Table 2. The most significant (>5% of the variance) principal components and their loadings.

	PC 1	PC 2	PC 3	PC 4
Eigenvalue	6.54	1.42	0.790	0.686
% variance	54.8	11.9	6.62	5.74
Loadings
Q3	−0.345	0.0665	0.229	0.140
Q4	−0.306	0.0742	0.182	0.118
Q5	−0.342	0.0843	0.244	0.210
Q6	−0.223	0.223	−0.440	0.727
Q12	0.277	0.472	0.0513	0.0996
Q13	0.288	0.482	−0.0455	−0.0468
Q14	0.295	0.458	0.226	−0.0606
Q15	0.250	−0.0769	0.593	0.498
Q16	0.314	−0.338	0.102	0.204
Q19	0.268	−0.241	0.0528	0.182
Q21	0.268	−0.298	−0.0487	0.115
Q22	0.264	−0.0432	−0.488	0.203

and Figure 8 display the results of the PCA. The first two principal components explain 66.7% of the total variation in the responses. As discussed above, public opinion is inherently noisy, so given the questionnaire-based nature of the dataset, covering two thirds of the variance with the first two principal components is reasonable. PC1 is strongly negatively loaded with Q3–Q5 (covering water use) and positively loaded with Q13, Q14, and Q16 (taste, odour, and presence of chemical substances), and accounts for 54.8% of the total variance. PC2 accounts for 11.9% of the variance and is strongly positively loaded with Q12–Q14 (all the organoleptic properties of water) and negatively loaded with Q16 and Q19, which cover the presence of chemicals in general and arsenic specifically. The clustering observed in Figure 7 is again evident in Figure 8, whereby PC1 is split roughly in half, with the left-hand cluster dominated by responses from Sombor, Subotica, and Bačka Palanka, and the right-hand cluster dominated by Kikinda, Temerin, and Zrenjanin. The biplots for Q12–Q14 are all together, demonstrating the strong correlation between the different organoleptic properties. These three biplots are also pointing directly to a group of responses from Kikinda, where the public perception of drinking water quality is evidently heavily influenced by direct sensory experiences.

The biplots for Q15 and Q22 (concerning microbiological quality and bottled water purchasing) are also closely located and point towards responses from Temerin, Zrenjanin, and Kikinda. Responses from Zrenjanin and Temerin are also evident in the directions indicated by Q19 (awareness of a local arsenic problem), Q21 (“Does the presence of arsenic impact your bottled water consumption?”), and Q16 (concern about chemical contaminants).

PC1 can be interpreted as a “trust versus concern” gradient. At the extreme negative end of PC1, high loadings for Q3 (use of tap water for consumption), Q4 (rating of tap water quality), and Q5 (use of tap water for drinking) reflect respondents who trust the quality and safety of their tap water and use it actively. On the positive side, we see high loadings for Q12–Q16 and Q21–Q22, i.e., reasons to avoid tap water (taste, odour, chemical and microbial concerns, and arsenic awareness) and higher consumption of bottled water, representing a concern or mistrust-based behaviour. Thus, PC1 clearly discriminates between trust in tap water and avoidance based on perceived dangers. PC2 is nearer to sensory concerns and beauty sensations. Q12–Q14 (colour, taste, and odour) have the highest positive loadings on PC2, suggesting that this axis represents concerns regarding tap water’s sensory quality. The moderate negative loadings of Q15–Q16 (microbiological and chemical contamination) on PC2 suggest that health-focused concerns can be distinguished somewhat from sensory-focused dissatisfaction. These factors combined illustrate how behaviours of domestic tap water are characterised by two overlapping but distinct axes: general trust vs. concern, and aesthetic vs. health-concerned perceptions. This account is useful in bridging statistical trends in the data with intuitive reasoning rationales in people’s homes.

One concerning aspect of the distrust in public water supplies manifests in the amount of bottled water consumed by our survey respondents (Q22: “How much bottled water do you purchase weekly?”). Outside of Novi Sad, the majority of people who took our survey purchase 5 to 15 L of bottled water every week, to meet the needs of an average of 3.4 people per household. However, 15 L of bottled water per week costs around EUR 7.5 per household per week, or about EUR 400 a year (according to the retail price of ten 1.5 L bottles at a supermarket in 2024). In comparison, for a private consumer in Novi Sad, 15 L of tap water per week would cost less than EUR 1 over an entire year. According to the Statistical Office of the Republic of Serbia [39], the average annual household income in Vojvodina was around EUR 9400 in 2024, so bottled water expenditure might represent as much as 4.3% of a households’ income. In 2024, Serbia’s GDP per capita was at 31% of the European average [40]. As such, bottled water expenses represent a significant burden to standards of living, with this bottled water expenditure on a similar level to total annual household expenditure on health (4.1%, according to the Household Budget Survey, Statistical Office of the Republic of Serbia) [41].

The reliance on bottled water is not only a huge personal economic expense; it also amplifies environmental pressures associated with plastic production, distribution, and waste. The life cycle of plastic bottles—from petroleum-based raw material extraction to energy-intensive manufacturing, transportation, and disposal—has a significant carbon footprint [42]. Moreover, in the absence of robust recycling infrastructure, much of this plastic ends up in open landfills or natural ecosystems, where it breaks down into microplastics that threaten both terrestrial and aquatic life [43]. The cumulative effect of this behaviour at a regional scale poses long-term environmental risks, particularly in a province like Vojvodina, where current estimates suggest that the recycling rates of municipal solid waste are as low as 15% [44]. In terms of public policy, significantly reducing bottled water consumption will require extensive public awareness campaigns in locations with good drinking water quality. These campaigns should focus on the safety of the tap water and the economic and environmental damage created by an overreliance on bottled water.

5. Conclusions

The results of our survey indicate that even where water suppliers implement treatment technologies to meet the regulatory water quality standards, public distrust in water quality often remains high. Small municipal water utilities, constrained by limited financial and human resources, face significant challenges in maintaining and modernising their infrastructure. As a result, public perception often diverges from the technical reality. In Vojvodina, distrust in tap water is largely driven by concerns with chemical water quality. Although these concerns are often valid, greater effort should be made to raise public awareness about the fact that although boiling water significantly increases the microbiological safety of tap water, its impact on chemical contamination is often limited.

Many utilities struggle to maintain quality levels as they are, and are not in a position to invest in significant infrastructure improvements. It is also debatable whether the local population would be willing to pay an increased price for water whilst distrust in the water supply is so high. A two-pronged approach is therefore needed: restoring public confidence in tap water quality can only be brought about by improved water treatment and transparency, while also promoting sustainable consumption patterns. Long-term sustained financial support from the state to improve treatment infrastructure should lead to significant enhancements in drinking water quality, with the public investments required entirely justified by the positive impact on household budgets and a significant reduction in poverty. Meanwhile, public awareness campaigns must highlight the environmental and economic costs of bottled water use, alongside clear communication on water safety and available filtration technologies, to shift consumer behaviour towards more sustainable options. Investing in trust is not only vital for public health but also essential for effective environmental stewardship in the face of growing global challenges.