Research on social norms has a long tradition in social psychology (Sherif, 1936; Asch, 1951; Sherif and Sherif, 1953). Social norms serve the purpose of reducing uncertainty within social contexts and providing behavioral guidance for both in-group and out-group members (Crandall et al., 2002). They are essentially “shared patterns of thought, feeling, and behavior” (Hogg and Reid, 2006, p. 8) that exist between members of a society or a group.

Despite the innate desire of individuals to comprehend and conform to the prevailing norms within their respective groups or communities, the perception of social norms can often be limited and prone to inaccuracies (Tankard and Paluck, 2016). Consequently, individuals form their understanding of norms based on observations and selective information obtained from their social environment. The motivation to align behavior with social norms originates from the need for group affiliation and the apprehension of potential rejection for deviating significantly from the established norm (Cialdini and Goldstein, 2004).

Early field studies conducted by Minard (1952), Reitzes (1953), and Kohn and Williams (1956) establishing the dependence of behavior on situational and local contexts. These studies posited that different locations create distinct normative contexts. In alignment with this notion, Crandall et al. (2002) suggested that “in certain circumstances, the expression of prejudice toward inappropriate targets is tolerated – in the locker room, at the poker table, or among close friends from the neighborhood where certain prejudices are tolerated” (p. 373). Building upon these insights, we contend that valuable insights into social norms and neighborhood dynamics can be drawn from the environment itself, in the here presented case specifically through the presence of political graffiti and environmental cues within urban spaces.

The condition of the environment has played a pivotal role in representing social norms in social science research (Kelling and Wilson, 1982; Cialdini et al., 1991). It inspired research that investigated the effect of the physical neighborhood conditions on various factors such as the perception of social cohesion (O’Brien and Wilson, 2011) or fear of crime. A recent review and meta-analysis however state that research methods in this field appear to be flawed and further research is needed (O’Brien et al., 2019). Our objective was to tackle this concern by conducting the present studies using established and controlled methodologies as well as pretested measurement scales in order to effectively address our research questions.

To address this concern, we use the thin slices paradigm (Ambady et al., 2000). This paradigm draws upon a wealth of research from various disciplines outside of psychology, demonstrating humans’ ability to form accurate judgments based on brief exposures to scenes or photographs. Notably, research has shown that people accurately assess other people’s emotions, intentions, or motives from photographs or short videos (Hall et al., 2016; Schmid Mast and Hall, 2018). In the context of neighborhood perception studies utilizing images conducted by O’Brien and Wilson (2011) or Salesses et al. (2013), participants evaluated the social dynamics of specific areas after examining photographs of neighborhoods. It was found that participants’ perceptions of social dynamics aligned with ratings provided by residents of the respective neighborhoods (O’Brien and Wilson, 2011). Furthermore, perceptions of safety and social status were found to correlate with actual crime rates and income levels within the areas (Salesses et al., 2013).

Finally, a recent replication of Cialdini et al. (1991) by Bergquist et al. (2021) showed that the amount of normative cues in a room increased the amount of intended behavior in participants, indicating a dose–response dependency of normative cues. To our knowledge, the theoretical framework provided by Cialdini et al. (1991) of normative cues in environments has not found much attention outside the realm of environmental psychology (Stok and de Ridder, 2019) and we aimed to apply it innovatively to political cues in the environment.

The concept of social cohesion has been frequently utilized by researchers to characterize social dynamics within neighborhoods over the past decades, and periodic reviews have been conducted to assess the current state of the field (Friedkin, 2004; Schiefer and van der Noll, 2017). However, its extensive use across various research disciplines and policy applications has resulted in a lack of a precise definition and operationalization of this construct. Consequently critiques contend that social cohesion is a multi-dimensional construct that refers to several aspects such as trust, reciprocity, social ties, solidarity, shared values, and the evaluation of the inhabitants (Breidahl et al., 2018). It is also discussed to represent a form of homogeneity within a community (Wickes et al., 2013). Notably, early research indicate a strong correlation between social control (the ability to enforce social norms and act against threats to safety and security) and the concept of social cohesion, often combining them into single scales (Sampson and Raudenbush, 1999; O’Brien and Wilson, 2011; Collins et al., 2017).

However, emerging research indicates the utility of considering control and cohesion as distinct constructs (Wickes et al., 2013; Collins et al., 2017). Social control plays an important in managing the diverse values and attitudes held by heterogeneous groups within a neighborhood, given that inhabitants may vary in their beliefs. Consequently, it remains an open question how important homogeneity is to social cohesion and if diversity (in its broadest sense) poses a potential challenge to social cohesion on a societal level (Putnam, 2007; Ariely, 2014).

In relation to political diversity within neighborhoods, research conducted in the United States has reported an increase in political segregation over the last decades (Motyl et al., 2014; Motyl, 2016). A recent US study by Motyl et al. (2020) revealed a growing trend of individuals self-segregating into politically homogenous communities and selecting neighborhoods based on social cues. Interestingly, participants demonstrated a preference for communities that aligned with their own political beliefs, even in the absence of explicit information about the political composition of the community’s inhabitants. The authors concluded that people’s perceptions of communities may be influenced by subtle cues, indicating the role of implicit factors in shaping community choices.

According to Festinger’s (1957) theory of cognitive dissonance, individuals experience a state of discomfort when confronted with messages or situations that contradict their expectations or attitudes. In response, they employ various strategies to alleviate this dissonance, such as selectively exposing themselves to confirming information, rejecting opposing viewpoints, or devaluating individuals holding opposing views (Stroud, 2014). This theory has found recent application in the study of political messages within news and media (Arceneaux et al., 2013; Camaj, 2019; Jean Tsang, 2019). It has been observed that individuals tend to prefer messages that align with their pre-existing political views, reinforcing their initial positions (Knobloch-Westerwick and Meng, 2011) and even leading them to adopt more extreme positions (Levendusky, 2013). A meta-analysis conducted on this topic provides further insights into the effects of political message exposure indicating that people prefer congenial over uncongenial information (d = 0.36) (Hart et al., 2009).

Further, political messages tap into intergroup dynamics, as defending specific political information becomes an expression of ingroup solidarity (Van Bavel and Pereira, 2018; Wojcieszak and Garrett, 2018; Dvir-Gvirsman, 2019). Consequently, when individuals encounter political information that aligns with their own political identity, it tends to reinforce political polarization and may even lead to negative sentiments toward the opposing group (Garrett et al., 2014; Gvirsman, 2014).

Similarly, studies on exposure to counter-attitudinal information suggests an even further political divide, as participants’ opinions become even more divergent after being exposed to information that challenges their pre-existing beliefs. This exposure activates their political identity, leading to increased ingroup solidarity and hostility toward the outgroup (Arceneaux et al., 2013; Nyhan and Reifler, 2015; Zhou, 2016). Nevertheless, recent research has raised doubts about the presence of this backlash effect (Guess and Coppock, 2020).

In light of these findings, we formulated the hypothesis that perceiving political graffiti that aligns with an individual’s political opinion would result in a more favorable evaluation of the neighborhood and its potential senders (i.e., neighborhood inhabitants), thereby reinforcing the initial opinion. Conversely, we proposed that encountering graffiti that challenges the perceiver’s opinion would lead to a devaluation of the potential senders but would also strengthen the individual’s initial position.

In the present series of studies, our objective was to examine the impact of political graffiti within urban neighborhoods on individuals’ perception of the neighborhood and their individuals attitudes. We propose that political graffiti can be regarded as a form of political messaging, representing prevailing social norms that can either align or conflict with the recipients’ own attitudes. Therefore, we hypothesized an interaction effect between political graffiti and participants’ political orientation on the perception of social dynamics within a neighborhood. In Study 1 we postulate that exposure to political graffiti should tap into intergroup relations and evoke a positive or negative evaluation of the potential sender depending on the participant’s political orientation. In Study 2, we developed a robust and reliable measurement scale to thoroughly investigate this effect. Furthermore, in Study 3 we explored the possibility of a dose–response relationship of the effects of right-wing graffiti. Specifically, we anticipated that as the number of graffiti increased, the effect on individuals’ perceptions would also intensify.

Additionally, we conducted an exploratory analysis to examine the influence of graffiti on attitudes toward foreigners, aiming to assess whether our stimulus material had the ability to impact political attitudes related to the content of the graffiti.

All data and Supplementary materials can be retrieved from the OSF repository (https://osf.io/2cmwr/). All studies were preregistered (Preregistrations for each study can be found in the osf repository). In each study, we report how we determined our sample size, all data exclusions (if any), all manipulations, and all measures in the study (Simmons et al., 2012).

Detailed analysis of each dependent variable can be found in the Supplementary materials S1. As noted in our preregistration, previous research has indicated a high correlation among our dependent variables, which sometimes led to the use of a composite measure of social cohesion (Sampson and Raudenbush, 1999; O’Brien and Wilson, 2011; Kim and Kawachi, 2017). An exploratory factor analysis (EFA) with oblimin rotation revealed single-factor solutions over all items for all pictures, explaining 58–67% of the variance. Accordingly, we collapsed all items into composite scales. The internal consistency of these scales was high, with Cronbach’s alpha ranging from 0.89 to 0.94.

Prior to our main analysis, we conducted an analysis of variance (ANOVA) to examine the effect of the condition (independent variable) on participants’ responses to the manipulation test item, which asked about the perceived political orientation of the people living in the depicted neighborhood. This was followed by a post-hoc Tuckey honestly significant difference test (HSD) to compare group means. Our analysis revealed significant differences between all conditions for 7 out of the 9 manipulated pictures, providing evidence that participants accurately interpreted our manipulations as either right-wing or left-wing graffiti (See Supplementary materials S1).

In our main analysis, we tested our hypotheses in a moderated regression analysis, using the PROCESS Macro for R (Hayes, 2022). The variable political orientation was z-transformed before the analysis and we used Helmert coding for our multicategorial moderator (Hayes and Montoya, 2017). Overall, model fits R² for all models ranged from 0.063 to 0.38 (MR² = 0.22, 95% CI [0.29, 0.16]). As expected, political graffiti influenced the evaluation of the neighborhood (H1), and we found the moderating effect of political orientation in 6 of the presented pictures. In three out of nine pictures the R²-Change for the interaction was significant (p < 0.05). Subsequently plotted the interaction for each picture to gain deeper understanding of the data. We observed the expected slope pattern in three additional pictures; however, the interaction did not reach statistical significance at convectional levels (all p < 0.08). In three pictures we did not find an interaction. We believe that this outcome can be attributed to a combination of factors, including the chosen pictures, failed manipulation tests, sample distribution, and sample size. Detailed visualization of the interaction plots are depicted in Figure 2. Notably, in those pictures where we did not find an interaction, the presence of graffiti alone had a negative effect on the overall evaluation, aligning with previous research on graffiti and its association with incivilities (Miethe and Venger, 2021).

A noteworthy limitation of our study, which bears significance for interpreting the results, is the non-normal distribution of our moderator variable, political orientation. The distribution showed positive skewness, indicating a greater number of participants identifying themselves as leftists. The relatively small number of right-leaning participants contributed to the observed results, whereby the right-wing conditions did not reach the same evaluation by rightist participants as the left-wing conditions did among leftist participants, thereby reducing the prominence of the interaction effect. Additionally, the generalizability of our findings is constrained by the relatively small magnitude of the interaction effects, stemming from the single exposure in each picture. This limitation could be attributed, for instance, to our deliberately chosen broad-range dependent variable.

Notwithstanding these limitations, our findings provide support for H1 across all pictures examined. Moreover, H2 and H3 were confirmed in 6 out of 9 pictures. Consequently, our results indicate that political graffiti holds appeal to those who align with the corresponding political ideas, resulting in a more positive perception of the neighborhood’s inhabitants and the potential senders of the conveyed messages within our experimental context. Building upon these initial findings, our subsequent studies aim to further refine our understanding of the impact of political graffiti.

First, an exploratory principal factor analysis (EFA) with promax rotation was performed for the variables of the first neighborhood to identify suitable items. Subsequently, a confirmatory factor analysis (CFA) was performed using the data from areas 1, 2, and 3 to validate the identified factor structure. Model fit was assessed using the comparative fit index (CFI), Tucker–Lewis Index (TLI), and the root mean square error of approximation (RMSEA) following recommendations (Hu and Bentler, 1999). Internal reliability and scale means were calculated. Additionally, we conducted a repeated measure ANOVA to examine whether the scales were capable of detecting differences among the presented neighborhoods.

Detailed analysis, factor loadings, and item wording are reported in the Supplementary materials S2. In the first step of the analysis, all 18 items were included in the exploratory factor analysis (EFA). We used principal axis factor analysis with promax rotation and parallel factor extraction. In the second step, we dropped items with low or cross-loadings and selected items most suitable to represent the underlying constructs. This process resulted in a final set of 13 items that loaded onto three distinct factors. Factor one captured evaluations of the inhabitants and accounted for 14.6% of the variance. Factor two represented the ability to enforce social control and explained an additional 11.2% of the variance. The third factor explained 8.3% of the variance and represented the preceded social cohesion in the neighborhood. Together, these three factors accounted for 34.2% of the total variance.

In the subsequent step, we conducted confirmatory factor analysis (CFA) to validate the structure identified in the EFA for neighborhood 1 with each of the three neighborhoods. We made some adjustments during the CFA process. Firstly, we removed one item from the social cohesion scale due to its reverse scoring and strong residual covariances with other items in the same block. Secondly, we introduced residual covariances among items within each scale to account for the method factor of positive and negative wording. Eventually, we settled on four items for each dimension.

Overall fit indices for each neighborhood indicated a good fit (CFI > 0.96 TLI ≥ 0.95, RMSEA 0.4–0.6; RMSEA Upper 90% CI ≤ 0.08). However, during the analysis we observed that two items of the social cohesion factor had factor loadings bellow 0.5. Despite this, we decided to retain these items in the scale to maintain conceptual coverage. The results of the CFA are reported in the Supplementary materials for Study 2 (S2) and a summary is provided in Table 1.

Cronbach’s Alphas for the scales ranged from 0.56 to 0.89 and are reported in Table 2. Overall, we observed that some of the reversed worded items presented a minor challenge to scale validity and may have contributed to the slightly lower reliabilities. In particular, the variable indented to measure social cohesion showed mixed results in the factor analysis. However, it is important to note that this discrepancy may not necessarily stem from a statistical issue, but rather to the limited information conveyed in the pictures themselves. Since the questions posed to participants pertained complex social behavior, they might be not have been easily answered based solely on the visual stimuli provided at the pictures. Table 2 reports the correlations between the scales for further examination.

Finally, we tested for perceived differences between the three neighborhoods conducting a repeated-measures ANOVA. Due to a violation of sphericity, we corrected degrees of freedom using Huynh-Feldt’s estimates of sphericity. The repeated measures ANOVA overall rated areas revealed significant results for all constructs: η² for social cohesion = 0.42, F(1.89, 398.69) = 266, p < 0.001; η² for evaluation of inhabitants = 0.15, F(1.98, 418.53) = 72, p < 0.001; η² for social control = 0.16, F(1.63, 344.25) = 68.6, p < 0.001. Post hoc tests indicated that all presented areas showed significant mean differences on our dependent variables except the evaluation of inhabitants between neighborhood 1 and neighborhood 2. Post-hoc tests can be found in the Supplementary materials S2. Overall, social cohesion exhibited the largest effect among the other dimensions in our repeated measure ANOVA, indicating that results were most clear on this facet of our dependent variable.

Taken together, the results from Study 2 indicated that our newly developed instrument exhibited favorable fit indices and yielded reliable scales. Additionally, it successfully captured participants’ varying perceptions of the social dynamics within three neighborhoods. Encouraged by these results, we decided to employ our measurement instrument in Study 3, where we sought to examine the impact of political graffiti on neighborhood perception with greater precision.

Detailed analyses are reported in the Supplementary materials S3 for a comprehensive understanding. To examine for the influence of the conditions and the dose dependency on our outcome measures (H4), we conducted separate ANOVAs for each dependent variable and Tukey Post Hoc HSD comparisons between the conditions. The means of the conditions were unequal according to a one-way ANOVA for the evaluation of inhabitants F(4, 549) = 90.171, p < 0.001, η² = 0.396, cohesion F(4, 549) = 20.967, p < 0.001, η² = 0.133 and social control F(4, 549) = 5.341, p < 0.001, η² = 0.037. However, the impact of the condition on the dependent variables exhibited variation, with the evaluation of inhabitants demonstrating the strongest effect and the perception of social control displaying the weakest effect. Detailed Tukey Post Hoc comparisons between the conditions are reported in Supplementary materials S3. Figure 3 illustrates the effects of the conditions on the three dependent variables.

The results support H4 and indicate that as the number of right-wing graffiti increases, there is a decrease in the perceived qualities of the neighborhoods across all measured constructs. This effect is particularly pronounced in the evaluation of inhabitants compared to cohesion and control. One possible explanation for this pattern is that it is easier for individuals to attribute the perceived information conveyed by political graffiti directly to the people living in the neighborhood (Festinger, 1957). In contrast, assessing cohesion and control might involve understanding the complex dynamics and relationships between the inhabitants – information, that might be difficult to assess from our stimuli material. From this perspective, right-wing graffiti would directly result in an intergroup-based perception of neighborhoods (Arceneaux et al., 2013) devaluating those who are presumed to be responsible for creating the political graffiti (Stroud, 2014).

Furthermore, serial comparisons of the means using Tukey’s Honestly Significant Difference (HSD) procedure revealed medium to large effects between the control group and condition 1. For evaluation of inhabitants it was t (549) = 6.86, p < 0.001, d = 0.92; for cohesion it was t (549) = 4.31, p < 0.001, d = 0.58. For control, the effect between the control group and condition 1 was not significant t (549) = 2.32, p = 0.139, d = 0.31. These findings suggest that even a small dose of rightwing graffiti has the potential to influence the perception of neighborhood qualities and their inhabitants. Interestingly, our results indicate a non-linear relationship, as there were no significant differences between conditions 3 and 4 in all dependent variables. This may be due to a saturation effect of the stimulus, right-wing graffiti, which aligns with previous research on this topic (Arendt, 2015). The results further demonstrate that, for individuals with a mean value in political orientation, additional graffiti beyond a critical number does not increase the effect.

To test the interaction hypothesis (H5) we conducted moderation analyses using the PROCESS Macro for R (Hayes, 2022). Social cohesion, evaluation of inhabitants, and social control were set as dependent variables, while condition served as the independent variable, and political orientation (z-transformed) was included as the moderator. Table 3 shows the regression coefficients, R² and ΔR² for the interaction. To compare each condition to the control group, we utilized simple dummy coding. Consequently, the group means in the conditions when the moderator is at zero can be calculated from the effect sizes of the control condition and condition.

The model ΔR² for the interaction between political orientation and condition was significant for the evaluation of inhabitants ΔR² = 0.03, p < 0.001 and social cohesion ΔR² = 0.03, p < 0.001, but not for social control ΔR² = 0.03, p = 0.273. The interactions were significant in all experimental conditions for the dependent variable evaluation of inhabitants. For social cohesion, the interaction only reached levels of significance in groups 3 and 4. For social control, the interaction was only significant in condition 3. The positive regression coefficients of the interaction terms provide evidence that individuals with a right-leaning political orientation tend to evaluate the neighborhoods and their inhabitants more positively compared to those with a left-leaning political orientation. These findings align with our moderation hypothesis and suggest that political orientation moderates the relationship between the conditions and the perception of neighborhoods.

However, it is important to note the moderation hypothesis did not hold across all measured constructs and conditions, which may be attributed to the effect sizes of the conditions.

Additionally, the distribution of political orientation in Study 3 was skewed toward the political left, which could have influenced the results. Considering this skew distribution, even individuals in our sample who represent rightists, our manipulation material may have found our manipulation material which included legally forbidden symbols such as the swastika, too extreme. Despite these considerations, the reported results of the moderation remain highly valuable. Hey indicate that even individuals with a moderate right-leaning political orientation experience less confrontation with the presented stimuli, leading to more favorable ratings of neighborhoods compared to those with a left-leaning political orientation.

In a final step, we conducted an exploratory analysis (E1) to examine the impact of our manipulation onto the attitudes toward foreigners living in Germany. We used a feeling thermometer to measure if the manipulation had the potential to influence more distal attitudes that were not directly related to the neighborhoods but connected to the overall manipulation (e.g., graffiti containing messages like “foreigners out”). However, the results of this exploratory analysis were not conclusive. We found a significant positive effect (b = 4.12, [0.05, 8.21]) in condition 4 where all pictures depicting right-wing graffiti, indicating that the group means between the control condition and condition 4 were significantly different. However, we observed no interaction with participants’ political orientation. The finding of a significantly higher group mean in the last condition might be the result of a backlash or resistance hypothesis (Arceneaux et al., 2013), suggesting that participants rejected the implied notion of xenophobia. However, upon closer examination, this explanation weakens on as one would expect an interaction effect, with more leftist participants being bolstered in their initial position (i.e., having more positive attitudes toward foreigners) compared to those oriented toward the right who might agree with the underlying statements. Our data did not indicate such an interaction in the manipulations. These initial results provide an interesting starting point for further studies and highlight the complexity of attitudes toward foreigners in the context of our manipulation.

In conclusion, the results of Study 3 not only replicated but also extended the initial findings from Study 1 regarding the impact of right-wing political graffiti. Through the use of a dose–response approach and precise measurement instruments, we demonstrated that even a small number of right-wing graffiti had a significant influence on the evaluation of neighborhood qualities, particularly the evaluation of its inhabitants. Moreover, we successfully replicated the moderation effect of political orientation, indicating that both attitudinal and counter-attitudinal graffiti affected the evaluation of the neighborhood accordingly.

It is important to acknowledge a limitation of our study, namely the focus exclusively on right-wing graffiti. Due to financial constraints, we were unable to include left-wing graffiti in this study, which would have necessitated a doubling of the sample size. However, we recognize the importance of investigating the effects of left-wing graffiti as well, and we recommend that future research addresses this question to provide a more comprehensive understanding of the topic.

The present series of studies possesses several notable strengths, contributing to a comprehensive and methodologically rigorous investigation of a previously overlooked social phenomenon in social psychological research: political graffiti (Bloch and Phillips, 2020; Alves Diniz and Stafford, 2021). This widespread and historically significant form of expression has received limited attention in the literature until now. The first strength of our research lies in the utilization of reliable measurement scales specifically tailored to assess the impact of political graffiti perception in our experiments. These scales were carefully tested in Study 2 to meet the specific requirements of our studies, ensuring the validity and robustness of our findings.

Another significant merit of our study is the experimental design employed, allowing for the establishment of causal relationships. Study 1 revealed a dual process, demonstrating that alignment with political attitudes results in positive evaluations of neighborhood dynamics and discordance with less positive evaluations. Building upon this initial finding, Study 3 successfully replicated the effect of politicized perception of right-wing graffiti across a comprehensive range of measures, thus confirming its robustness and providing compelling evidence for neighborhood workers, municipalities, and policymakers seeking to foster more positive neighborhood norms.

It is worth emphasizing that political graffiti holds the potential to either harm or enhance neighborhoods and should be approached with caution. Our findings suggest that it could even be utilized as a tool for promoting desirable neighborhood norms, if formulated in the right way (Tankard and Paluck, 2016). These insights are of practical significance and can inform interventions and strategies aimed at creating and maintaining positive community dynamics. By recognizing the influence of political graffiti, stakeholders can better understand its implications and work toward fostering inclusive and cohesive neighborhoods.

While our studies have provided valuable insights into the interactive effects of political attitudes and political graffiti on shaping neighborhood perceptions, they also highlight certain limitations and raise new questions for further research. First and foremost, it is important to acknowledge that graffiti is considered a misdemeanor and as such can be perceived as vandalism or representing a minority opinion. This aspect might diminish the overall effect of the underlying political message as the prevailing attitude might not be perceived as dominant among the neighborhood inhabitants. Consequently, the ability of political graffiti to effectively communicate social norms may be compromised. This limitation could be overcome by exploring alternative forms of political messaging that are perceived as more legitimate and intentional by the community (Hogg and Reid, 2006). For instance, the use of murals, posters, signs displaying voting intentions (as commonly observed in the United States), flags or other symbols carrying political messages could be more strongly associated with formalized representations of the underlying opinions in the neighborhood. This shift toward more official and accepted forms of political expression may enhance the association between political messages and the perception of social norms in neighborhoods.

Furthermore the saturation of official political representations within a neighborhood could provide an even stronger basis for interpreting these cues as indication of a prevailing social norm with greater majority support (Hogg and Reid, 2006; Morris et al., 2015). Ideally, future research should aim to investigate the effects of more official and diverse forms of political messaging on neighborhood perceptions, as well as explore the dose–response dependency of these messages. Again, these studies could help to inform policies and interventions aimed at promoting positive social norms and community cohesion.

Furthermore, it would be valuable to replicate Study 3 by incorporating left-wing graffiti to gain additional insights and we appreciate further research into this direction. If aligning political graffiti can indeed lead to even more positive evaluations of inhabitants compared to the control condition, as suggested by our preliminary findings in Study 1, it would further support the notion that political messages, whether in the form of graffiti or other representations, have the potential to foster a stronger sense of community and cohesion within neighborhoods.

Ideally, future (field) experiments should be conducted to investigate the effects we have presented in real, existing neighborhoods. While self-reported methods have their merits, they also have limitations. Therefore, incorporating other methodological techniques, such as observations of behavior and interactions (Cialdini et al., 1991), would provide a more comprehensive understanding of local neighborhood life. By observing actual behaviors and interactions within the community with the presence of graffiti, we could gain deeper insights into the complex processes involved in community building among inhabitants.

Our findings add to the growing jet, inconclusive discussion of the relationship between various forms of diversity and sense of community (Neal and Neal, 2014; Neal, 2017) by extending the debate on political diversity within communities. Our findings highlight the need to consider diversity on a much broader spectrum when investigating community dynamics. Thus, conducting further studies with diverse neighborhoods would allow for a more comprehensive exploration of the effects of political graffiti on neighborhood perceptions. Different neighborhoods may have unique characteristics and dynamics that could influence the impact of political messaging. Therefore, a broader range of neighborhoods should be included in future research to capture the nuances and variations in the relationship between political graffiti and community perceptions.

Lastly, the preliminary findings of our exploratory analysis on attitudes toward foreigners highlight the need for further investigation on the effects of political graffiti on more distal political attitudes. Specifically, it raises important questions about how the representation of a predominant political attitude shapes intergroup attitudes and intergroup contact toward other groups. This line of inquiry aligns with the Normative Theory in Intergroup Relations proposed by Pettigrew (1991) and builds upon early research conducted in the US, which found that discriminatory behavior of white individuals toward black individuals varied depending on the prevailing social norm in a given social environment (Minard, 1952; Pettigrew, 1958).

To advance our understanding of intergroup relations in everyday life, future research should revive and expand upon the concept of situational and normative influence on socially desired behavior. By investigating the interplay between political attitudes, social norms, and intergroup dynamics, we can gain valuable insights into the mechanisms underlying intergroup relations and potentially identify strategies to promote positive intergroup interactions.