Isn’t the glass ceiling thick enough? Mapping perceptions on women’s economic empowerment in North-West Region of Romania

Abstract Gender-biased perceptions can influence the amplitude of vertical segregation within a region, as women face covert social obstacles and patriarchal attitudes in their struggle for career advancement. This study proposes a mixed methodology for analysing the underlying causes of the existing glass ceiling in North-West Region of Romania. Qualitative data collection techniques are mixed with statistical and spatial analysis in order to identify spatial clusters of communities with high levels of gender prejudice. Using a CATI cross-sectional survey, a total of 1007 respondents expressed their perception on gender equity and women’s economic empowerment opportunities. The identified biases are then included in Geographically Weighted Regressions that can highlight other objective causes for patriarchal mentalities by analysing a series of explanatory variables (unemployment, education rates, local investments, LHDI, etc.). Results show the impact local development levels can have on the upsurge of gender inequalities, despite the pre-existing cultural and psychological features.


Introduction
Occupational segregation is a long-established concept referring to economic sectors or positions dominated by women or men. The phenomenon has been divided into two categories: (1) horizontal segregation, defined as the over/under representation of one gender in certain domains (Bettio et al., 2008), and (2) vertical/hierarchical segregation, referring to the over/under representation of men or women in top positions (Maron & Meulders, 2009). The latter is frequently caused ABOUT THE AUTHORS Elena-Manuela Bîrsănuc is a geographer with experience in statistical and spatial analysis. Her research is focused on feminist and gender studies, urban studies and regional planning. She has used her GIS expertise in several national and international projects, including energy poverty, urban planning and socio-economic studies. Titus Cristian Man is an associate professor in Babeș-Bolyai University, Faculty of Geography. His major research areas include the application of GIS, spatial analysis, statistics, geostatistics, and remote sensing for service analysis and planning, particularly within the context of social services, public health, hydrology, environment, education, and urban studies. The GIS expertise included two Fulbright scholarship awards (2013 and 2020) to teach and research at the University of Wyoming and TexasTech. He published more than 50 articles on the topic of GIS. by a series of artificial barriers and covert obstacles that prevent women from occupying senior management roles (EIGE European Institute for Gender Equality, 2018), thus forming the glass ceiling. Despite being quantitatively measured by means of several indices and statistical descriptive methods (Dijkstra et al., 2021;Hardy et al., 2015;Maron & Meulders, 2009), the qualitative causes for the phenomenon are also enquired through the analysis of perceptions on gender equality in management roles (Costa et al., 2016;Dennis & Dennis Kunkel, 2004;Liff & Ward, 2001;Newman, 1993).
Most studies that are focused on analysing gender inequalities in management positions approach the subject from a series of different perspectives: business/management (Atwater et al., 2004;Gallo et al., 2021;Oakley, 2000;Ross-Smith & Huppatz, 2010), social (Costa et al., 2016;Dennis & Dennis Kunkel, 2004;Liff & Ward, 2001), or psychological (Paustian-Underdahl et al., 2014;Robertson et al., 2011). However, there is a shortage of geographical approaches that involve spatializing individual perceptions on gender equality for managerial roles. Conducting such an analysis allows researchers to identify spatial clusters of gender discrimination that form communities characterized by a set of similar attitudes towards the empowerment of women.

Objective
This paper aims to address two key aspects. The first refers to quantitatively measuring the level of gender prejudice regarding female economic empowerment through qualitative data depicting individual perceptions on gender inequities. Furthermore, besides descriptive statistics, the novelty of the approach consists in mapping these perceptions and, thus, giving a spatial perspective on the matter. Secondly, the study intents to analyse the relationship between gender prejudice levels and the level of local socio-economic development in order to highlight how individual perceptions and attitudes can be shaped by objective territorial data.
A case study set in North-West Region, Romania is presented in order to address the following research hypotheses: H1. Men encourage female domesticity more than women.
H2. There is a major discrepancy regarding gender prejudice levels between rural and urban settings.
H3. High levels of gender prejudice are located in areas with low local development levels.

Gendered leading abilities
Gender norms (Evans, 2019) that impose different social and economic roles for men and men, are known to dictate the social dynamics of local communities (Horne & Mollborn, 2020). Thus, a traditional attitude towards gender roles (Helgeson, 2012) has implications in the division of space, keeping women in the private (domestic responsibilities, childcare, housework) and men in the public (employment, financial independence, freedom of movement) (Day, 2001). Managerial and leading abilities are, therefore, associated with public spaces as they imply economic empowerment and professional development.
Perceptions on leading abilities are known to be gendered (Peachey & Burton, 2011;Saint-Michel, 2018), with two underlying causes. First, after decades of the male breadwinner model prevalence and despite the increased participation of females in the labour market (Garcia & Tomlinson, 2020;Kolpashnikova & Kan, 2021), women's best capabilities are still associated with the private sphere where femininity (as defined by gender norms) implies nurturing, caring and submissive behaviours (Sa'ar & Simchai, 2023). This aspect is connected to the second cause of gendered perceptions on leading abilities: masculinity defined as aggressive, competitive and powerful. The prevalence of male CEOs, directors and managers caused a biased perception on what senior management positions require in terms of professional abilities and personal aptitudes (Faizan et al., 2018).
Thus, as women are perceived as soft, nurturing and docile, on one hand, and management roles being associated with aggressive, dominant and assertive attitudes, on the other, this vertical segregation allowed for the glass ceiling to be formed and to harden female economic empowerment and professional development in terms of occupying management positions. Taking all of this into consideration, we address the following research question:

RQ1:
Are women perceived as equal to men in terms of leading abilities?

Work-life balance
Similar gender norms contribute to an unequal division of domestic labour (DDL). Studies show that inequalities within the domestic sphere persist despite of increasing female economic empowerment (Windebank & Martinez-Perez, 2018), suggesting that gender inequalities at home are induced less by professional status, and more by perceptions on masculinity and femininity (Offer, 2016;Rehel, 2014). For example, intensive motherhood (Gomez-Ortiz et al., 2022) encourages total dedication to childcare for mothers, despite her employment status. Gender socialization and transgenerational patriarchal attitudes embedded in the common mind leads to social stigmatization when traditional roles are not respected.
Consequently, in terms of work-life balance, women face more risks than men (Fujimoto et al., 2013). While the public discourse highlights the importance of women's participation in the labour market and professional development, perceptions embedded in the common mind encourages traditional feminine roles. Pursuing economic and professional empowerment, while violating traditional gender norms, women face social stigmatization that leads to a low quality of life (Foster & Stratton, 2019;Preisner et al., 2019). As individual behaviours are shaped by perceptions in the common mind of local communities, the study addresses a second research question: RQ2. Is female domesticity encouraged in Romania?

Mapping perceptions
Traditional methods of place study and evaluation include socio-cultural and geographical analysis. When representing location-based information, spatial analysis employs geographic mapping and information systems, whereas socio-cultural analysis uses techniques like focus groups, interviews, and public participation. Thus, a greater integration of theoretical frameworks and methodologies from both domains is required for a more comprehensive approach to regional analysis (Müller-Eie & Alvarez, 2019). Qualitative research, which incorporates the collection and examination of non-numerical data, including interview transcripts, focus group discussions, and observational fieldnotes, has also been used to map human attitudes in occupational segregation. The focus is frequently on how a community is seen and referred to, but it is equally important to consider how a community views itself, as well as how it uses and engages with space (Obermeyer, 1998). However, these sociocultural evaluations are typically conducted without much consideration for the more objective aspects of the physical environment or references to specific locations (Müller-Eie & Alvarez, 2019).
Subjective mapping is a mixed method that incorporates qualitative data of local users with geographic information to reconcile physical and geographic information with socio-cultural and subjective data. However, despite the emerging appreciation of the spatial component of social processes, there are only a few techniques for subjective mapping (Müller-Eie & Alvarez, 2019). On the other hand, behaviour mapping is used to observe and record activities, actions, and behaviours in human geography and environmental psychology. Different approaches of behaviour mapping depend only on collecting and clustering individuals and activities within a given area (Francis, 1984;Francis et al., 2010); more recent methods, however, have made it possible to plot behaviours more precisely, even in terms of connections (Golicnik, 2007;Golicnik Marusic, 2011;Müller-Eie et al., 2018). This effort of illustrating and evaluating the connection between behaviour and place is extremely helpful, especially when it comes to the examination of narrow settings.
Place identity, which refers to cognitive perceptions and connections, place attachment, which refers to emotional and psychological interactions, and place dependence, referring to cognitive and perceived behavioural possibilities, are the three factors Jorgensen (2010) identifies as connecting social systems to space. As social interactions and behaviours influence spatial dynamics (Massey, 2005), scholars highlight the importance of place-based analysis of perceptions, taking into consideration the local cultural context (Jalalkamali & Doratli, 2022).
In terms of gender bias, perceptual mapping allows the identification of prejudices within a region. The method brings new information in terms of differentiating between gender-bias perceptions distributed randomly in space and compact, spatially continuous communities characterized by a tendency towards gender prejudice. Furthermore, besides spatializing different levels of gender bias, the usage of perceptual mapping allows correlations with other spatial data regarding objective indicators of local development (LHDI, economic development, social status, etc.) that might actually cause favour and support the upsurge of gender prejudicious perceptions.

Women's economic empowerment in Romania
A balanced work-life relationship is essential for career advancement. In Romania, however, the female unemployed population (aged 20-64) due to family care and unpaid housework increased from 27% in 2011 to almost 40% in 2020, while men constantly represent under 5% of the total unemployed population due to domestic responsibilities (NIS 2011(NIS -2020. Additionally, between 1996 and 2020 the Romanian entrepreneurs are represented by under 30% females (NIS 1996(NIS -2020. Thus, the concept of work-life balance is systematically neglected at a national level due to an unequitable distribution of household responsibilities and the perpetuation of biased attitudes towards women's economic empowerment. In terms of gender equality, the United Nations report on Gender Equality Sustainable Development Goal indicates a total of 31.6% of managerial positions and 30.1% of senior and middle management positions held by women in Romania in 2019 (United Nations, 2019). Based on the 2021 Global Gender Gap Report of the World Economic Forum, Romania ranks 88/156, with a 0.723 index score for economic participation and opportunity, showing the greatest decline within Eastern European and Central Asian countries (WEF World Economic Forum, 2021). Likewise, Romanians register high disadvantage and low achievement scores for the female population, ranking last within Europe (Dijkstra et al., 2021). Furthermore, the 2022 Gender Equality Index presented by the European Institute of Gender Equality (EIGE, 2022) places Romania second to last within the EU, with a decrease of 2.1 points regarding power (gender inequalities dominating political and economic decision-making). Consequently, women share only 17.2% of the members of supervisory boards or boards of directors in largest quoted companies. However, gender gaps regarding FTE employment rate, duration of working life and Career Prospects Index are almost inexistent. Moreover, Romanian women spend more time caring for family members (46%) and doing household chores (75%) than men (25% and 41%, respectively). Consequently, ranking last in the EU at gender-parity for economic power is not that surprising for Romania.
Analysing these international rankings depicting economic gender inequalities in Romania, alongside with results presented in Barometrul de Gen (Grunberg, 2018), the existence of a glass ceiling in Romania is undeniable. While the former was presented above, the latter highlights biased attitudes within the domestic sphere (the man is the head of the family, women are better at raising children, DDL) and towards female economic empowerment (male breadwinner model, difficulty of full-time employment for mothers, work-life balance). Thus, given the persistence of gender inequalities in management positions and biased perceptions on female economic empowerment, the glass ceiling in Romania does not only exist, but is continuously maintained by a perpetuating patriarchal mindset.
Given the country's place within European gender equality ranking, this paper focuses on identifying gender-biased attitudes in North-West Region (NWR) regarding equitable opportunities for occupying managerial positions. NWR groups 6 counties and a total of 446 territorialadministrative units (90% rural localities, 28 cities and 15 municipalities). Out of the 8 major regions in Romania (NUTS2 level), it ranks third based on income rates, and second with lowest rates for poverty risk and school dropout (NIS National Institute of Statistics, 1996-2020). Thus, given its socio-economic development level (EUROSTAT, 2021), the proposed methodology aims to highlight the gender bias-local development nexus by analysing perceptions and prejudice levels within the local communities of the region.

Materials and methods
The study is based on primary data for measuring the level of gendered prejudice regarding equitable opportunities in occupying managerial positions within the 446 territorialadministrative units (TAU, corresponding to LAU 2 level units according to NUTS by Eurostat). A total of 1007 responses were collected through a CATI cross-sectional survey with a statistical margin of error of ± 3% with a 95% confidence (covering 52.5% of TAUs). Besides collecting demographic information (address, sex, age, level of education, marital status, number of children, occupation, income, ethnicity, religion), the main focus of the survey was to identify the perception on equitable opportunities for occupying a managerial role through a gender lens. Moreover, to ensure data representativity, a stratified random sampling was applied based on the demographic size of the localities (Table 1).
Respondents were asked to express their level of agreement/disagreement on a scale from 0 to 4 (0-total disagreement; 1-disagreement; 2-neither/nor; 3-agreement; 4-total agreement) regarding the following statements: (1) Men are more competent in a managerial position than women; (2) Women can have a managerial position only if they do not neglect their household responsibilities; (3) For a woman is more important to have a family/child than to have a managerial role; (4) Answering to a female director/manager would make me feel uncomfortable; (5) A woman should earn less than a man assigned to the same position/job. Thus, based on the responses collected, the aggregated (mean values) level of gender bias within a territorial-administrative unit would be placed in one of the classes described in Table 2.
Furthermore, secondary data (Table 3) were used in order to (1) establish the local level of development (gender wise) and to (2) analyse the statistical and spatial relationship between the level of local development and the level of gender prejudice in the area.
Multivariate clustering (Zhang et al., 2016) was applied in order to assess the local level of development through non-spatial clusters of selected data. By using the k-means method, we identified four statistical groups with a low data variance within each cluster, while data variation between clusters is high. In other words, TAUs are being grouped in four different clusters based on similar values of the dataset. On the other hand, each cluster has different data association, resulting in four clusters with specific characteristics. The standardised data variation within each statistical grouping allows the identification of areas characterized by a similar pattern within the dataset, regardless of spatial restrictions. Thus, the tool allows territorial diagnosis through the analysis of specific characteristics at TAU level.
In order to interrogate the relationship between the local development level and the level of gender prejudice, we applied the Geographically Weighted Regression (GWR) that takes the form of a locally applied linear regression for modelling the variability of spatial association between the analysed elements (Fotheringham et al., 2002). The results would express the way the gender prejudice level, expressed as the prejudice score measured through the collected responses (dependant variable) can be explained by other territorial data (explanatory variables). GWR creates an equation for each explanatory variable entered into the analysis, thus resulting in a variety of spatial layouts depending on the selected dataset. Therefore, the combinations of data considered relevant must reach a determination coefficient of at least 60% (R 2 ≥ 0.6). The methodology is based on progressively increasing the number of explanatory variables so as to identify the elements that have a greater weight in explaining the gender-bias levels, and to exclude data that decreases the aggregated value of R 2 . Subsequently, TAUs where the R 2 value exceeds 0.6 were selected for highlighting spatial patterns within the region. The analysis was conducted using ArcGIS Pro v.3.0.0 (ESRI, Redlands, CA).

Results
A patriarchal mindset could be the reason behind the gender inequities perpetuating over time within a society, despite implementing dedicated legislative measures. Thus, although there is a set of laws in Romania referring to gender discrimination (Law 202/2002 regarding equal opportunities for men and women and OUG 137/2000 prohibiting discrimination) (Monitorul Oficial, 2013Oficial, , 2014, these do not have the power to reduce gender inequities generated by the gender-biased mentalities transferred from one generation to the next through formal and informal education. For a broader and more comprehensive view upon the magnitude of patriarchal behaviours, qualitative aspects must be interrogated, such as gender perceptions and gender prejudice levels. As this study focuses on the gendered perception on equal opportunities for occupying a managerial position, the analysis is based on identifying the levels of prejudice regarding women's leading abilities (compared to men), division of household responsibilities, work-life balance, and attitudes towards women empowerment.
Based on the responses collected and analysed by gender subsamples (Figure 1), a greater inclination towards gender bias is highlighted among male respondents compared to females. The greatest discrepancy between the two refers to leadership abilities, over 25% of men claiming they possess better managerial skills than women (an opinion shared by cca. 15% of the female respondents). Regarding revenues, gender equality is argued for by 94% of women and 77% of men, both advocating for equal pay for similar positions, regardless of gender. However, men's tendency towards patriarchal attitudes are regularly highlighted within the subsamples, with 80% of male respondents respecting the authority of a female manager, compared to over 90% of the surveyed women. Based on these results, regarding this study's first research question (RQ1) we can argue that perceptions on leading abilities are biased and gendered. However, even though the inclination towards gender prejudice is higher for men than for women, this does not seem to impact their attitude towards a woman occupying a managerial position.
Furthermore, from a general perspective, both men (44%) and women (40%) consider having a family or children more important for a woman than being successful professionally. Lastly, gender inequities generated by a prejudiced distribution of household responsibilities are overlooked by 23% of women and 30% of men who support women's professional aspirations as long as domestic unpaid work is not neglected. Consequently, it can be argued that female domesticity is encouraged in Romanian communities (RQ2) by both men and women, even though female economic empowerment (equal revenues) is seen with less prejudice. Additionally, the first research hypothesis of the study (H1) is only partially confirmed: men do encourage female domesticity more than women; however, high shares of both male and female respondents contribute to the persistence of intensive motherhood (40.2% females, 33.5% males) and gendered DDL (23.1% females, 30.9% males).
Men's higher levels of gender prejudice (compared to women) are also highlighted spatially (Figure 2), with almost 60% of the analysed TAUs characterised by a low or medium level of prejudice for men, while women lack prejudice in over 75% of the region. The significant differences between the spatial extent of the gender perceptions of male population compared to those of the female population suggest a discriminatory, patriarchal attitude (mostly held by male respondents) towards women's authority and socio-economic role. There is a significant difference of perceptions between rural and urban communities, most municipalities being characterized by the lack of prejudiced perceptions on gender equality.
The average prejudice score for both sexes reflects the general level of gender bias within the region (Figure 3). Thus, over 75% of TAUs register a low level of prejudice, while medium levels characterize almost 10% of the study area. Cluj-Napoca and its metropolitan area are the only contiguous urban/ periurban spaces lacking prejudices. Consequently, the overall prejudice score illustrates insignificant discrepancies between urban and peri-urban rural areas, with no clear spatial pattern.

Figure 2. Gendered prejudice levels for equitable opportunities in occupying managerial roles based on individual perceptions of women (up) and men (down).
Based on these results, our second hypothesis (H2) is also only partially confirmed: while results by subsamples indicate a clear distinction of prejudice levels between rural and urban localities (urban centres with no gender prejudice, rural areas with low and medium levels of prejudice), the general prejudice score highlights similar prejudice levels within municipalities and their (rural) neighbouring areas.
The application of multivariate analysis techniques (Figure 4) revealed that urban areas grouped in the first statistical cluster within the local development dimension show the highest spatial variability. Thus, the 32 TAUs in this cluster are characterized by above average values of LHDI, share of women with tertiary education, and income per capita, while the remaining variables record average or low values (part-time employees, unpaid domestic workers, number of firms in the trade sector). Clusters 3 (33.40%) and 4 (37.44%) are the most spatially extensive. Showing contiguous spatial concentrations in Bihor and Satu Mare counties, cluster 4 is characterized by above average values of part-time employed women and unpaid domestic workers. On the other hand, cluster 3 describes large areas of Bistrița-Năsăud County, as well as local spatial contiguities in Maramureș and Sălaj counties, defined by below-average values of part-time employed women and above-average values of unemployed female population.
A distinct aspect is highlighted in the second statistical cluster (spatially concentrated in the north, west and east of Cluj County) in terms of below-average values of LHDI and unemployment rate values, associated with high values of investments in public services and number of firms in industry and trade.
The results of Geographically Weighted Regressions (GWR) carried out to determine the extent to which existing biases can be explained by socio-economic and gender-differentiated indicators, reveal several spatial patterns ( Figure 5). Male population' gender bias is determined in areas with declining economic sectors between 2011 and 2020: low per capita income, percentage decrease in the number of employees in retail (−6%), industrial manufacturing (−72%) and services (−57%), while the agricultural and construction sectors show increases in the number of employees. These are associated with low incomes and low levels of literacy among the female population, thus explaining the tendency towards patriarchal thinking through the embracing of gender discriminatory perceptions. The perception of women is explained by a determination factor of over 60%, with a majority in Cluj and Sălaj counties, and locally at the administrative border between Maramureș -Bistrița-Năsăud, Bihor-Satu Mare counties. The results suggest low and average levels of bias due to the low share of the male population with higher education, average per capita income ranging from 0.06% to 12.8%, and an average increase in employees in agriculture, retail and construction, associated with decreases in industry and services, over the period 2011-2020.
At regional level (Figure 6), a total of 24 TAUs have a coefficient of determination greater than 0.6 explained by the low share of the male population with higher education (between 3.2% and 24.6%) and a very low per capita income. These localities are mainly concentrated in Bihor, Cluj and Sălaj counties, and locally in Maramureș and Bistrița-Năsăud. These results confirm our third hypothesis (H3), indicating that low incomes/inhabitant, over-development of agriculture and constructions settings, under-development of the tertiary sector, together with low shares of highly educated population shape a social environment beneficial for the advancement of gendered-biased perceptions.

Discussion
The process of mapping perceptions creates a favourable context to correlate them with other social, economic or political data that may enhance or explain the amplitude of vertical segregation within a region. By applying a mixed methodology, this study highlights the influence of the local development level on forming gender biased perceptions within local communities.

Figure 5. Gendered prejudice level of women (up) and men (down) explained by local development level by means of GWR.
A series of attitudes towards gender equitable opportunities for occupying managerial roles were analysed in order to highlight the amplitude of patriarchy within North-West Region in Romania. Results show that the highest level of gender prejudice for both sexes refers to women's work-life balance, household responsibilities and their managerial abilities. Restricting women to the private sphere, alongside an unequal distribution of unpaid domestic work are the main reasons for limited economic empowerment of women in the region, in accordance with previous findings (Helgeson, 2012;Offer, 2016). These are a series of gender inequities leading to the inequalities that form the glass ceiling and which are expressed by national and European statistics regarding vertical professional segregation due to the gendergap in top positions.
The fact that both men and women demonstrate biased attitudes towards DDL and childcare can be explained by previous studies that connect the perpetuation of patriarchal mindsets with the social stigmatization women face when gender norms are violated (Foster & Stratton, 2019;Rehel, 2014). Furthermore, as demonstrated in this paper and similar to Garcia and Tomlinson's findings (Garcia & Tomlinson, 2020), female economic empowerment is encouraged by both sexes, as long as the domestic responsibilities perceived as feminine are not neglected by the female partner within the household.
On the other hand, besides the cultural and psychological background, other local indicators are proven to influence the spread of gender prejudices. These define the local development level through rates of unemployment, percentage of highly educated population, local investments in public services and education, unequal revenues, etc. Consequently, the results highlight higher levels of gender biased attitudes in underdeveloped communities (especially rural localities with inadequate educational and economic infrastructure), whilst periurban and urbanized areas have surpassed the traditional, patriarchal mentalities that restrict women to the private/domestic space.
However, some unexpected results surfaced regarding our third hypothesis. The hypothesis was confirmed by the results by subsamples. While female prejudice levels have a random spatial distribution and are spatially restricted, the spatial extension of the male prejudice levels gives room for spatial clusters to be shaped within rural areas, whereas municipalities and their close peri-urban neighbours form a different pattern characterized by the lack of gender prejudices. Thus, the subsamples results indicate a clear differentiation between rural and urban settings, as expected. However, the aggregated scores of the female and male prejudice levels indicate a different territorial distribution. The general level of prejudice in the region depicts large areas characterized by low levels of prejudice, with two major findings: first, when the level of prejudice is low within a municipality, its periurban and neighbouring rural areas show similar values; secondly, when a municipality lacks gender-biased attitudes (e.g., Cluj-Napoca, Carei, Turda), the surrounding localities also highlight no gender prejudice. This can be explained by the polarizing role that urban settings have on neighbouring communities. The proximity to the city may explain the spatial distribution of these results within adjacent rural areas, as large communities are formed based on similar social behaviours.
Using Geographically Weighted Regressions, we were able to identify spatial patterns reflecting high levels of dependency between the variables in the central part of the region (in Sălaj County and its neighbourhood). The perception mapping technique presented in this paper introduces a new approach for analysing the causes of gender inequities that lead to the formation of the glass ceiling, in accordance with fellow researchers that highlight the importance of analysing perceptions in their own cultural context (Jalalkamali & Doratli, 2022).
These correlations indicate that communities that are economically under-developed are more likely to maintain traditional gender-role attitudes, based on low income, low economic development, low shares of population with higher education, confirming our third research hypothesis. However, GWR results indicate that women's biased perceptions are explained differently by the local territorial context compared to men.
This surfaces as an extremely valuable data-based proof that men and women behave, interact and are influenced by the local environment differently. The areas where women's levels of prejudice are best explained by the explanatory variables differ from those where men's perceptions are more likely to be shaped by local data. Furthermore, when analysing the aggregated score of prejudice, a different spatial distribution of GWR results is depicted. This demonstrates the imperative need to design methodological frameworks that take into account local data, specific cultural contexts, and that assumes that men and women are impacted differently by the socioeconomic environment that they reside in.

Conclusions
By mixing qualitative data collection methods with quantitative analysis techniques, researchers can look beyond standard quantitative indicators that explain the effects, and identify the underlying causes of vertical occupational segregation at a local scale. Although the subject has been approached from a series of different perspectives, the geographical input gives room for analysing neighbouring communities and the existing territorial relationships by means of spatial analysis.
Our study presents some weaknesses in terms of data, as the last available national census dates back in 2011. However, the 2021 Census data will allow us to design a comparative study in order to highlight the development of gender-based socio-economic discrepancies in Romania. Given the novelty of the methodological framework presented in this paper, future research should aim to extend the analysis to other NUT2 level regions by adapting it to the local cultural and economic context.
The complexity of the research subject also requires further correlations with other social and demographic data (such as marital status, number of children, monthly income, level of education, etc.) in order to better explain the existing prejudice levels and to highlight the complex dynamics that are embedded in the common mind at a local level.