Female entrepreneurship in Brazil: how scientific literature shapes the sociocultural construction of gender inequalities

The number of female-led enterprises in Brazil is increasing, leading to a rise in related research. These studies are crucial for informing public policy and funding initiatives to support women entrepreneurs. However, in Brazil, women face significant disadvantages compared to men, including fewer funding opportunities and a tendency to start businesses out of necessity rather than opportunity. Their businesses are often concentrated in gender-stereotyped sectors like fashion, food, and beauty. Scientific discourse, despite not being directly accessed by most micro and small entrepreneurs, is powerful enough to shape societal perceptions and realities, potentially reinforcing gender inequalities. The authors argue that analyzing the language used in academic research on female entrepreneurship in Brazil is crucial for understanding how gender inequalities are (re)produced and potentially mitigated. Grammatical sexism in Portuguese (and other Latin languages) uses masculine forms as the default, potentially erasing women's experiences from the narrative. This research critically examines how scientific papers describe female entrepreneurship in Brazil, aiming to understand the implications of this language on the social construction of gender inequalities.

The study reviewed 88 articles (English and Portuguese) on female entrepreneurship in Brazil, published between 1999 and 2019. These articles were sourced from ProQuest, Web of Science, and Lilacs databases. The review aimed to identify recurring themes and representations of female entrepreneurship within the scientific literature.

The research employed a qualitative approach, analyzing the abstracts of 88 articles on female entrepreneurship in Brazil. Abstracts were chosen because they are typically written in English (for wider international visibility) and provide a concise summary of the research. The articles were divided into two corpora: English and Portuguese, as Iramuteq's lexicometric tools require a single language per analysis. The software was used to perform two key analyses: Hierarchical Descending Classification (HDC) and Word Cloud Analysis (WCA). HDC reveals links between lexical classes based on frequency and proximity, while WCA organizes lexical forms based on their overall frequency. Preprocessing removed grammatical elements like articles, pronouns, and prepositions, focusing on the core lexical units. The analysis considered several factors like total word counts, unique word counts, and average occurrences per text, performing these analyses separately for the English and Portuguese corpora.

The analysis revealed striking differences in the representation of female entrepreneurship between the English and Portuguese corpora. In the English corpus, the word “woman” appeared more frequently than “female,” reflecting a distinction between sex and gender. The tree diagram showed “woman” linked to terms such as “gender,” “future,” and “innovation,” suggesting a focus on women's roles in modernizing Brazilian entrepreneurship. “Female” was linked to “economic,” “entrepreneurship,” and “country,” implying a focus on women's contributions to economic development. The word cloud further supported this, placing “woman” centrally and dividing the cloud into two sections, one emphasizing gender and economic factors and the other focusing on female activity, business, and personality. The Portuguese corpus showed a strong masculine bias. The masculine term “empreendedor” (entrepreneur) dominated the word cloud, dividing the discourse into two parts: one focused on “mulher” (woman), “pesquisa” (research), and “objetivo” (goal) and the other on “empresa” (business), “feminino” (feminine), and “estudo” (study). This suggests that even in the context of female entrepreneurship, the default language remains masculine, highlighting the continued challenge of achieving gender equality. The hierarchical classification in the English corpus showed that the most representative class was class 4, focused on gender differences and personality. In contrast, the Portuguese corpus lacked explicit references to gender and sex in its most representative classes, with class 6 primarily focusing on methodology. This suggests a methodological focus that overshadows broader considerations of gender. The Portuguese analysis indicated a lack of coherence between methodology and research subject, with a priority given to formal and organizational variables over gender-specific issues.

The study's central finding is the consistent underrepresentation of female-specific terms in the literature on female entrepreneurship in Brazil, regardless of the language of publication. This demonstrates how deeply ingrained gender biases are within the scientific discourse, often unintentionally reinforcing existing inequalities. The preference for masculine terms as a default, even when the research focuses on women, perpetuates the invisibility of women's unique experiences and contributions. This underscores the need for a conscious effort to create a more inclusive and gender-sensitive vocabulary in academic discussions of female entrepreneurship. The authors highlight the importance of including the voices and experiences of women entrepreneurs themselves, especially those from marginalized backgrounds, to foster a more nuanced and representative understanding of the challenges and opportunities they face.

This study reveals a significant gap in how female entrepreneurship is represented in Brazilian scientific literature. The persistent use of masculine-coded language, even when discussing women's experiences, masks gender inequalities. Future research should prioritize including the lived experiences of women entrepreneurs, particularly from diverse socioeconomic backgrounds, to create more accurate and impactful knowledge. This requires developing a gender-sensitive vocabulary and research methodologies that center the voices and perspectives of women.

The study's focus on abstracts might limit the depth of analysis compared to examining the full texts of the articles. The reliance on available databases could also introduce biases related to publication patterns and article selection. The research primarily analyzed abstracts, which might not fully capture the complexities of the entire articles.