Abstract
Words carry connotations: signaling warmth or ambition, care or confidence, which consequently shapes how crowdfunding pitches are perceived. Building on debates about bias in entrepreneurial finance, the paper examines the interplay of gendered language and founder gender in reward-based crowdfunding. A dataset of 671 technology-based reward crowdfunding projects provides the empirical basis. The Linguistic Inquiry and Word Count (LIWC) software is used to quantify communal (feminine) and agentic (masculine) expressions in campaign texts. Logistic and linear regression models assess their associations with funding access and depth of funding, while the founder’s gender serves as a moderating variable. The results reveal that communal language is generally associated with lower probabilities of funding access and lower depth of funding. However, for female-led campaigns, communal language is related to a reversal of the previously negative relationship with funding access probability, while the disadvantage in depth of funding; i.e., lower amounts pledged, persists. In contrast, the use of agentic language by women remains related to lower probabilities of funding access and lower depth of funding in technology-oriented crowdfunding campaigns. This paper extends crowdfunding research by highlighting how language interacts with founder gender in shaping performance. It also applies role congruity theory to an online entrepreneurial finance context, showing that the alignment or misalignment between linguistic style and gender expectations influences funding outcomes. The insights are relevant for scholars examining bias in entrepreneurial contexts, as well as practitioners crafting campaign narratives. The evidence underscores the role of language-based stereotypes in shaping perceptions of female founders and the need for greater awareness of such biases on crowdfunding platforms.
Introduction
Access to early-stage entrepreneurial finance remains a central challenge for founders, particularly in technology-driven ventures characterized by high uncertainty, intangible assets, and limited performance histories (Gompers & Lerner, 2001; Hall & Lerner, 2010; Shneor & Munim, 2019; Shneor & Vik, 2020). In recent years, crowdfunding has emerged as an alternative financing instrument that complements established forms of entrepreneurial finance such as angel investment, venture capital, and accelerator-based funding (Cumming et al., 2021; Drover et al., 2017). By enabling entrepreneurs to directly solicit capital from a large and geographically dispersed crowd of backers, crowdfunding is often portrayed as a more open and inclusive funding channel, potentially lowering barriers for groups that have traditionally faced disadvantages in entrepreneurial finance (Serwaah & Shneor, 2021). At the same time, a growing body of evidence suggests that gender disparities persist in crowdfunding markets, raising questions about the mechanisms through which gendered evaluations continue to shape funding outcomes even in alternative finance settings (Cowden et al., 2021; Greenberg & Mollick, 2017; Johnson et al., 2018; Liao, 2021; Lin & Pursiainen, 2023).
Prior research on gender and crowdfunding provides a mixed picture: some studies document advantages for female founders in certain crowdfunding contexts (Greenberg & Mollick, 2017; Lin & Pursiainen, 2023), while others identify persistent disadvantages or no systematic differences, depending on platform characteristics, project categories, and outcome measures (Cowden et al., 2021; Elitzur & Solodoha, 2021; Johnson et al., 2018; Liao, 2021). To reconcile these inconsistencies, recent work increasingly emphasizes the role of observable cues that shape evaluators’ judgments under uncertainty (Anglin et al., 2018; Parhankangas & Renko, 2017). Among these cues, language plays a central role in crowdfunding. Campaign descriptions constitute a core element of crowdfunding pitches, providing backers with signals related to founders’ intentions, competence, and identity. Prior research demonstrates that linguistic styles vary systematically along gendered dimensions, most prominently communal expressions emphasizing warmth, cooperation, and care, and agentic expressions emphasizing assertiveness, ambition, and competence (Pietraszkiewicz et al., 2019).
Drawing on role congruity theory (Anglin et al., 2022; Eagly & Karau, 2002), prior research argues that such gendered cues influence evaluations depending on their alignment with prescriptive gender norms. When observable signals align with gendered role expectations, evaluations tend to be more favorable, whereas misalignment might give rise to penalties or skepticism (Rudman & Phelan, 2008). Within this study, we examine how gendered language functions as the primary observable basis on which backers assess gender role congruity in entrepreneurial fundraising contexts.
Entrepreneurship research has applied this perspective to explain gendered evaluation patterns across a range of contexts, including venture capital, accelerators, and other early-stage financing environments (Butticè et al., 2023; Eddleston et al., 2016). In crowdfunding, gendered language constitutes an observable linguistic signal embedded in campaign narratives through which gender role expectations might be implicitly activated during backers’ evaluations, without requiring the backers’ conscious intent or explicit assessment.
Despite these advances, existing research leaves two issues insufficiently addressed. First, crowdfunding performance is often conceptualized as a single outcome, typically operationalized as campaign success or failure. This unidimensional view obscures the possibility that gendered cues might affect different funding performance dimensions in distinct ways. In this study, we explicitly distinguish between access to funding—whether a campaign reaches its funding threshold—and its depth of funding, capturing the amount of capital that backers pledge. While closely related, these dimensions reflect qualitatively different aspects of entrepreneurial finance. Access to funding captures entry and approval, whereas depth of funding reflects the scale of financial commitment and the extent of crowd support. Distinguishing between these outcomes enables more precisely assessing how gendered evaluations shape not only whether female founders obtain funding, but also the level of financial resources they ultimately attract.
Second, while role congruity theory has been extensively applied in professional investment contexts, less is known about how these mechanisms manifest in reward-based crowdfunding and how they compare to other early-stage financing instruments. Crowdfunding differs from professionalized finance settings in several respects. Investment decisions are decentralized, individual pledges are typically small, and backers strongly rely on narrative information rather than formal due diligence (Allison et al., 2017; Anglin et al., 2018; Mollick, 2014). These characteristics make crowdfunding a theoretically informative context for examining whether gendered role expectations persist in alternative finance markets and how they are reflected across different dimensions of funding performance. Moreover, crowdfunding allows researchers to observe both access to funding and depth of funding within the same setting, providing a unique opportunity to examine whether gendered cues translate similarly—or differently—across these dimensions.
Against this background, we examine how gendered language in crowdfunding campaigns interacts with founder gender to shape funding outcomes in technology-based reward crowdfunding. We focus on technology projects as a domain in which masculine stereotypes are particularly salient and gender disparities in entrepreneurial finance are well documented (Brooks et al., 2014; Greenberg & Mollick, 2017; Seigner et al., 2022). In this context, gendered language serves as an observable signal embedded in campaign narratives, while role congruity theory provides the theoretical framework for understanding how such signals might shape evaluative responses under uncertainty. Accordingly, this study addresses the following research question: How does the use of gendered language interact with founder gender in shaping both access to funding and the depth of financial support in female-led technology crowdfunding?
To answer this question, we analyze a dataset of 671 US-based technology campaigns on Kickstarter. Using dictionary-based text analysis, we quantify communal and agentic language in campaign descriptions and relate these measures to funding outcomes.
Our results reveal a differentiated pattern. First, the results show that communal language is generally associated with lower probabilities of funding access and lower depth of funding in technology-oriented crowdfunding campaigns. However, the second main result indicates that communal language is associated with a higher probability of funding access for female-led campaigns. While communal language is penalized overall, female founders appear more likely to obtain funding access when their language aligns with traditionally feminine, communal expectations. This suggests that evaluators respond differently to the same linguistic signals depending on the founder’s gender. Third, despite this higher likelihood of funding access, female-led campaigns that use communal language achieve a significantly lower depth of funding. Thus, while gender-congruent language is associated with higher chances of obtaining funding access, it does not translate into a greater depth of funding, as the benefits associated with communal language for female founders appear limited to funding access rather than financial scale. Finally, the results show penalties when female founders use agentic language. Female-led campaigns emphasizing agentic traits such as assertiveness, dominance, or ambition experience lower probabilities of funding access and reduced depth of funding. These penalties indicate that deviations from expected gender roles are evaluated negatively, even in decentralized, crowd-based financing environments.
This study contributes to the literature in several ways. First, by distinguishing between access to funding and depth of funding, we extend prior research that has predominantly focused on binary success outcomes (Cumming et al., 2021; Greenberg & Mollick, 2017; Lin & Pursiainen, 2023; Mollick, 2014) and demonstrate that gendered evaluations operate differently across these dimensions. Second, we advance role congruity research (Eagly & Karau, 2002) by showing that role-congruent linguistic signaling can facilitate access to funding for female founders while simultaneously constraining the scale of financial support they receive, resulting in unequal payoffs despite apparent approval. Third, by situating our analysis in the context of reward-based crowdfunding, we contribute to the literature on alternative finance (Belleflamme et al., 2014; Cumming et al., 2021; Greenberg & Mollick, 2017) by demonstrating that decentralized funding mechanisms do not eliminate gendered evaluation patterns, but instead redistribute them across distinct funding performance dimensions.
Theoretical Background and Hypotheses Development
Gendered Language and Funding Success From a Social Role Congruity Perspective
Social role theory, as proposed by Eagly & Kite (1987), suggests that individuals’ behavior is largely influenced by the social roles they occupy. These roles are based on culturally established norms and expectations, which dictate appropriate behaviors for men and women. In a conventional context, gender roles allocate men to positions of high status and leadership (termed as “the breadwinner”), whereas women are predominantly linked with lower-status, subordinate positions (referred to as “the caretaker”) (del Carmen Triana et al., 2024; Eagly & Karau, 2002; Merluzzi & Phillips, 2022). More specifically, females are often attributed with communal qualities such as being affectionate, sympathetic, supportive, compassionate, gentle, and nurturing, whereas males are frequently identified with agentic characteristics such as being ambitious, aggressive, dominant, independent, self-confident and inclined to assume a leadership role (Heilman et al., 1989; Schein, 1975, 2001).
Building on social role theory, role congruity theory (Anglin et al., 2022; Eagly & Karau, 2002) posits that people are evaluated more favorably when their behaviors align with the gendered expectations of their social roles. When individuals deviate from these expected roles, they might face social penalties (Hoobler et al., 2009). For example, women who display agentic traits, such as assertiveness or competitiveness, might be viewed negatively because these behaviors conflict with the communal attributes typically associated with the female gender role (Merluzzi & Phillips, 2022). Conversely, men who display communal traits might also face backlash for not adhering to traditional masculine roles (del Carmen Triana et al., 2024; Eagly & Karau, 2002). This theory underscores the importance of role congruity in shaping perceptions and outcomes in both personal and professional contexts (Carli & Eagly, 2011; Eagly & Chin, 2010; Paustian-Underdahl et al., 2014).
Language plays a crucial role in expressing and reinforcing these social roles by providing observable cues through which evaluators infer role congruity in financing decisions (Kerber, 1988; Lakoff, 2004; Pennebaker et al., 2003; Um et al., 2022). In the crowdfunding context, gendered language can function as a subtle linguistic cue, reflecting socially learned expectations. One manifestation of role congruity is the use of gendered language, where certain words and phrases are associated with either femininity or masculinity. For example, words such as “collaborative,” “supportive,” and “nurturing” are often perceived as feminine, while terms such as “competitive,” “dominant,” and “independent” are seen as masculine (Brownlow et al., 2003; Gaucher et al., 2011; Leaper & Ayres, 2007). Therefore, language choice can signal conformity to or deviation from socially prescribed gender roles. In this study, we define gendered language as the systematic use of linguistic expressions that are culturally associated with femininity or masculinity and convey communal or agentic communication orientations. Communal language emphasizes warmth, cooperation, care, and relational focus, whereas agentic language emphasizes assertiveness, ambition, autonomy, and goal orientation (Eagly & Karau, 2002; Pietraszkiewicz et al., 2019). Throughout the manuscript, the terms gendered language, gender-role congruence, and alignment with gendered expectations refer to different analytical levels of the same underlying role congruity mechanism.
In entrepreneurial finance, gender-congruent language can influence how capital providers perceive competence and leadership ability and, ultimately, a venture’s success (Um et al., 2022). Studies in venture capital and pitching settings demonstrate that linguistic cues embedded in pitch narratives and presentations provide investors with inferences about entrepreneurs’ competence, growth orientation, and leadership potential (Balachandra et al., 2017; Brooks et al., 2014; Brush et al., 2018; Huang et al., 2021). This line of research suggests that gendered language can activate role-based expectations, shaping evaluators’ judgments to advantage gender-congruent communication while penalizing deviations from stereotypical norms, particularly in male-typed domains such as technology entrepreneurship (Eddleston et al., 2016; Um et al., 2022).
In the context of crowdfunding, the importance of language is particularly salient and represents a critical campaign feature (Adamska-Mieruszewska et al., 2021). Unlike traditional funding environments, where face-to-face interactions allow entrepreneurs to convey their vision, passion, and competence personally, crowdfunding relies on written communication and online communication tools such as videos (Moritz et al., 2015). The language that project descriptions use becomes a tool for entrepreneurs to establish credibility, build trust, and persuade individuals to support their projects financially. Building on role congruity theory, this study aims to examine how gender-congruent and gender-incongruent language relates to access to funding and depth of funding as two dimensions of crowdfunding performance.
Technology Crowdfunding as a Gendered Entrepreneurial Finance Context
The technology sector represents a gender-typed entrepreneurial context in which masculine norms strongly shape expectations regarding founders and ventures. Prior research consistently associates technology entrepreneurship with agentic attributes such as assertiveness, risk-taking, growth orientation, and technical competence, which are culturally coded as masculine (Eagly & Karau, 2002; Heilman et al., 1989; Seigner et al., 2022). As a result, women entrepreneurs in technology not only face numerical underrepresentation but also heightened scrutiny regarding their perceived fit with prevailing role expectations.
These gendered expectations extend to financing decisions. Studies on entrepreneurial finance show that investors tend to evaluate technology ventures through lenses of scalability, ambition, and growth potential, criteria that might disadvantage founders who do not conform to the stereotypical image of the technology entrepreneur (Brooks et al., 2014; Kanze et al., 2018). In this context, communication cues in crowdfunding become salient, as backers often lack the technical expertise to assess project feasibility directly and instead rely on narrative signals to infer competence and credibility (Ahlers et al., 2015; Parhankangas & Renko, 2017).
For female founders, this setting creates a tension between conforming to gendered expectations of communal behavior and meeting the agentic ideals associated with technology entrepreneurship. Prior research suggests that such role incongruity might lead to penalties when women display stereotypically masculine behavior, while conformity to feminine norms might enhance perceived legitimacy but simultaneously constrain evaluations of growth potential (Eagly & Karau, 2002; Seigner et al., 2022). Accordingly, the technology sector provides a relevant context for examining how gendered language is related to funding outcomes in crowdfunding, as linguistic signals might activate both industry-specific and gender-based stereotypes in backers’ evaluations.
Hypotheses Development
Impact of Communal Language on Crowdfunding Performance
Communal language emphasizes relational orientation, cooperation, care, and social connectedness (Brownlow et al., 2003; Gaucher et al., 2011). In technology-oriented reward-based crowdfunding, such language might conflict with dominant expectations that associate entrepreneurial success with assertiveness, ambition, and risk-taking (Heilman et al., 1989; Seigner et al., 2022). When backers evaluate whether to support a campaign at all, they strongly rely on narrative cues to assess feasibility, leadership capability, and strategic orientation under uncertainty (Parhankangas & Renko, 2017). In this context, communal language might be interpreted as signaling risk aversion or a lack of growth orientation, thereby reducing perceived legitimacy and lowering the likelihood that a campaign reaches its funding threshold (Brownlow et al., 2003; Newman et al., 2008; Obi, 2023). On this basis, we derive the following hypothesis:
In reward-based crowdfunding campaigns, communal language is negatively related to access to funding.
Beyond the binary decision of whether a campaign reaches its funding goal, communal language might also affect the depth of funding. In reward-based crowdfunding, higher funding amounts are closely associated with perceptions of scalability, ambition, and long-term growth potential (Belleflamme et al., 2014; Mollick, 2014). While communal language can foster goodwill and a sense of social connection, it might fail to convey urgency, assertiveness, or strong leadership signals that motivate backers to increase their financial commitment beyond the minimum threshold. As a result, even when campaigns using communal language achieve funding access, they might attract lower total funding amounts. This leads us to our next hypothesis:
In reward-based crowdfunding campaigns, communal language is negatively related to depth of funding.
Impact of Agentic Language on Crowdfunding Performance
Agentic language focuses on individual achievement, autonomy, and goal-driven actions (Brownlow et al., 2003; Gaucher et al., 2011; Leaper & Ayres, 2007). These attributes closely align with prevailing stereotypes of successful entrepreneurs, particularly in technology-driven contexts where innovation, decisiveness, and leadership are highly valued (Eagly & Karau, 2002; Nielsen & Binder, 2020). When backers decide whether to support a campaign, agentic language might function as a positive signal of competence and execution capability, increasing the likelihood that a campaign reaches its funding goal (Ahlers et al., 2015; Parhankangas & Renko, 2017). This leads to the following hypothesis:
In reward-based crowdfunding campaigns, agentic language is positively related to access to funding.
Agentic language might also influence the depth of funding. In reward-based crowdfunding, larger pledges are often motivated by expectations regarding product delivery, scalability, and future success (Steigenberger, 2017). Campaign narratives that emphasize determination, confidence, and strategic clarity might reassure backers that the entrepreneur is capable of delivering on promises, thereby encouraging higher individual contributions. Prior research shows that stereotypically entrepreneurial and leadership-oriented traits are positively associated with crowdfunding performance (Maurer et al., 2024). From this, we derive our next hypothesis:
In reward-based crowdfunding campaigns, agentic language is positively related to depth of funding.
Impact of Female Founders on the Relationship Between Communal Language and Crowdfunding Performance
As discussed in H1a and H1b, communal language is generally assumed to be negatively associated with crowdfunding performance. At the same time, prior research in entrepreneurial finance indicates that women are often disadvantaged in funding contexts and tend to secure less capital than their male counterparts (Geiger, 2020). Role congruity theory provides a framework for understanding how founders’ behavior aligns with prescriptive gender role expectations to mitigate such disadvantages (Anglin et al., 2022; Eagly & Karau, 2002). According to this perspective, individuals who behave in ways that are consistent with stereotypical gender norms are evaluated more favorably than those who violate these expectations.
With respect to funding access, female founders’ use of communal language in reward-based crowdfunding might therefore be particularly beneficial. Prior research suggests that communal communication closely aligns with societal expectations of female behavior and is perceived as more authentic, sincere, and trustworthy when used by women (Gorbatai & Nelson, 2015; Johnson et al., 2018; Wesemann & Wincent, 2021). In crowdfunding contexts, where backers strongly rely on narrative cues and often make threshold-based decisions about whether to support a project, such perceived authenticity might increase backers’ willingness to provide initial support. Moreover, in the male-dominated technology sector, female-led campaigns that emphasize communal qualities might positively stand out by offering a relational and cooperative framing that differentiates them from more stereotypically agentic campaigns (Seigner et al., 2022). Together, given that these mechanisms suggest that communal language might improve female founders’ likelihood of gaining access to funding, we propose the following hypothesis:
In female-led reward-based crowdfunding campaigns, the relationship between communal language and access to funding is less negative than in male-led reward-based crowdfunding campaigns.
Beyond the decision of whether a campaign reaches its funding goal, communal language might also positively shape the depth of funding for female-led campaigns. By fostering emotional connection, trust, and a sense of shared purpose, communal language might encourage backers not only to support the campaign but also to contribute more generously (Gorbatai & Nelson, 2015; Wesemann & Wincent, 2021). Because reward-based crowdfunding often involves relational and symbolic motivations alongside economic considerations, gender-congruent communal language might strengthen backers’ identification with the campaign and increase their willingness to commit higher capital amounts. In this sense, communal language might enhance both the likelihood and the intensity of support for female-led campaigns, offsetting the generally negative association between communal language and funding performance. Therefore, we hypothesize the following:
In female-led reward-based crowdfunding campaigns, the relationship between communal language and depth of funding is less negative than in male-led reward-based crowdfunding campaigns.
Impact of Female Founders on the Relationship Between Agentic Language and Crowdfunding Performance
In H2a and H2b, we assumed that agentic language—emphasizing assertiveness, ambition, autonomy, and goal orientation—is generally associated with more favorable crowdfunding outcomes. However, role congruity theory suggests that the effects of agentic language critically depend on the gender of the individual using it (Eagly & Karau, 2002). Agentic traits are culturally associated with masculinity, whereas women are expected to display communal characteristics such as warmth, cooperation, and care (Brownlow et al., 2003; Gaucher et al., 2011; Leaper & Ayres, 2007; Pietraszkiewicz et al., 2019). When female entrepreneurs employ agentic language, this might therefore create a perceived incongruence between their communication style and prescriptive gender role expectations.
With respect to funding access, this role incongruity might lead to negative evaluations and reduce the likelihood that female-led campaigns reach their funding goal. Role congruity theory predicts that behaviors violating gender norms trigger skepticism and backlash, particularly in evaluative contexts characterized by ambiguity and limited information (Eagly & Karau, 2002). In crowdfunding, where backers strongly rely on narrative cues to form impressions of founders, agentic language used by female entrepreneurs might be perceived as inauthentic, overly aggressive, or norm-violating (Rudman & Phelan, 2008). Such perceptions can undermine trust and legitimacy, reducing backers’ willingness to provide initial support. As a result, we hypothesize that the positive association between agentic language and funding access observed on average might be attenuated or reversed for female-led campaigns.
In female-led reward-based crowdfunding campaigns, the relationship between agentic language and access to funding is less positive than in male-led reward-based crowdfunding campaigns.
Beyond the decision of whether a campaign reaches its funding goal, role incongruity is likely to be particularly consequential for the depth of funding. Larger financial commitments involve higher perceived risk and require stronger confidence in the entrepreneur’s competence, leadership, and execution capability. Prior research shows that women who display agentic traits such as dominance or assertiveness often face stronger backlash when higher levels of authority, influence, or resource control are at stake (Eagly & Karau, 2002; Rudman & Phelan, 2008). In crowdfunding contexts, this dynamic might translate into backers’ reluctance to make larger pledges to female-led campaigns that employ agentic language, even if they are willing to provide minimal support.
Because depth of funding reflects assessments of scale, ambition, and upside potential rather than mere acceptability, gender-incongruent agentic language might disproportionately reduce backers’ willingness to escalate their financial commitment. Agentic language used by female entrepreneurs might thus amplify perceptions of norm violation precisely in the high-stakes dimension of crowdfunding performance, resulting in lower total funding amounts relative to comparable male-led campaigns. This leads to the following hypothesis:
In female-led reward-based crowdfunding campaigns, the relationship between agentic language and depth of funding is less positive than in male-led reward-based crowdfunding campaigns.
Figure 1 illustrates our research model, including the key constructs and hypotheses. Research model and hypotheses. The diagram illustrates the research model used in our analysis. It shows the independent variables (communal and agentic language) and the dependent variables (funding access and funding depth), with female founder serving as a moderator. The model is divided into two segments, one for communal language and another for agentic language, each analyzed separately to assess their relationship with crowdfunding outcomes. The model also reflects the interaction effects between the language types and the gender of the project founder, aiming to capture the influence of gendered language on crowdfunding outcomes
Data, Variables, and Sample Description
Data Collection
Our study focuses on technology-based crowdfunding campaigns from Kickstarter in 2023. Kickstarter is one of the most established and widely used crowdfunding platforms globally, and is also widely used for empirical research (e.g., Calic & Shevchenko, 2020; Chan et al., 2021; Lins et al., 2018; Tafesse, 2021). We focus on technology-based projects because—as discussed above—this category is strongly male-typed, making it a particularly suitable setting for examining role congruity and potential backlash against role-incongruent behavior in entrepreneurial finance. Technology is one of the largest categories on crowdfunding platforms and is commonly used to study how signals, gender, and language shape funding outcomes (Kleinert & Mochkabadi, 2022; Liao, 2021).
Moreover, we focus on the US because it leads globally in transaction value within reward-based crowdfunding, with an expected $464.80 million in 2024. To ensure linguistic consistency for LIWC-22, we restrict the sample to US-based Kickstarter campaigns in the technology category from 2023. We collected these campaigns using web scrapers, yielding 895 projects. Restricting the analysis to a single year helps hold constant time-varying conditions (e.g., economic conditions, platform policies, and technological trends).
Given the observational and unobtrusive nature of our data, we examine associations between gendered language use and crowdfunding outcomes and whether these associations differ by founder gender. We focused only on ‘failed’ and ‘successful’ projects (Chan et al., 2021), excluding campaigns marked as ‘canceled’ or ‘submitted’, as outcome status constitutes a defining characteristic of campaign performance and is essential for analyzing differences in funding dynamics. We further removed entries from the dataset that did not list the US as their country of origin, or did not use USD as their currency. Subsequently, we found that 15 campaigns included no story text, possibly focusing on images or videos. Since the story description is crucial for our linguistic analysis, we excluded these projects from our dataset. Next, we applied the “BigTwo” dictionary via LIWC to generate the variables of communal and agentic language. In alignment with previous research (e.g., Altmeier & Fisch, 2023; Block et al., 2023; Gorbatai & Nelson, 2015), we set a minimum word count of 50 for the analyzed texts. After these refinements, the final dataset comprises 671 projects.
Variables
Dependent Variables: Funding Access and Funding Depth
We used two variables to measure the performance of a reward-based crowdfunding campaign on Kickstarter. Our first dependent variable is funding access, a dummy variable that captures whether a project reached its funding goal at the end of the campaign. To gain additional insights on crowdfunding performance, we also consider the money a project raised as our second dependent variable funding depth. The variable captures the total amount of money pledged to a project at the end of the campaign. We used the natural logarithm of funding depth to address its skewness in our analysis (e.g., Calic & Shevchenko, 2020; Ryu, 2024; Yang, Li, et al., 2020).
Independent Variables: Communal and Agentic Language
We apply a linguistic perspective to reward-based crowdfunding and analyze the impact of role-specific language in the story description of a project. Unlike prior research relying on visual stimuli such as pitch videos (e.g., Cowden et al., 2021), we focus on written campaign narratives to examine how language might shape perceptions of founder characteristics. Therefore, our two independent variables, communal language and agentic language, capture communal and agentic expressions in project descriptions. These variables are calculated by LIWC based on the “Big Two (Agency & Communion) Dictionary” created to capture agentic and communal expressions in natural language 1 . The dictionary was created using the LIWC methodology and underwent rigorous validation through four distinct studies. These studies confirmed the dictionary’s validity through frequency-based analyses, semantic similarity measures, and subjective ratings. Notably, the dictionary strongly aligned with relevant LIWC categories and demonstrated excellent discriminant validity by exhibiting minimal overlap with unrelated categories.
For the linguistic analysis, we use LIWC, which has been increasingly utilized in recent entrepreneurship studies (e.g., Block et al., 2023; Moradi & Badrinarayanan, 2021; Tajvarpour & Pujari, 2022). LIWC facilitates the analysis of large text datasets and enables researchers to assess specific linguistic and psychological dimensions within a text. As a closed-dictionary software, LIWC calculates the percentage of the text composed of words found in its predefined dictionaries (Pennebaker et al., 2003). In line with this convention, Pietraszkiewicz et al. (2019) define their agency and communion scores as the percentage of words in a text that belong to the respective dictionaries. Accordingly, our agentic and communal language variables are LIWC percentage scores indicating the share of words in the campaign description that match the Big Two dictionary.
Moderator Variable: Female Founder
Our moderator variable, female founder, is a dummy variable that captures whether a female or male founder or CEO leads the project. We hand-coded the most prominently presented founder based on information and images in the project description and related creator materials (e.g., “meet our team”; “about us”), to mirror how campaigns signal founder gender to backers. Unlike approaches relying only on creator names (e.g., Gorbatai & Nelson, 2015; Wesemann & Wincent, 2021), this procedure accounts for cases where the listed creator is not the venture founder. Our approach does not capture the gender composition of the broader entrepreneurial team, which might also influence project design, signaling strategies, and funding outcomes (Bi et al., 2017). Future studies could incorporate team-level data to provide a more comprehensive assessment.
Moreover, we focus on the binary distinction between female and male founders because this is where prior research on role congruity theory and gender stereotypes in entrepreneurship provides the clearest theoretical grounding. While other gender identities are relevant and merit scholarly attention, the available data did not consistently allow us to identify or code them reliably, particularly given that self-presentation on crowdfunding platforms is often limited and strategically curated.
Control Variables
We also incorporated a set of control variables recommended in the literature to minimize the risk of omitted variable bias. First, we controlled for the financing goal (USD), as lower targets are associated with higher success likelihood, while also being potentially indicative of lower confidence to backers (Block et al., 2018; Prokop & Wang, 2022). Our dummy variable team captures whether an entrepreneurial team or a solo entrepreneur creates a project, as team-based ventures can signal greater credibility (Kim et al., 2016).
Research on this topic also denotes other characteristics of the campaign promotion, such as the number of words in the campaign description (measured as word count and log-transformed for skewness) and media use (e.g., the number of images or the availability of a video) as quality indicators (e.g., Bi et al., 2017; Mollick, 2014; Moritz et al., 2015; Shneor et al., 2022).
Additionally, we controlled for entrepreneur-related factors that are visible on the project’s page on Kickstarter (e.g., Moradi & Badrinarayanan, 2021; Tafesse, 2021). The dummy variables facebook, website, and collaborators capture whether the Facebook page, the creator’s website, and any collaborators are linked to the campaign page. Previous crowdfunding experience is measured as created projects and backed projects.
Description of Variables. This table presents a description and information about the data source for the variables used in the analysis
Descriptive Statistics
Summary Statistics of all Campaigns. This Table Summarizes the Key Variables Used in the Analysis, Including the Number of Observations (N), Means, and Distribution Statistics Such as the Minimum, Maximum, and Standard Deviation
The success rate of campaigns is 35.3%, indicating variability in outcomes, while the mean funding amount pledged is $73,505.12. However, this widely varies, with the highest funding amount of $4.12 million. 16.7% of the campaigns (overall, 112 out of the 671 projects in the dataset) are led by female founders, reflecting the general underrepresentation of female entrepreneurs in the economy. The language used in the campaign stories shows an average communal language score of 4.502 and an agentic language score of 6.183, reflecting diverse communication styles. The average word count in campaign stories is 750.735, with considerable variability, ranging from 54 to 4,031 words.
Summary Statistics by Campaign Leadership. The Table Reports Summary Statistics Separately for Female-Led and Male-Led Campaigns. Differences are Assessed Using Two-Sided T-tests, With T-values Reported for (Female-Led−Male-Led). Only Continuous Covariates are Reported; Binary (Dummy) Controls are Omitted
**p < 0.05.
Results
Impact of Communal Language on Crowdfunding Performance
Regression Analysis on the Impact of Communal and Agentic Language on Funding Access. This Table Reports Logistic Regression Results for Funding Access (Reaching the Funding Goal). Models 1–6 Introduce Communal and Agentic Language Separately and Jointly and Include Interactions With Female-Led Campaigns to Assess Whether These Associations Differ by Founder Gender. All Models Control for Campaign Characteristics and Creator Attributes, Including Campaign Goal (log), Visual Content (Number of Images and Video Availability), Environmental and AI Indicators, Story Length (log Word Count), Team Status, Social Media and Website Presence, Creators’ Prior and Supported Projects, and Collaborators. Standard Errors are Reported in Parentheses. ***p < 0.01, **p < 0.05, *p < 0.10
Regression Analysis on the Impact of Communal and Agentic Language on Funding Depth. This Table Reports OLS Regression Results for Funding Depth (Log-Transformed). Models 1–6 Introduce Communal and Agentic Language Separately and Jointly and Include Interactions With Female-Led Campaigns to Examine Whether These Associations Differ by Founder Gender. All Models Control for Campaign and Creator Characteristics, Including Campaign Goal (log), Visual Content (Number of Images and Video Availability), Environmental and AI Indicators, Story Length (log Word Count), Team Status, Facebook and Website Presence, Creators’ Prior and Supported Projects, Collaborators, and Category. Robust Standard Errors are Reported in Parentheses. ***p < 0.01, **p < 0.05, *p < 0.10
Regarding H1a, the logistic regression results show that communal language is significantly and negatively associated with the probability of funding access (β = −0.125, p < 0.05 Table 4/Model 2). To facilitate interpretation, we translate this estimate into predicted probabilities using predictive margins. The predicted probability of funding access is 34.9% at the mean level of communal language (4.50). At one standard deviation above this mean level (7.39), the predicted probability is 31.9%, corresponding to a difference of about 3 percentage points, thus supporting H1a.
To test H1b, we estimate a linear regression model with funding depth as the dependent variable. The results show that communal language is negatively associated with the depth of funding (β = −0.099, p < 0.01 in Table 5/Model 2). This finding is consistent with H1b, indicating that communal language is associated with lower total depth of funding. To facilitate interpretation, we translate this estimate into economic effects using predictive margins. The predicted funding amount is $40,314 at the mean level of communal language (4.50). At one standard deviation above this mean level (7.39), the predicted funding amount is $30,304, corresponding to a difference of about $10,009 (p = 0.003).
Impact of Agentic Language on Crowdfunding Performance
In H2a, we hypothesize that agentic language is positively linked to funding access. However, we reject H2a because Model 2 in Table 5 shows a lack of statistical significance (β = −0.095, p > 0.1).
Contrary to H2b, the analysis also does not find a significant positive association between agentic language and the depth of funding. The coefficient for agentic language is not significant (β = −0.056, p > 0.1 in Table 5/Model 3). Thus, H2b is not supported.
Impact of Female Founders on the Relationship Between Communal Language and Crowdfunding Performance
We further investigated whether having a female founder is related to differences in the relationship between communal language and crowdfunding performance. The interaction between female founder and communal language is positive and significant in Model 4 in Table 4 (β = 0.222, p < 0.05), which supports H3a. This indicates that for female-led campaigns, communal language is associated with a higher probability of funding access.
To facilitate interpretation, we translate the interaction effect into predicted probabilities using predictive margins. For male-led projects, the predicted probability of funding access is 33.3% at the mean level of communal language (4.50). At one standard deviation above this mean level (7.39), the predicted probability is 29.4%, corresponding to a difference of about 3.9 percentage points. For female-led projects, the predicted probability of success is 40.9% at the mean level of communal language and is 42.2% at one standard deviation above the mean, corresponding to a difference of about 1.3 percentage points. Accordingly, a one standard deviation increase in communal language is associated with a 5.3 percentage-point larger difference in funding access for female-led compared to male-led projects (p = 0.071). Figure 2 visualizes these predictive margins across standardized levels of communal language, illustrating that the predicted probability of funding access is lower at higher levels of communal language for male-led projects, whereas it is slightly higher for female-led projects. Adjusted predictions of funding access across levels of communal language for female-led and male-led campaigns. This figure visualizes the interaction between communal language and female-led status on the predicted probability of success.
However, contrary to our H3b, we expected the relationship between communal language and depth of funding to differ by founder gender, such that the association would be attenuated for female-led campaigns. Our results do not provide evidence consistent with this expectation, outlining that female founders negatively moderate the relationship between communal language and funding depth (Table 5/Model 4, β = −0.167, p < 0.1). This finding indicates that for female-led campaigns, communal language is associated with lower depth of funding. Therefore, we reject H3b.
To facilitate interpretation, we translate this interaction into economic effects using predictive margins. Consistent with the negative moderation reported above, the predicted funding amount for male-led campaigns is $37,854 at the mean level of communal language (4.50) and $30,514 at one standard deviation above the mean (7.39), implying a difference of approximately $7,340. For female-led campaigns, the predicted funding amount is $56,771 at the mean level of communal language and $28,272 at one standard deviation above the mean, corresponding to a difference of about $28,499. Accordingly, a one standard-deviation increase in communal language is associated with a $21,159 larger negative difference in funding amount for female-led compared to male-led campaigns (p = 0.059), highlighting gender differences in the relationship between communal language and funding depth. Figure 3 visualizes these adjusted predictions across standardized levels of communal language, indicating that predicted funding amounts are lower at higher levels of communal language, with the difference being more pronounced for female-than for male-led campaigns. Adjusted predictions of funding depth across levels of communal language for female-led and male-led campaigns. This figure visualizes the interaction between communal language and female-led status on the predicted funding depth.
This observed phenomenon, where female-led campaigns utilizing communal language achieve higher probabilities of funding access but significantly lower depth of funding, suggests that backers might be more inclined to support these campaigns with smaller contributions. Communal language, which emphasizes themes of collaboration, community, and relationships, fosters a sense of collective support among backers, whereby social norms and altruism further shape this dynamic. Backers might thus support female entrepreneurs based on a sense of social responsibility (e.g., promoting gender equality), which can increase contributions but in a more symbolic way, resulting in smaller average amounts.
Impact of Female Founders on the Relationship Between Agentic Language and Crowdfunding Performance
Finally, we investigated whether the relationship between agentic language and crowdfunding performance differs by founder gender. In H4a, we expected that the association between agentic language and funding access would be weaker for female-than for male-led campaigns. The interaction between female founder and agentic language is indeed significant and negative in Model 5 in Table 4 (β = −0.250, p < 0.1), supporting H4a. This result shows that female-led campaigns are associated with a change in direction of the relationship between agentic language and funding access, with higher levels of agentic language associated with a lower probability of funding access when high levels of agentic language are used. This indicates that female leadership might be associated with differences in how agentic language is perceived in crowdfunding contexts.
To facilitate interpretation, we translate this estimate into predicted probabilities using predictive margins. For male-led projects, the predicted probability of funding access is 34.0% at the mean level of agentic language (6.18) and 33.0% at one standard deviation above the mean (8.60), corresponding to a difference of about 1 percentage point. For female-led projects, the predicted probability of success is 40.9% at the mean level of agentic language and 34.6% at one standard deviation above the mean, corresponding to a difference of about 6.3 percentage points. Accordingly, a one standard deviation change in agentic language is related to a 5.3 percentage-point more negative association with funding access for female-than for male-led projects (p = 0.061). Figure 4 visualizes these adjusted predictions across standardized levels of agentic language, indicating that the predicted probability of funding access declines as agentic language increases, with a notably steeper decline for female-than for male-led campaigns. Adjusted predictions of funding access across levels of agentic language for female-led and male-led campaigns. This figure visualizes the interaction between agentic language and female-led status on the predicted probability of success.
A similar pattern emerges in the interaction of female founder on the relationship between agentic language and funding depth as the interaction effect of female founder and agentic language is negative and significant in Model 5 in Table 5 (β = −0.256, p < 0.05), supporting H4b. Thus, female founders using agentic language face lower probabilities of funding access and lower depth of funding regarding the amounts raised, consistent with the idea that adopting a more traditionally masculine communication style might be penalized when it conflicts with gender expectations.
To facilitate interpretation, we translate this estimate into economic effects using predictive margins. For male-led projects, the predicted funding amount is $49,452 at the mean level of agentic language (6.18). At one standard deviation above this mean level (8.60), the predicted funding amount is $47,559, corresponding to a difference of about $1,894 (p = 0.712). For female-led projects, the predicted funding amount is $70,151 at the mean level of agentic language and $36,419 at one standard deviation above the mean, corresponding to a difference of about $33,732 (p = 0.041). Accordingly, a one standard deviation increase in agentic language is related to $31,838 lower funding amount for female-led than for male-led projects (p = 0.060). Figure 5 visualizes these adjusted predictions across standardized levels of agentic language, indicating that predicted funding amounts are lower at higher levels of agentic language, with the difference being substantially more pronounced for female-than for male-led campaigns. Adjusted predictions of funding depth across levels of agentic language for female-led and male-led campaigns. This figure visualizes the interaction between agentic language and female-led status on the predicted funding depth.
Robustness of Results
To test the fit of our logistic regression model, we used the Hosmer–Lemeshow goodness-of-fit test. The results indicate that the model fits the data well (chi-square statistic = 4.75; p = 0.7836). Next, we tested the assumption of homoscedasticity for our linear regression using the Breusch-Pagan test. The results indicate that heteroskedasticity is present (chi-square statistic = 10.98; p < 0.001). Therefore, we use robust standard errors throughout. Finally, we examined multicollinearity using correlation coefficients and variance inflation factors. As Table 6 shows, the correlation matrix and variance inflation factors (VIFs) suggest no multicollinearity concerns; VIF values range from 1.032 to 2.073 and remain well below conventional thresholds (O’brien, 2007).
To address the concern that our funding depth results might primarily reflect differences in campaigns’ goal-setting, we conduct an additional robustness check based on relative funding performance. Specifically, we operationalize performance as the log ratio of funding to the funding goal, log (funding depth/goal; equivalently, funding depth (log) – goal (log)), thereby normalizing realized funding by the campaign’s target. The results are consistent with our main analyses: communal language remains negatively associated with funding performance, and this pattern holds across model specifications with and without additionally controlling for goal (log). 2
To account for potential differences between female- and male-led campaigns, we used propensity score matching with multiple matches per treated observation. Each female-led campaign was matched to up to four nearest male-led campaigns based on the funding goal. We included funding goal as a matching criterion, following Cumming et al. (2021) and Duan et al. (2020), who argue that entrepreneurs set campaign goals and that goals, which are therefore correlated with intrinsic project characteristics.
Variance Inflation Factor (VIF) and Correlation Matrix. This Table Presents the Variance Inflation Factors (VIF) and Correlation Matrix for the Variables Used in the Analysis. The VIF Values (Column 2) Indicate That Multicollinearity is Not a Concern, as all Values are Below the Critical Threshold of 10. The Correlation Matrix (Columns 3–16) Displays the Pairwise Correlations Between Key Independent Variables, Such as Female Founders, Communal Language, Agentic Language, and Various Control Variables Like Goal (log), Images, and Team. Significant Correlations are Marked With Asterisks (***p < 0.01, **p < 0.05, *p < 0.1), Highlighting Relationships Between Variables That May Influence the Regression Outcomes. The Correlation Between the Variables Remains Moderate, Further Indication That the Independent Variables do Not Exhibit Problematic Interdependence
***p < 0.01, **p < 0.05, *p < 0.1.
Regression Analysis on Funding Access (Post-matching). This Table Presents Regression Results Estimated After the Propensity Score Matching Procedure, Examining Associations Between Communal and Agentic Language and Funding Access. The Key Independent Variables Include Communal and Agentic Language as Well as Interactions With Female-Led Campaigns to Assess Whether These Associations Differ by Founder Gender. Control Variables Such as the Campaign Goal, Use of Images and Videos, and Team Presence are Included to Account for Other Campaign Characteristics. The Analysis is Based on a Matched Sample of 386 Observations, of Which 112 are Female-Led Campaigns. Standard Errors are Reported in Parentheses, and Significance Levels are Indicated as Follows: ***p < 0.01, **p < 0.05, *p < 0.10
Regression Analysis on Funding Depth (Post-matching). This Table Presents Regression Results Estimated After the Propensity Score Matching Procedure, Examining Associations Between Communal and Agentic Language and Funding Depth (Log-Transformed). The Key Independent Variables Include Communal and Agentic Language as Well as Interactions With Female-Led Campaigns to Assess Whether These Associations Differ by Founder Gender. Control Variables Such as the Campaign Goal, Use of Images and Videos, and Team Presence are Included to Account for Other Campaign Characteristics. The Analysis is Based on a Matched Sample of 386 Observations, of Which 112 are Female-Led Campaigns. Standard Errors are Reported in Parentheses, and Significance Levels are Indicated as Follows: ***p < 0.01, **p < 0.05, *p < 0.10
Discussion
Discussion of Findings in Relation to Prior Literature
This study’s findings both support and extend prior research on gender bias and role congruity in entrepreneurial finance, while also challenging optimistic interpretations of reward-based crowdfunding, at least in male-typed categories such as technology.
First, the finding that communal language is generally associated with lower probabilities of access to funding and lower depths of funding supports earlier work showing that entrepreneurship, particularly in technology-oriented contexts, is culturally associated with agentic traits, whereas communal orientations are often evaluated less favorably. Prior research documents that relational and care-oriented behaviors tend to conflict with dominant expectations of entrepreneurial competence and leadership (Brownlow et al., 2003; Gaucher et al., 2011). Consistent with Eddleston et al. (2016), our results point to systematic links between gendered expectations about “appropriate” founder behavior and venture evaluations. They are also in line with evidence from venture capital (Balachandra, 2020; Balachandra et al., 2017; Butticè et al., 2023) and early-stage investor settings, e.g., accelerators (Yang, Kher, & Newbert, 2020) showing that role-incongruent signals are associated with less favorable evaluations and lower investment-related outcomes. Extending these patterns to reward-based crowdfunding, our findings suggest that communal orientations tend to be evaluated less favorably in entrepreneurial finance.
Second, the finding that communal language is associated with a higher probability of access to funding for female-led campaigns extends earlier work on role congruity theory. Prior studies suggest that women benefit when their entrepreneurial behavior aligns with traditional gender role expectations by enhancing perceived legitimacy and reducing social penalties (Eagly & Karau, 2002; Eddleston et al., 2016). Our results are consistent with this logic and extend it to a digitally mediated crowdfunding context, where evaluative judgments are made by a dispersed crowd rather than professional investors. Similar to Butticè et al. (2023), who show that role congruity is associated with investment success in venture capital, we demonstrate that role-congruent behavior is related to a higher likelihood of approval in crowdfunding. This extension suggests that role congruity mechanisms are not confined to professionalized or hierarchical financing contexts but also operate in crowd-based environments.
Third, the finding that female-led campaigns using communal language raise significantly lower funding amounts challenges earlier research suggesting that reward-based crowdfunding might systematically advantage female entrepreneurs. Studies such as Greenberg and Mollick (2017) and Lin and Pursiainen (2023) argue that women might perform equally well or better than men in reward-based crowdfunding. Our results qualify this conclusion by demonstrating that gendered advantages operate differently across distinct dimensions of funding performance. Specifically, apparent advantages are limited to access to funding, whereas disadvantages persist with respect to depth of funding. While female entrepreneurs might be rewarded for conforming to gender norms by gaining access to funding, this reward remains confined to crossing the funding threshold and does not extend to attracting larger volumes of capital. Conceptually, this distinction suggests that access to funding reflects evaluative judgments related to legitimacy and acceptability, whereas depth of funding captures assessments of scale, growth potential, and the perceived scope of the venture. One explanation might be that gender-congruent behavior is associated with higher perceived legitimacy and a greater willingness to provide initial support, facilitating approval and entry into the funded set, yet simultaneously reinforces stereotypes that constrain perceptions of growth potential or scalability, particularly in technology-oriented projects (Butticè et al., 2023; Eddleston et al., 2016).
Finally, the significant penalties observed when female founders use agentic language provide evidence consistent with prior research on backlash and role incongruity. Role congruity theory predicts negative evaluations when women display behaviors that violate prescriptive gender norms (Eagly & Karau, 2002). Empirical evidence from entrepreneurial and investment contexts shows that women who exhibit agentic traits often face social and economic penalties (Balachandra, 2020; Rudman & Phelan, 2008). Our findings are consistent with Butticè et al. (2023), who demonstrate that role-incongruent gender behavior reduces investment success. Moreover, we extend this work by showing that such penalties persist in crowdfunding contexts. Importantly, these penalties manifest across both access to funding and depth of funding, indicating that role-incongruent signaling not only reduces the likelihood that female-led campaigns gain initial approval but also constrains the scale of financial support they receive when funded. This extension suggests that the absence of traditional gatekeepers does not eliminate role incongruity penalties but instead relocates them to decentralized, crowd-based evaluations.
Theoretical Contributions
Our study contributes to the existing literature in three distinct ways. First, our findings extend the crowdfunding literature by highlighting the role of gender-specific language as a critical factor influencing campaign outcomes. While previous research has primarily focused on founders’ or backers’ demographic characteristics (e.g., Cumming et al., 2021; Greenberg & Mollick, 2017), our analysis introduces a new dimension by demonstrating that the language used in project descriptions, specifically communal language, is systematically associated with different dimensions of crowdfunding performance rather than with performance as a single, undifferentiated outcome. When applied by female founders, communal language increases the likelihood of funding access yet lowers the amount raised. This divergence underscores that access to funding and depth of funding capture conceptually distinct evaluative processes: whereas access reflects threshold-based judgments related to legitimacy, acceptability, and initial support, depth of funding reflects assessments of scale, ambition, and the magnitude of financial commitment. We hereby show that linguistic strategies might facilitate entry into the funded set while simultaneously constraining the scale of financial resources allocated, illustrating how gendered language and founder gender interact to produce unequal payoffs across different dimensions of crowdfunding performance.
Second, we contribute to role congruity theory by testing its relevance in online entrepreneurial finance. We show that gendered language is linked to crowdfunding performance via its (in)congruence with gender-role expectations (del Carmen Triana et al., 2024; Eagly & Karau, 2002). Because crowdfunding pitches are digitally mediated and evaluated by dispersed non-professional backers, our findings extend role congruity theory beyond traditional organizational settings (e.g., Uzuegbunam et al., 2021; Yang & Triana, 2019) and inform research on the role of perceived masculine and feminine traits in the evaluation of nascent entrepreneurs (Cowden et al., 2021).
Third, we contribute to the literature on stereotype management by illustrating that female entrepreneurs might navigate gendered expectations in crowdfunding through strategic language use. Our findings suggest that while the use of communal language is related to differences in a campaign’s likelihood of reaching its funding goal, it is often associated with smaller financial contributions, possibly due to heightened perceptions of risk associated with female-led ventures (Brooks et al., 2014; Kanze et al., 2018). One further possible explanation is a divergence between initial support decisions and actual contribution levels, with gendered preferences becoming more visible in higher-stakes commitments (Hao et al., 2024). This insight adds to the existing literature on gender and female entrepreneurship by suggesting that gender-specific language is a double-edged sword in crowdfunding, highlighting the need for a balanced approach to stereotype management (Malmström et al., 2020).
Our findings also resonate with related research that documents gender-asymmetric effects in crowdfunding. For instance, Serwaah and Shneor (2024) find that homophily positively predicts female backers’ funding intentions, but no such effect exists for men. This parallel suggests that backers’ reactions are not gender invariant but rather reflect selective psychological mechanisms that operate only in the case of women. Our study adds to this broader pattern by showing that female entrepreneurs who employ agentic language are penalized, while male entrepreneurs using communal language do not face equivalent disadvantages.
In sum, this study supports existing evidence that social role congruity shapes entrepreneurial evaluations, extends this insight to digitally mediated crowdfunding environments, and challenges assumptions that crowdfunding mitigates gender bias. Instead, our findings indicate that crowdfunding reshapes—rather than eliminates—gendered evaluation patterns by rewarding role conformity symbolically while maintaining material constraints on female entrepreneurs.
Limitations and Future Research Opportunities
As with any empirical study, this research is subject to several limitations. First, the dataset is restricted to campaigns from a single reward-based crowdfunding platform and focuses exclusively on technology projects. Although Kickstarter is the largest crowdfunding platform worldwide, its scale, global reach, and relatively mature institutional setting make it an outlier within the broader crowdfunding landscape. Consequently, the findings might not readily generalize to smaller platforms that primarily operate at the national or regional level. In addition, because our analysis is limited to technology projects, as a male-typed and innovation-intensive industry, the observed patterns might not extend to other industries with different gender compositions, norms, or evaluative criteria. Future research could therefore examine whether similar relationships emerge on other platforms with different governance structures and market scopes, as well as across industries that vary in their degree of gender-typing and technological orientation.
Second, while all campaigns originate from the US, the US context itself is marked by substantial heterogeneity in cultural norms and economic development levels across states and local communities. Such variation might affect both entrepreneurial behavior and backers’ evaluations. Comparative analyses across US states or communities, as well as cross-country studies, would help to assess the robustness of the findings in more diverse institutional and cultural settings.
Third, a further limitation concerns the absence of data on backer characteristics. In particular, information on the gender of individual backers is not available. This constraint is not unique to the present study but applies broadly to empirical crowdfunding research, as major platforms do not disclose systematic backer-level demographic information. Consequently, our analysis focuses on aggregate campaign outcomes and cannot assess whether female-led campaigns employing communal language attract different types of backers or benefit from gender-based homophily effects. Future research could address this limitation by using experimental designs, surveys, or alternative data sources that allow for a more direct examination of backer demographics and targeting strategies.
Fourth, the study does not directly address the mechanisms underlying gender-asymmetric evaluation patterns observed in crowdfunding. Prior research suggests that such asymmetries are persistent across platforms and contexts (Serwaah & Shneor, 2024). While our findings document systematic associations among gendered language, founder gender, and funding outcomes, the underlying psychological and behavioral processes driving these patterns remain unobserved. Further research is therefore necessary to examine the psychological or behavioral processes that drive these effects. In particular, future studies could employ methods that directly capture how backers attend to and process different elements of the campaign page. Approaches such as eye-tracking or other attention-based techniques commonly used in marketing research could provide valuable insights into which textual, visual, or narrative cues are salient in backers’ decision-making and how gendered signals are implicitly processed during evaluation. In this context, campaign videos represent a particularly promising avenue for future research, as they convey additional verbal and nonverbal cues such as tone, facial expressions, and delivery style that might interact with textual language to shape gendered evaluations. Integrating textual and video-based measures would thus provide a more comprehensive understanding of how gendered signals operate across multiple communication channels in crowdfunding.
Implications for Practice
From a practical standpoint, entrepreneurs must be aware of investor biases in crowdfunding contexts. Given that casual backers in crowdfunding are not professional investors but individuals who invest in a business or product idea, they might react differently to language emphasizing risk-taking or aggressive promotion. Our research indicates that for female-led campaigns, using communal language that aligns with traditional female stereotypes can significantly increase the likelihood of funding access in reward-based crowdfunding, even in male-dominated industries such as technology. This suggests that female entrepreneurs might benefit from emphasizing communal traits in their campaign narratives, as such language seems to resonate positively with backers and might enhance perceived authenticity.
However, our findings also indicate that this advantage is primarily confined to access to funding and does not translate into greater depth of funding. As a result, female founders should be aware that while communal language might facilitate initial approval by the crowd, it might simultaneously constrain the scale of financial resources available for subsequent growth and scaling. While female-led campaigns already tend to set higher funding targets on average than male-led campaigns, they might benefit from strategically combining communal language with even more ambitious funding goals to ensure that successfully reaching the funding threshold translates into a level of capital that supports meaningful project development and scaling. Being aware of this trade-off between access to funding and depth of funding is particularly important when designing campaign narratives and funding targets, especially for technology projects that require substantial resources to scale.
Conclusion
This study examines how gendered language in reward-based crowdfunding campaigns relates to access to funding and depth of funding, as well as how these associations depend on founder gender. Using 671 US-based technology projects on Kickstarter, we find gender-contingent evaluation patterns. Communal language is generally associated with lower crowdfunding performance, yet for female founders it is associated with a higher likelihood of access to funding, along with lower depths of funding. In contrast, female founders using agentic, stereotypically masculine language are associated with lower access to funding probabilities and lower amounts raised.
Together, these findings point to a pronounced double bind for women in technology-based reward crowdfunding: alignment with gender stereotypes appears linked to greater symbolic support but constrained financial upside, whereas deviation from these stereotypes is linked to lower financial outcomes. While observational, our results suggest that gendered language is a salient channel through which differential evaluations might emerge in crowdfunding contexts. By extending role congruity theory to digitally mediated entrepreneurial finance, we show that such patterns can persist even in reward-based crowdfunding environments often viewed as more accessible, raising questions about whether platforms mitigate or reproduce gendered dynamics in entrepreneurial finance.
Footnotes
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The paper uses public data or commercial data available for purchase or via subscription. The authors will provide instructions regarding how to obtain all data, as well as all code necessary to reproduce the paper’s results.
Use of AI Statement
During the preparation of this manuscript, the authors used ChatGPT (OpenAI) to assist with phrasing and refinement of the text. After employing this tool, the authors thoroughly reviewed and edited the content to ensure accuracy and take full responsibility for the final version of the manuscript.