Sage Journals: Discover world-class research

Abstract

The domain of finance is stereotypically associated with men, and these stereotypes can permeate into the language used by consumers. This study examines whether women and men employ language differently within the finance domain, aiming to better understand the gender gap in consumer finance and inform interventions to mitigate it. Using interdisciplinary approach that integrates machine learning, statistics, and linguistics, we analyzed three distinct language corpora produced by non-expert women and men and centered around 10 key terms relevant to consumer finance. Our analyses revealed notable gender-based differences in language use, manifested in both surface and deep structures of language. These differences were observed in word frequency, metaphor usage, professionalization of language, and conversational strategies, confirming patterns known from previous research in other subject domains. A novel contribution is the identification of a semantic distinction: men’s language more frequently signals agency (active semantic value), whereas women’s language tends to adopt an “experiencer” stance (passive semantic value). We discuss the implications of these findings, emphasizing the need for financial education initiatives that challenge stereotypes and empower both men and women by addressing their distinct financial perspectives and needs.

Keywords

consumer finance financial literacy gender gender gap language Polish

Introduction

Studies have shown that women perform worse than men on financial literacy tests, report lower financial confidence, and are less likely to participate in financial markets (Al-Bahrani et al., 2020; Almenberg & Dreber, 2015; Arellano et al., 2018; Cwynar et al., 2019; Fonseca & Lord, 2020; Grohmann & Schoofs, 2021; Ooi, 2020; Preston & Wright, 2024; Cwynar et al., 2025). This phenomenon – commonly referred to as the gender gap in consumer finance – perpetuates gender inequality and remains largely unexplained by observable differences between women and men such as age, education or income (Bucher-Koenen et al., 2017; Cupák et al., 2018; Potrich et al., 2024; Robson & Peetz, 2020). Researchers suggest that it may result from less observable cultural factors related to social gender roles and the ensuing stereotypes (Aristei & Gallo, 2022; Bottazzi & Lusardi, 2021; Driva et al., 2016; Preston et al., 2023; Rink et al., 2021; Tinghög et al., 2021).

In this article, we examine one such factor, which is a primary manifestation of culture – language. Its role in the gender gap in consumer finance is not well understood. Language encodes socially imposed gender roles and stereotypes (Goodhew et al., 2022; Johansson, 2005; Shoham & Lee, 2018). Extensive literature indicates that the language used by women differs from that used by men due to cultural rather than biological factors (Meier et al., 2020; Pennebaker et al., 2003; Piersoul & Van de Velde, 2023). Male linguistic expression tends to be more direct, instrumental, impersonal and socially distant, often focusing on object characteristics and neutral topics. In contrast, female linguistic expression is generally more elaborate, warmer, more interpersonal, and frequently involves references to psychological and social processes, with a tone that is more polite, complimentary, and considerate (Biber & Burges, 2000; Mulac et al., 2001; Newman et al., 2008; Park et al., 2016; Thomson et al., 2001). Additionally, there are gendered expressions and words tied to stereotypes about social gender roles (e.g., leader, competitive, dominant vs. support, understand, interpersonal) (Gaucher et al., 2011).

Although the role of language in shaping consumer financial choices has been recognized, particularly through framing studies (Bayer & Ke, 2013; Florence et al., 2022; Jin et al., 2017), the impact of language on the gender gap in consumer finance is only beginning to emerge (Boggio et al., 2015, 2020). Language, and words specifically, are considered “strategic instruments of influence” (Farrow et al., 2018, p. 580), capable of shaping cognition, attitudes, and behavior by enhancing identity and social belonging (Farrow et al., 2018; Phillips & Boroditsky, 2003). By using specific words or invoking specific metaphors, language signals what is compatible with a person’s gender and the resulting social norms, encouraging or discouraging a particular activity. Nowadays, finance is still stereotyped as a masculine domain, associated with active, sometimes aggressive, pursuit of livelihoods and material prosperity (Allen & Gervais, 2017; Tinghög et al., 2021; Von Hippel et al., 2015). The domain-specific masculine language might contribute to the gender gap in consumer finance by not giving women a sense of belonging and resulting in their withdrawal from the domain of finance.

Identifying the linguistic causes of this gap is necessary for developing expedient solutions to narrow or even eliminate it. Confirming language as a factor in this detrimental phenomenon could significantly improve financial education programs. Goodhew et al. (2022) empirically show that gender stereotypes are pervasive in general language use, that is even when communication is aimed at large and diffuse audiences, and does not assume interaction – as is the case with many financial education materials (book content, websites, leaflets, podcasts etc.). Ample empirical evidence indicates that even minor linguistic alteration can lead to significant behavioral changes (see the literature indicated in Gaucher et al., 2011). Thus, altering the linguistic design of financial education materials could be a cost-effective strategy for policymakers to reduce gender inequality (Farrow et al., 2018), offering a low-cost solution to financial education, which is considered expensive and economically inefficient (Willis, 2008).

The lack of research on gender differences in language use in consumer finance, and the potentially serious socioeconomic implications of the existence of such differences, prompted us to fill this gap. Therefore, the purpose of this article is to investigate whether women use written language differently than men when referring to financial issues. In undertaking this study, we posed three questions: (Q1) Does the written language of women differ from that of men in the domain of consumer finance? (Q2) If it does, what are the differences? (Q3) Are these possible differences in line with what is known about women’s and men’s language use from wider trends observed in other content (subject) domains, or are they specific to the consumer finance domain? To answer these questions, we conducted three studies using language corpora (Polish) on 10 financial terms (cost, credit, insurance, interest, investment, money, pension, profit, saving, tax) commonly used in formal and informal consumer finances communication. These corpora were analyzed separately using machine learning (natural language processing), statistical, and linguistic methods to identify gender differences.

Our study makes an original contribution to two strands of literature. First, it advances the use of textual and machine learning analysis in financial phenomena. While such analyses have been applied in corporate finance (see an overview in Li, 2020), their application to consumer finance is emerging (Levantesi & Zacchia, 2021). Our study is the first to use machine learning to examine gender-specific language in consumer finance contexts.

Second, it contributes to the literature on gender differences in financial language. While general gender differences in language use are well-documented, little research has explored these differences within specific domains, such as consumer finance, which possess distinct linguistic characteristics. Previous research by Boggio et al. (2015) showed that metaphors on retail investor websites predominantly draw from masculine conceptual domains (such as physical activity, game, war/conflict, etc.). However, this study focuses exclusively on metaphors, using a very small (totaling around 15,000 words) and undifferentiated linguistic corpus (topics related only to investing) and very simple analytical methods (expert identification of metaphors and links between gender and the use of these metaphors).

The second study which identifies gender differences in financial language use (Ben-Shmuel et al., 2024), although rigorously designed and conducted, uses only a qualitative evaluation of a linguistic corpus from only one source (social media) created by individuals with (presumably) above-average financial expertise (finfluencers).

In contrast, our study uses three language corpora, which are large and diverse – both in terms of the nature of communication (e.g., parsimonious and concise statements, constrained by word limits, on Twitter vs. the unrestricted and unlimited written expression in a questionnaire survey) and in terms of subject matter (the full range of issues representing consumer finance: from day-to-day cash management to savings to taxes and insurance). Second, the content we analyzed was produced largely by non-expert women and men (ordinary consumers of both sexes). Third, our study uses a rich array of analytical methods, both statistical and linguistic, as well as those representing natural language processing. As a result, we were able to study very different aspects of language, both those manifested in the surface and deep structures of the language: formal, semantic, related to pragmalinguistics, etc. Hence, our findings offer new insights into the gender gap in consumer finance and suggest potential avenues for addressing this disparity, including the need to rethink how financial literacy and behavior indicators should be developed.

Literature Review

Theoretical Framework

The study of language’s role in finance is an emerging interdisciplinary field lacking a well-established theoretical foundation. Farrow et al. (2018, p. 558) suggest that behavioral economics is uniquely positioned to explore this area, emphasizing various psychological processes through which language can shape cognition and behavior in the domain of finance. One key process is the impact of language on identity. By focusing on identity, we can better understand how language influences financial cognition and behavior, offering a unified theoretical framework that integrates existing partial approaches. Accordingly, we adopt social identity theory as the primary theoretical foundation for our analysis.

Social identity theory, developed by social psychologists (Tajfel & Turner, 1979), has been effectively applied in economics to explain economic behavior (Akerlof & Kranton, 2000). The theory suggests that individuals are categorized as either “female” or “male,” each associated with specific behavioral norms and socially prescribed gender roles, which give rise to stereotypes – generalized and simplified beliefs about different social groups. These stereotypes are not merely descriptive but often prescriptive or proscriptive, and deviating from them can result in negative consequences such as anxiety, discomfort, conflict, and reduced well-being. To avoid these costs, individuals tend to conform to gender-specific behavioral norms.

Language, the central focus of this article, both reflects and reinforces social categorization (Newman et al., 2008). Social identity theory posits that individuals derive a sense of self from their group memberships, which strengthens adherence to group norms, including linguistic markers of social categories. In a socially “group-structured world” (Goodhew et al., 2022), distinct groups – such as females and males – use language to differentiate between in-group and out-group members, much like subcultures employ slang or jargon (Farrow et al., 2018). Social gender roles were likely encoded in language in the ancient past and began to be expressed in language by human ancestors due to inherent gender differences in reproduction and the resulting division of labor and specialization (Johansson, 2005). Because linguistic structures are highly durable (Mavisakalyan & Weber, 2018), language as a vehicle for cultural expression has preserved social norms and belief systems associated with gender roles until today.

Since finance remains stereotyped as a male-dominated field (Allen & Gervais, 2017; Tinghög et al., 2021; Von Hippel et al., 2015), the language used in this domain is likely to reflect this socially resonating gendered division. For instance, men’s language in financial contexts may convey confidence, expertise, and a sense of agency, reinforcing their perceived affiliation with (belongingness to) the financial world, in line with the mechanisms discussed by Gaucher et al. (2011) and theoretical expectations stemming from social identity theory. Conversely, associating finance with men may contribute to women’s feelings of alienation and lack of belonging, diminishing their confidence and perceived influence in financial matters (Gaucher et al., 2011; Kricheli-Katz & Regev, 2021).

Empirical Evidence

Gender and Language Use in General

Research on the relationship between gender and language use dates back to Lakoff (1973). The literature on this topic indicates statistically significant, though often small, differences in linguistic production between women and men, observed across various linguistic levels, including phonetic, lexical, and morphosyntactic. Contemporary linguistic, sociological, and discourse studies confirm that women and men use language differently in both speech and writing, albeit with modest distinctions (Meier et al., 2020; Newman et al., 2008; Park et al., 2016; Piersoul & Van de Velde, 2023; Plug et al., 2021; Van der Velde et al., 2015).

Empirical evidence suggests that women tend to use more affiliative speech, whereas men are more likely to adopt assertive speech patterns (Leaper & Ayers, 2007). Women’s language is generally warmer, more sympathetic, and polite (Thomson et al., 2001), while men’s language tends to be colder, more hostile, and less personal (Park et al., 2016). Additionally, women use more words related to psychological and social processes, whereas men focus more on object properties and impersonal topics, despite having a similar overall word count (Newman et al., 2008).

Women are also more likely to use language relationally, whereas men use it primarily for informational purposes (Biber & Burges, 2000; Newman et al., 2008; Park et al., 2016; Tannen, 1994), a distinction first noted by Lakoff (1973). Since emotions play a key role in relationship-building, women’s language is often described as more tentative, attentive to the interlocutor, and rich in features that foster emotional connection (Leaper & Robnett, 2011; Mulac et al., 2001; Newman et al., 2008; Park et al., 2016). This so-called “affective” or “emotional” style is reflected in various linguistic markers, including sentence-initial adverbs, uncertainty verbs, modal auxiliary verbs, emotion words, and intensifying adverbs.

One example of men’s greater emphasis on information is their more frequent references to quantities and locations (Mulac & Lundell, 1986), as well as their focus on objects, in contrast to women’s tendency to focus on people (Newman et al., 2008; Park et al., 2016). More broadly, women’s language can be characterized as personal and interpersonal, whereas men’s language is more impersonal and socially distant.

Gender and Language Use in Finance

Boggio et al. (2015) examined a linguistic corpus (English, Italian, and Dutch) from websites targeting retail investors, focusing on the frequency of metaphors representing different (feminine and masculine) conceptual domains. They found that the vast majority of the metaphors were drawn from masculine source domains such as, for instance, physical activity, war, game, and farming.

In a series of studies, Gaucher et al. (2011) showed that job advertisements for male-dominated occupations (including accounting and financial management) contained more masculine words and that inserting more masculine language into job advertisements made them less appealing to women and led participants to perceive the jobs as male-dominated. They concluded that “masculine wording in job advertisements leads to less anticipated belongingness and job interest among women, which (…) likely perpetuates gender inequality in male-dominated fields” (Gaucher et al., 2011, p. 119).

In an experimental study, Boggio et al. (2020) found that using language that women can identify with in describing the financial tasks increased attention, interest, and consistent responses between participating girls. A recent study by Ben-Shmuel et al. (2024) found that financial advice varies by gender. Male finfluencers tend to emphasize quantitative aspects, utilizing numbers and graphs to communicate their messages, whereas female finfluencers are more inclined to incorporate narratives and personal stories, connecting financial topics to broader life goals and personal experiences.

The empirical findings presented in this section are promising; however, this research has certain limitations that our study successfully addresses (see the Method section for details). The study of Boggio et al. (2015) focuses solely on metaphors, relying on a small and undifferentiated linguistic corpus of approximately 15,000 words, limited to investment-related topics. The analysis employs simple methods, involving expert identification of metaphors and their associations with gender. In turn, the study of Ben-Shmuel et al. (2024) is based on a qualitative analysis of linguistic data from a single source – social media – produced by individuals presumed to have above-average financial knowledge (finfluencers).

Nevertheless, building on the theoretical framework outlined in section 2.1 and the emerging empirical evidence discussed in section 2.2, we propose the following main hypothesis for this study:

H1: In the domain of consumer finance, women and men use language differently.

Drawing on the findings of Ben-Shmuel et al. (2024), which demonstrate that gendered language use in finance aligns with broader trends reported in linguistic and discourse studies (e.g., more personal, experience-based, and narrative style in women vs. more technical, argumentative, and informational style in men), we propose the following additional hypothesis:

H2: Gender differences in language use within the consumer finance domain reflect the general patterns observed across other subject areas.

The Present Research

General Information

Three distinct studies were conducted to examine gender differences in language use in consumer finance: (a) a computer-assisted web interview (CAWI) survey on a sample of 1,068 individuals representative for the population of adult Poles in terms of sex, age, size of residence and region (Study 1), (b) a study of over 77,000 tweets retrieved from platform X (Twitter) (Study 2), and (c) an online survey on a purposive sample of 358 students (Study 3). In each study, language corpora were collected and the analyses were conducted in a uniform manner.

The studies were approved by the WSEI University Bioethics Committee. They were carried out in accordance with relevant guidelines and according to the principles expressed in the Declaration of Helsinki. Except Study 2 (in which we collected anonymized data from the X platform), before data collection all participants were informed about the study protocol and gave their consent to take part in the study. With the exception of Study 1, which took place on a paid research panel, participants did not receive compensation.

We chose to utilize a diverse linguistic corpora to capture universal patterns, independent of the medium’s characteristics, rather than focusing on patterns specific to one medium. The linguistic corpus derived from Twitter (Study 2) is peculiar, primarily due to the limitations on the number of words one can use. It also serves as a forum for exchanging opinions and expressing views. The same features are shared by the language corpus collected in Study 3, which was stylized on Twitter. In contrast, the survey questionnaire we prepared for the CAWI survey (Study 1) allowed for unrestricted and unlimited written responses, without the involvement of interaction with other individuals. For these reasons, the use of language in these distinct media can be (and certainly is) at least somewhat different. We were keen to take this into account.

The language corpora from Study 1, Study 2, and Study 3 were analyzed using the same set of 10 keywords (henceforth referred to as 10-KW) covering the thematic range of consumer finance issues indicated in the relevant literature (Dew & Xiao, 2011; Huston, 2010; Remund, 2010). This consistent vocabulary ensured that we were examining the same research objects across studies. Our goal was to investigate the (written) language used by consumers of both sexes around these keywords.

The word selection was a four-stage process with six research team members from various disciplines (economics and finance, psychology, sociology), including two women: (a) literature review; (b) collecting literature-based proposals from team members; (c) checking the frequency of proposed words in Google search engine; (d) selecting the top 10 words from those proposed. We considered the frequency of proposed words in Google searches as a measure of each word’s importance in consumer communication. The final selected keywords are: cost, credit, insurance, interest, investment, money, pension, profit, saving, tax.

Figure 1 synthesizes the set-up of our research. First, we selected 10 keywords for the consumer finance domain (10-KW). We then used them in three studies of written language from two media that differ strongly in the constraints imposed on communication mode and language (upper and lower parts of Figure 1). Studies 2 and 3 were related, as we explain in detail in Section 3.4.1. Study 3 was designed to verify the accuracy of artificial intelligence classification algorithms for recognizing differences in financial language between men and women in Study 2 (in a linguistic corpus extracted from platform X, which is devoid of gender tags). For this reason, in Figure 1 we present them at the same level, that is, in the lower part of the figure.

Figure 1.

Study design.

To make this article easier to navigate, we briefly introduce here the structure of the Method and Results section. The Method section continues with the following:

1) In Sections 3.2, 3.3, and 3.4, we present the assumptions of each of our three studies (Study 1, Study 2, and Study 3) in more detail (each of these sections is divided into two smaller sections: Study design and Data)

2) In Section 3.5, we explain the nature of the analyses performed, dividing this section into sub-sections on the domains represented by the methods used (machine learning, statistics, linguistics).

In turn, the Results section is divided into three parts presenting the results of Study 1, Study 2, and Study 3. Each of these three parts in turn consists of the results of applying statistical and linguistic analyses (machine learning analyses were used as pre-processing to enable statistical and linguistic research).

Study 1

Study Design

In Study 1, the 10-KW set prompted respondents to generate free-form written statements regarding: (i) associations with the 10-KW (e.g., What are your first three associations with …?); (ii) definitions of the 10-KW (e.g., What is …?); (iii) their synonyms (e.g., What terms could be used to replace the word “…”?); and (iv) related references (e.g., What kind of thing is …?). These statements helped reconstruct each respondent’s conceptual framework regarding financial terms and their designata. Additionally, respondents rated the importance of these10-KW in their daily lives, gauging their ability to “sort out” financial matters in relation to each of these keywords (You feel that you are finding your way well in finance when you sort out the following things…; the 10-KW were listed here; in Polish, we used a very popular recently and also colloquial verb “ogarnąć coś” which is difficult to translate into English – with the verb “to sort out” being a poor equivalent).

Data

The survey was conducted on a quota-based sample (1,068 respondents) representative of the population of adult Poles in terms of sex, age, size of the place of residence, and region (voivodeship). Data were collected on one of the most populous research panels in Poland (Ariadna Panel –https://panelariadna.com/) between February 14 and 22, 2024 using computer-assisted web interviews (CAWI). Participants comprised 561 females (52.5%) and 507 males (47.5%), with a balanced age distribution, primarily 55 years and older (40.8%) and 35 to 44 years (20.9%) (Table 1).

Table 1.

Participants of the Study 1.

	Sex
	Female		Male		Total
Variable	N	%	N	%	N	%	χ²
Age
18–24	48	8.6	35	6.9	83	7.8	2.55
25–34	91	16.2	70	13.8	161	15.1
35–44	114	20.3	109	21.5	223	20.9
45–54	84	15.0	81	16.0	165	15.4
55+	224	39.9	212	41.8	436	40.8
Education level
Primary/lower secondary	9	1.6	10	2.0	19	1.8	4.68
Upper secondary/Post-secondary/Associate	305	54.4	256	50.5	561	52.5
Bachelor’s degree	59	10.5	72	14.2	131	12.3
Master’s degree	175	31.2	161	31.8	336	31.5
Doctorate or professional degree	13	2.3	8	1.6	21	2.0
Field of education
Engineering, manufacturing and construction	50	8.9	139	27.4	189	17.7	120.93***
Business, administration and law	113	20.1	67	13.2	180	16.9
General	66	11.8	61	12.0	127	11.9
Services	51	9.1	48	9.5	99	9.3
Other	58	10.3	36	7.1	94	8.8
Education	44	7.8	19	3.7	63	5.9
Agriculture, forestry, fisheries, and veterinary	28	5.0	31	6.1	59	5.5
Information and communication technologies	13	2.3	45	8.9	58	5.4
Health and welfare	46	8.2	9	1.8	55	5.1
Arts and humanities	33	5.9	19	3.7	52	4.9
Social sciences, journalism and information	31	5.5	20	3.9	51	4.8
Natural sciences, mathematics and statistics	28	5.0	13	2.6	41	3.8
Marital status
Single	79	14.1	105	20.7	184	17.2	34.15***
Cohabitation	90	16.0	68	13.4	158	14.8
Married (first time)	260	46.3	255	50.3	515	48.2
Married (reconstructed family)	31	5.5	34	6.7	65	6.1
Divorced/legally separated	51	9.1	35	6.9	86	8.1
Widowed	48	8.6	10	2.0	58	5.4
Other	2	0.4	0	0.0	2	0.2
Employment status
Full-time employee	237	42.2	250	49.3	487	45.6	51.04***
Retired	171	30.5	146	28.8	317	29.7
Self-employed	29	5.2	35	6.9	64	6.0
Part-time employee	18	3.2	35	6.9	53	5.0
Homemaker	39	7.0	7	1.4	46	4.3
Student	32	5.7	13	2.6	45	4.2
Unemployed	20	3.6	21	4.1	41	3.8
Parental leave	15	2.7	0	0.0	15	1.4
Size of place of residence
Village	214	38.1	198	39.1	412	38.6	0.36
Town <20 thousands	69	12.3	62	12.2	131	12.3
City 20–100 thousands	115	20.5	97	19.1	212	19.9
City 101–500 thousands	95	16.9	86	17.0	181	16.9
City >500 thousands	68	12.1	64	12.6	132	12.4
Total	561	100.0	507	100.0	1,068	100.0

p < .05. **p < .01. ***p < .001.

Educational attainment varied, with most participants having an upper secondary or post-secondary education (52.5%). Engineering, manufacturing, and construction were indicated as the field of education by 17.7%, and 16.9% pursued business, administration, and law. Nearly half were married for the first time (48.2%), while singles made up 17.2%. In terms of employment, 45.6% were full-time employees, and 29.7% were retired. Residence was predominantly in villages (38.6%), followed by towns with populations of 20 to 100 thousand (19.9%), and cities with 101 to 500 thousand inhabitants (16.9%).

Pearson chi-square tests revealed significant differences between women and men in educational background (χ² = 120.930, df = 11, p < .001), marital status (χ² = 34.154, df = 6, p < .001), and employment status (χ² = 51.042, df = 7, p < .001), consistent with the structure of the general population in Poland. No significant differences were found for age, education level, or residence.

Study 2

Study Design

We collected textual data via API of the platform X (still Twitter at that time) from September 7, 2022 to January 6, 2023. We gathered 343,709 tweets using 10 hashtags (equivalent to our 10-KW): cost, credit, insurance, interest, investment, money, pension, profit, saving, tax. After removing duplicates and short messages (less than three words), we ultimately worked with a corpus of 77,088 tweets. The distribution of hashtags in the cleaned corpus is as follows: credit – 10,537; money – 9,919; profit – 9,589; tax – 9,371; cost – 8,919; investment – 8,060; pension – 7,489; insurance – 5,246; interest – 4,425; and saving – 3,542.

However, the language corpus extracted from platform X has the property that posts (tweets) do not have gender tags. Therefore, in the next step we created a dictionary of gender-specific linguistic endings in Polish, where the conjugation of verbs and adjective endings often indicate the writer’s gender. Using the endings, we identified 828 tweets by women and 2,588 tweets by men.

Having labeled the training material by gender endings, we used artificial intelligence methods to identify a larger number of male and female posts (though not all, as identification by word endings is not exhaustive in determining the author’s gender). To this end, we employed three Polish language models: HERBERT, Polish_BERT, and XLM. After quantitative and qualitative verification, the Polish_BERT returned the best results. In total, we identified exactly 10,000 posts by women and 11,997 posts by men. We then conducted statistical and linguistic analyses on these data.

To verify the accuracy of the classification model, we tested it on data collected in Study 3 (see section 3.4), which was designed to replicate the communication properties of the X platform, but at the same time the gender of the post authors was known to us. As indicated in section 3.4, in Study 3 we collected data from 268 females and 85 males. It should be emphasized that it is quite difficult to achieve a good classification result on such an imbalanced dataset, hence the necessity for class weighting. It turned out that the best model is Polish_BERT with an accuracy of 77% (Herbert 37%, XLM 18%).

Study 3

Study Design

Study 3 aimed to verify the accuracy of the artificial intelligence classification algorithms to recognize financial language differences between women and men in the linguistic corpus extracted from the X platform which is devoid of gender tags (see Study 2 in section 3.3; as we indicated there, the highest accuracy – provided by the Polish_BERT model – was 77%).

To achieve this, we developed a survey questionnaire to gather content similar to posts from platform X. The process consisted of several stages:

Statement creation: Two individuals (one male and one female) each formulated 10 short statements (stylized as platform X posts) using a different word from the pool of 10-KW.

Psychological verification based on the patters indicated in Newman et al. (2008).

Linguistic cleansing: A linguist/communicologist on the team removed any gender-marking characteristics from the statements.

Survey preparation: The 20 statements (10 produced by a male and 10 by a female) were added to a survey questionnaire, along with other questions about consumer finances, prepared by two team members. This questionnaire was distributed online to students from several Polish universities.

Participants answered demographic questions and formulated responses (re-tweets) to the statements. The automated online questionnaire displayed two pools of 10 statements: one from a woman, one from a man, in random order, without indicating the author’s gender. Participants responded to at least three tweets from each pool. The choice of tweets, the number of responses (more than three), and the length of each response were up to the participants, who could stop responding at any time without explanation.

Data

Data were collected in November 2023 using the platform of the market research company BEW.EDU.PL. The questionnaire was distributed to students through the universities’ educational platforms. The survey link was opened by 1,260 individuals. Out of these, 902 provided their sociodemographic information (gender, age, and native language skills), and 358 native speakers completed the questionnaire fully. We rejected five responses due to unintelligible random entries, leaving us with data from 353 participants (268 females and 85 males).

Table 2 reports distribution of the sample in terms of the key sociodemographic variables. The mean age of the participants was 28.19 years (SD = 10.26). Most of participants were 6 to 7 ISCED level students (58.7%), majoring in business and administration (57.2%), human sciences (14.1%), and other fields (28.7%).

Table 2.

Participants of the Study 3.

	Sex						χ²
Variable	Female		Male		Total
	N	%	N	%	N	%
Age
18–24	171	63.6	47	56.0	218	61.8	4.28
25–34	31	11.5	16	19.0	47	13.3
35–44	36	13.4	12	14.3	48	13.6
45–54	30	11.2	8	9.5	38	10.8
55+	1	0.4	1	1.2	2	0.6
Size of place of residence
Village	126	46.8	30	35.7	156	44.2	8.27
Town <20 thousands	23	8.6	6	7.1	29	8.2
City 20–100 thousands	39	14.5	14	16.7	53	15.0
City 101–500 thousands	65	24.2	32	38.1	97	27.5
City >500 thousands	16	5.9	2	2.4	18	5.1
Education level
Upper secondary/Post-secondary/Associate	169	62.8	56	66.7	225	63.7	0.41
Bachelor’s degree	100	37.2	28	33.3	128	36.3	0.41
Total	269	100.0	84	100.0	353	100.0

p < .05. **p < .01. ***p < .001.

The initial pool of words from these 353 individuals totaled 61,289. The shortest re-tweet was a single one word, indicating the respondent stopped participating; the longest was 557 words, with a median length of 155 words. Our final language corpus included 588 frequently repeated words, each used at least twice, after filtering out stop-words and non-dictionary or nonsensical terms.

Analysis

In each Study (Study 1–3) collected data underwent the same set of analyses: machine learning processing, followed by statistical analysis, and then linguistic analysis.

Machine Learning Analysis

To analyze the text data (linguistic corpora) collected in Studies 1 to 3, we employed a range of natural language processing techniques using Python libraries including sklearn, nltk, gensim, and spacy. The analysis consisted of several stages. Initially, we cleaned the data and lemmatized the text corpus using the publicly available Polish language model, “pl_core_news_lg” (GNU GPL 3.0), which has 500,000 unique vectors across 300 dimensions. Given that Polish has highly inflectional nature, lemmatization reduced words to their roots. We then removed common stop-words and punctuation, which carry little meaning (equivalent to “the,”“at,”“on” in English). Graphical elements such as emojis were excluded from the analysis to maintain consistency across data sources.

Next, we counted the frequency of words in the lemmatized corpus, providing insights into the differences in language usage between women and men regarding financial matters. We then conducted n-gram analyses, generating lists of bigrams (pairs of words) and trigrams (triples of words) that appeared most frequently in the text corpus. These n-grams help capture the structure of statements in the textual material gathered.

The identified bigrams and trigrams were visualized as networks using Python libraries such as networkx, matplotlib, and seaborn. Although n-grams are not presented in this article, they were instrumental in understanding the linguistic structures used by women and men.

Statistical Analysis

We identified the words most frequently used by women and by men – both in separate corpora produced by women and men, and in one aggregate corpus combining the two. We then determined which words showed statistically significant frequency differences between the sexes, examining the individual sub-corpora and the combined corpus.

We analyzed 10 most common words in each sub-corpora, setting this cut-off threshold (10 words) for statistical reasons, as less frequent words did not provide sufficient statistical differentiation power. We used the chi-square coefficient to compare the frequency of use of each word by women and men. With the significance level of .05, the minimum acceptable test power of 0.8, and the expected effect size of 0.5, the minimum sample size was N = 44 (N = 36 for the power of 0.7). The minimum sample size range between 36 and 44 included up to 10 words from the analyzed corpora.

In Study 1, in addition to strictly linguistic questions, we asked respondents a question to learn how important the chosen financial terms (10-KW) are to their everyday functioning (see the section 3.2.1 for the wording of the question). A five-point Likert scale ranging from Definitely no to Definitely yes was applied in this question. To compare female and male respondents’ responses to the question, we used the Mann-Whitney coefficient Normal Approximation Z due to the non-normality of the distributions of the responses.

Linguistic and Communicative Analysis

To analyze biolects, that is, the way of speaking that is characteristic of a particular biological group, often based on gender or age and their interpretation of the world, including finance, we first applied traditional generative grammar tools to identify similarities and differences in the deep and surface structures of the content produced by women and men (Wołkowski, 2010, pp. 145–161). We focused on: (i) thematic-remedial structures, examining syntagmatic connections within sentences and whole responses, (ii) valences and semantic roles of the statement components (Fillmore, 1969), and (iii) how generative models in the deep structure influence surface linguistic constructions (Chomsky, 1982, 2010).

Secondly, we analyzed the extracted connections, constructions and linguistic units using textological, ethnolinguistic (referring mainly to the methods of cultural semantics), and cognitive linguistics methods (de Beaugrande & Dressler, 1990). This comprehensive approach allowed us to reconstruct the linguistic image of finance highlighting similarities and differences between women and men in key interpretative categories such as categorization, axiology, onomastics, and metaphorization (Bartmiński, 2012; Nowak, 2002; Tokarski, 2014). The methods of ethnolinguistics/cultural and cognitive linguistics helped us interpret conceptual metaphors and contexts (Lakoff & Johnson, 2003) and assess whether only masculine metaphors appear (Boggio et al., 2015).

Third, from a pragmatic perspective, we evaluated the intentions and effects of participants’ language production, assigning single sentences and whole statements to different types of speech acts (Searle, 1987). We described differences and similarities in the verbal responses of women and men regarding finance, influencing the differentiation of the locus (formal structure) and illocution (intentionality) of these acts. This analysis allowed us to categorize statements into types such as assertions, directives (command or request), commissives (committing to future actions or expressing intentions to perform certain actions), expressives (expression or evocation of emotional states), or, finally, constatives (stating facts, beliefs, or truths).

The Polish Language

Studies 1 to 3 were conducted in Polish, which influenced the study design and might have impacted the findings as well. The Polish language, belonging to the West Slavic group, has evolved over 1,000 years, influenced by Proto-Slavic and other Slavic languages, distinguishing it from English, French, and German. Notable differences lie in its intricate grammar, rich grammatical categories, and forms, and heterogeneous semantics shaped by both intra-linguistic factors and external contexts such as culture, politics, and society.

Contemporary Polish is characterized by:

Grammatical and semantic gender markers, more abundant than in English.

A deliberate mid-20^th-century decision of the Polish authorities to neutralize masculine gender leading to fewer distinctions between genders in terms like professions’ names, unlike English.

Patriarchal cultural norms and administrative choices favoring linguistic masculinity in public discourse.

Preference for masculine forms in conveying professionalism and seriousness.

Consequently, the “femininity” of Polish language manifests predominantly in the underlying structures of communication (public, official, didactic, scientific, etc.), emphasizing valuation, emotionalization, outcome description over processes, and less abstractness and typicality/ordinariness in narrative.

Results

Study 1

Statistical Analysis

Women used more words (Me = 37) than men (Me = 35; Z = −3.071, p = .002). Words significantly more frequent in women’s sub-corpora varied by the keyword (10-KW). Additionally, women associated the 10-KW with specific life situations, for example “insurance” with “accident” (χ² = 4.577; p = .032) and “costs” with “to make a living” (χ² = 6.623; p = .010). The adjective “free” was more frequent in the “pension” sub-corpus for women (χ² = 4.473; p = .029). For men, the verb “to know” appeared more often across eight of the 10-KW (chi-square value between 6.137 (p = .013) for “credit” and 47.644 (p = .000) for “costs”). Men also frequently used words like “profit” (in the “interest” sub-corpus; χ² = 6.818; p = .009) and “loss” (in the “costs” sub-corpus; χ² = 6.082; p = .014). In men’s sub-corpus for “tax,” the noun “ (forced) tribute/exaction” (“haracz” in Polish) was more common (χ² = 4.018; p = .045). The word “bank” was significantly more common in men’s “saving” sub-corpus (χ² = 4.018; p = .045) and in women’s “credit” sub-corpus (χ² = 4.931; p = .026). Across the entire linguistic corpus, gender differences were notable for three words: “to know” (more common in men; χ² = 53.274; p = .000), “a purchase” (more common in women; χ² = 5.400; p = .019), and “oldness (old age)” (more common in women; χ² = 5.743; p = .017).

As mentioned in section 3.5.2, we asked respondents a question allowing us to learn how important the chosen financial terms (10-KW) are to their everyday functioning. Table 3 presents the importance of these terms based on the responses to the survey question “You feel that you are finding your way well in finance when you sort out the following things …” (response options from 1 –Definitely no to 5 –Definitely yes).

Table 3.

Respondents’ Answers to the Question “You Feel That You Are Finding Your Way Well in Finance When You Sort Out the Following Things ….”

Keyword (subject domain)	Sex						Mann-Whitney test statistic normal approximation
	Female		Male		Total		Mann-Whitney test statistic normal approximation
	M ^a	SD	M ^a	SD	M ^a	SD	Z
Money	3.81	0.96	3.81	0.96	3.81	0.96	−0.143
Saving	3.81	0.98	3.80	1.04	3.81	1.01	−0.165
Cost	3.72	0.95	3.76	0.96	3.74	0.96	−0.772
Interest	3.52	1.05	3.72	1.00	3.61	1.03	−3.017**
Insurance	3.58	0.98	3.61	0.99	3.60	0.99	−0.489
Profit	3.53	0.99	3.65	1.01	3.59	1.00	−2.160*
Pension	3.41	1.11	3.60	1.12	3.50	1.12	−2.802**
Tax	3.42	1.07	3.46	1.08	3.44	1.08	−0.517
Investment	3.05	1.10	3.38	1.15	3.20	1.14	−4.905***
Credit	3.09	1.14	3.28	1.14	3.18	1.14	−2.774**

Mean in range between 1 –“Definitely no” and 5 –“Definitely yes.”

p < .05. **p < .01. ***p < .001.

Respondents most often reported that they felt that they were finding their way well in finance when sorting out money, saving, cost, and interest (Table 3). Conversely, the least often reported were credit, investment, tax, and pension. The most pronounced gender differences were noted for five tested terms: “investment” (Z = −4.905, p < .001), “interest” (Z = −3.017, p = .003), “pension” (Z = −2.802, p = .005), “credit” (Z = −2.774, p = .006), and “profit” (Z = −2.160, p = .031), all favoring men. No gender differences were found for money, saving, cost, insurance, and tax.

Linguistic Analyses

Differences in financial language between women and men appear across almost every keyword studied.

Money

Both genders identify money primarily a means of payment. However, men showed a need for detailed categorization referring to cash, currency, banknotes, coins, change, and the slang terms like dough, rocks, dust, and many more. Women do not use such terms and focus on money’s presence, associating it with wealth, whereas men discuss the lack of money, that the need for credit, and insufficiency (generally, in female language production, “money is there,” while in male language production, “money is missing”).

Additionally, women discuss money in the context of everyday life, usually positively aligning with common sense and in accordance with a communication strategy known as the natural order of things. The strategy refers to the colloquial understanding of reality. However, they also use scripts and cognitive frameworks that prioritize understanding over emotionality, metaphorization, and literariness/aestheticization/figurativeness (Nowak, 2020, p. 234). Figurative language, such as Even Solomon cannot pour from an empty, is also used by women. Men, instead, structure and categorize their interpretation, emphasizing expertise, resulting in neutral or pessimistic assessments.

Pension

Both genders emphasize importance of knowing what their lives will be like upon retirement. The verb “to know” is highly ranked in the frequency for both sexes, reflecting a rational, specialized approach. In men, the verb “to know” ranks first, while in women – second. Women’s most frequent term is “oldness (old age),” combined with the noun “time” and the verb “to rest.” In contrast, men’s frequent terms include “work” and “money,” introducing financial categories and scaling pensions with words like “minimum” or “average,” emphasizing a specialized paradigm and competences. Women do not scale pension, but instead create an axis of valuation using terms like “small”/“low” and “high.” This reflects the difference in perspectives: male linguistic production is more process-oriented while female – more result-oriented. Men also connect pension to government actions and contextual, macroeconomic factors, often pointing out the low pensions of women. Women’s economic contextualization is more private (daily) focused, without indicating the higher pensions of men.

Cost

Men approach costs rationally and are process-oriented as reflected in their frequent use of the verb “to know” and other imperfective verbs like “to spend” or “to grow.” Women often use nouns like “a cost” and “an expense,” along with perfective verbs like “to spend,”“to incur,” and “to increase.” Once again, men tend to categorize and detail costs precisely, for example, auto, fuel, VAT, energy, gas, etc., while women discuss the experienced effects of costs, using terms like “pricey.”

Investment

The linguistic richness of women’s statements about investments is notable, with 114 communicatively independent words compared to men’s 97. Women often evoke their affective state, linking investment to values and emotions (e.g., responsibility, security, uncertainty, good). This emotional dimension is largely absent in men’s statements. As with the other 10-KW, men tend to categorize investment types in detail, from deposits to cryptocurrencies, listing 23 types in total.

Credit

Both men and women associate credit with negative concepts. Women link credit to risk, stress, and negativity, while men associate it with fear, exploitation, and deception. Men view credit more fearfully than women, using phraseologisms like “crutch” (“ball and chain”) and portraying credit institutions negatively as exploitative or usurious. These associations are absent in women’s statements. Women also see credit as a means of achieving goals and as a financial boost, whereas men often see it as last resort and a tool for consumer exploitation (fraud, thievery).

Interest

Both genders view interest negatively, akin to credit. Women perceive interest more negatively than credit, while men see both equally negatively, though distinguishing between harassment (interest) and usury (credit). Both genders recognize the connection between interest and state institution, as indicated by references to Marek Belka (former Prime Minister of Poland and Minister of Finance) who introduced the capital gains tax (so-called Belka tax).

Saving

The interpretation of saving differs the least between genders due to its common usage and metaphorical richness (imagery) in general language. According to the Modern Dictionary of the Polish Language (www.wsjp.pl), “saving” has four meanings, including financial saving, saving time (declaratively important for women), sparing pain and sparing life. Women tend to save to spend, showing a result-oriented perspective, while men save for the future, indicating process-oriented perspective.

Tax

Tax, along with credit and interest, forms the most negatively interpreted triad of the 10-KW. Both genders interpret tax negatively, but women focus on paying tax, while men categorize and explain taxation system. Men mention “CIT” (corporate income tax) and other specific terms, while both genders refer to “VAT” (value added tax), “PIT” (personal income tax), and excise taxes. Only women mention “toll” (“myto” in Polish). Women also connect tax to the budget, and see it as a societal contribution, whereas men view tax as “coercion,”“fraud,”“deception,” and an individual burden.

Insurance

Women frequently mention “life insurance” and “casualty insurance,” while men focus more on the institutional nature of insurance (e.g., “PZU”– Powszechny Zakład Ubezpieczeń– Poland’s largest insurance company), “peace of mind” and associated costs. Men also mention “MTPL” (motor third party liability) policy more often than women. Women’s language associates insurance with “fear” and “illness.” Men also view insurance as a “scam” (this term also appears in women’s corpus), “(forced) tribute,”“coercion,” and a “thief.” Once again, men’s language about insurance is self-centered and express belief in their invulnerability, while women’s language is sociocentric and narrative, emphasizing responsibility, which appears in their corpus but not in men’s.

Profit

Men focus on the process of calculating profits, using words like “positive,”“result,”“difference,”“plus,”“minus,” and “balance.” Women associate profit with more emotional and random concept, such as “reward,”“surplus,”“win,”“good,”“joy,” and “satisfaction.” Women describe the result of profit with terms like “effect” and “commodity,” which are absent from men’s language.

Study 2

Statistical Analysis

The three recurring words that were used significantly more often by women in the sub-corpora centered around the 10-KW were the verbs “to want” (chi-square ranging from 4.167 (p = .041) to 4.428 (p = .035)), “to be able to” (chi-square ranging from 4.042 (p = .044) to 8.867 (p = .003)), and especially “to have” (chi-square ranging from 10.684 (p = .001) to 25.247 (p = .000)). The pool of words used more frequently by men was much more diverse. Men used the verb “to know” significantly more often in the sub-corpus related to “tax” (χ² = 6.925; p = .008) and “investment” (χ² = 6.668; p = .010), while women used it more in relation to “money” (χ² = 10.623; p = .001).

Compared to women, men significantly more frequently used: (i) professional language terms such as “net” (in the sense of the opposite of “gross”; in the sub-corpus related to “profit”; χ² = 18.591; p = .000), “premium” (related to “insurance”; χ² = 11.636; p = .001), “digital” (related to “money”; χ² = 6.739; p = .009); (ii) exact sciences terms such as “plus” (related to “pension”; χ² = 5.407; p = .020), and “energy” (related to “costs” and “saving”; chi-square ranging from 4.724 (p = .030) to 18.712 (p = .000)); (iii) activity-related terms including “to buy” (related to “investment”; χ² = 6.871; p = .009), “to sell” (related to “profit”; χ² = 10.397; p = .001), “to take” (related to “credit”; χ² = 11.025; p = .001), “to pay” (related to “tax”; χ² = 4.447; p = .035), and “to enter” (in related to “tax”; χ² = 7.755; p = .005).

Linguistic Analyses

In analyzing the language corpus extracted from platform X, we faced two challenges. First, the platform itself is a communicologically and sociologically specific medium. In Poland, its users are primarily aged 18 to 34 (55.5%) and predominantly male (74%). Tweets about economics and finance rank 10th in popularity, viewed by 41% of users (IAB Polska, 2023). Second, identifying gender differences in the language of tweets on financial topics is complicated by technical constraints and Polish language rules. The 280-character limit on tweets encourage concise and direct communication – a style traditionally attributed to men. This limitation often forces women to adopt a similar style, reducing the visibility of gender-specific linguistic features. Additionally, Polish language rules allow to use masculine forms as generic, further blurring gender distinctions in language.

Despite these limitations, we identified key linguistic differences in how women and men linguistically interpret financial phenomena. These differences manifest not necessarily in the vocabulary but in the use of word ambiguity highlighting and hiding a different ordering of semantic features of the same lexical unit. For instance, men frequently use the verb “to grow” in tweets about money and saving, appearing 12 times and related terms (e.g., “growing,”“growth,”“increase,”“stop” and “decrease” or “fall”) appearing 24 times. This reflects a view of finance as a dynamic process influenced by external factors (politics, economy). Men’s language sees financial activities as processes with time varying outcomes, emphasizing growth and decline.

In contrast, women rarely use the verb “to grow” in the same context. When they do, it often refers to outcomes (again, result-oriented perspective), rather than processes, aligning finance with the metaphor Finance is a plant. Women’s financial language rarely use terms related to downward movement, instead it comprises references to missing, wasting, and losing (money) as well as saving it from such consequences (e.g., through insurance). This language is more focused on humanistic and social aspects, and everyday life, with less emphasis on technical precision. For example, women often refer to the adage Money doesn’t give happiness.

Men’s statements are agentive, focusing on action and instruments leading to financial goals. In contrast, women’s language is experiential and dative, highlighting how financial activities affect them emotionally and personally. Safety and health are important themes in women’s financial communication, almost non-existent in men’s. Additionally, the genres of speech differ: women’s posts are often narratives about everyday life, while men’s are mini-lectures, news stories, or instructions, aiming to showcase expertise.

Study 3

Statistical Analysis

The analysis of word frequency in the language corpora revealed relatively few differences between women and men. In some sub-corpora such as those related to “profit,”“tax,”“cost,”“insurance,” and “investment,” no differences were noted. High frequencies of both “money” and “happiness” in both women and men were found. However, it must be noted here that some part of the linguistic corpus resulted from automatic re-tweeting by the study participants. For example, when our post mentioned that money does not give happiness, participants overwhelmingly expressed their opinions whether money gives happiness or not. This could be a reason for high prevalence of these two terms.

Interestingly, no words were significantly more frequently used by men in the sub-corpora produced in relation to various 10-KW. However, several words occurred significantly more often in women’s posts. The verb “to have” was more frequent in the women’s corpus for “pension” (χ² = 6.587; p = .010) and “insurance” (χ² = 8.425; p = .004). With “pension,” women also used the verb “to want” (χ² = 26.115; p = .000) and the noun “state” (χ² = 12.948; p = .000) more frequent. In connection with “money,” women used the verb “must” (χ² = 3.937; p = .047). Additionally, women also used “self” (χ² = 3.882; p = .049) and “worth” (χ² = 4.714; p = .030) significantly more often.

In the combined corpus of all responses, only one significant difference emerged: men used the verb “to have” more frequently than women (χ² = 3.892; p = .049). Notably, in relation to the noun “money,” an inverse but insignificant trend was observed, with women using this term more frequently than men (χ² = 3.609; p = .057).

Linguistic Analyses

From a grammatical perspective, generative linguistics connects the formal structures of language with its semantics. Beyond the obvious differences due to the gender inflection in Polish, we observed other characteristics of gendered language use. Women’s statements often include suppositional mode and conditional sentences (if…, then…). This implies that women are more likely to use conditional and situational language structures, framing their statements with conditions and contextual circumstances more frequently than men. There are more emotional terms in women’s bigrams and trigrams, with fewer professional terms (e.g., the term “interest,” common in men’s language, is almost non-existent among women). Passive semantic valences in predicate-argument structures are more frequent in women’s statements, indicating women undergo or are affected by actions rather than initiating them.

Men’s statements, in contrast, frequently feature indicative mode and causal or manner sentences (“therefore,”“as”/“this way”). There are more professional terms and fewer emotional terms in men’s bigrams and trigrams (e.g., “stress,” common among women, is rare among men). Active semantic valences are more frequent in men’s statements, with men assigning themselves agentive roles (agens), portraying themselves as initiators of actions or events.

From the perspective of cultural linguistics and cognitivism, women’s interpretation of the financial world utilizes different conceptual metaphors than in men – these are mainly the typical metaphors such as “money/finances are something valuable” (“lose,”“compound”) and “credit is something dangerous” (“be afraid of credit,”“be careful with credit”); centers around the individual rather than finance itself, evident in the frequent use of the noun “person”; treats communication about finances primarily as communication, indicated by verbs of speak and think: “say,”“suppose,”“assume.”

In contrast, men’s interpretation includes ordering and rational spatial metaphors, with idealized cognitive models (ICMs) like center-periphery, source, way, and space (“even”–“uneven,”“high”–“low,”“source of income,”“something will be nibbled on the side” (“nibble” as a metaphor for food/eating/consumption), “option,”“catch on to something”); conceptualizes finances as essential and varying by context; with metaphors related to movement and parameters (e.g., “arrives,”“flows down” (“streams down”), “everything goes to the current,”“I did not move anything in this direction,”“stable,”“trimmed,”“weigh” (“burden,”“load”), “went 1,480 zloty,”“additional burden”); and views finances not just as communication but as concrete actions and tasks to be accomplished (e.g., “to solve,”“to cope,”“to know”).

Pragmalinguistic analysis reveals that women’s speech acts are propositional, less categorical, and often contain questions. Their assertions emphasize the individuality and subjectivity rather than universal or objective judgments. Men’s speech acts, however, are more constatives, aiming to establish or define reality, with greater illocutionary force, confirming the rationality and expertise of the speaker.

Discussion

The purpose of this research was to examine whether the language used by women regarding financial matters differs from that used by men. The results indicate that such linguistic differences do exist, positively answering our first research question (Q1) and confirming H1 hypothesis, and likely have implications for consumer finance phenomena such as the gender gap in finance. These findings align with previous studies that have shown that financial language is gendered (Ben-Shmuel et al., 2024), particularly at the level of metaphors (Boggio et al., 2015), and that the language may be linked to the gender gap in financial literacy (Boggio et al., 2020).

The differences in language use identified between women and men are numerous. Some of these differences are known from broader trends observed in other content (subject) domains. However, we have identified some novel differences that are likely unique to the language used in the finance domain, which answers our third research question (Q3). Gender differences in language production identified in our studies offer opportunities for designing policy interventions. These interventions could target language change in measuring consumer finance phenomena (e.g., financial literacy and behavior) and developing financial education materials. However, before discussing these potential applications, we will outline the most important differences in how women and men use language when dealing with financial matters, which answers our second research question (Q2).

First, men perceive financial phenomena as processes they both understand and control, presenting their statements as informative and authoritative while positioning themselves as experts. This tendency is reflected in their frequent use of professional terminology (or at least stylized as such). In contrast, women are more likely to focus on financial states rather than processes, conveying personal experiences and specific cases through colloquial language, with specialized terms appearing less frequently.

Second, women tend to use more figurative, narrative, sociocentric, and humanistic language, often referring to social aspects and expressing a higher degree of emotional intensity. In contrast, men are more likely to use technical and impersonal language, prioritizing precision and striving to establish structure, order, and categorical distinctions between phenomena (objects).

Third, women and men tend to ascribe slightly different meanings to key terms in consumer finance (saving, investing, and borrowing). For women, credit is often perceived as a means of achieving goals and providing financial support, whereas men are more likely to view it as a last resort for financing and a tool for consumer exploitation. Similarly, women more frequently associate saving with purchasing and day-to-day financial management, while men link it to future financial planning and the transfer of resources across time periods. Regarding investment, women tend to interpret the concept more broadly, often extending beyond the financial domain, whereas men typically adhere to the professional product categorizations established within the investment industry.

Fourth, men often perceive financial instruments and related institutions as exploitative, expressing feelings of being used, manipulated, cheated, or robbed. This is reflected in their frequent use of terms such as “(forced) tribute” or “exaction” when referring to taxes and their association of “government” with “costs.” In contrast, women tend to view manipulation and fraud as inherent aspects of life experience, assessing financial matters through a moral lens of good and evil without expressing surprise.

Fifth, men position themselves at the center of financial processes, viewing themselves as active agents who control or evaluate financial outcomes – often framed as winning or losing, in line with the game metaphor. Their focus is primarily on performance and results, as reflected in the higher frequency of “bottom-line”-oriented nouns such as “profit” and “loss” in men’s linguistic corpora. In contrast, women are more likely to perceive themselves as experiencers within the financial domain, viewing themselves as “inhabitants” subject to its rules and events rather than as agents shaping financial processes. Women tend to assess financial matters through the lens of emotions and values, as evidenced by the more frequent use of language related to personal experiences. For instance, words such as “accident” in the context of “insurance” and phrases like “to make a living” in relation to “cost” appear more frequently in women’s linguistic corpora.

Many of the differences we identified align with findings from previous studies on language corpora outside the financial domain, confirming our H2 hypothesis. Our study corroborates that women tend to use language in a more personal and relational manner, often referencing social and psychological phenomena as well as their own experiences. In contrast, men are more likely to use language in an informational, instrumental, technical, argumentative, and impersonal way. These patterns are well-documented in prior linguistic, discourse, and sociological research (Biber & Burges, 2000; Newman et al., 2008; Park et al., 2016; Thomson et al., 2001). In this regard, our findings are also consistent with the recent study by Ben-Shmuel et al. (2024, p. 1), which examined social media content produced by finfluencers (“Men emphasize quantitative aspects, whereas women incorporate narratives and personal stories”). These findings are also consistent with psychological research on gender differences in financial beliefs, attitudes, and behaviors. A recent scoping review by Sesini et al. (2023) found that men tend to perceive money as a symbol of success, status, power, and prestige. In contrast, women are more likely to view money in a way that reflects their tendency to consider not only their own needs but also those of others, such as household members (relational and personal aspect).

However, our study yielded several new and original findings. A particularly notable result, which is not widely documented in the literature on gendered language use, is the pronounced tendency of men to use language that signals agency (active semantic value), whereas women’s language more often reflects an “experiencer” stance (passive semantic value). In addition, men exhibit a greater propensity to use professional terminology and linguistically construct themselves as experts in finance. This patterns may stem from deeply ingrained stereotypes, as we hypothesized, that frame finance as a male domain, creating social pressure for men to identify with and demonstrate efficacy in financial matters – hence the strong linguistic manifestation of agency observed in our data. How the (stereotypical) beliefs in implied ability, activity, and agency of women and men are used in the language of advertising messages that shape children’s future choices and perpetuate gender stereotypes and inequities has been convincingly shown by Hourigan (2021).

These novel findings confirm that the language used in the domain of finance reflects the stereotypical perception of this field as male-dominated. As a result, men’s language in financial contexts conveys greater confidence, expertise, and a sense of agency, reinforcing their perceived affiliation with the financial world. This pattern aligns with the mechanisms outlined in social identity theory and discussed by Gaucher et al. (2011).

Implications and Future Research

There are many important implications from our study: for financial institutions, financial educators and advisors, and individual consumers. As for financial institutions, banks and other financial intermediaries should tailor their communications regarding financial products (e.g., loans, saving plans, investment instruments) to different audiences. Women may be more responsive to relational, narrative-based communications, while men may prefer precise, technical language. Women may also be more receptive to communications that emphasize security, while men may engage more with content that emphasizes control, efficiency, and expertise.

Regarding financial educators and advisors, institutions should ensure that financial training does not perpetuate or reinforce stereotypes, but rather empowers both men and women, taking into account their different financial perspectives and needs. Financial advisors should recognize gender differences in financial perceptions. Men may prefer expert, analytical advice, while women may appreciate a more personalized, experience-based approach.

When it comes to individual consumers, understanding that gender norms shape language can help individuals critically evaluate their financial attitudes and behavior and avoid the influence of stereotypes. Consumers can become more aware of the way they use language in the finance domain, increasing their chances of avoiding the pitfalls and cognitive biases resulting from language. Recognizing that finance is often perceived as male-dominated, women can proactively build confidence in financial matters, while men can benefit from incorporating relational and social considerations into their approach to finances.

Our results also point to many interesting and important directions for future research. First, this study underscores the need for further investigation into how language shapes financial behavior, particularly in digital spaces such as social media and financial education. The next logical step is to investigate whether manipulating language in financial education and diagnostic tests (assessing financial literacy and behavior) leads to differential outcomes for women and men. Building on our findings presented in this article, we have developed a financial literacy test, a scale to measure sound financial behaviors, and financial education materials in three language versions: feminized, masculinized, and gender-neutral. A forthcoming longitudinal randomized controlled trial (Cwynar et al., 2025) will assess the impact of these language versions. We hypothesize that reactions will vary based on language compatibility with gender, which could inform the development of unbiased financial literacy and behavior scales, as well as the content of financial education programs.

Second, future research should examine how other factors (e.g., culture, socioeconomic status, education) interact with gender in financial discourse. Third, our findings can help future researchers develop gender-sensitive financial policies, educational materials, and intervention programs to reduce gender gaps in financial confidence, literacy, and decision-making. Future researchers will face the challenge of “de-masculinizing” (or more universally: “de-genderizing”) financial language. Although Boggio et al. (2020) have shown that linguistic alignment of educational material with gender yields positive results, we believe in the need of universal language that avoids gender biases. The key question for future research is how to achieve this effectively.

Limitations

Future research should also aim to overcome the limitations inherent in our study. First, our research was conducted exclusively on Polish, so its findings may not directly apply to other languages. While we believe the patterns we identified are universal, further research in different linguistic contexts is necessary to confirm this. It is crucial to investigate whether our findings apply to other languages, particularly those with different level of gender-marking, such as Spanish (heavily gendered) on the one hand, and Finno-Ugric languages (genderless) on the other – see Corbett (2008).

Second, our study focused solely on written language, lacking data on spoken language. Therefore, we cannot confirm or refute our findings in relation to spoken discourse. Third, our linguistic corpus was built around 10 selected financial keywords (10-KW). Exploring the sensitivity of such a corpus to changes in keyword selection would be an interesting avenue for future research.

We used three research samples with varying distributions of socio-demographic characteristics. In Study 1, we used data from a quota-based sample representative of the adult Polish population, which allows us to generalize the conclusions of this study. The adult population, however, is characterized by a preponderance of older individuals. The design of our study was, however, to balance this predominance with data mainly from the younger part of the population (Studies 2 and 3). Nevertheless, the results from these research samples with distinctly different age structures converge. However, we consider it necessary and interesting to check in future studies whether age does not significantly differentiate the linguistic production of women and men in the financial domain.

Last, the reliability of gender identification in a linguistic corpus extracted from platform X (Twitter) poses significant challenges. Utilizing machine learning methods for gender labeling warrants further investigation, which we intend to explore in detail in a forthcoming article (Błaszczykowski et al., 2025).

Footnotes

ORCID iDs

Andrzej Cwynar

Kamil Filipek

Paweł Nowak

Robert Porzak

Dorota Weziak-Bialowolska

Ethical Considerations

The studies demonstrated in our article were approved by the WSEI University Bioethics Committee. They were carried out in accordance with relevant guidelines and according to the principles expressed in the Declaration of Helsinki.

Consent to Participate

Except Study 2 (anonymized data from the X platform),before data collection all participants were informed about the study protocol and gave their consent to take part in the study.

Author Contributions

Andrzej Cwynar: conceptualization,funding acquisition,investigation,methodology,project administration,resources,supervision,writing – original draft,writing – review & editing. Kamil Filipek: data curation,formal analysis,investigation,software,validation,visualization,writing – original draft. Paweł Nowak: formal analysis,writing – original draft. Robert Porzak: data curation,formal analysis,investigation,software,validation,visualization,writing – original draft. Dorota Weziak-Bialowolska: formal analysis,writing – original draft,writing – review & editing.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This research was funded by the National Science Centre,Poland (grant number 2021/43/B/HS4/00401).

Declaration of Conflicting Interests

The authors declare no competing interests. The authors are not aware of any actual or potential conflict of interest that could inappropriately influence,or be perceived to influence,our work. The material included in the manuscript will not infringe upon any statutory copyright.

Data Availability Statement

Data & materials used in this research have been made available in the OSF repository at

References

Akerlof

G. A.

Kranton

R. E.

(2000). Economics and identity. Quarterly Journal of Economics, 115(3), 715–753. https://doi.org/10.1162/003355300554881

Al-Bahrani

Buser

Patel

(2020). Early causes of financial disquiet and the gender gap in financial literacy: Evidence from college students in the southeastern United States. Journal of Family and Economic Issues, 41, 558–571. https://doi.org/10.1007/s10834-020-09670-3

Allen

Gervais

S. J.

(2017). The femininity–money incongruity hypothesis: Money and femininity reminders undermine women’s cognitive performance. Psychology of Women Quarterly, 41(4), 407–419. https://doi.org/10.1177/0361684317718505

Almenberg

Dreber

(2015). Gender, stock market participation and financial literacy. Economics Letters, 137, 140–142. https://doi.org/10.1016/j.econlet.2015.10.009

Arellano

Cámara

Tuesta

(2018). Explaining the gender gap in financial literacy: The role of non-cognitive skills. Economic Notes, 47(2–3), 495–518. https://doi.org/10.1111/ecno.12113

Aristei

Gallo

(2022). Assessing gender gaps in financial knowledge and self-confidence: Evidence from international data. Finance Research Letters, 46, 102200. https://doi.org/10.1016/j.frl.2021.102200

Bartmiński

(2012). Językowe podstawy obrazu świata. (Linguistic foundations of the worldview). UMCS.

Bayer

R.-C.

(2013). Discounts and consumer search behavior: The role of framing. Journal of Economic Psychology, 39, 215–224. https://doi.org/10.1016/j.joep.2013.08.003

Ben-Shmuel

A. T.

Hayes

Drach

(2024). The gendered language of financial advice: Finfluencers, framing, and subconscious preferences. Socius: Sociological Research for a Dynamic World, 10, 1–17. https://doi.org/10.1177/23780231241267131

10.

Biber

Burges

(2000). Historical change in the language use of women and men: Gender differences in dramatic dialogue. Journal of English Linguistics, 28(1), 21–37. https://doi.org/10.1177/00754240022004857

11.

Błaszczykowski

Filipek

Cwynar

(2025). An iterative algorithmic approach to gender identification in financial discourse on Polish Twitter. Natural Language Processing, In review.

12.

Boggio

Coda Moscarola

Gallice

(2020). What is good for the goose is good for the gander?: How gender-specific conceptual frames affect financial participation and decision-making. Economics of Education Review, 75(C). https://doi.org/10.1016/j.econedurev.2019.101952

13.

Boggio

Fornero

Prast

Sanders

(2015). Seven ways to knit your portfolio. Is investor communication neutral? (DP 10/2015-030; Netspar Academic Series). Netspar, Tilburg.

14.

Bottazzi

Lusardi

(2021). Stereotypes in financial literacy: Evidence from PISA. Journal of Corporate Finance, 71, 101831. https://doi.org/10.1016/j.jcorpfin.2020.101831

15.

Bucher-Koenen

Lusardi

Alessie

van Rooij

(2017). How financially literate are women? An overview and new insights. Journal of Consumer Affairs, 51(2), 255–283. https://doi.org/10.1111/joca.12121

16.

Chomsky

(1982). Zagadnienia teorii składni (Aspects of the theory of syntax). Ossolineum.

17.

Chomsky

(2010). O naturze i języku (On nature and language). AXIS.

18.

Corbett

G. G.

(2008). Number of genders. In Haspelmath

Dryer

M. S.

Gil

Comrie

(Eds.), The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Retrieved from https://wals.info/chapter/30

19.

Cupák

Fessler

Schneebaum

Silgoner

(2018). Decomposing gender gaps in financial literacy: New international evidence. Economics Letters, 168(C), 102–106. https://doi.org/10.1016/j.econlet.2018.04.004

20.

Cwynar

Bialowolski

Weziak-Bialowolska

Porzak

(2025). Language and gender differences in financial literacy: The role of feminized, masculinized, and neutral question phrasing. International Journal of Consumer Studies, In press. https://doi.org/10.1111/IJCS.70083

21.

Cwynar

Wais

(2019). Debt literacy and debt literacy self-assessment: The case of Poland. Journal of Consumer Affairs, 53(1), 24–57. https://doi.org/10.1111/joca.12190

22.

de Beaugrande

R.-A.

Dressler

W. U

. (1990). Wstęp do lingwistyki tekstu (Introduction to text linguistics). PWN.

23.

Dew

Xiao

J. J.

(2011). The financial management behavior scale: Development and validation. Journal of Financial Counseling and Planning, 22(1), 43–59.

24.

Driva

Lührmann

Winter

(2016). Gender differences and stereotypes in financial literacy: Off to an early start. Economics Letters, 146(C), 143–146. https://doi.org/10.1016/j.econlet.2016.07.029

25.

Farrow

Grolleau

Mzoughi

(2018). What in the word! The scope for the effect of word choice on economic b4ehavior. Kyklos. International Review for Social Sciences, 71(4), 557–580. https://doi.org/10.1111/kykl.12186

26.

Fillmore

C. J.

(1969). The case for case. In Bach

Harms

(Eds.), Universals in linguistic theory (pp. 1–88). Holt, Rinehart & Winston.

27.

Florence

E. S.

Fleischman

Mulcahy

Wynder

(2022). Message framing effects on sustainable consumer behaviour: A systematic review and future research directions for social marketing. Journal of Social Marketing, 12(4), 623–652. https://doi.org/10.1108/JSOCM-09-2021-0221

28.

Fonseca

Lord

(2020). Canadian gender gap in financial literacy: Confidence matters. Hacienda Pública Española/Review of Public Economics, 235(6), 153–182. https://doi.org/10.7866/HPE-RPE.20.4.7

29.

Gaucher

Friesen

Kay

A. C.

(2011). Evidence that gendered wording in job advertisements exists and sustains gender inequality. Journal of Personality and Social Psychology, 101(1), 109–128. https://doi.org/10.1037/a0022530

30.

Goodhew

S. C.

Reynolds

Edwards

Kidd

(2022). The content of gender stereotypes embedded in language use. Journal of Language and Social Psychology, 41(2), 219–231. https://doi.org/10.1177/0261927X211033930

31.

Grohmann

Schoofs

(2021). Financial literacy and intra-household decision making: Evidence from Rwanda. Journal of African Economies, 30(3), 225–250. https://doi.org/10.1093/jae/ejaa007

32.

Hourigan

K. L.

(2021). Girls try, boys aim high: Exposing difference in implied ability, activity, and agency of girls versus boys in language on McDonald’s Happy Meal boxes. Sex Roles, 84, 377–391. https://doi.org/10.1007/s11199-020-01173-7

33.

Huston

S. J.

(2010). Measuring financial literacy. Journal of Consumer Affairs, 44(2), 296–316. https://doi.org/10.1111/j.1745-6606.2010.01170.x

34.

IAB Polska. (2023). Przewodnik po social mediach w Polsce [A guide to social media in Poland]. https://www.iab.org.pl/wp-content/uploads/2023/06/Przewodnik-Social-Media-2023-IAB-Polska.pdf

35.

Jin

Zhang

Chen

(2017). How consumers are affected by product descriptions in online shopping: Event-related potentials evidence of the attribute framing effect. Neuroscience Research, 125, 21–28. https://doi.org/10.1016/j.neures.2017.07.006

36.

Johansson

(2005). Origins of language: Constraints on hypotheses. John Benjamins BV.

37.

Kricheli-Katz

Regev

(2021). The effect of language on performance: Do gendered languages fail women in maths? Npj Science of Learning, 6(1), 9. https://doi.org/10.1038/s41539-021-00087-7

38.

Lakoff

R. T.

(1973). Language and woman’s place. Language and Society, 2(1), 45–80.

39.

Lakoff

R. T.

Johnson

(2003). Metaphors we live by. University of Chicago Press.

40.

Leaper

Ayers

M. M.

(2007). A meta-analytic review of gender variations in adults’ language use: Talkativeness, affiliative speech, and assertive speech. Personality and Social Psychology Review, 11(4), 328–363. https://doi.org/10.1177/1088868307302221

41.

Leaper

Robnett

R. D.

(2011). Women are more likely than men to use tentative language, aren’t they? A meta-analysis testing for gender differences and moderators. Psychology of Women Quarterly, 35(1), 129–142. https://doi.org/10.1177/0361684310392728

42.

Levantesi

Zacchia

(2021). Machine learning and financial literacy: An exploration of factors influencing financial knowledge in Italy. Journal of Risk and Financial Management, 14(3), 120. https://doi.org/10.3390/jrfm14030120

43.

(2020). When financial literacy meets textual analysis: A conceptual review. Journal of Behavioral and Experimental Finance, 28, 100402. https://doi.org/10.1016/j.jbef.2020.100402

44.

Mavisakalyan

Weber

(2018). Linguistic structures and economic outcomes. Journal of Economic Surveys, 32(3), 916–939. https://doi.org/10.1111/joes.12247

45.

Meier

Boyd

R. L.

Mehl

M. R.

Milek

Pennebaker

J. W.

Martin

Wolf

Horn

A. B.

(2020). Stereotyping in the digital age: Male language is “ingenious”, female language is “beautiful”– And popular. PLoS One, 15(12), e0243637. https://doi.org/10.1371/journal.pone.0243637

46.

Mulac

Bradac

J. J.

Gibbons

(2001). Empirical support for the gender-as-culture hypothesis. Human Communication Research, 27(1), 121–152. https://doi.org/10.1111/j.1468-2958.2001.tb00778.x

47.

Mulac

Lundell

T. L.

(1986). Linguistic contributors to the gender-linked language effect. Journal of Language and Social Psychology, 5(2), 81–101. https://doi.org/10.1177/0261927X8652001

48.

Newman

M. L.

Groom

C. J.

Handelman

L. D.

Pennebaker

J. W.

(2008). Gender differences in language use: An analysis of 14,000 text samples. Discourse Processes, 45(3), 211–236. https://doi.org/10.1080/01638530802073712

49.

Nowak

(2002). SWOI i OBCY w językowym obrazie świata (INSIDERS and OUTSIDERS in the linguistic worldview). UMCS.

50.

Nowak

(2020). Naturalny porządek rzeczy w języku (The natural order of things in language). KSAP.

51.

Ooi

(2020). Give mind to the gap: Measuring gender differences in financial knowledge. Journal of Consumer Affairs, 54(3), 931–950. https://doi.org/10.1111/joca.12310

52.

Park

Yaden

D. B.

Andrew

S. H.

Kern

M. L.

Eichstaedt

J. C.

Kosinki

Stillwell

Ungar

L. H.

Seligman

M. E. P.

(2016). Women are warmer but no less assertive than men: Gender and language on Facebook. PLoS One, 11(5), e0155885. https://doi.org/10.1371/journal.pone.0155885

53.

Pennebaker

J. W.

Mehl

M. R.

Niederhoffer

K. G.

(2003). Psychological aspects of natural language use: Our words, our selves. Annual Review of Psychology, 54, 547–577. https://doi.org/10.1146/annurev.psych.54.101601.145041

54.

Phillips

Boroditsky

(2003). Can quirks of grammar affect the way you think? Grammatical gender and object concepts. In Alterman

Kirsch

(Eds.), Proceedings of the 25th Annual Cognitive Science Society (pp. 928–933). Psychology Press.

55.

Piersoul

Van de Velde

(2023). Men use more complex language than women, but the difference has decreased over time: A study on 120 years of written Dutch. Linguistics, 61(3), 725–747. https://doi.org/10.1515/ling-2021-0022

56.

Plug

Stommel

Lucassen

P. L. B. J.

olde Hartman

T. C.

van Dulmen

Das

(2021). Do women and men use language differently in spoken face-to-face interaction? A scoping review. Review of Communication Research, 9, 43–79. https://doi.org/10.12840/ISSN.2255-4165.026

57.

Potrich

A. C. G.

Matheis

T. K.

Vieira

K. M.

(2024). Gender differences in financial literacy between and within generations. Journal of Financial Counseling and Planning, 35(3), 381–399. https://doi.org/10.1891/JFCP-2023-0058

58.

Preston

A. C.

Qiu

Wright

R. E.

(2023). Understanding the gender gap in financial literacy: The role of culture. Journal of Consumer Affairs, 58(1), 146–176. https://doi.org/10.1111/joca.12517

59.

Preston

A. C.

Wright

R. E.

(2024). When does the gender gap in financial literacy begin. Economic Record, 100(328), 44–73. https://doi.org/10.1111/1475-4932.12785

60.

Remund

D. L.

(2010). Financial literacy explicated: The case for a clearer definition in an increasingly complex economy. Journal of Consumer Affairs, 44(2), 276–295. https://doi.org/10.1111/j.1745-6606.2010.01169.x

61.

Rink

Walle

Y. M.

Klasen

(2021). The financial literacy gender gap and the role of culture. The Quarterly Review of Economics and Finance, 80, 117–134. https://doi.org/10.1016/j.qref.2021.02.006

62.

Robson

Peetz

(2020). Gender differences in financial knowledge, attitudes, and behaviors: Accounting for socioeconomic disparities and psychological traits. Journal of Consumer Affairs, 54(3), 813–835. https://doi.org/10.1111/joca.12304

63.

Searle

J. R.

(1987). Czynności mowy: rozważania z filozofii języka. Speech acts: An essay in the philosophy of language. PAX.

64.

Sesini

Manzi

Lozza

(2023). Is psychology of money a gendered affair? A scoping review and research agenda. International Journal of Consumer Studies, 47(6), 2701–2723. https://doi.org/10.1111/ijcs.12975

65.

Shoham

Lee

S. M.

(2018). The causal impact of grammatical gender marking on gender wage inequality and country income inequality. Business & Society, 57(6), 1216–1251. https://doi.org/10.1177/0007650317696231

66.

Tajfel

Turner

(1979). An integrative theory of intergroup conflict. In Austin

W. G.

Worchel

(Eds.), The social psychology of intergroup relations (pp. 33–57). Wadsworth.

67.

Tannen

(1994). Gender and discourse. Oxford University Press.

68.

Thomson

Murachver

Green

(2001). Where is the gender in gendered language? Psychological Science, 12(2), 171–175. https://doi.org/10.1111/1467-9280.00329

69.

Tinghög

Ahmed

Barrafrema

Lind

Skagerlund

Västfjäll

(2021). Gender differences in financial literacy: The role of stereotype threat. Journal of Economic Behavior and Organization, 192, 405–416. https://doi.org/10.1016/j.jebo.2021.10.015

70.

Tokarski

(2014). Światy za słowami. Wykłady z semantyki leksykalnej (Worlds behind words: Lectures on lexical semantics). UMCS.

71.

Van der Velde

Tyrowicz

Siwinska

. (2015). Language and (the estimates of) the gender wage gap. Economics Letters, 136, 165–170. https://doi.org/10.1016/j.econlet.2015.08.014

72.

Von Hippel

Sekaquaptewa

McFarlane

. (2015). Stereotype threat among women in finance: Negative effects on identity, workplace well-being, and recruiting. Psychology of Women Quarterly, 39(3), 405–414. https://doi.org/10.1177/0361684315574501

73.

Willis

L. E.

(2008). Against financial-literacy education. Iowa Law Review, 94, 197–285.

74.

Wołkowski

(2010). Biologiczne aspekty teorii zdolności językowej Chomsky’ego (Biological aspects of Chomsky’s theory of language ability). Diametros, 23, 145–161.