Teachers play a key role in supporting and encouraging egalitarian attitudes at school. This study aimed to develop and validate the Feminist Attitudinal and Behavioral Scale for Teachers (FABS-T), a measure that comprises both attitudinal and behavioral domains. To do this, we conducted two studies of Childhood and Primary Education teachers and pre-service teachers. Study 1, qualitative (N = 309), asked participants about situations in which they had been portrayed as feminist. Six themes emerged, for which we developed an initial 60-item version. Study 2 (N = 451) aimed to evaluate the factor structure, reliability and validity. Exploratory and confirmatory factor analysis supported a 42-item version of the FABS-T with a final structure composed of five dimensions (56.7% of the variance explained). For both the attitudinal (Public Egalitarian Support and Private Egalitarian Support) and behavioral domain (Sexism Confrontation, Activism and Teaching Practices), reliability was good. Most dimensions were positively linked with feminist self-identification, and negatively with ambivalent sexism and social dominance orientation. Gender differences were observed in the behavioral dimensions. We highlight the need to distinguish attitudes from behaviors, where the latter was revealed as a core aspect of self-identifying as feminist.

Gender equality is one of the main priorities that the Education 2030 Agenda emphasizes in its Sustainable Development Goals (SDGs), i.e., “women and men (…) are empowered equally”. Feminism addresses gender inequality in society (Swirsky & Agelone, 2015), while Sustainable Development Goal 4 (SDG 4) in Agenda 2030 highlights the importance of inclusive and quality education. Therefore, schools are often the best scenario in which changes to obtain gender equality can be made. Recent progress, exemplified by coeducation, aims for equal student development, free from gender stereotypes (Aragonés-González et al., 2020). The LOMLOE, Spain's current Education Law, recognizes and promotes gender equity at all educational levels from a fully inclusive perspective. Teachers can play a crucial role in fostering positive gender roles and promoting gender equality (Chinen et al., 2017). They play a key role in this process, as their attitude towards gender equality greatly influences the way students learn about and understand gender equality. This is significantly influenced by the teacher training in gender equality, which is still limited. Previous research concluded that primary school teachers unconsciously perpetuate gender stereotypes (Gray & Leith, 2004). Gender stereotypes are entrenched conceptions about the characteristics, priorities, and needs that are assumed to be possessed by individuals of each gender (Ellemers, 2018; Fabris et al., 2020). It should be noted that, for childhood and primary school teachers, detecting gender inequality remains arduous because they report difficulties with both implicit and explicit forms of inequality based on gender (Pinedo et al., 2018). A study conducted by Romero and Lugo (2014) with university teachers found that they reported positive attitudes toward gender equality, which were slightly more frequent among women. Díaz and Anguita (2017) indicated that gender stereotypes are still rooted in teachers, but today we make more efforts to dissimulate these mental representations because explicit forms of sexism are not socially tolerated. Thus, one of the main strategies to challenge the endorsement of gender stereotypes is education.

The endorsement of feminist attitudes and feminist self-identification are not necessarily paired. Intuitively, one would consider an individual to self-identify as feminist based on their attitudes and behaviors. However, this has not been demonstrated. Research suggests that some women support certain feminist behaviors, but seem ambivalent about others (Rollero et al., 2022) or are even reluctant to be defined as feminist (Aronson, 2003). This inconsistency opens debate about what conceptualization of feminism people endorse because negative opinions of some aspects of feminism are frequently held (Hoskin et al., 2017). Therefore, instead of focusing on a unique definition of feminism, it would be interesting to further probe everyone’s definition of feminism and have them recall specific situations in which they behaved like feminists.

According to the Royal Spanish Language Academy (n.d.), feminism is defined as the “principle of equality of rights for both men and women”. However, the term feminism still leads to some confusion. In line with this, several factors that underlie feminist self-identification have been explored. The feminist identity seems to be better explained by positive attitudes toward feminist prototypes (Redford et al., 2018) based on endorsing feminist values and perceived discrepancy (Meijs et al., 2017). On the contrary, holding a negative prototype of feminists or one that is dissonant from their own self-concept would lead individuals to refuse being self-labeled as feminists (Gundersen & Kunst, 2019; Moradi et al., 2012). In fact, a negative view of the term feminism has somehow become widespread in several spheres in which feminists are portrayed as irrational, man-hating, unfeminine, and fanatical (Calder-Dawe & Gavey, 2016), or are even represented by terms like “feminazi” (Cohen, 2015). For some, feminism is considered the antagonist of “sexism” -“machismo” in Spanish.

Many scales designed to assess feminism have psychometric limitations or are derived from outdated models and do not capture modern attitudes in our changing cultural context (Siegel & Calogero, 2021). The following factors must also be taken into account: a) none of the scales analyzed in the review study evaluate both attitudinal and behavioral aspects of feminism, except that of Sudkämper et al. (2020) developed for men (although it clustered items, i.e., does not distinguish the two aspects, b) a more detailed qualitative study in which women´s and men´s opinion are considered is needed, and c) no specific measure of teachers’ views has been developed (Siegel & Calogero, 2021). Therefore, scales in which attitudes and behaviors are clearly distinguished based on sounded psychometric properties are lacking.

Men are less likely to identify as feminist and report lower acceptance of activism oriented toward feminism (Wietholter & Hillard, 2021). This is probably due to several reasons, such as the heteropatriarchal society (Arvin et al., 2013). Men are in a position of strength and dominance, and are more likely to be aggressor (Sedgwick, 2006). In this regard, it is necessary to highlight the influence of the biologically assigned sex at birth on the construction of gender roles and expectations (Clarke, 2022; van Anders et al., 2017). Sex refers to the biological and physiological characteristics that differentiate between males and females, while gender pertains to the roles, behaviors, activities, and expectations that society deems appropriate for men and women (van Anders et al., 2017).

Other factors, such as ideologies (sexism or social dominance orientation [SDO]; see Glick & Fiske, 1996; Sidanius & Pratto, 1999), may be at the root of men´s non-endorsement of the feminist label. Individuals who report more feminist attitudes also report less benevolent sexism (Kunst et al., 2019) and hostile sexism (Estevan-Reina et al., 2020). Furthermore, feminism is related to certain social and political attitudes, where witnessing gender discrimination predicts greater willingness to participate in collective action for gender justice (Uluğ et al., 2023). SDO (Pratto et al., 1994) is an ideology that predicts several forms of prejudice, such as sexism (Duckitt & Sibley, 2010), and is therefore negatively related to feminism (Smith & Winter, 2002). Based on social dominance theory (SDT) (Sidanius & Pratto, 1999), SDO indexes individual differences in the preference for group-based hierarchy and inequality. As such, it is one of the most powerful predictors of intergroup attitudes and behaviors (Asbrock et al., 2010). Therefore, SDO can stratify dominant and subordinate groups by gender, where men have more power than women (Sidanius & Pratto, 1999).

The present research arises to develop and validate a questionnaire on feminism for use in the education field. We conducted two studies:

Study 1 aimed to develop a measure of feminism, considering both attitudinal and behavioral aspects. Due to its exploratory nature, no hypotheses were established.

Study 2 had the main goals of examining the factorial structure and assessing the reliability and validity of the developed scale through comparison with related measures.

The specific goals and hypotheses of Study 2 were as follows:

Quota convenience sampling was used. We collected data from 321 individuals. Sample size was estimated considering the number of teachers in Spain, at a confidence level of 95%. Participants who did not provide an answer to the open question were excluded (n = 12). Data from 309 individuals were analyzed. Inclusion criteria were: (a) aged ≥ 18 years, (b) Spanish nationality, (c) enrolled in a Childhood and/or Primary Education Degree and experience working with children during their formation or professionally for at least 1 year as a practice teacher, or (d) teaching in Childhood and/or Primary Education. A main exclusion criterion was considered: Not to have being enrolled in a previous teaching experience. Participants were mostly women (n = 209; 67.6%) ranging in age from 20 to 69 years (M = 41.64; SD = 10.74). Most participants had a university degree.

The URL of the questionnaire, which was hosted on the University of BLINDED LimeSurvey Platform, was distributed by a news service of the universities involved in the study, a social mailing list maintained by educational centers, and among university students at the University of BLINDED. Once participants clicked on the link to access the survey, they provided informed consent; the form explained that the study had the goal of analyzing attitudes, opinions and behaviors related to gender issues. Questionnaires were completed in approximately 20 to 30 min. Anonymity and confidentiality were guaranteed. No compensation was provided. This study was approved by the Research Ethics Committee of the Autonomous Community of Aragon: CEICA (No. 08/2019) from the University of BLINDED.

Feminist attitudes and behaviors. Participants were asked: Do you identify as a feminist? They had the following options: “Yes”, “No” and “Uncertain”. They were also asked to answer the following: If you have marked the “yes” option, please describe at least one situation in which you have thought/behaved like a feminist.

Most answers related to one situation that reflected a main theme. In a few cases, the response was divided into two “units” (24 responses out of 309) corresponding to two different situations. Ultimately, we analyzed 338 units of information. MAXQDA 12 was used both for transcription and analysis. Two trained researchers reviewed all collected answers independently and organized them into themes. Raters met to discuss and agree on themes. They found that Rater 1’s identified themes were easily collapsible into Rater 2’s broader theme categories. Ultimately, there were seven distinct themes. Raters agreed about 263 units (kappa index of .77). Most of the discrepancies in the number of themes between the two raters were because Rater 2 collapsed answers into broader themes, while Rater 1 used more stringent criteria to constitute a theme.

Most women (n = 124; 59.6%), and almost half of the men (n = 48; 48%), self-identified as feminist, while 26% of women (n = 54) and 31% of men (n = 31) chose the option “uncertain about feminist”. The remaining participants responded “No”.

Six main themes relating to situations in which participants felt they behaved as feminists emerged and were divided into two broader domains: Attitudes and Behaviors. For Attitudes, the themes were: 1) Public Egalitarian Support, 2) Private Egalitarian Support and 3) Working Egalitarian Support. For Behaviors, the themes were: 1) Sexism Confrontation, 2) Activism and 3) Teaching Practices.

Among the themes, the most reported were Sexism Confrontation (26.8%), Teaching Practices (21.2%), and Activism (10.4%) from the Behavioral domain; and Private Egalitarian Support (17.5%), Working Egalitarian Support (7.8%) and Public Egalitarian Support (6.7%) from the Attitudinal domain. Roughly 9.7% of the responses were ambiguous or nonspecific i.e., did not belong to any category. In the supplementary material, definitions and examples for each theme are shown according to domain.

No significant gender differences emerged in the reported themes [2 (3, 269) = 4.34, p = .820].

Two researchers independently devised a pool of 10-15 items for each dimension. To do this, the researchers followed the recommendations set out in previous research (Muñiz & Fonseca-Pedrero, 2019), including definition of the construct, item development, pilot study, selection of other measures, test application and analysis of the scale´s psychometric properties to develop the final version of the test.

Second, the researchers selected the 10 items that best fit the definition of each dimension. If two items were duplicated or contain redundant information, one was removed. Ultimately, we devised a 60-item version of the FABS-T. Participants indicated their agreement with each item on a 6-point Likert scale (1 = Strongly disagree, 6 = Strongly agree). The Spanish version can be provided upon request.

The initial preliminary scale was sent to five experts in Psychometry and/or Feminism for them to judge the representativeness, clarity and precision of its items. The experts were provided with the definitions and corresponding items. They all agreed about the items and only slight modifications related to specific aspects of their wording were made according to the comments. No conceptual changes were made.

Following the same sampling method, estimation of sample size and inclusion criteria to Study 1, we collected data from 482 individuals; however, due to missing values in more than 25% of the FABS-T items, the final sample comprised 451 participants. About 82.7% of the sample were women who ranged in age from 18-66 years (M = 39.25; SD = 13.31). The sample was randomly divided into two sub-samples, as follows: a) The sub-sample for EFA comprised 225 individuals, who were mostly women (76%) and ranged in age from 18-62 years (M = 37.43; SD = 14.29) while b) the sub-sample for the CFA was composed of 226 individuals (82.7% women) who ranged in age from 18-66 years (M = 39.25; SD = 13.31).

Identical to study 1.

Feminist attitudes and behaviors. We used the Feminist Attitudinal and Behavioral Scale (FABS-T) developed in Study 1.

Feminist Self-Identification. We used the FSI (Morgan, 1996) scale, which asks the following: “To what extent do you consider yourself a feminist?”. As there is no available Spanish version, we conducted a direct or forward translation of the original version into Spanish. Responses were provided via a Likert response scale (1, I do not consider myself a feminist at all; 7, I am a committed and active feminist in the women’s movement).

Sexism. We used the Ambivalent Sexism Inventory (ASI; Glick & Fiske, 1996) in its Spanish form (Expósito et al., 1998). This 22-item inventory assesses sexist attitudes by distinguishing between hostile (11 items) and benevolent sexism (11 items) using a Likert response scale (0, Completely disagree; 5, Completely agree). Hostile sexism is understood as discrimination against women as a group simply because of their gender (e.g., Women seek to gain power by getting control over men). Benevolent sexism is defined as a more subtle form of discrimination that hides unequal and degrading treatment of women associated with their “weaker sex” status (e.g., Many women have a quality of purity that few men possess). Higher summed scores indicate greater support of hostile or benevolent sexism. Cronbach's alpha values range from .80 to .92 for hostile sexism and from .73 to .85 for benevolent sexism. In the present study, Cronbach's alpha was .92 for hostile sexism and .83 for benevolent sexism.

Social dominance orientation. We used the Spanish version of the SDO (Pratto et al., 1994; Silván-Ferrero & Bustillos, 2007). This scale assesses the degree to which someone wants their own group to dominate another that they consider inferior (e.g., It’s probably a good thing that certain groups are at the top and other groups are at the bottom). It consists of 16 items and uses a Likert response scale (1, Strongly oppose; 7, Strongly favor). After reversing the scores for some items, all scores are summed. Cronbach's alpha was reported as .90; in this study, it was .75.

Prior to further analysis, item´s descriptive analysis was extracted in which corrected item-total correlation was analyzed. Values under .30 were not yielded indicating weakness for any of the items.

The adequacy of the data to perform factorization was verified by the Kaiser-Meyer-Olkin test (KMO = .902) and Bartlett’s sphericity test (χ2 = 6,252.11; p <.001). EFA was performed via principal component analysis (PCA) with the Oblimin rotation method. A parallel analysis confirmed the EFA's findings. As shown in Table 1, a six-factor structure was tested. Factor loadings were significant and ranged between .31 (item 17) and .89 (item 28). Some of the items shared factor loadings on more than one factor, although the difference between both factor loadings exceeded .10 for all of them except items 12, 31, 38 and 46. For all factors, eigenvalues ​​exceeded 1. The six-factor structure explained 56.7% of the variance: the first factor explained 27.29% of the variance, followed by Factor 2 (14.59%) and Factors 3-6 (4.62%, 3.90%, 3.36% and 2.92, respectively). Item distributions across factors are shown in Table 1.

Likewise, the corrected item-total correlations for each factor were examined (≤ .30 values, item was discarded), as was reliability after eliminating any item (if significantly improved, item was discarded). Six items from Factor 1 were eliminated by this procedure (items 7, 9, 17, 24, 26 and 48), two from Factor 2 (1 and 22), two from Factor 4 (23 and 29) and one (38) from Factor 5. In all cases, the eliminated items corresponded to those with a lower factor loading.

We performed CFA using AMOS v. 24. The maximum likelihood score was calculated and the following goodness of fit indices were used: the root mean square error of approximation (RMSEA), comparative fit index (CFI), Tucker-Lewis index (TLI) and normed fit index (NFI). RMSEA values less than .06 indicated a good fit, and values below .10 were considered acceptable. For the CFI, TLI and NFI, values above .90 indicate acceptable fit. χ2 values and their corresponding degrees of freedom were also reported, where lower values indicate a better fit (Kline, 2023).

Initially, the following models were tested: (1) Model 1: six-factor model that emerged from the EFA, (2) Model 2: one-factor model in which only one latent factor clustered all items and (3) Model 3: two-factor model in which Attitudinal and Behavioral were the two latent dimensions.

Table 2 shows the goodness of fit indices for all models. When testing Model 1, the goodness of fit indices were inadequate; therefore, we modified some indices and found that the model would improve if some items from Factor 6 were deleted, as errors should be correlated with errors from items in other factors. Therefore, and based on the previous EFA in which Factor 6 comprised a mix of several items that conceptually should belong to other factors, we decided to eliminate this factor. Once this change was made, Model 1(b) was constructed and had a five-factor structure. This model was significantly better: all goodness of fit indices were good. Finally, for Models 2 and 3, no goodness of fit indices had adequate values.

Figure 1 shows the results of the CFA, which yielded a robust five-factor model with 42 items confirming the model equation. Standardized factor loadings are shown.

As seen in Table 3, all factors obtained high mean values, except for Sexism Confrontation. Gender differences emerged only for Sexism Confrontation and Activism, for which women had higher scores. Finally, Cronbach’s alpha values were high and ranged between .81 for Factor 5 (Teaching Practices) and .96 for Factor 1 (Public Egalitarian Support).

Significant correlations emerged among factors, except between Activism with Private Egalitarian Support and with Public Egalitarian Support (Table 4). The lowest correlation was shown between Public Egalitarian Support and Sexism Confrontation (r = .26; p < .01), and the highest was between Private Egalitarian Support and Public Egalitarian Support (r = .66; p < .01), with both factors slightly overlapping.

There were significant correlations between Sexism Confrontation, Activism, Teaching Practices and Public Egalitarian Support and the degree of feminist self-identification. All FABS-T dimensions negatively correlated with ambivalent sexism. Finally, a negative correlation was observed between the dimensions of the FABS-T and SDO.

Our study aimed to develop and validate a measure of feminism for use in the educational field, comprising both attitudinal and behavioral domains, since “feminist” is better understood using a multifaceted measure (McLaughlin & Aikman, 2020). The FABS-T is a 42-item measure of two attitudinal dimensions, Public Egalitarian Support and Private Egalitarian Support, and three behavioral dimensions: Sexism Confrontation, Activism and Teaching Practices. While the Attitudinal domain refers to supporting ideas related to equality (rights, responsibilities, opportunities) between women and men, the Behavioral dimension reflects the willingness to actively behave as a feminist (e.g., activism, sexism confrontation against sexism and the use of egalitarian forms of teaching). Reliability and validity are supported.

In Study 1, feminism was expressed by several diverse behaviors and attitudes. The most reported feminist behavior was related to Sexism Confrontation. In prior definitions of feminism, this is presented as a social movement to end the sexist oppression that hurts everyone (Hooks, 2000) and emphasizes robust social action toward gender equality. This theme was followed by Public Egalitarian Support, which aligns with the classical definition of feminism based on equality between men and women and has commonly been reported in other studies asking participants to define “feminist” (Ogletree et al., 2017).

Individuals also endorse feminist attitudes in the private sphere, particularly their home, indicating that domestic tasks should be distributed. This reflects current trends where men more often play a care-giving role (Langner & Furstenberg, 2018). However, in Spain, egalitarianism in division of roles has not been reached, and only the most advantaged women are in couples with a relatively egalitarian division of time (García-Román, 2023). However, some legal changes have been undertaken by the Spanish government to promote gender equality in caring (e.g., paternity leave was increased in 2021). It seems that while some progress has been made toward gender equality in public spheres (England et al., 2020), this has not spread to the private sphere to the same extent (Hopcroft & McLaughlin, 2012).

Another dimension of feminism is Activism (Conlin et al., 2019), which is among the main means through which women can increase their feminist identification (Conlin & Heesacker, 2018) by becoming actively involved in feminist groups or sharing information. Finally, the only dimension specifically related to the educational field was Teaching Practices, in which teachers encourage gender equality behaviors in the classroom. Teaching strategies at schools are a crucial starting point for promoting gender equity, as gender-stereotyped beliefs may substantially influence gender differences in students (Gunderson et al., 2012).

Our findings indicate gender differences in the behavioral dimensions, but not the attitudinal ones, with women being more willing to behave as feminists than men, as shown previously (e.g., Wietholter & Hillard, 2021). The lack of gender differences in the attitudinal domain could be explained by both the public and domestic spheres being broad domains in which individuals can easily support gender equality, where society is now more aware that gender equality should be defended due to the influence of mass media, public and private dialog, and popular culture (Renegar & Sowards, 2009). As indicated by Díaz and Anguita (2017) gender stereotypes are deep-rooted, but today we make more effort to dissimulate these mental representations because explicit forms of sexism are not socially tolerated. Therefore, the core message of gender equality seems clear in society, but is not always been translated into action, especially for men; they may be unwilling to drive social change due to their privileged status that discourages involvement in feminist activism (Burrell & Flood, 2019). Other plausible interpretations include a weakening of traditional gender role attitudes and strengthening of feminist ones (Díaz & Sellami, 2014), where a 10-country United Nations pilot study (2019) indicated that 84% of individuals believe that it is essential for society to treat women as equal to men.

Attitudinal and behavioral domains, although part of the same construct, should be distinguished because endorsing feminist attitudes may not be sufficient to identify someone who actually performs feminist behaviors. Furthermore, positive attitudes toward equality are not necessarily linked with feminist self-identification, which again emphasizes that attitudes per se are independent of self-identification as feminist. Instead, behaviors are key aspects of self-identifying as feminist. According to our study, although egalitarian attitudes are part of feminist self-identification, they have very little to do with the strength thereof; feminist identification is more clearly represented by specific actions. This is consistent with previous research underlining that feminist identification is a key correlate of feminist collective action (e.g., Yoder et al., 2011). However, attitudes should be understood as a first step to change behaviors: improving attitudes toward feminist prototypes may promote feminist identification (Weis et al., 2018).

All FABS-T dimensions were negatively associated with benevolent and hostile sexism and SDO. Hence, in line with previous studies, a link between endorsement of sexist attitudes and reluctance to identify as feminist is apparent (Estevan-Reina et al., 2020). Additionally, individuals who support the superiority of some social groups also refuse to support gender equality (Sudkämper et al., 2020), to maintain their privileged social position. Ideologies seem important drivers of the degree of support of feminist actions (e.g., the #metoo campaign) (Kunst et al., 2019).

Finally, feminism is aligned with human rights. The statement “Women's rights are human rights” (Amnesty International, 2023) has been praised by international women's rights activists and experts for its rights-based perspective on gender equality. SDGs are amenable to analysis that considers the interaction between social identities and related inequalities and can transform rhetoric into practice through national plans and policy changes, for example (Stuart & Woodroffe, 2016).

Our study has several limitations. First, individuals mainly described one situation, likely that most relevant or important to them, when describing their feminist behavior. However, whether these behaviors were performed is uncertain, as is the extent of their other forms of feminism. Second, the number of men and women in each study differed, although this reflects the gender bias in teaching. Third, although anonymity and confidentiality were guaranteed, our findings were based on self-report data, and could be affected by social desirability and recall bias.

Several lines of research could be pursued using the FABS-T. First, the variables and motivations underlying each domain of feminism (i.e., attitudinal and/or behavioral) should be investigated, where people may or may not actively engage in behaviors fostering social change. Second, both individual and social outcomes of attitudinal and behavioral domains of feminism could be explored. Several studies concluded that feminist identification provides some individuals with benefits, such as greater well-being (Saunders & Kashubeck-West, 2006). Studies should address the outcomes of social commitment to change, where feminist ideology and behaviors could be related to collective action favoring women (Girerd & Bonnot, 2020). Third, a more comprehensive study of the diverse feminist identities could improve understanding of the attitudes and behaviors that characterize different forms of feminism, such as non-labelers and feminists, and shed light on the different levels of feminist consciousness (Duncan et al., 2021).