• Miriam Liss and Mindy J. Erchull, Department of Psychology, University of Mary Washington.

Address correspondence and reprint requests to: Miriam Liss, Department of Psychology, University of Mary Washington, 1301 College Avenue, Fredericksburg, VA 22401-5300. E-mail: mliss@umw.edu


Research findings raise questions about whether the feminist identity development model provides information about women's social identification as a feminist. Specifically, the penultimate stage, Synthesis, has been theorized to capture when feminist identity formation coalesces and women take on the feminist label. However, available data have suggested this stage may not be related to feminist self-labeling, calling for a better understanding of the variables associated with identifying oneself as a feminist. An online questionnaire was administered to 653 female self-identified feminists and nonfeminists in order to investigate the association between feminist self-labeling and Synthesis scores and to better understand what it means to take on the social identity of a feminist. Feminist self-labeling was not associated with Synthesis; however, women who self-labeled as feminists were more likely to acknowledge the existence of sexism, view the current gender system as unjust, and believe that women should work together in order to enact change. Synthesis was related to a combination of feminism- and conservatism-related constructs. Women high in Synthesis viewed the current gender system as just yet also believed that women should work together to enact change. We discuss the paradox represented by this combination of beliefs as well as their implications for the feminist identity development model and the women's movement in general.

What does it mean when a woman has fully embraced a feminist identity? Presumably, this woman would hold beliefs about gender that could be classified as liberal, or even radical, rather than conservative. She would be aware of and intolerant of sexism and would believe that women should work together to achieve feminist goals. This woman would both be aware of structural inequities and unhesitatingly call herself a feminist.

The feminist identity development model was developed by Downing and Roush (1985) in order to explain the process women go through on their path toward a fully integrated feminist identity. They postulated that women move through a series of five stages. Women begin in Passive Acceptance wherein they accept traditional gender roles and do not question established gender inequity. Next, women experience Revelation, where they become aware of inequality and feel angry and disillusioned. Women then transition to Embeddedness-Emanation, a stage where they surround themselves with like-minded women. Women gradually emerge from this seperationist perspective and enter the stage of Synthesis, where they are able to integrate their sense of themselves as women with their sense of themselves as individuals. It is at this stage that Downing and Roush (1985) hypothesized that women have achieved a fully formed feminist identity and, presumably, would take on the characteristics described above. The final stage, Active Commitment, involves working on behalf of women's rights because the feminist identity achieved in the previous stage compels individuals to act to change society for all women.

Although Downing and Roush (1985) conceptualized the feminist identity development model as a series of progressive stages, some scholars conceptualize these stages as attitudinal dimensions (e.g., Hyde, 2002; Moradi, Subich, & Phillips, 2002). The measurement of these dimensions has been operationalized in three different ways. The Feminist Identity Development Scale (FIDS; Bargad & Hyde, 1991) measured each stage separately. The Feminist Identity Scale (FIS; Rickard, 1989) was an alternate operationalization in which the Synthesis and Active Commitment dimensions were combined. The Feminist Identity Composite (FIC; Fisher et al., 2000) was developed to address the poor reliability found within some subscales of both the FIDS and the FIS. This composite instrument was developed using the most reliable items from each subscale. All five stages were measured separately in this composite measure.

Although the development of the FIC resulted in improved reliability, a number of problems remained, particularly with the Synthesis stage. Conceptually, the Synthesis stage is the time when women fully articulate their feminist identity. Indeed, Bargad and Hyde (1991) specifically noted that feminist self-labeling was the behavioral manifestation of Synthesis. Although Bargad and Hyde (1991) intended to explicitly measure self-labeling as one aspect of the Synthesis stage, these items were removed from the final Synthesis subscale to achieve an acceptable level of reliability. The items that remained dealt with women's relationships with men. The Synthesis items in the FIS (Rickard, 1989) reflected the idea that women felt strong, proud, competent, and able to integrate their individuality with their sense of femininity. These items are striking in their individualistic tone because they focus on the woman herself rather than on her relationships with other women or with the feminist movement. This individualism makes sense in the context of Downing and Roush's (1985) model because women entering the Synthesis stage have just emerged from Embeddedness, where they had been focused on a community of women rather than on themselves. Finally, the Synthesis subscale of the FIC consists entirely of items from the FIS Synthesis subscale (without the Active Commitment items) and thus maintains the themes of individualism and of integrating one's femininity with one's sense of self.

The Synthesis subscale has been conceptually problematic from the beginning (for a review of these issues, see Moradi et al., 2002). Women enrolled in women's studies classes reported no change in Synthesis scores over the course of the semester (Bargad & Hyde, 1991). In studies that have attempted to differentiate feminists from women who endorsed feminist beliefs but did not take on the label, Synthesis has consistently been the only stage that did not differentiate the groups (Erchull & Rubin, 2004; Zucker, 2004). One study did find a small relationship between feminist self-labeling and Synthesis; however, this study utilized the FIS, which operationalizes Synthesis as including Active Commitment (Liss, O'Connor, Morosky, & Crawford, 2001). Furthermore, research has consistently found high Synthesis scores for all women (e.g., Erchull et al., 2009; Liss et al., 2001).

One group of researchers has postulated that the Synthesis dimension, as measured by the FIC, should no longer be conceptualized as the apex of the feminist identity development process (Erchull et al., 2009). Many young women today may feel strong, proud, competent, and independent, but they have not necessarily gone through previous stages or labeled themselves as feminists. These researchers asked women who scored high in Synthesis whether or not they had ever experienced previous stages, and for the most part, the participants reported that they had not (Erchull et al., 2009). Thus, Synthesis may actually be a starting point for young women today. These women may not have developed their sense of empowerment through identifying as a feminist; rather, they may have a sense of empowerment simply because they live in a society where women presumably can now do anything they wish. In this way, young women with high Synthesis scores may actually be somewhat unaware of the continued existence of gender discrimination.

We know of no research specifically investigating variables that are related to Synthesis scores. Given that the link between Synthesis and feminism is ambiguous at best, better understanding the construct of Synthesis is particularly important. Some items in the Synthesis subscale (e.g., “I am proud to be a competent woman”) contain ideas about being a strong and powerful woman, which may be related to feminist constructs. However, other items stress individualism (e.g., “As I have grown in my beliefs I have realized that it is more important to value women as individuals than as members of a larger group of women”) and an ability to integrate one's femininity with one's sense of self (e.g., “I have incorporated what is female and feminine into my own unique personality”). These ideas may be less related to core feminist beliefs. Indeed, in one study, an inexplicable small positive correlation was found between Synthesis and Passive Acceptance (Liss et al., 2001). Thus, Synthesis, as currently operationalized, may be related paradoxically to both feminism- and conservatism-related constructs.

Given that Synthesis may not be related to holding a feminist social identity, especially among younger women, what does a woman with a feminist social identity look like? There is a considerable body of research looking at variables related to feminist identity (e.g., Nelson et al., 2008; Williams & Wittig, 1997). Research has suggested that women who self-label as feminists are more likely to have liberal attitudes and less likely to have conservative attitudes (Liss et al., 2001; Nelson et al., 2008; Peltola, Milkie, & Presser, 2004; Reid & Purcell, 2004). Because, over time, liberal beliefs have become more common in American society (Twenge, 1997), some researchers have begun to look at the role of more radical feminist ideologies in understanding feminists (Nelson et al., 2008). Research has also consistently shown that life experiences, such as personal experience of sexism, play a role in developing a feminist identity (Henderson-King & Stewart, 1997; Nelson et al., 2008; Reid & Purcell, 2004).

These frequently studied attitudes and life experiences are not the only markers of a well-developed feminist identity. Feminists have a sense of common fate with other women (Duncan, 1999; Rowland, 1986). They believe that women need to work together to achieve goals and thus have a sense of gender collectivity (Liss et al., 2001). Feminists also have an awareness of social injustice and believe that the current gender system can be detrimental to women and thus requires change. Young women today, who individually may indeed feel strong and empowered, may not be aware of, or assign different meanings to, continued gender inequity. For example, they may notice that a woman is receiving a lower salary than a man but attribute that to individual competence rather than systemic gender inequity. Thus, these women may not feel any need to work together with other women for social change. This pattern would mean that these women could be high in gender system justification, a specific conservative belief structure in which individuals believe that the current gender system is fair and does not require change (Jost & Kay, 2005).

Additionally, investigating more general constructs of individualism and collectivism may be important. Although individualism and collectivism may appear to be opposite ends of a spectrum, they have more recently been operationalized as distinct dimensions (e.g., Triandis & Gelfand, 1998). Given that feminists within the United States are influenced by the dominant individualistic culture in which they live (Markus & Kitayama, 1991), feminists may be both individualistic and collectivistic. For example, one study found no difference between feminists and nonfeminists on individualism, but feminists were significantly higher in gender collectivity, a type of collectivism associated with gender (Liss et al., 2001). It is unclear whether feminists would also score higher on a more general measure of collectivism.

In sum, the present study had several goals. Our first goal was to determine the extent to which Synthesis would be associated with feminist self-labeling. We then sought to determine what other variables might be associated with feminist self-labeling above and beyond Synthesis in order to better understand what constitutes a well-developed feminist identity. Another goal was to understand which variables we related to Synthesis scores above and beyond feminist self-labeling in order to gain a better understanding of the potential contradictions inherent in this construct. Synthesis may have different meanings for different cohorts of women (Erchull et al., 2009), and we were particularly interested in the meaning of Synthesis for young women who have been raised in the wake of the gains made by second-wave feminists.

In order to achieve these goals we undertook two hierarchical regression analyses: one predicting feminist self-labeling and one predicting Synthesis scores. We hypothesized that Synthesis scores would not be related to feminist self-labeling but, when the rest of the FIC dimensions were entered, Passive Acceptance, Revelation, and Active Commitment scores would be. We predicted that liberal ideologies would not be related to feminist self-labeling because we expected little variability in these scores, but we expected that more radical feminist ideologies would be. Finally, we predicted that gender system justification, gender collectivity, and the perception that one experienced sexist events would be associated with self-labeling.

Our expectations about predicting Synthesis were less specific because these analyses were primarily exploratory. We did, however, propose specific hypotheses about three variables: We believed that feminist self-labeling would not be related to Synthesis scores whereas liberal beliefs and individualism would be. Because the Synthesis items contain a theme of empowered womanhood, liberal beliefs about gender equality may be related to Synthesis. As described above, individualism is also an important theme in these items; therefore, we anticipated that individualism would be positively related to Synthesis scores.



All women 18 to 25 years of age (n = 653) who participated in a larger online study about women's issues were selected as the sample for the present study. Twenty-four women did not provide information about feminist self-labeling, resulting in a final working sample of 629 women. These women were either undergraduates enrolled in introductory psychology courses at a liberal arts college or were recruited using a snowball sampling technique via online listservs, discussion groups, and blogs targeting feminists. The majority of the undergraduate sample self-identified as nonfeminists (84%); therefore, the online sample was specifically recruited to oversample feminists because we were interested in comparing feminists with nonfeminists. Students from the liberal arts college (n = 195) received partial course credit for their participation; participants from the online sample (n = 434) received no compensation.

The average age of participants was 19.25 (SD = 2.38). The undergraduate sample was significantly younger (M = 18.72, SD = 1.12, range 18–25) than the online sample (M = 21.81, SD = 2.16, range 18–25), F(1, 651) = 387.40, p < .001. Thus, we decided to control for age differences throughout our analyses. There were no significant differences between the two recruitment methods in terms of race/ethnicity when comparing White/Caucasian participants to participants of color, χ2(1, N = 618) = 1.01, p = .32. The majority of participants (86.1%) identified as White/Caucasian, 1.5% as African American, 4.7% as Asian/Pacific Islander, 1.9% as Latina, and 5.9% as “other.” The most common self-identified socioeconomic status (SES) was middle class (49.9%), and this did not differ by recruitment method testing middle class versus other SES groups, χ2(1, N = 627) = .05, p = .82. One percent of the sample identified as living in poverty, 16.6% as working class, 31.3% as upper middle class, and 1.3% as wealthy. Overall, the majority of the sample (74.7%) identified as heterosexual, 4.9% as lesbian, 16.4% as bisexual, and 4% as “other.” Online recruitment resulted in significantly more sexual minority participants, χ2(1, N = 629) = 60.91, p < .001; 97% of the undergraduate sample identified as heterosexual in contrast with 65% of the online sample. We did not choose to control for sexual orientation because, in our data, it was conflated with feminist self-labeling, a variable more central to the current investigation. Indeed, only 7 of the 159 (4.4%) sexual minority participants identified as nonfeminists in our sample.

Overall, 411 women (65.3%) self-labeled as feminists and 218 (34.7%) identified as nonfeminists. As intended, significantly more feminists were recruited online, χ2(1, N = 629) = 305.09, p < .001, than through our college recruitment.


Demographics and feminist self-labeling Participants answered basic demographic questions about age, race/ethnicity, sexual orientation, and SES. Feminist self-labeling was assessed by asking the participants to respond yes (coded 1) or no (0) to the question, “Do you consider yourself a feminist?”

FIC This measure is an operationalization of the Downing and Roush (1985) feminist identity development model (Fischer et al., 2000). Passive Acceptance (e.g., “I think that men and women had it better in the 1950s when married women were housewives and their husbands supported them”), Revelation (e.g., “I feel angry when I think about how I'm treated by men and boys”), Embeddedness-Emanation (e.g., “I am very interested in women musicians”), Synthesis (e.g., “I am proud to be a competent woman”), and Active Commitment (e.g., “It is very satisfying to me to be able to use my talents and skills in my work in the women's movement”) were assessed using a 5-point scale ranging from 1 (disagree strongly) to 5 (agree strongly). As support for the validity of this measure, correlations in the expected directions were found for ego identity status, perceived experience of sexist events, and involvement in women's organizations (Fischer et al., 2000). Acceptable Cronbach's alphas from the original investigation provided evidence for the reliability of this measure, although the Cronbach's alpha for Synthesis was low (Fischer et al., 2000). Cronbach's alphas from the original investigation and our study were .75 and .86, respectively, for the Passive Acceptance items, .80 and .71 for the Revelation items, .84 and .77 for the Embeddedness-Emanation items, .68 and .58 for the Synthesis items, and .77 and .86 for the Active Commitment items.

Attitudes Toward Women Scale (AWS) The 25-item short form of the AWS was used to measure liberal gender attitudes (Spence, Helmreich, & Stapp, 1973). A sample item from this measure is “Women should assume their rightful place in business and all the professions along with men.” Participants responded on a 4-point scale ranging from 1 (disagree strongly) to 4 (agree strongly). The original investigation of the short form of the AWS did not include validity or reliability information. In other research using the short form of the AWS, it was related to negative attitudes about women and feminists and yielded a Cronbach's alpha of .79 (Swim & Cohen, 1997). Cronbach's alpha was .90 in the current investigation.

Feminist Perspective Scale (FPS) We used the short form of this scale to examine more radical perspectives of feminism (Henley, Spalding, & Kosta, 2000). We combined the 5-item radical (e.g., “The workplace is organized around men's physical, economic, and sexual repression of women”) and 5-item socialist (e.g., “It is the capitalist system which forces women to be responsible for child care”) subscales into one scale. Participants were asked the extent to which they agreed with statements on a scale ranging from 1 (strongly disagree) to 7 (strongly agree). In the original investigation, both the radical and socialist subscales were correlated with, yet distinct from, liberal attitudes and self-rated degree of feminism. Cronbach's alphas from the radical and socialist subscales (r = .86, p < .001) in the original investigation of the short form were .73 and .59, respectively. Our Cronbach's alpha derived from the combined items was .91 (.85 for the radical and .83 for the socialist items).

Gender System Justification This measure assesses beliefs regarding the fairness and legitimacy of the current gender system (e.g., “In general, relations between men and women are fair”; Jost & Kay, 2005) on a 9-point scale ranging from 1 (strongly disagree) to 9 (strongly agree). The validity of this measure was not specifically assessed in the initial study, but it consisted of items from a more general system justification measure that had been reworded to focus on gender inequality (Kay & Jost, 2003). Correlations between the original measure with belief in a just world and the Protestant work ethic provided evidence for the parent measure's validity. In the original investigation, the Cronbach's alpha coefficient was .65, and it was .88 in the current investigation.

Gender Collectivity Scale This scale assesses women's attitudes about the need for collective action with other women (e.g., “Women need to work together in order to create an equal society”; Liss et al., 2001). Reponses to the 11 items were on a scale ranging from 1 (disagree strongly) to 7 (agree strongly). Correlations in the expected direction with other measures of feminist beliefs and attitudes provided evidence for the validity of this measure in the original investigation. Cronbach's alpha for these items from the original investigation was .81, and it was .86 in the current investigation.

The Schedule of Sexist Events The lifetime subscale of this measure was used to assess personal experience of sexism (Klonoff & Landrine, 1995). It asks individuals about the frequency with which various sexist events have occurred throughout their entire lives (e.g., “How many times have you heard people making sexist jokes, or degrading sexual jokes?”). Participants responded to each of 20 items on a scale ranging from 1 (never) to 6 (almost all the time). Correlations with other measures of stressful life events (Klonoff & Landrine, 1995) provided evidence for the validity of this measure. The Cronbach's alpha derived from these items in the original investigation was .92; it was .94 in the current study.

Individualism and Collectivism Scale This scale measures individualistic and collectivistic orientations on a 9-point scale ranging from 1 (strongly disagree) to 9 (strongly agree) (Triandis & Gelfand, 1998). The horizontal subscales, which emphasize relationships to people of similar levels of rank and power, were used in the present study to measure individualism (e.g., “Being a unique individual is important to me”) and collectivism (e.g., “I feel good when I cooperate with others”). Support for both convergent and divergent validity was provided through correlations in the expected directions with other established measures of individualism and collectivism (Triandis & Gelfand, 1998). The Cronbach's alphas for these two subscales were .67 for individualism and .74 for collectivism for the original investigation; they were .71 and .72, respectively, for the current investigation.


Participants attending the liberal arts college completed the survey in computer labs where dividers were placed between each computer to ensure privacy. Participants from the internet sample followed links posted on various feminist listservs, discussion groups, and blogs and completed the questionnaire at their convenience. A wide variety of these sources were used, most targeting young women (e.g., college age and 20s). Suggestions for possible online recruiting locations were solicited from feminist colleagues, students, and friends. All suggested locations were queried, and moderators and bloggers were contacted and asked to make a brief description of our study and a link available on their Web site.

After reading an online informed consent in which they were told the study concerned women's beliefs and experiences, all participants were prompted to enter their e-mail address so a secure link to the survey could be provided. E-mail addresses were not linked to participants' data. The questionnaire took approximately 30 minutes to complete; participants were then taken to an online debriefing page.


Descriptive Statistics and Group Differences

Descriptive statistics separated by group (feminist and nonfeminist) are presented in Table 1. Overall, scores on Embeddedness-Emanation, Synthesis, and Active Commitment were relatively high; means for both groups were above the midpoints. AWS scores were high for both groups (higher scores indicated more liberal attitudes). Gender collectivity, individualism, and collectivism were also generally high, with average scores for both groups above the midpoint.

Table 1. 
Descriptive Statistics and Group Differences for Feminists and Nonfeminists
 Possible RangeFeminists M (SD)Nonfeminists M (SD)F (1, 597)pη2
  1. FPS = Feminist Perspective Scale.

Passive Acceptance1–51.52 (.50)2.86 (.70)414.42<.001.41
Revelation1–53.21 (.65)2.75 (.75)56.45<.001.09
Embeddedness-Emanation1–54.22 (.66)3.33 (.74)146.21<.001.20
Synthesis1–54.11 (.56)3.97 (.60)1.21 .27.00
Active Commitment1–54.28 (.50)3.34 (.57)246.14<.001.29
Attitudes Toward Women1–43.85 (.17)3.36 (.37)268.52<.001.31
FPS Radical/Socialist1–7 4.57 (1.14) 2.64 (1.09)233.52<.001.28
Gender System Justification1–9 2.98 (1.08) 5.53 (1.21)403.52<.001.40
Gender Collectivity1–75.84 (.65)4.98 (.80)131.49<.001.18
Schedule of Sexist Events20–120 53.21 (14.55) 39.43 (11.50)61.60<.001.09
Individualism1–9 7.31 (1.04) 6.98 (1.07)5.77 .02.01
Collectivism1–9 6.29 (1.00)6.68 (.94)8.83 .003.02

A multivariate analysis of covariance (MANCOVA) was conducted to test for differences between the groups on all study variables controlling for age. Levene's Test of homogeneity of variance indicated that five variables violated this assumption: Passive Acceptance, F(1, 598) = 31.12, p < .001; Revelation, F(1, 598) = 7.15, p = .008; Attitudes Toward Women, F(1, 598) = 160.74, p < .001; Gender Collectivity, F(1, 598) = 4.68, p = .03; and the Schedule of Sexist Events, F(1, 598) = 13.34, p < .001. These results indicated that variances of these dependent variables differed between feminist and nonfeminist identified women. For all of these variables, except for the Schedule of Sexist Events, feminists were more similar to each other than were nonfeminists. This makes sense because feminists typically share similar views about women, which may lead them to take on this label. Nonfeminists are a more heterogeneous group in terms of their views of women's issues. Additionally, this analysis violated the assumption of homogeneity of variance-covariance matrices, Box's M = 336.01, p < .001. Although MANOVA is robust to violations of these assumptions, statistically significant group differences should be interpreted with caution.

The MANCOVA was significant, F(12, 586) = 56.54, p < .001, η2= .54 (results based on untransformed data are reported because they were unchanged when transformations to correct for homogeneity violations were used). Univariate analyses of covariance, controlling for age, indicated that the two groups were different on all variables except for Synthesis scores. Results for all of the gender-related variables were in the expected direction (i.e., feminists expressed more liberal attitudes, had lower Passive Acceptance scores, and reported greater experience with sexism). Nonfeminists had lower individualism and higher collectivism scores than did feminists (see Table 1). It should be noted that a number of our effect sizes were in the very small range. For example, the effect sizes for individualism and collectivism were .01 and .02, respectively.

Predicting Feminist Self-Labeling

Correlations among all variables can be seen in Table 2. A hierarchical binary logistic regression analysis was conducted to predict feminist self-labeling, which was coded so that 0 represented a nonfeminist self-labeling and 1 represented feminist self-labeling (see Table 3). In a logistic regression, one estimates the odds of being placed in one of two groups: feminist or nonfeminist in the present analysis. The overall fit of the model is assessed with a chi-square goodness of fit test, and changes in fit with the addition of variables in subsequent steps can be assessed with a chi-square change test. Additionally, we report pseudo-R2 values, which allow researchers to draw conclusions about the overall strength of the model, although they are calculated differently from a traditional R2 value in a linear regression.

Table 2. 
Intercorrelations Among Study Variables
    1   2   3   4   5   6   7   8   9   10   11   121314
  1. Note. n = 600; feminist self-labeling was coded such that 0 = nonfeminist and 1 = feminist. FPS = Feminist Perspective Scale.

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

 1. Age –             
 2. Feminist Labeling .55*** –            
 3. Passive Acceptance−.53***−.74*** –           
 4. Revelation .11** .30***−.25*** –          
 5. Embeddedness-Emanation .31*** .52***−.49*** .36*** –         
 6. Synthesis .14** .14**−.03−.04 .26*** –        
 7. Active Commitment .45*** .65***−.62*** .31*** .65*** .35*** –       
 8. Attitudes Toward Women .49*** .67***−.80*** .25*** .52*** .16*** .64*** –      
 9. FPS Radical/Socialist .42*** .63***−.61*** .52*** .57*** .08* .62*** .54*** –     
10. Gender System Justification−.50***−.73*** .77***−.40***−.55***−.02−.62***−.68***−.75*** –    
11. Gender Collectivity .29*** .50***−.39*** .42*** .61***−29*** .67*** .46*** .53***−.48*** –   
12. Schedule of Sexist Events .36*** .43***−.46*** .44*** .44*** .11** .51*** .40*** .60***−.59*** .41*** –  
13. Individualism .12** .15***−.21*** .07 .15*** .22*** .24*** .23*** .11**−.11** .08 .19*** 
14. Collectivism−.16***−.19*** .29***−.09*−.003 .16*** .04−.16***−.19*** .29*** .22***−.13**.01
Table 3. 
Hierarchical Logistic Regression Analysis Predicting Feminist Self-Labeling
VariableStep 1 B (S.E.)Step 2 B (S.E.)Step 3 B (S.E.)Step 4 B (S.E.)Step 5 B (S.E.)
  1. Note. N = 600; chi-square values reported in the table represent values from the omnibus goodness of fit test. FPS = Feminist Perspective Scale.

  2. *p < .05. **p < .01. ***p <.001 based on Wald statistics.

Age.72 (.06)***.71 (.06)***      .32 (.08)***      .27 (.08)**      .25 (.09)**
Synthesis .15 (.18)     −.32 (.31)     −.32 (.32)     −.25 (.35)
Passive Acceptance      −2.11 (.28)***    −1.70 (.33)***    −1.47 (.39)***
Revelation        .59 (.24)*      .29 (.27)      .16 (.30)
Embeddedness-Emanation        .39 (.27)      .26 (.29)      .08 (.32)
Active Commitment       1.41 (.36)***     1.14 (.38)**      .99 (.45)*
Attitudes Toward Women         .93 (.76)      .39 (.83)
FPS Radical/Socialist         .49 (.16)**      .25 (.19)
Gender System Justification         −.75 (.21)***
Gender Collectivity          .97 (.33)**
Schedule of Sexist Events         −.05 (.02)**
Individualism         −.004 (.17)
Collectivism         −.26 (.21)
χ2(1) = 216.84***χ2(2) = 217.54***χ2(6) = 490.77***χ2(8) = 501.80***χ2(13) = 530.03***
Cox & Snell R2= .30Cox & Snell R2= .30 Cox & Snell R2= .56 Cox & Snell R2= .57 Cox & Snell R2= .59
Nagelkerke R2= .42Nagelkerke R2= .42Nagelkerke R2= .77Nagelkerke R2= .78Nagelkerke R2= .81
 inline imageinline imageinline imageinline image

As indicated in the diagnostics for the parallel linear regression analysis presented below, we did not deem multicollinearity to be problematic. In order to control for age in all further analyses, we entered age in the first step of the model. At this step, age was related to feminist self-labeling and correctly classified 78.8% of participants. We added Synthesis at our second step because our goal was to determine the extent to which Synthesis was uniquely associated with feminist self-labeling above and beyond other variables. Synthesis did not enhance the ability of the model to classify women as feminists or nonfeminists. At the third step, we entered all other FIC variables to see the extent to which these other dimensions of feminist identity development were related to feminist self-labeling. The overall model was improved at this step, and the model was significant. This step correctly classified 92.3% of the participants: 93.4% of feminists and 90.2% of nonfeminists.

In the fourth step, feminist beliefs (AWS scores and the radical/socialist composite score from the FPS) were entered to assess the extent to which more commonly assessed attitude variables aided in the understanding of feminist self-labeling above and beyond the prior variables. The change between models was again significant as was the overall model, given the increased ability to explain the variance in self-labeling. This step, however, had identical rates of correctly classifying participants as compared to step three.

In the final step, five additional variables were entered, Gender System Justification, Gender Collectivity, Schedule of Sexist Events, Individualism, and Collectivism, to assess the extent to which these less traditionally studied variables aided in the understanding of feminist self-labeling above and beyond those variables more commonly studied and entered in prior steps. The model was improved with the addition of these variables, and significantly more variance was explained. At the final step, 92.5% of the participants were correctly classified. This small improvement in classification was due to the improved ability to correctly classify feminists (94.2%), although the ability to correctly classify nonfeminists (89.2%) declined.

Predicting Synthesis Scores

Correlations among all variables are presented in Table 2. Synthesis scores were positively correlated with feminist self-labeling such that feminists had higher Synthesis scores. Synthesis was unrelated to Passive Acceptance and Revelation. It was positively correlated with Embeddedness-Emanation and Active Commitment. Synthesis scores were correlated with AWS scores such that higher Synthesis scores were related to holding more liberal attitudes. Synthesis scores were also related to having more radical/socialist feminist beliefs; however, this was an extremely small effect. Women high in Synthesis were more likely to report having experienced sexist events. Synthesis scores were positively correlated with both individualism and collectivism; interestingly, Synthesis was negatively correlated with gender collectivity. Synthesis was uncorrelated with gender system justification.

A hierarchical linear regression analysis was conducted to predict Synthesis scores (see Table 4). Variables were entered in steps parallel to the above logistic regression analysis with the exception of step two, where feminist self-labeling replaced Synthesis scores. Given the strength of the correlations among some variables, we tested for multicollinearity using the variance inflation factor (VIF). The largest VIF, for Passive Acceptance, was 4.5, well below the value of 10, suggested as a level for concern by Myers (1990). Thus, we did not consider multicollinearity to be a concern for our analyses.

Table 4. 
Hierarchical Linear Regression Analysis Predicting Synthesis Scores
VariableStep 1Step 2Step 3Step 4Step 5
  1. Note. N = 600; standardized regression coefficients are reported in table; feminist self-labeling was coded such that 0 = nonfeminist and 1 = feminist. FPS = Feminist Perspective Scale.

  2. *p < .05. **p < .01. ***p < .001 based on t-tests.

Age.14** .11*        .08        .08        .08
Feminist Self-Labeling .05       −.04       −.03        .002
Passive Acceptance          .33***        .42***        .34***
Revelation         −.16***       −.14**       −.16***
Embeddedness-Emanation          .17**        .16**        .14**
Active Commitment          .48***        .47***        .37***
Attitudes Toward Women           .18**        .15*
FPS Radical/Socialist          −.08       −.03
Gender System Justification            .19**
Gender Collectivity            .15**
Schedule of Sexist Events            .04
Individualism            .15***
Collectivism           −.02
R2= .02R2= .02R2= .21R2= .23R2= .27
Adjusted R2= .02Adjusted R2= .02Adjusted R2= .21Adjusted R2= .22Adjusted R2= .25
F(1, 598) = 11.81**F(2, 597) = 6.51**F(6, 593) = 26.74***F(8, 591) = 21.59***F(13, 586) = 16.28***
 inline imageinline imageinline imageinline image
 FΔ(1, 597) = 1.21FΔ(4, 593) = 36.09***FΔ(2, 591) = 5.03**FΔ(5, 586) = 6.25***

At step one, age was significantly related to Synthesis and explained 2% of the variance. Feminist self-labeling did not contribute additional unique variance in Synthesis scores at step two. The remaining four FIC dimensions were entered in step three and represented a significant improvement, explaining an additional 19% of the variance. At this step, age was no longer significantly associated with Synthesis. Feminist beliefs (AWS scores and radical/socialist beliefs from the FPS) were added in the fourth step, and these variables significantly accounted for an additional 1% of the variance. We added Gender System Justification, Gender Collectivity, reported experience of sexist events, Individualism, and Collectivism in the final step, and an additional 4% of the variance was explained.


The goals of this study were to better understand what is associated with feminist self-labeling and whether or not the Synthesis stage of the feminist identity development model captures these associations. We also hoped to understand what measures the Synthesis stage is assessing because we believed that Synthesis may represent a construct that is distinct from feminism, at least among young women. As we hypothesized, Synthesis was not related to feminist self-labeling. Furthermore, our results were consistent with previous research that has pointed to the problematic nature of Synthesis as a construct (Liss, Crawford, & Popp, 2004; Moradi et al., 2002). Although the stage was originally conceptualized as one where women self-labeled as feminists (Downing & Roush, 1985), Synthesis scores, at least as operationalized in the FIC, were not useful in classifying women as feminists or nonfeminists.

Given that Synthesis was not related to feminist self-labeling, we wanted to understand what constructs were associated with taking on this label. Consistent with our hypotheses, Passive Acceptance, Revelation, and Active Commitment were all uniquely associated with feminist self-labeling when first entered into the equation. When feminist beliefs were entered, radical/socialist feminist beliefs were uniquely associated with feminist self-labeling, but liberal beliefs were not, which was consistent with our hypothesis. At this step, Passive Acceptance and Active Commitment remained related to self-labeling whereas Revelation did not. Consistent with our hypotheses, Gender System Justification, Gender Collectivity, and the perceived experience of sexist events were all uniquely related to feminist self-labeling at the final step of the model. In the final model, Passive Acceptance and Active Commitment remained associated with self-labeling, but radical/socialist beliefs were no longer related.

It is interesting to note that the first and last stages (i.e., Passive Acceptance and Active Commitment) of the Downing and Roush (1985) model remained uniquely associated with feminist self-labeling throughout every step of the regression analysis. There may well be a developmental aspect to feminist self-labeling that these presumably endpoint stages are assessing. On the other hand, Passive Acceptance and Active Commitment may not represent developmental processes at all; rather, they may be capturing opposite ends of a single continuum of feminist attitudes. Regardless, the middle stages of the model appear less useful in differentiating young feminists and nonfeminists, and more research is needed to understand how young women move from Passive Acceptance to Active Commitment, if indeed there is a developmental progression.

Additionally, these feminists were less likely to see the current gender system as equitable. They were also more likely to value working collectively together with other women (also see Liss et al., 2001); however, they did not endorse overall collectivism. Feminists were also more likely to report having experienced sexist events, a finding that is consistent with previous research (e.g., Henderson-King & Stewart, 1997; Nelson et al., 2008; Reid & Purcell, 2004).

Overall, these young feminists were more likely to believe that there are problems facing women in society and that women need to work together to fix them. This underlying idea appeared more central to feminist self-labeling than belief measures that assessed specific ideological branches of feminism (e.g., liberal and radical/socialist beliefs). Previous research examining correlates of feminist self-labeling has generally focused on radical, liberal, and conservative ideologies (e.g., Liss et al., 2001; Nelson et al., 2008). The present study suggests that feminist self-labeling may be better understood by focusing on beliefs about the fairness of the current gender system and the need to work collectively for change. This conclusion is consistent with the ideas of theorists who have claimed that feminists are “united by a belief that unequal and inferior social status of women is unjust and needs to be changed” (Jaggar, 1983, p. 322).

Theoretically, women high in Synthesis should share the qualities detailed above as this stage or dimension was understood to be when women accepted the feminist label and developed a feminist identity (Downing & Roush, 1985). Given that Synthesis was unrelated to feminist self-labeling in our data, an important goal was to better understand Synthesis as a construct. Synthesis, as it is measured by the FIC, represents a seemingly strange combination of empowerment, individualism, and comfort with one's own femininity. Our analyses were primarily exploratory, although we did have hypotheses about three variables included in the regression equation. Consistent with our hypotheses, Synthesis scores were not uniquely associated with feminist self-labeling, but they were uniquely associated with liberal beliefs and individualism.

Additionally, the four remaining dimensions of the feminist identity development model were uniquely associated with Synthesis, but in curious directions. On the one hand, Synthesis scores were associated with greater Passive Acceptance and lower Revelation scores, indicating that Synthesis includes a conservative dimension consistent with previous research (Liss et al., 2001). It should be noted though that at the bivariate level there was no relationship between Passive Acceptance and Synthesis. Thus, the significant relationship in the regression equation may indicate that the association between Passive Acceptance and Synthesis is only evident when the variance associated with the other FIC variables is removed. On the other hand, Synthesis was associated with higher Embeddedness-Emanation and Active Commitment scores, indicating that women high in Synthesis feel connected to other women and believe that women should work together for women's rights. In sum, Synthesis is a confusing construct in relation to the rest of the feminist identity development model.

Synthesis was also uniquely related to a mix of feminist and nonfeminist beliefs outside the FIC dimensions. In addition to liberal attitudes mentioned above, Gender Collectivity was related to Synthesis. On the other hand, women high in Synthesis justified the current gender system. Finally, as noted above, women high in Synthesis were more individualistic.

The construct of Synthesis is a paradox among young women. It appears to represent a state where women feel empowered but remain unaware of social injustice. These women feel strong and able to achieve life's goals but retain an acceptance of traditional gender roles, perhaps because of their lack of awareness of continued gender inequity. Thus, they do not see the necessity to change the current gender status quo as do self-labeled feminists. Women high in Synthesis do hold a pro-woman stance (represented by high scores in Embeddedness-Emanation). However, this pro-woman stance may be related to valuing femininity because items on the FIC Synthesis scale specifically mention comfort with one's femininity. Future research should explore this possibility more closely, perhaps by looking at the construct of cultural feminism (Henley, Meng, O'Brien, McCarthy, & Sockloskie, 1998). Cultural feminism represents a stance in which traditionally feminine roles, such as mother, are seen as valuable and important (Budig, 2004). In fact, one study found that Synthesis was positively correlated with cultural feminism but not other types of feminist ideologies among lesbian and bisexual women (Szymanski, 2004).

Overall, Synthesis scores were generally high, a finding that is consistent with previous research (Erchull et al., 2009; Liss et al., 2001). Respondents' high level of agreement with this construct indicates that it is a commonly endorsed position among young women today and, as such, is worth understanding. For example, Synthesis may serve as a starting point for feminist identity development rather than as an ending point (Erchull et al., 2009). If Synthesis is a starting point, the entire feminist identity development model may need to be reconceptualized. Although some young women may start in Passive Acceptance, it may be more common for women to start in a stage where they feel as though they have infinite opportunities but are unaware of continued gender discrimination. Young women in this stage may still experience Revelation where they realize that gender discrimination continues to exist. However, Revelation may be more likely to be initiated by an educational experience (e.g., a women's studies class) than through a personal experience of discrimination (Erchull et al., 2009). Whether young women today experience a stage similar to Embeddedness-Emanation remains an open question because women-focused consciousness-raising groups no longer hold a central place in the women's movement. How an experience of Revelation ultimately translates into a feminist identity and feminist activism needs to be further understood. Given that much feminist organization happens through the Internet, rather than through marches and rallies, and the feminist movement is generally less centralized than it was at the height of the second wave (Taylor & Whittier, 1997), the stage of Active Commitment also merits reconceptualization.

Above and beyond our suggestions for recasting the Downing and Roush (1985) model, we recognize that the utility of a developmental model of feminist identity itself has been questioned (e.g., Gertsmann & Kramer, 1997; Moradi & Subich, 2002). There are three potential problems with understanding feminist identity through a stage model. The first, as we have suggested above, is that Downing and Roush's (1985) model may be outdated and a new model needs to be developed. It is also possible that Downing and Roush's (1985) model remains relevant, but that the operationalizations do not adequately assess the stages (e.g., Moradi & Subich, 2002; Szymanski, 2004). Finally, it is possible that a stage model of feminist identity development is an inadequate way to understand feminism because women may follow many different, idiosyncratic paths toward a feminist identity.

There were several limitations to the current study. Although our sample was strengthened by including a large number of feminists, our convenience sampling does not ensure that it is representative of young women in the United States. In addition to our oversampling of self-labeled feminists, our sample was largely Caucasian and middle class. Thus, our results should be generalized with caution to more diverse populations. It is possible that the phenomenon of women who are empowered, yet remain unaware of continued gender inequity, is specific to young, middle-class, White women. Furthermore, the majority of our feminists were recruited online, and thus, recruitment source was confounded with feminist self-labeling. Whereas a goal of our study was to recruit feminist participants, the feminists in the sample were likely more invested in the topic than were the nonfeminists, who were largely recruited through an on-campus subject pool. It should also be noted that our feminists were, on average, 4 years older than our nonfeminists, although we controlled for age throughout our analyses. Finally, the feminists in our sample were considerably more likely to be lesbian or bisexual than were the nonfeminists. Prior work has suggested that feminism is a particularly useful ideology for sexual minority women because it provides a framework that might better allow them to cope with heterosexism (Szymanski & Chung, 2003). Thus, our conclusions should be interpreted in light of the fact that feminist self-labeling was confounded with sexual orientation.

It is also important to note that the reliability of the Synthesis subscale was low in this study (.58) as in other published studies (e.g., .67 in Erchull et al., 2009, and .68 in Fischer et al., 2000). In our sample, one particular item, “As I have grown in my beliefs I have realized that it is more important to value women as individuals than as members of a larger group of women,” was uncorrelated with all but one of the other items in the subscale. When reliability was re-calculated without this item, it was improved to .69, a level comparable to previous research. All analyses were run using the entire Synthesis scale so that our results could be compared to previous research. Future researchers may want to investigate the utility of including this item, although our results were essentially unchanged when this item was excluded. In general, the reliability of the Synthesis scale has been low, so researchers should be particularly cautious about using it. It is also possible that the construct of Synthesis actually represents multiple ideas (e.g., empowerment and comfort with one's own femininity). The fact that these potentially distinct ideas are assessed in a single measure may contribute to its low reliability.

It should be noted that we measured feminist self-labeling with a dichotomous, forced-choice measure. We believe this measure provides a clear indication of self-described identity in the same way asking someone about their political affiliation would. However, there are many other ways to measure feminist self-labeling, with both single- and multiple-item measures (e.g., Williams & Wittig, 1997; Zucker, 2004). Nevertheless, we strongly believe that a yes/no self-labeling measure results in the least ambiguous data that best allow researchers to differentiate people who are and are not willing to take on the feminist label.

Our data suggest that almost all young women feel empowered, but self-labeled feminists are most likely to see the need for continued societal change. Why would women join a movement designed to create more opportunities for women when they feel that they have all the opportunities they need already? Consequently, feminists need to remain focused on raising women's awareness of continued gender inequity in order to motivate young women to believe that work still needs to be done. Empowerment, although positive, can lead to complacency. Challenging this complacency is an important goal for the future of the women's movement.