Introduction

One major consequence of the restructuring of the world economy that began in the 1980’s and 1990’s—or globalization—is that it maintains or even exacerbates violations of women’s human rights. Consequences of this economic and social restructuring include the feminization of labor and/or poverty whereby women have expanded their unpaid labor to compensate for the increase in poverty and loss of local resources, resulting in an exacerbated sexual division of labor (Naples and Desai 2002). The current global economic crisis has worsened this situation with a projected 200 million new working poor earning less than two dollars a day, the majority of whom are women (ODI 2010). As processes of globalization continue to intensify, women’s human rights and empowerment become ever more relevant.

Empowerment is a key concept in both community psychology and international development (Kabeer 2005; Rappaport 1987). Within each discipline, empowerment is central to the work of enhancing well-being and improving human lives (Zimmerman 1995; Sen 1999). Across disciplines, it is widely agreed upon that empowerment processes encompass material resources and inequities in the environment, strengths of the individual and a sense of personal control, and the enhancement of well-being (Cattaneo and Chapman 2010; Zimmerman 1995).

In recent decades the international development field has begun a concerted effort to broadly address women’s empowerment with organizations ranging from the World Bank, to grassroots non-governmental organizations, to the United Nations (UN), all advocating for women’s human rights and empowerment in an effort to address UN Millennium Development Goal 3—“to promote gender equality and empower women.” Yet, what remains unclear in this growing global awareness is what is meant by women’s empowerment. When policymakers and interventionists include empowerment as a development goal, what are they trying to accomplish and how do they determine whether it has been achieved (Malhotra and Schuler 2005; Mosedale 2005)? The gap in empirical support for empowerment processes in this context may stem from the fact that investigation in this area does not have a recognized niche in any one academic field. Several reviews have demonstrated that most investigations surrounding women’s empowerment have emerged from economics, demography, sociology, and public health (e.g., Mosedale 2005; Narayan 2005). Although there are long-established models of empowerment theory within community psychology, psychologists have yet to apply them to the investigation of women’s empowerment in the context of globalization. Thus, despite the proliferation of empowerment rhetoric in the championing of global social interventions, the explicit connections between empowerment research and program development are in many cases tenuous (Perkins 1995). A transdisciplinary analysis, whereby an integration of perspectives and methods drawn from both disciplines is necessary to close this chasm (Christens and Perkins 2008).

Why, to date, have we not begun to more accurately conceptualize and investigate the processes surrounding women’s empowerment in a globalized context (Christens and Perkins 2008; Mosedale 2005; Narayan 2005)? In the 1970’s women’s empowerment was first invoked to facilitate a struggle for social justice and women’s equality through a transformation of political structures (Mosedale 2005). However, by the 1990’s many agencies began using the term empowerment in association with a wide variety of strategies in isolation from women-centered agendas and applied it, instead, to the context of broad-based neo-liberal economic development strategies (Perkins 1995). In the service of one-size-fits-all development models, buzzwords like “empowerment” have been spun into an apoliticized form with a nearly unimpeachable moral authority (Cornwall and Brock 2005). In this manner, international agencies have appropriated concepts once used by progressive social movements. The popularity and subsequent ambiguity in the use of the term “empowerment” has created an even greater need for assessment in the applied context (Perkins 1995).

The current study outlines a comprehensive approach to measuring and analyzing components involved in processes of women’s empowerment in the context of international development. This study is the first step toward building communication between psychologists and development researchers. It aims to develop a shared framework bridging the gap between disciplines so that we may deepen our understanding of empowerment by testing a model that can inform practice and policy-relevant guidelines. What follows is a discussion of the conceptualization of empowerment as well as a detailed review of the three main components that comprise the theoretical model of empowerment being tested in this paper: structural inequities, agency, and well-being outcomes (see Fig. 1).

Fig. 1
figure 1

The hypothesized model. Land ownership and organizational participation are hypothesized to predict gender role ideology, which in turn, is hypothesized to influence decision-making, relationship power and control, which are expected to predict women’s agency, and, finally, women’s subjective well-being

Full size image

Conceptualization of Empowerment

Similar conceptual definitions of empowerment are offered from the fields of psychology and development. Within psychology empowerment has been defined as a sense of personal control and freedom, whereby individuals’ gain agency and mastery over issues of concern to them and are supported by access to and control over resources (Rappaport 1987; Zimmerman 1990, 1995). Empowerment theory also explicitly links subjective well-being with larger social and political contexts and integrates a critical understanding of the sociopolitical environment (Perkins and Zimmerman 1995; Zimmerman 1995). Similarly, in the international development literature, empowerment has largely come to refer to the expansion of freedom of choice and action to shape one’s life, yet it is recognized that for many marginalized groups that freedom is severely curtailed by lack of opportunity determined by structural inequities (see Mosedale 2005; Narayan 2005, for reviews). In perhaps the most widely used definition of empowerment in the development literature, Kabeer describes empowerment as “a process of change during which those who have been denied the ability to make strategic life choices acquire such an ability” (Kabeer 1999, p. 437). According to Kabeer, three interrelated components—resources, agency, and achievements—are critical to understanding the concept and process of empowerment (1999). Resources may be construed as material or social and are considered the medium through which agency, or the ability to exercise choice, is carried out. Achievements refer to the outcomes of agency. This approach is set apart from a singular focus on agency and highlights the profound importance of social context in empowerment processes.

Although an abundant literature suggests that empowerment is a process whereby multiple components influence each other, much empirical research does not identify the multiple components or the links among them (Cattaneo and Chapman 2010; Kabeer 1999; Zimmerman 1995). Early conceptualizations and investigations of empowerment within psychology focused primarily on individual psychological components, such as perceptions of personal control, thereby giving limited attention to context and social structures (Perkins 1995; Riger 1993). The current study extends the investigation of empowerment out of the individual psychological realm by identifying various components of empowerment and specifying relationships among them in a manner that enables the formulation of a measurement model.

Structural Inequities: Resources and Ideology

Resources

Structural power, whereby dominant individuals have more control over resources than subordinates, is one of the major contributors to social inequalities. A growing body of evidence suggests that institutionalized inequities in the distribution of resources contribute to power imbalances and gender-based norms that create an environment that legitimizes and perpetuates women’s subordinate status (e.g., Connell 1987; Glick and Fiske 1999). Because the economic policies of the 1980’s and 1990’s introduced or exacerbated several structural factors that have contributed to rising levels of gender inequity, rapidly changing conditions in the restructuring of resources in “developingFootnote 1” countries provide a perfect context for an analysis of social structures that reflect dominance and power (Grabe 2010a; Naples and Desai 2002). Inequities have been especially visible within the area of property rights, with pervasive gender inequities in land ownership, in particular, being recognized as a violation of women’s human rights (Deere and Leon 2001; Pena et al. 2008). In the current study, land ownership will be examined as a sociocultural inequity that perpetuates the unequal distribution of power.

In recent years, a small body of literature has emerged examining processes surrounding women’s land ownership. In the first published study in this area, authors found that in Kerala, India women’s receipt of long-term physical violence was related to owning land, a house, or both (Panda and Agarwal 2005). Since that initial publication, investigators have expanded on this research and demonstrated links between property ownership and women’s negotiating power within the marital relationship, financial decision-making, and receipt of physical and sexual violence in West Bengal, Nepal, and Nicaragua, respectively (Grabe 2010b; ICRW 2006; Pandey 2010). Collectively, these studies put forth a framework for investigating land ownership as a potential resource that may be related to women’s well-being.

Although in many countries women’s property rights have improved (FAO 2004), deeply entrenched social barriers still prohibit women from taking advantage of opportunities to effectively exercise their right to own property (Narayan 2005). In this manner, women’s organizations and social movements, where women can collectively address injustices, have an important role to play in creating the conditions for change (Kabeer 1999). Specifically, women’s organizations have emerged in Nicaragua, and throughout Latin America, to assist women in obtaining formal titles to land. Although strategic gender interests may be met through women’s land rights, it also has been argued that ownership is meaningless if women are not also assisted in utilizing their assets (e.g., workshops on farming) and made aware of their rights through some level of organizing (Pena et al. 2008). As Freire (1970) suggested, collective organizing and raising awareness of one’s own social reality is a vital means to initiating action and creating social change. Thus, although some evidence exists for the direct relation between land and well-being, it is hypothesized that greater levels of participation in an advocacy organization might have more impact than land ownership alone. As such, this study examines land ownership as facilitated by an organization that viewed women’s ownership and control over land as a significant advancement in women’s rights and empowerment.

Ideology

Although resources may provide the material conditions through which gender inequities are produced, cultural ideology i.e., social rules, norms, and values that govern gender roles, plays a critical role in how they are sustained (Glick and Fiske 1999). Because of cultural ideologies surrounding gender, women’s lack of access to institutional resources are a central locus of disempowerment in a way that is not true for other disadvantaged groups (Malhotra and Schuler 2005). Thus, considering women, as a group, involves questions about what women have in common that warrants the interest in women’s empowerment, in particular. According to Social Dominance Theory, ideologies are functional and serve to promote or maintain group inequality (Pratto et al. 1994). Thus, women’s empowerment is likely to involve an awareness of the sociopolitical environment that can lead to an emergence of new beliefs about the right to exercise capabilities and take advantage of opportunities in one’s community. As such, this study examines how land ownership and organizational participation relate to gender ideology. It is believed that a combination of land ownership, organizational participation, and more progressive ideology, will relate to a greater sense of women’s agency.

Agency

Agency, the second component involved in empowerment processes, is defined across disciplines as the capacity of actors to define their own goals and to take purposeful action, a function of both individual and structural opportunities (Bandura 2006; Kabeer 1999; Sen 1999). Although work in community psychology highlights the importance of expanding our understanding of empowerment beyond the individual level of analysis, the challenge for researchers is not to ignore one level of analysis in the interest of another, but to integrate levels of analysis, and how they relate to each other, for an understanding of empowerment in its entirety (Zimmerman 1990). Indeed, psychologists argue that while external conditions are necessary for empowerment, material resources alone do not inevitably lead to empowering processes if people do not have internal feelings of competence (Diener and Biswas-Diener 2005).

Although including individual level variables in empowerment research is not itself limiting, an overemphasis on individual agency in the investigation of women’s empowerment has become problematic for two primary reasons. First, within the international development literature, several reviews have found that the vast majority of research in this area has used a narrow and limited range of indicators of empowerment by routinely assessing household decision-making as the primary indicator of women’s empowerment because it is meant to reflect agency (Hill 2003; Malhotra and Schuler 2005). However, because women’s decisions tend to fall in the areas of household consumption and decisions related to child health, evidence that women play a role in making decisions based on pre-existing gender roles tells us little about their ability to define their own goals and take purposeful action (Kabeer 1999). In addition, a lack of contextual analysis does not allow an understanding that defines the ways that agency is arrived at or, alternatively, what outcomes agency may influence. Secondly, many authors argue that, in addition to individual agency, the investigation of women’s empowerment requires an analysis of gender relations (i.e., the ways in which power relations between women and men are constructed and maintained; Malhotra and Schuler 2005; Riger 1993). However, the limited number of studies that have attempted to measure relationship power focus rather on proxies such as relative education and income levels (e.g., Hill 2003). Indirect measures of women’s marital power overlook the elements of empowerment that are psychological in nature or that index how power operates in the relationship (Greig and Koopman 2003). Moreover, the absence of a dyadic assessment leaves unanswered questions surrounding how gender ideology and power relations in the home may relate to women’s individual agency.

In methodological terms there exists a need for greater care in selecting and quantifying components of the empowerment process that index agency and fit within a larger, contextual understanding of empowerment. This study will include dyadic measures of gender relations (decision-making and relationship power and control) as well as women’s individual autonomy and mastery. The more commonly used decision-making scales from the development literature will be compared to the additional measures included in this study. A comprehensive model of empowerment would explain not only women’s ability to act with agency, but also how women’s belief surrounding their capacity to exert control over their lives relates to well-being.

Outcomes

Although empowerment is conceptualized as an iterative process whereby relationships between components are likely reciprocal, there is also evidence to suggest that there are pathways to empowerment—in other words, changes in resources and agency—that may lead to positive outcomes (Cattaneo and Chapman, 2010; Kabeer 1999). Specifically, there is a wealth of evidence in psychology that individuals’ beliefs in their abilities to exercise control over events that affect their lives is related to human achievements and well-being outcomes, in particular to one’s experience of self-worth and depression (Bandura 1989). It is also reasonable to suggest that women’s perceived autonomy and competence may be related to lower levels of coercion or intimidation from an intimate partner, thereby reflecting a healthier psychosocial environment.

Within psychology, empowerment theory identifies subjective well-being as a fundamental part of empowerment (Zimmerman 1990). Therefore it is argued that well-being indicators, rather than more traditional outcomes measures (e.g., education or income levels; Hill 2003; Malik and Lindahl 1998), are critical to assess as components of empowerment. Because subjective well-being heightens individuals’ probability of taking action, a certain level of well-being is necessary to pursue a range of options—such as employment or civic engagement—that may have been previously denied to women. In contrast, it is also possible that a woman may be empowered, but choose not to participate in the domains that are typically assessed as empowerment outcomes (e.g., political). Therefore, this study explicitly assesses subjective well-being (i.e., self-esteem, depression, psychological violence) as a primary components of the empowerment process. Unless a broad view of empowerment is adopted that recognizes the importance of psychological variables, development efforts can fail even though adequate material resources or opportunities have been provided.

The Current Study

This paper responds to calls from social scientists in the area of globalization and women’s empowerment (e.g., Mosedale 2005), as well as from within community psychology (Cattaneo and Chapman 2010; Riger 1993), to test a model that incorporates an investigation of both structural and individual factors to address the egregious gaps in our understanding and evaluation of women’s empowerment. Although it has been argued that investigating empowerment will be most useful when done in a framework that examines pathways that may lead to positive outcomes (Narayan 2005), prior investigation of women’s empowerment has failed to empirically examine the relations between resources, agency, and well-being as multiple components of the empowerment process (Kabeer 1999). The current assessment employs a multilevel, culturally relevant approach that incorporates not only structural factors, but also cultural norms and attitudes (e.g., gender ideology), decision-making, power and control within the relationship, autonomy and mastery, and subjective well-being (see Fig. 1).

The investigation emerged out of a partnership between a social scientist and a community-based organization active in the women’s movement in Nicaragua. The study combines the goals shared by community and liberation psychology by focusing on the science of psychology as an instrument for informing social action (Grant et al. 2003). Specifically, the study uses a feminist liberation psychology perspective by taking into account global structural inequities, international human rights discourse, and women’s activism within social movements to contribute to a growing body of work that identifies processes that aid in transforming oppression through social and psychological mechanisms (Lykes and Moane 2009). An analysis with this level of complexity is necessary to lend scientific merit to the understanding of empowerment and increase the acceptability for empowerment approaches among policy makers.

Methods

Sample and Procedure

The data were collected in 2007 in the municipality of Malpaisillo/Larreynaga in the state of León, Nicaragua. Household surveys were administered to two different groups of women. Because customary practices still largely prohibit women from owning land, this research was conducted in collaboration with a women’s organization legally facilitating women’s land ownership to obtain a sufficient number of land-owning women for the first group. The second group of women was selected from neighboring communities in the same municipality that were not actively involved in the organization. Sampling for these groups resulted in 124 landowners and 114 non-landowners. These groups allow for direct comparison of women involved in land resource allocation aimed at empowerment and women who did not receive this intervention.Footnote 2 The total sample size was 238 women. Field procedures recommended by the World Health Organization (WHO) in conducting violence research in developing countries were followed to hire and train a local research team (Ellsberg and Heise 2005). After oral consent was obtained, data were collected in private interviews conducted in Spanish.

Measures

The questionnaires were developed in partnership with the research team, translated into Spanish by a member of the team, and then back-translated with a local Nicaraguan speaker to ensure the meanings were conveyed properly before the survey was piloted. As has been demonstrated in prior work in remote areas where literacy rates are low, we learned during the pilot phase that the complexity of a scaled response was difficult for respondents to understand (Ellsberg and Heise 2005). Therefore, the scales assessing Gender Ideology, Relationship Power, and Partner Control were all modified for dichotomous responses as indicted below.

Demographic Characteristics

Sociodemographic data included age, number of children, education, occupation, earnings, employment status, relationship status, and duration of relationship. Participants also reported data on their current partner: partner’s age, work status, and earnings in relation to the respondent. Because it has been repeatedly demonstrated that male alcohol use is related to power and control in the marital relationship (Malik and Lindahl 1998), participants were also asked about their partners’ use of alcohol and drugs.

Structures and Ideology

Organizational Participation

In order to assess organizational participation, participants were asked how regularly they participate in workshops and seminars aimed at women’s empowerment. Responses were coded on a 4-point likert scale (1 = less than 1 day a week, 2 = 1–2 days a week, 3 = 3–4 days a week, and 4 = most days of the week).

Land Ownership

Questions assessing land acquisition and land ownership were adapted from assessments used by the International Center for Research on Women. Women were asked whether or not they owned land, how much land, how the land was titled (individual, joint, other), how the land was acquired (e.g., inheritance, agrarian reform, NGO intervention), and who controlled the land (self, partner, self and partner equally).

Gender Ideology

Respondents were asked to agree or disagree with eight items that were chosen from the 25-item short version of the Attitudes Towards Women Scale (Spence et al. 1973) based on cultural relevance. Lower scores reflect more subordinate views of women (more traditional Gender Ideology). Internal consistency for this scale was .67.

Agency

Decision-making

In order to compare the most commonly used indicator of women’s agency to the additional measures included in the current study, participants completed two subscales that were designed by the International Center of Research for Women (2006) to measure decision-making within the marital relationship. There were eight items that assessed who had the final decision in both household expenditure decisions (e.g., “Buying food items”) and financial decision-making (e.g., “Putting money into savings”). Internal consistency for these scales were .80 and .89, respectively.

Relationship Power

Respondents were asked to agree or disagree with eleven items from the Relationship Control Subscale of the Sexual Relationship Power Scale (Pulerwitz et al. 2000). Three items from the original scale that assessed condom use, as well as one item that did not translate well were not included. Higher scores reflect greater levels of power within the relationship for the respondents. Internal consistency for this scale was .86.

Partner Control

Respondents were asked whether or not their partners generally prohibit or control their ability to carry out everyday activities or exhibit controlling behavior or jealousy with seven items from the World Health Organization (Ellsberg and Heise 2005). Three additional items were added to assess whether partners prevented women from working outside of the home, studying, or using contraceptives. Affirmative responses in each category were summed. Higher scores reflect greater levels of partner control. Internal consistency for this scale was .89.

Agency

Participants completed two of the six subscales from Ryff’s Scales of Psychological Well-Being (Ryff 1989). There were nine items each on the Autonomy and Mastery subscales. The Autonomy scale assesses self-determination and independence of thought. Mastery reflects a sense of competence and control over one’s environment. Internal reliabilities for these scales were low at .53 and .55, respectively.Footnote 3

Outcomes

Self-esteem

Self-esteem was assessed with ten items from a Spanish version of the Rosenberg Self-Esteem Scale (Baños and Guillén 2000). Respondents were asked to agree or disagree with items such as, “At times, I think I am no good at all.” Internal consistency was .68.

Depression

Depressive symptomotology was assessed with eleven items from the Center for Epidemiologic Studies—Depression Scale (CES-D), which has been validated in Spanish-speaking samples (Grzywacz et al. 2006). One of the original twelve items (“I felt depressed”) was deleted because the translation of the item caused confusion in pilot assessments. Internal consistency for this scale was .73.

Psychological Violence

Psychological violence was assessed with four items from the Conflict Tactics Scale (CTS) indexing insults, humiliation, intimidation, and threats from an intimate partner in the past 12 months (Straus et al. 1996). Because the scale was indexed as a count score, internal consistency was not computed.

Results

Sample Profile

Differences between the groups of women were tested to ascertain the need to control for demographic variables in subsequent analyses (see Table 1). The average ages of the respondents were in the early-mid 40’s and the majority of women had three or more children. Approximately three-quarters of the sample were in relationships that were between 6 and 10 years in duration, though the landowners reported longer relationships. Most of the respondents reported being literate, although approximately a quarter of the sample never received formal schooling and the landowners reported higher levels of secondary schooling. A significantly higher percentage of women in the land-owning group also reported current employment.Footnote 4 Because age differences explained the differential duration of women’s relationships, only age and education were controlled in subsequent analyses.

Table 1 Sample profile
Full size table

Preliminary Analyses

Descriptive analyses of land ownership revealed that 99% of the landowners held individual titles to their land and 82% reported making an income off of the land. Women owned, on average, seven and a half manzanas (12.65 acres) and reported acquiring the land, on average, 14 years after they were married. The majority of these women (58%) reported that they alone made decisions regarding the land, whereas 36% reporting that they made decisions equally with their husbands (only six percent reported that their husbands controlled the land).

Table 2 presents group differences in the proposed components of empowerment. As can be seen from the table there are significant differences on nearly all of the variables, with landowners reporting more progressive gender-role ideology, greater say in household and financial decision-making, more relationship power, less partner control, and higher levels of autonomy and mastery than their non landowning counterparts. The findings also suggest that land owners reported significantly higher levels of self-esteem and lower levels of depression.

Table 2 Mean differences among study variables
Full size table

Given that the current study was the first comprehensive investigation of empowerment that included multiple indicators of agency, the relations between the indicators were examined (see Table 3). The aim was to evaluate the most commonly used measure of individual empowerment in the literature, decision-making, relative to the additional indicators that were included in this study. The results suggest that neither measure of decision-making was strongly correlated with women’s autonomy or mastery nor with women’s relationship power or partner control, despite that decision-making is often used as a proxy of either agency or relative power within the relationship. In addition, the decision-making scales were not related to any of the well-being measures. In contrast, relationship power and control, autonomy, and mastery were consistently related to each other and to the well-being measures.

Table 3 Correlations between components of the empowerment process
Full size table

Results for Proposed Pathways

The proposed model was estimated using EQS Maximum Likelihood estimation procedures with variance–covariance matrices serving as input. To test the hypothesized model, a path diagram was constructed that details the pathways between structural resources (i.e., land ownership, organizational participation, and gender-role ideology), agency (i.e., decision-making, relationship power, partner control, and women’s autonomy and mastery), and outcomes (i.e., subjective well-being) (see Fig. 1). Based on significant relationships, age was controlled when predicting land ownership and psychological violence, and education was controlled when predicting land ownership, participation, ideology, relationship power, partner control, and self-esteem and depression. Because prior studies have demonstrated that alcohol use by male partners is one of the strongest correlates of violence (Coker et al. 2000), partner alcohol use was assessed and controlled for with nearly all of the study variables with which it was significantly correlated: organizational participation, relationship power, partner control, autonomy, mastery, depression, and psychological violence.

As predicted, both land ownership and participation level were related to more progressive gender-role ideology which was, in turn, related to higher levels of household, but not financial, decision-making, greater relationship power, and less partner control. Neither of the decision-making measures were associated with either indicator of women’s individual agency. In contrast, relationship power was related to higher levels of autonomy and mastery, and partner control was related to lower levels of women’s mastery. Autonomy and mastery were each associated with higher self-esteem and lower depression. Mastery was associated with less psychological violence. The hypothesized model provided a reasonably good fit to the data (i.e., χ2 = 175.82, df = 61, χ2/df = 2.88, NFI = .81, CFI = .86, RMSEA = .09, AIC = 53.82); however, given that neither of the decision-making variables served as both significant outcomes or predictors in the hypothesized process, they were dropped from the model and a trimmed model was re-run (see Fig. 2). Fit statistics from this model indicate a slight improvement (χ2 = 110.34, df = 39, χ2/df = 2.83, NFI = .87, CFI = .91, RMSEA = .09, AIC = 32.34). The trimmed model was compared to the initial model by evaluating the Akaike Information Criterion (AIC) (Bozdogan 1987). The model with the lowest AIC is preferable. The trimmed model provides a better fit to the data and yields the smallest of the AIC.

Fig. 2
figure 2

Empirically supported model. Values are standardized beta weights. *p < .05, **p < .01, ***p < .001

Full size image

Next, in order to help explain the mechanisms by which land ownership and organizational participation are related to agency and well-being, product of coefficients tests were used to test for indirect effects (Sobel 1990). A test of the indirect relation of structural factors on gender relations suggests that land ownership is indirectly related to relationship power (t = 3.00, p = .002) and partner control (t = − 1.96, p = .050), because it is related to more progressive gender-role ideology. Participation level was also indirectly related to relationship power (t = 2.06, p = .039), but not partner control. Thus, it seems that although the role of the organization may be important in terms of facilitating women’s roles as land owners, actual ownership of land is a more robust predictor of altered gender relations.

To examine whether gender relations played a role in the link between gender ideology and women’s levels of individual agency, tests of the indirect relation of gender ideology on autonomy and mastery were conducted. Gender-role ideology was significantly indirectly related to autonomy, but not mastery, through relationship power (t = 2.21, p = .027), but not partner control. Next, the indirect effects of gender relations on subjective well-being were examined. Relationship power was indirectly related to women’s self-esteem and depression via autonomy (t = 1.98, p = .047) and (t = − 2.00, p = .046), respectively, but not mastery. Partner control was indirectly related to depression via mastery (t = − 2.05, p = .04). Neither of the gender relations variables indirectly explained levels of psychological violence.

Discussion

This study not only contributes to a growing body of literature that helps us better conceptualize women’s empowerment, but supports a theoretical model that suggests social structures may be associated with social and psychological factors often associated with subordination and oppression. The findings—namely, that contexts where power may be unevenly distributed (i.e., women owning land in a context whereby female land ownership defied social roles) are related to ideology and patterns of personal control and freedom—provide support for the suggestion that multiple components of empowerment relate to each other and are critical to our understanding of the processes surrounding empowerment.

Identifying structural patterns of domination has long been the task of political and social theorists (e.g., Bartky 1990); however as Martín-Baró et al. (1994) argued, psychologists can and should reframe standard methods to consider that the root causes of oppression lie in the structures—political and cultural—and ideologies that underlie oppressive social conditions. Although power has been a key concept in feminism, it has been largely ignored by psychologists, especially as it occurs at an institutional or structural level (Yoder and Kahn 1992). The current study suggests that control over resources that systematically privilege certain societal members over others may be an integral part of the process of empowerment. Although, the lack of random assignment and longitudinal design limits the extent to which notions of causality can be drawn, the study findings allow for a discussion of how the various components of empowerment may relate to each other in a manner that can most effectively create change.

The findings suggest that while it may be possible that resources serve as catalysts for empowerment, empowering women requires a contextualized understanding of power in different dimensions. For example, the results highlight a relation between structural factors and gender ideology. Although the direction of these effects cannot be discerned from this data, the findings support the notion that resources may provide the material conditions through which inequalities are produced, but cultural ideology plays a critical way in how they are sustained (Glick and Fiske 1999). In particular, it was found that women with more progressive gender ideology reported having greater relationship power and received less partner control. It is possible that women with greater awareness of their sociopolitical environments, who hold beliefs about their rights to exercise their capabilities, exert greater influence in their marital relationships. Perhaps not surprisingly then, higher levels of interpersonal agency were related to women’s greater individual agency as reflected in measures of autonomy and internal feelings of competence, which, in turn, were related to higher levels of subjective well-being. Although the empowerment process demonstrated in Fig. 2 is likely iterative, not linear, the findings suggest that manifestations of power between men and women may not be static, but rather may be malleable under certain conditions. The results also suggested that the most widely relied upon measure of empowerment in the development literature to date, decision-making, was not a robust or reliable component in the empowerment process.

Several limitations in the current study are important to consider when interpreting the results and making suggestions for future research. First, the low internal consistencies reported for autonomy and mastery raise questions about whether the scales may be measuring more than one construct. However, the scales had strong face and construct validity in this sample and the demonstrated relations were as predicted. Nevertheless, future research should aim to construct measures that will more accurately capture women’s individual agency. Perhaps of greater import, based on the demonstrated model we cannot account for women’s initial willingness to join the organization, nor ascertain if more progressive women, or women with already higher levels of empowerment, were more likely to become involved in the first place. However, correlational findings suggest that relationship power and control were not related to women’s organizational participation, suggesting the sample of women involved were not simply the women whose husbands would allow it. Similarly, women from both groups reported comparable levels of lifetime domestic violence suggesting that women from the intervention group were not from relationships that were more progressive in this regard. In addition, qualitative interviews with both organization leaders and women beneficiaries suggest that women, by and large, were reluctant to join a “feminist” organization. Moreover, rather than reporting nontraditional attitudes regarding women’s roles, they reported pulling “double-duty” at the outset in order to attend the organization (e.g., preparing meals for their husband ahead of time to plan for their absence). The women reported that their participation was motivated by changes in their sense of confidence and self-worth. Regardless, priorities for future research should include longitudinal studies to investigate the causal effects of structural changes. Sound methodology surrounding the investigation of empowerment is imperative to understanding women’s well-being and effecting interventions that can contribute to social change in a globalized context.

The importance of understanding the process of empowerment is not just academic. First, the findings suggested that organizational intervention may provide an important and effective means to achieving change. Organizational participation was related not only to land ownership, but also to more progressive gender ideology. Again, although the direction of this relation cannot be discerned from the model, this link lends support to Freire’s (1970) theory of consciousness raising through group forums as a means to bring about empowerment. Similarly, we know that even when self-selection processes may be at play (e.g., students who enroll in Women’s or Ethnic studies classes), that identity development and consciousness unfold through variously more progressive stages due to organized participation (Bargad and Hyde 1991). Second, the study demonstrates a synergistic relationship between a women’s organization and a social psychologist whereby the group of women developed their own strategies for action and the psychologist, in the words of Ignacio Martín-Baró et al. (1994), used the discipline in the service of social justice by focusing on the oppressive reality of social structures. This study demonstrates that successful collaborations between community-based organizations and activist scholars may be critical in the struggle for social justice.

In addition, the study findings support a number of guiding principles and strategies for interventions that can advance the international empowerment agenda and contribute to the aims of social justice articulated in the Beijing Platform for Action (UN 1995). First, the findings suggest that development practitioners should not confuse practical interventions (i.e., those based solely on resource distribution) with strategic (i.e., those with transformative potential). Given the demonstrated importance of the organization in the current study, the findings suggest that it may be most useful to think about resources as ‘enabling factors’ that may be critical catalysts in the empowerment process, rather than an end in themselves (Malhotra and Schuler 2005). In other words, the design of empowerment programs should be based on the potential for transformative change and on outcomes that suggest a greater ability on the part of women to act on the structures of power that constrain their lives, and not on buzzwords that are in favor of a neo-liberal globalized economy. These designs may be most effective when policy makers and interventionists work with women’s organizations to combine equity in the distribution of resources with a sense of personal power and control to optimally impact well-being.