Role-specific Self-efficacy as Precedent and Product of the Identity Model

Abstract

This article examines empirically ideas initially proposed in speculative work by Ervin and Stryker dealing with what could be understood about human social behavior and interaction by bringing together self-esteem and identity theories. Necessary to that task is distinguishing between two key and often conflated concepts of identity theory, salience and prominence. We argue that role-specific self-efficacy, embedded in self-esteem theory, is a precedent and a product of the identity theory model and global self-efficacy is a link from role-identities to the self-concept through the impact of global self-efficacy on prominence. Findings support a hypothesized feedback loop from role-specific self-efficacy to prominence to salience and back to role-specific self-efficacy.

Keywords

identity self-esteem self-efficacy

Introduction

Identity theory (Stryker 1968, [1980] 2003) was intentionally minimal in its inception, explaining role behavior as a product of commitment to and consequent salience of an identity. Also since its inception, the theory has been refined and extended by distinguishing and differentiating new concepts from existing ones and incorporating concepts from proximate theories. Each extension, expansion, and refinement of the theory has been in pursuit of its original goal of understanding the relationship of identity and behavior.

The current research focuses on a concept from a proximate theory, namely, self-efficacy, itself basic to a broader concept, self-esteem, to better understand the relationship between identity and behavior. It takes advantage of data from a longitudinal study of science students to address this central goal of identity theory while attending to an important topic of educational inequality: racial, ethnic, and gender disparities in science, technology, engineering, and mathematics (STEM) degrees and occupations.

While the United States tops many rankings of scientific production, including scientific article counts, number of patents, gross domestic expenditure on research and development, and PhDs awarded (Organization for Economic Co-operation and Development [OECD] 2013), it lags far behind other countries in science education. The United States ranks 20th of the 34 OECD countries in science education and near the bottom (27 of 34) in math education (OECD 2012). Maintaining American scientific production requires facilitating the achievement of science students at many points during the educational trajectory, including enhancing science education in primary and secondary schools and investing in postsecondary education to improve access and affordability (National Academy of Sciences 2011). Providing academic and social support for students from underrepresented groups—women and racial and ethnic minorities—to support their motivation to pursue science (National Academy of Sciences 2011; National Science Foundation 2015) is an important element in this effort. Notably, these students are already motivated: Research demonstrates that underrepresented racial and ethnic minorities are just as interested in pursuing STEM education and careers as white and Asian students (Higher Education Research Institute 2010). The problem is that, at some point, these students disproportionally “leak” from the proverbial pipeline.

The current research focuses on this last piece of the puzzle: What encourages and sustains students’ motivation to pursue science degrees and careers? We approach this question by considering the relationship between students’ identities as science students and their sense of self-efficacy as science students. We examine the science student identity as students’ understandings of themselves as science students, the extent to which they think of themselves as science students, and how important this identity as a science student is to their self-concept. We examine science student self-efficacy as students’ sense that they are capable of successfully performing as science students. Our primary research question asks how encouraging their science student self-efficacy affects their connection to a science student identity. In pursuing an answer to this question, we analyze four waves of data from a longitudinal study of underrepresented racial and ethnic minority students engaged in STEM programs.

At a more fundamental level, this research question allows us to pursue a goal more foundational to the disciplines of social psychology. By investigating the relationship between self-efficacy (and thereby a dimension of the larger construct of self-esteem) and the key constructs of identity theory, we importantly address a key concern of social psychology: linking the multiple “faces” (House 1977) of the two houses of psychology, the sociological and the psychological (Oishi, Kesebir, and Snyder 2009; Stryker 1989).

Identity Theory

The original purpose of identity theory is understanding the connection between individuals’ sense of who they are, their behavior in interactional situations, and the effect of social structure on this relationship (Stryker 1968, [1980] 2003). Identity theory is concerned with the relationship between the self-concept and individual positional components of the self called identities which, when taken altogether, constitute the self-concept. While Sheldon Stryker’s (1968, [1980] 2003) original formulation of identity theory focused on one type of identity, role-identities, identity theorists today include two other forms of identity in their conception of the self: group-identities and person-identities (Burke and Stets 2009; Stets and Burke 2000; Stets and Serpe 2013).

We focus here on role-identities which we understand as the meanings attributed to the roles individuals fill in interaction with others in counter-roles. In modern societies, the individual’s number of role-identities may be large, determined by the roles the individual performs in interaction with others in counter-roles (James [1890] 1950). The self is comprised by these role-identities (as well as group-identities and person-identities), each contributing to the individual’s sense of self (Stets and Serpe 2013). We can understand the dynamic nature of the individual’s multiplicity of identities from an ecological perspective. As individuals move through social situations and the social world, they reweight the importance placed on particular identities and alter their propensity of enactment (Serpe and Stryker 1987; Smith-Lovin 2003; Stryker and Serpe 1982). This perspective illuminates the original motivation driving identity theory: understanding the decision to call up a specific role-identity in the context of multiple identities. That is, how does the individual decide which role or roles to perform in a given situation (Stryker 1968, [1980] 2003)?

Since its original formulation (now referred to as identity theory’s structural version), identity theory has been complemented by a perceptual control version (Stryker and Burke 2000). Where the structural version focuses on behavior in role-identities, the perceptual version of identity theory focuses on the consequences of identity enactment: the verification of the identity by interactional partners in counter-roles and the outcomes of identity verification (or the lack of verification) on the identity and the self (Burke 1991; Stets and Burke 2005).¹ One such outcome, as will be discussed shortly, is a positive or negative effect on self-esteem.

The structural version of identity theory proposes two key concepts to account for the multiplicity of identities that comprise the self and influence when and how they are enacted: identity salience and identity prominence (Brenner, Serpe, and Stryker 2014; Stryker and Serpe 1994). Identity salience is defined as the probability that a given identity will be invoked in social interaction (Stryker 1968, [1980] 2003) or the propensity to define a situation in a way that provides an opportunity to perform that identity (Stryker and Serpe 1982). Identities can be thought of as existing in a salience hierarchy in which each identity is ranked according to its propensity of being called up. As this suggests, a highly salient identity is very likely to be enacted in situations that are relevant to it and may even be invoked in tangentially relevant or even irrelevant interactional situations. For example, an individual with a highly salient science student identity will be very likely to enact it not only in chemistry and biology courses but also in conversations with others with whom the individual is linked through his or her other identities. In contrast, identity prominence is affective (McCall and Simmons 1978), the subjective value the individual places on an identity. The concept is highly related to or synonymous with the concepts of identity importance (Ervin and Stryker 2001; Marsh 1986) and psychological centrality (Rosenberg 1979). Much like the salience hierarchy, identities are also placed in hierarchy according to their prominence.

George J. McCall and J. L. Simmons (1978) proposed a causal relationship between these two constructs, from prominence to salience, mirroring the relationship between affect and behavior in cognitive psychology (see Lazarus and Folkman 1984; Zajonc and Markus 1984). A developing empirical research program has supported this hypothesis, finding that higher placement of an identity in the prominence hierarchy increases the likelihood that the identity will be invoked. Sheldon Stryker and Richard T. Serpe (1994) included both concepts in structural equation models estimating their correlation using a single wave of data from a longitudinal study. Unfortunately, and similar to earlier work by Larry Nuttbrock and Patricia Freudinger (1991), inadequate sample size prevented a robust test of the causal relationship posited by McCall and Simmons. Critiquing the use of hypothetical situations in measuring salience, Philip S. Brenner (2011, 2012) operationalized salience as a retrospective measure of the frequency of recent behavior. While supportive of the theorized prominence-to-salience causal relationship, his cross-sectional data were also unable to provide a robust test. However, more recent work using multiple waves from a large-sample longitudinal study estimated a series of structural equation models testing the hypothesized and plausible alternative relationships. Results strongly support the hypothesized relationship: Prominence is a strong and significant cause of salience (Brenner, Serpe, and Stryker 2014).

This relationship implies that many prominent identities, for example, familial and professional identities, are also highly salient and, conversely, that disvalued identities are relatively unlikely to be performed. However, these concepts need not be in alignment. For example, the religious identity, highly prominent for many Americans, may be relatively unlikely to be performed (Brenner 2011). Obversely, a role that is very likely to be performed (e.g., the role of a prisoner for an inmate in interaction with other inmates and corrections officers) may not be valued or may even be actively disvalued (Asencio and Burke 2011).

Operationalizations of salience reflect its cognitive focus on role-relevant behavior and the propensity for its enactment. Salience has been most commonly measured prospectively by asking survey respondents how they would present themselves in situations in which they meet someone new, like a friend of a close friend (Stryker and Serpe 1994), or the likelihood that they would invoke an identity across situations (Merolla et al. 2012; Owens and Serpe 2003). In contrast, prominence is focused on affect, conventionally measured as either a single or a series of attitudinal survey questions asking a respondent to evaluate the strength of the sentiment he or she has for an identity.

While self-esteem has been well theorized and empirically examined as an outcome of identity verification in identity theory’s perceptual control version (Asencio 2013; Burke and Stets 1999; Cast and Burke 2002; Stets and Burke 2014; Stets and Cast 2008), its role in the structural version of identity theory is underresearched. We use Ervin and Stryker’s theoretical development as our starting point to understand how one part of the self-esteem construct, specifically self-efficacy, can influence key concepts of the structural version, namely, identity salience and identity prominence.

As will be discussed, self-efficacy is posited by Laurie H. Ervin and Sheldon Stryker (2001) as a clear pathway by which self-esteem (of which self-efficacy is a dimension) and identity theory can be integrated. Moreover, given the role-identity focus of identity theory’s structural version, we propose a primary function for role-specific self-efficacy. Given the postulates of Ervin and Stryker, we hypothesize role-specific self-efficacy as both a precedent and product of the structural identity model. Global self-efficacy, or mastery, also has an important role. Its function as a primarily subjective assessment of the self-concept provides a linkage for role-identity to fit into the self-concept through the impact of global self-efficacy on prominence. Through this relationship, global self-efficacy moderates the relationship of prominence on role-specific self-efficacy. As self-efficacy is an important part of the theory and research on self-esteem, we now turn our attention to it.

Self-esteem: Types and Dimensions

William James’ ([1890] 1950:310) original conception defines self-esteem as the “ratio of our actualities to our supposed potentialities,” or more succinctly as the ratio of successes to pretensions. He posits a self-esteem that arguably transcends role-specific operationalizations like identity salience and prominence. Following James, Morris Rosenberg’s (1965) conceptualization describes self-esteem as an evaluation of the worth or effectiveness, positive or negative, of the self as a whole. His self-esteem scale follows this approach, evaluating the self-concept in its entirety without reference to any particular subset of the self. Thus, Rosenberg’s self-esteem is an omnibus construct covering the individual’s evaluation of the global self rather than focusing on any specific identity.

While much research treats self-esteem as unitary, early operationalizations propose multiple dimensions (Coopersmith 1967). Current operationalizations typically divide self-esteem into two levels: global and specific (Hoelter 1986; Marsh 1986, 1990; Marsh and Shavelson 1985; Rosenberg et al. 1995). Global self-esteem, which evaluates the worth and adequacy of the self-concept overall, appears more commonly in the extant literature. Although it is less common, role-specific self-esteem is important for a complete understanding of self-esteem (Rosenberg et al. 1995). Role-specific self-esteem can be traced back to self-esteem’s original conceptualization. James’ ([1890] 1950) example of self-esteem compares his own successes and pretensions within his specific role-identity as a psychologist. Indeed, facets of the self-concept at the specific level, including individual role-identities, contribute to the maintenance of global self-esteem, especially if the role-identity is highly prominent (Marsh 1986; Pelham and Swann 1989).

Both the global and specific levels of self-esteem are comprised of two dimensions, self-worth and self-efficacy.² Self-efficacy is the individual’s perceived ability to influence outcomes of his or her life. Global self-efficacy is arguably synonymous with Pearlin’s sense of mastery, defined as “the extent to which people see themselves as being in control of the forces that importantly affect their lives” (Pearlin et al. 1981:340). His mastery scale measures a person’s global sense of his or her ability to overcome problems, affect change, and exert control without reference to particular roles or situations (Pearlin and Schooler 1978). In contrast to the generalized sense of mastery, the specific level of self-efficacy is rooted in the person’s belief in his or her capabilities to successfully perform particular actions in specific contexts (Bandura 1986, 1995).

Self-worth is defined as the affective self-evaluation of the self-concept measured on a global level as an overall assessment of one’s qualities. Pearlin distinguishes global self-efficacy/mastery from self-esteem, measured using Rosenberg’s scale, which he defines as equivalent to self-worth (Pearlin et al. 1981). While Pearlin’s theoretical perspective may not agree, in light of more recent theoretical developments and empirical findings our conception casts global self-worth as a component of self-esteem. Alicia Cast and Peter J. Burke (2002) used seven items from the Rosenberg Self-Esteem Scale, in addition to a few other questions, to measure global self-worth, including “I feel that I’m a person of worth, at least on an equal plane with others” and “On the whole, I am satisfied with myself.”

At a specific level, self-worth is invested in specific facets of the self-concept (Crocker 2002; Crocker and Knight 2005; Crocker and Wolfe 2001). The self-worth attributed to each of the specific facets of the self-concept contributes to global self-worth. Role-identities are one such specific-level facet. Thus, specific role-identities contribute to global self-worth in varying degrees, depending on the extent to which one’s self-worth is “staked” or “contingent upon” on a particular identity. In this way, some role-identities have a higher influence on global self-worth than do others.

More recently, Jan E. Stets and Peter J. Burke (2014) have argued for a third dimension of the self-esteem construct—authenticity—in addition to self-efficacy and self-worth. Gecas (1982, 1991) theorized authenticity as a motive for behavior as individuals are prompted by the desire to act and be seen in a way that is congruous with their view of their “real” self. Stets and Burke (2014:411) synthesized Gecas’ and others’ conceptualizations of authenticity, defining it as “involving one’s internal or personal standards as to who one really is.” They measured the construct using a series of seven items tapping into the respondent’s belief that he or she is able and typically does act in ways that reflect his or her “real self” and that others accept that identity as authentic.

Bringing Together Identity Theory and Self-efficacy

An important line of the research program aimed at integrating these theories has focused on the role of self-esteem, particularly its self-efficacy dimension, in the identity verification process of identity theory’s perceptual control version. Cast and Burke (2002) conceptualized self-esteem both as a motivation for and outcome of the identity verification process. The verification of a role-identity performance maintains or increases self-esteem through its self-efficacy dimension (Burke and Stets 1999; Stets and Burke 2014). Engaging others in a role-identity (e.g., participating in science classes for science students) provides opportunities for verification of that identity (e.g., tests and exams, classroom discussions). Assessing the alignment of self-concept with the three bases of an identity—role, person, and group—Stets and Burke (2014) made the case for self-efficacy’s conceptual alignment with role-identities and further argued that self-worth is aligned with group-identities, and authenticity is aligned with person-identities (Stets and Serpe 2013).

Moreover, self-esteem can protect the individual when interactions go wrong. While a single disconfirming interaction (e.g., a poor performance on a single test) may cause momentary negative emotions, the individual can rely on the self-esteem built up from previous interactions that yielded successful verification (e.g., previous tests earning high scores) of the role-identity (Cast and Burke 2002). However, repeated, disconfirming interactions in a role-identity can drain the self-esteem reservoir and lead to a reassessment of the role-identity (e.g., changing one’s major to a non-STEM field). Alternatively, Asencio (2013) argued that self-esteem moderates, rather than buffers against, the effect of feedback from others about the self depending on the source of the feedback.

This perspective on the role of self-esteem, and particularly self-efficacy, based on the perceptual control version of identity theory, may benefit from bringing it together with concepts from the structural version of the theory. Similar to the moderating effect of self-esteem theorized and tested by Asencio (2013), McCall and Simmons (1978) argued that the effect of a disconfirming interaction is moderated by the prominence of that role-identity given the extent to which self-esteem is staked on that identity. A disconfirming interaction, like a failed chemistry exam, would be more threatening if the individual’s self-concept is heavily staked on the science student role-identity being enacted, and less threatening if the disconfirmed identity is relatively low in the individual’s prominence hierarchy. Moreover, repeated disconfirmatory interactions may cause the individual to reduce the prominence of the role-identity, thus protecting self-esteem. However, this moderating role of prominence has not been directly tested. Timothy J. Owens and Richard T. Serpe (2003) modeled self-esteem as a predictor of identity salience, describing but not testing the moderating role of prominence preventing damaging identities from negatively affecting the self-concept.³

Of the three dimensions of self-esteem, we focus on self-efficacy given strong theoretical and empirical connections between it and the original focus of identity theory: role-identities (Stets and Burke 2014; Stryker 1968, [1980] 2003). Parallels can be readily drawn between self-efficacy and identity constructs. First, both global self-efficacy and identity prominence subjectively evaluate their referents, the self-concept and a role-identity, respectively, at the level of the individual’s awareness. As the sense of one’s capability to control his or her life, global self-efficacy is the antithesis of fatalism. Accordingly, such an evaluation of overall capability cannot be based on an objective assessment of one’s roles in all situations; one simply cannot expertly perform all his or her roles nor will he or she value all roles equally and highly. Rather, global self-efficacy has a subjective component that is arguably strongly influenced by prominent identities. Thus, prominence affects the identity’s contribution to self-esteem through self-efficacy (Hoge and McCarthy 1984; Marsh 1986). The higher the prominence of a role-identity, the more that role-identity will matter to global self-efficacy.

Like identity salience, role-specific self-efficacy is primarily a cognitive construct reflecting one’s perceptions of his or her feelings. While identity salience addresses the propensity for role performance, role-specific self-efficacy addresses the individual’s confidence in that performance. Where these concepts differ is in one’s awareness of them. Awareness of the salience of an identity is not a necessary feature of its existence (Serpe and Stryker; Stryker and Serpe 1994). However, by definition, persons are aware of the confidence they place in their ability to successfully perform a role.

Correspondence between these concepts is more than definitional; self-efficacy theory and identity theory are complementary and may even be considered interdependent. Ervin and Stryker (2001) proposed the integration of these theories into a single model, noting that such an expansion is intuitive and theoretically and empirically justifiable. They posited causal relationships between key constructs of each, two of which are pertinent to the current research: (1) unidirectional causation allowing role-specific self-efficacy to directly influence identity salience and prominence, and (2) bidirectional causation between role-specific and global self-efficacy (as a part of global self-esteem) allowing them to reciprocally influence each other. Hoelter’s (1986) findings partially support these propositions as the role-specific level influences the global level, although he finds that the prominence of the identity does not influence its contribution to global self-esteem. Noting the role-based focus of identity theoretical constructs and the self-concept focus of global self-esteem, Ervin and Stryker do not hypothesize a direct relationship between global self-esteem and identity salience and prominence.

Current Research

Prior work (Brenner et al. 2014; Nuttbrock and Freudinger 1991; Stryker and Serpe 1994) highlights the measurement of and relationship between prominence and salience. A causal pathway from prominence to salience has been hypothesized in the extant work; the higher the value placed on an identity, the more likely that identity is to be enacted (McCall and Simmons 1978). As this proposition has been supported empirically (Brenner 2011, 2012; Brenner et al. 2014; Nuttbrock and Freudinger 1991; Stryker and Serpe 1994), it is taken as a premise of the model tested here. Thus, science student identity prominence is included at time t and salience at time t + 1 with a causal path from the former to the latter (see Figure 1). Given the findings from Brenner et al. (2014) using these same data, as well as the findings of other prior research, we assume that identity prominence is positively and directly associated with science student identity salience.

Figure 1.

Heuristic model including identity salience, identity prominence, and global and role-specific self-efficacy.

Given its connection with identity theory constructs, we include role-specific self-efficacy in this model at two times preceding and following the identity constructs. Ervin and Stryker (2001) hypothesized that a strong desire to competently perform a role-identity will increase the subjective value placed on that identity, implying a strong causal relationship from role-specific self-efficacy to prominence.

Hypothesis 1: Science student role-specific self-efficacy is positively and directly associated with subsequent prominence of the science student identity.

Science student role-specific self-efficacy and salience of the science student identity are likely to be highly associated as both concern role-related behavior . First, role-specific self-efficacy is hypothesized to influence identity salience. However, this pathway is likely indirect, operating through prominence. In line with Ervin and Stryker (2001), confidence in one’s ability to successfully perform a role increases the subjective value placed on that role-identity. Increased prominence in turn increases the propensity to define interactional situations as role-relevant. Second, role-specific self-efficacy is also argued to be an outcome of salience: Increased propensity to perform a role encourages confidence in the individual’s ability to perform it. In short, practice makes perfect.

Hypothesis 2: Science student role-specific self-efficacy is positively associated with subsequent salience of the science student identity indirectly through prominence.

Hypothesis 3: Salience of the science student identity is positively and directly associated with role-specific self-efficacy as an outcome.

In addition to a hypothesized large and significant stability coefficient from role-specific self-efficacy at Time 1 to Time 4, these constructs and paths comprise the hypothesized associations between identity constructs and role-specific self-efficacy.

Global self-efficacy (rather than global self-esteem or global self-worth) is included here for theoretical consistency with the focus on role-specific self-efficacy. (Unfortunately, authenticity cannot be considered as it was not measured in the data to be analyzed here.) Moreover, prior work links role-identities to the efficacy dimension of the self-esteem construct rather than the worth dimension (Burke and Stets 1999; Cast and Burke 2002; Stets and Burke 2014).⁴ Two paths between self-efficacy specific to the science student identity and global self-efficacy are hypothesized. First, exogenous role-specific self-efficacy is included as a predictor of global self-efficacy. The former contributes to the latter, as mastery of one role can lend to a more general sense of ability to control one’s life. Thus, role-specific self-efficacy at time t is hypothesized to influence global self-efficacy at time t + 1. Second, global self-efficacy is included as a predictor of role-specific self-efficacy. The general sense of mastery in one’s life both influences and is influenced by (as hypothesized above) the mastery of specific roles. Following Morris Rosenberg et al. (1995), the path from role-specific to global self-efficacy is hypothesized to be stronger than the reverse relationship.

Hypothesis 4: Exogenous science student role-specific self-efficacy is directly and positively associated with global self-efficacy.

Hypothesis 5: Global self-efficacy is directly and positively associated with science student role-specific self-efficacy as an outcome.

Global self-efficacy is measured at the same time as identity prominence. Because both concepts are more subjective assessments of their components of the self (Ervin and Stryker 2001), we expect a relationship to emerge between global self-esteem and prominence in the science student identity. As global self-esteem refers to the overall self-concept and identity prominence is role-specific, we expect that the effect of the former on the latter will be only moderate in size. Given evidence for the stability of the overall self-esteem construct (Cast and Burke 2002; Marcus and Kunda 1986; Owens and Samblanet 2013; Owens and Serpe 2003; Serpe 1987; Swann, Chang-Schneider, and McClarty 2007; Trzesniewski, Donnellan, and Robins 2003), of which global self-efficacy is a dimension, it is hypothesized to operate as a force in the model.

Hypothesis 6: Global self-efficacy is directly and positively associated with prominence in the science student identity.

Ervin and Stryker did not posit a connection between global self-efficacy and salience, and recent research by Peter J. Burke and Jan Stets (1999; Stets and Burke 2014) found a strong connection between role-identities and self-efficacy at the role-specific rather than the global level. Given the connection between global self-efficacy and prominence, we hypothesize that the relationship between global self-efficacy and identity salience will be indirect through prominence.

Hypothesis 7: Global self-efficacy is positively, but only indirectly, associated with salience of the science student identity through prominence.

This discussion implies a set of hypothesized null relationships between constructs. These hypothesized nonsignificant relationships are illustrated in Figures 1 and 2 as dashed lines. These hypotheses will be tested, fitting a series of path models to data from a longitudinal study of underrepresented racial and ethnic minority students studying STEM at the university level. Respondents’ identities as “science students” are the focal identities of the model.

Figure 2.

Structural model including identity salience, identity prominence, and role-specific self-efficacy.

Data and Method

A series of structural equation models are fit to four waves of panel data from The Science Study, a national panel study following underrepresented college students in STEM fields. Study participants were recruited from universities with one or more of the following programs: Research Initiative for Scientific Enhancement (RISE), Minority Access to Research Careers (MARC), Undergraduate Student Training in Academic Research (U-STAR), and Bridges programs geared toward increasing minority students’ engagement and success in these disciplines (Merolla et al. 2012; Schultz et al. 2011). The sample consisted of 1,420 students who were enrolled in one of these science enhancement programs as well as students from each of the campuses who were not enrolled in a science enhancement program but exhibited the same characteristics of major, academic performance, interest in becoming a scientist, and so on, as the students who were participating in science enhancement programs. Eligibility was limited to students enrolled in an undergraduate, graduate, or professional (e.g., medical school) program at the time of the screening interview in fall 2005. While panel attrition was relatively low, models are estimated using full information maximum likelihood to retain cases with missing data.

Panel members were asked to complete one web survey each semester.⁵ The screening interview (Wave 0) was fielded in the fall of 2005 with panel members resurveyed biannually (fall and spring semesters) beyond the initial wave. Four waves of data, collected in spring 2006, fall 2006, spring 2007, and spring 2008, are used in these analyses.⁶ At the time of the screening interview, 10 percent were freshmen, 15 percent sophomores, 29 percent juniors, 31 percent seniors, and 16 percent graduate or professional students. The panel is heavily female (72 percent) and comprised largely of racial and ethnic minority students: Most respondents are either black (49 percent) or Latino/a (39 percent) (see Table 1).

Table 1.

Descriptive Statistics.

Measures	M/%	SD	Range	Median
Prominence
Self-image	3.40	1.20	1-5	3
Reflection	3.44	1.07	1-5	3
“Scientist”	3.70	1.10	1-5	4
Community	3.58	1.15	1-5	4
Salience
Coworker	7.56	2.57	0-10	8
Opposite sex	7.19	2.86	0-10	8
Friend of friend	7.28	2.84	0-10	8
Friend of family	7.46	2.82	0-10	8
Role-specific self-efficacy (Time 1)
Research question	3.80	0.90	1-5	4
Data collection	3.85	0.83	1-5	4
Analyze data	3.93	0.83	1-5	4
Explain results	3.95	0.80	1-5	4
Role-specific self-efficacy (Time 5)
Research question	3.95	0.96	1-5	4
Data collection	4.03	0.88	1-5	4
Analyze data	4.03	0.88	1-5	4
Explain results	4.05	0.86	1-5	4
Global self-efficacy
Do things well	3.51	0.57	1-4	4
Feel useless^a	3.10	0.87	1-4	3
Feel like failure^a	3.60	0.64	1-4	4
Race
Black	48.8%
Latino	39.0%
Sex (female = 1)	71.5%
Income	4.01	2.62	1-9	3
Parental education (college = 1)	45.3%

Indicates reverse-coded item.

The diversity of this sample is a strength for the purpose of testing theory. We do not say these data allow inference to a wider population, but they do allow a strong test of the theory presented. Respondents to this survey have good reason to be engaged in the focal identity: Mote than 90 percent of respondents are women and ethnic or racial minorities pursuing a college degree (or more) in STEM fields in which they are underrepresented.

Models include 15 manifest variables operationalizing the four latent constructs of interest, each measured using a preexisting scale adapted for the focal identity (Serpe 1987; Serpe and Stryker 1987; Stryker and Serpe 1982, 1994). The first latent construct, the salience of a science student identity, is measured using a four-question scale. Each item asks students to reflect how certain they would be to mention their desire to be a scientist upon meeting a person for the first time who fits into one of each of four categories: (1) a coworker, (2) a person of the opposite sex, (3) a friend of a friend, or (4) a friend of a family member.⁷ Each situation is rated on an 11-point scale measuring the likelihood that respondents would tell a new person about their desire to be a scientist from (0) “certain I would not” to (10) “certain I would.” The items demonstrate excellent internal consistency: Cronbach’s alpha is .95.

The second latent construct, prominence of a science student identity, is operationalized following prior work (Estrada et al. 2011; Hernandez et al. 2013). We use a four-item battery of questions that tap into the respondents’ perceptions of themselves and what is central to their self-concept in this identity. Each item uses a 5-point Likert agreement scale to measure the subjective value placed on the identity. The four items are as follows:

Being a scientist is an important part of my self-image.

Being a scientist is an important reflection of who I am.

I have come to think of myself as a “scientist.”

I have a strong sense of belonging to the community of scientists.

The prominence scale is strongly internally consistent: Cronbach’s alpha is .85.

The third latent construct, role-related self-efficacy, is operationalized using a modified version of a science student self-efficacy scale (Chemers 2006; Estrada et al. 2011). We use four questions tapping the respondents’ confidence in their ability to function as a burgeoning scientist. Respondents are asked to rate their level of confidence on a 5-point scale (from “not at all confident” to “absolutely confident”) in their ability to (1) generate a research question, (2) figure out what data to collect and how to collect them, (3) analyze what data/observations mean, and (4) to explain study results. The self-efficacy scale demonstrates strong internal consistency at both Time 1 (α = .87) and Time 4 (α = .92).

The fourth latent construct, global self-efficacy, is measured following Cast and Burke (2002), using three Likert agreement items from the Rosenberg Self-Esteem Scale (Rosenberg 1965) with the greatest focus on behavior without reference to specific identities:

I am able to do things as well as most other people.

All in all, I am inclined to feel that I am a failure.

I certainly feel useless at times.

Negative Items 2 and 3 are reverse coded to match the positive direction of Item 1. The scale demonstrates acceptable internal consistency: Cronbach’s alpha is .71.

Each model controls for race and ethnicity (black [reference category], Latino/a, and other race), gender (female = 1), household income (in nine categories from “less than $10,000” to “$80,000 or more”), and parental educational attainment. The manifest variables for each latent construct are constrained to load only on that latent construct in the same wave. Manifest variable errors are allowed to correlate with the errors of other manifest variables loading on the same construct in the same wave. No other factor loadings or error correlations are estimated, and no changes in the error covariance structures are made between models. Model fit is evaluated using root mean square error of approximation (RMSEA; Browne and Cudeck 1993) and alternative models are compared using Bayesian information criterion (BIC; Raftery 1995), in addition to the likelihood ratio chi-square, although fitting a model to four waves of panel data with relatively large sample sizes will likely lead to statistically significant chi-square values (Loehlin 2004).⁸

Results

Manifest variables load on their intended constructs as anticipated. Factor loadings for role-specific self-efficacy are similar in value and range at Time 1 (from .75 to .80) and Time 4 (from .79 to .89). Factor loadings for prominence are similar in value, ranging from .72 to .80. The loadings of the manifest variables on the salience construct are somewhat higher, ranging from .92 to .94. These loadings change negligibly when global self-efficacy is added to the model. Loadings for the global self-efficacy measurement model range from .53 to .74 (see Table 2).

Table 2.

Model Estimates.

	Model 1			Model 2
Parameters/Variables	Coefficient	SE	p	Coefficient	SE	p
Structural model
Role-specific efficacy 1 → prominence	0.342	0.043	***	0.301	0.047	***
Global self-efficacy → prominence				0.137	0.044	***
Black → prominence	−0.039	0.053		−0.063	0.052
Latino → prominence	0.044	0.051		0.027	0.050
Female → prominence	−0.043	0.034		−0.049	0.034
Income → prominence	−0.023	0.037		−0.029	0.036
Parented → prominence	0.031	0.033		0.024	0.033
Prominence → salience	0.531	0.032	***	0.525	0.033	***
Role-specific efficacy 1 → salience	0.043	0.044		0.034	0.040
Global self-efficacy → salience				0.038	0.040
Black → salience	−0.022	0.045		−0.029	0.045
Latino → salience	0.030	0.043		0.025	0.043
Female → salience	0.007	0.030		0.005	0.030
Income → salience	−0.073	0.030	*	−0.074	0.030	*
Parented → salience	0.005	0.029		0.003	0.029
Black → role-specific efficacy 1	0.166	0.056	**	0.159	0.056	**
Latino → role-specific efficacy 1	0.093	0.056		0.086	0.056
Female → role-specific efficacy 1	−0.122	0.035	***	−0.122	0.035	***
Income → role-specific efficacy 1	0.024	0.039		0.018	0.039
Parented → role-specific efficacy 1	0.002	0.035		0.000	0.035
Prominence → role-specific efficacy 5	0.102	0.050	*	0.089	0.049
Salience → role-specific efficacy 5	0.096	0.043	*	0.095	0.042	*
Role-specific efficacy 1 → role-specific efficacy 5	0.522	0.037	***	0.487	0.042	***
Global self-efficacy → role-specific efficacy 5				0.122	0.042	**
Role-specific efficacy 1 → global self-efficacy				0.311	0.046	***
Black → global self-efficacy				0.189	0.057	***
Latino → global self-efficacy				0.139	0.055	*
Female → global self-efficacy				0.041	0.038
Income → global self-efficacy				0.054	0.040
Parented → global self-efficacy				0.052	0.037
Measurement model
Self-image → prominence	0.745	0.019	***	0.742	0.019	***
Reflection → prominence	0.750	0.018	***	0.751	0.018	***
“Scientist” → prominence	0.726	0.020	***	0.724	0.020	***
Community → prominence	0.801	0.017	***	0.802	0.016	***
Coworker → salience	0.924	0.006	***	0.924	0.006	***
Opposite sex → salience	0.942	0.005	***	0.942	0.005	***
Friend of friend → salience	0.937	0.006	***	0.937	0.006	***
Friend of family → salience	0.933	0.005	***	0.933	0.005	***
Research question → role-specific efficacy 1	0.752	0.020	***	0.747	0.021	***
Data collection → role-specific efficacy 1	0.804	0.019	***	0.805	0.019	***
Analyze data → role-specific efficacy 1	0.780	0.022	***	0.789	0.022	***
Explain results → role-specific efficacy 1	0.772	0.020	***	0.774	0.020	***
Research question → role-specific efficacy 5	0.840	0.012	***	0.839	0.012	***
Data collection → role-specific efficacy 5	0.892	0.011	***	0.893	0.010	***
Analyze data → role-specific efficacy 5	0.791	0.016	***	0.794	0.016	***
Explain results → role-specific efficacy 5	0.833	0.013	***	0.835	0.013	***
Do things well → global self-efficacy				0.531	0.030	***
Feel useless^a → global self-efficacy				0.653	0.029	***
Feel like failure^a → global self-efficacy				0.744	0.029	***
Variance
Prominence	0.873	0.029		0.855	0.029
Self-image	0.445	0.028		0.450	0.028
Reflection	0.438	0.027		0.436	0.027
“Scientist”	0.473	0.029		0.475	0.029
Community	0.358	0.026		0.357	0.026
Salience	0.690	0.030		0.689	0.030
Coworker	0.146	0.011		0.146	0.011
Opposite sex	0.112	0.010		0.112	0.010
Friend of friend	0.122	0.011		0.122	0.011
Friend of family	0.129	0.010		0.129	0.010
Role-specific efficacy 1	0.977	0.011		0.977	0.011
Research question	0.435	0.031		0.441	0.031
Data collection	0.353	0.030		0.353	0.030
Analyze data	0.391	0.034		0.378	0.035
Explain results	0.404	0.031		0.400	0.031
Role-specific efficacy 5	0.638	0.035		0.625	0.035
Research question	0.294	0.021		0.297	0.021
Data collection	0.204	0.019		0.203	0.019
Analyze data	0.374	0.025		0.369	0.025
Explain results	0.306	0.021		0.303	0.021
Global self-efficacy				0.877	0.029
Do things well				0.718	0.032
Feel useless^a				0.573	0.037
Feel like failure^a				0.447	0.043
Covariance
Self-image ↔ “scientist”	0.306	0.038	***	0.310	0.037	***
Opposite sex ↔ friend of friend	0.234	0.041	***	0.234	0.041	***
Friend of friend ↔ friend of family	0.432	0.035	***	0.433	0.035	***
Data collection 1 ↔ analyze data 1	0.253	0.047	***	0.240	0.049	***
Analyze data 1 ↔ explain results 1	0.296	0.043	***	0.282	0.045	***
Data collection 5 ↔ analyze data 5	0.315	0.040	***	0.309	0.041	***
Analyze data 5 ↔ explain results 5	0.352	0.033	***	0.348	0.034	***
Model fit
N	1,420			1,420
LR χ² (df)	245.5 (156)			359.8 (210)
RMSEA	0.020			0.022
BIC	−887			−1,164

Note. LR = likelihood ratio; RMSEA = root mean square error of approximation; BIC = Bayesian information criterion. All statistical tests are two-tailed.

Indicates reverse-coded item.

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

Paths are estimated from each control to exogenous role-specific self-efficacy, prominence, and salience in Model 1, and in addition to global self-efficacy in Model 2. The first control variable, gender, reaches conventional levels of statistical significance only for its association with role-specific self-efficacy; women report 12 points lower role-specific self-efficacy than do men. Gender fails to influence the other constructs in the model. The second control, income, is associated with salience, although somewhat counterintuitively, negatively. Each unit increase in income is associated with a .07 point decrease in salience but is not associated with the other constructs. Neither control is altered when global self-efficacy is included in the model.

Race and ethnicity are included as a set of two dummy variables (Latino/a and other race/ethnicity) compared with a third (Black). Coefficients for the latter rise to conventional levels of statistical significance when predicting role-specific self-efficacy. Black students report .17 points higher role-specific self-efficacy than do students of other races and ethnicities, while Latino students fail to differ from other nonblack students. As these findings are in line with prior research indicating that blacks tend to score higher on self-esteem scales (which include measures of global self-efficacy) than whites (Bachman and O’Malley 1984), they will not be discussed further. Race and ethnicity fail to influence the other constructs.

Structural Model

The first model tests the first three hypotheses, focused on the relationship between role-specific self-efficacy, salience, and prominence. Prominence in the science student identity is measured at Time 2 as an outcome of role-specific self-efficacy at Time 1 (spring 2006). Salience of the student identity is included at Time 3 as a product of both role-specific self-efficacy at Time 1 and prominence at Time 2. Role-specific self-efficacy at Time 4 (spring 2008) is included as a product of role-specific self-efficacy at Time 1, prominence at Time 2, and salience at Time 3. In line with Ervin and Stryker’s conjecture, role-specific self-efficacy is hypothesized to be positively associated with prominence, though self-efficacy’s direct relationship with salience is theorized to be weak or nonexistent. Results show a causal chain from role-specific self-efficacy to prominence to salience and back to role-specific self-efficacy (see Figure 2). The first path, from exogenous role-specific self-efficacy to prominence, is positive and statistically significant (.34, p < .001). This finding supports Hypothesis 1.

The strength and significance of the previous two paths, from role-specific self-efficacy to prominence to salience, in combination with a lack of a direct effect of role-specific self-efficacy on salience, suggest that the effect of role-specific self-efficacy on salience is mediated by prominence. Mediation analyses (Sobel 1982) demonstrate that a large and statistically significant proportion, 81 percent, of the total effect of self-efficacy on salience is indirect through prominence (t = 7.17; p < .001). This finding supports Hypothesis 2.

Both of the science student identity variables, prominence and salience, have small but significant effects on role-specific self-efficacy at Time 4 (both equaling .10, p < .05). This finding supports Hypothesis 3, although the strength of the association is somewhat weaker than expected. The extended time period between salience (measured at Time 3 in spring 2007) and self-efficacy (measured at Time 4 in spring 2008) for this sample of students whose science student identities are being developed may have reduced the strength of this relationship.

The final two paths of Model 1 are the stability coefficient from role-specific self-efficacy at Time 1 to Time 4, which is large and statistically significant as hypothesized (.52, p < .001), and the path from prominence to salience, which is large and significant (.53, p < .001), in line with previous research. Although the likelihood ratio chi-square is highly statistically significant, Model 1 appears to fit these data well (RMSEA = .020) with a large negative value of BIC (−887).

The second model adds global self-efficacy, included at Time 2.⁹ In addition to the paths estimated in the previous model, this model also includes paths from exogenous role-specific self-efficacy to global self-efficacy, and from global self-efficacy to prominence and salience in the science student identity and science student role-specific self-efficacy.

Two paths are estimated between global and role-specific self-efficacy: from role-specific self-efficacy (at Time 1) to global self-efficacy, and from global self-efficacy to role-specific self-efficacy (at Time 4). The path from exogenous role-specific self-efficacy to global self-efficacy is of moderate size (.31) and is highly statistically significant (p < .001; see Figure 3). Global self-efficacy has small but highly statistically significant effects on role-specific self-efficacy at Time 4 (.12, p < .01). Thus, as predicted, exogenous role-specific self-efficacy influences global self-efficacy, which in turn influences role-specific self-efficacy as an outcome of the model. These results support Hypotheses 4 and 5.

Figure 3.

Structural model including identity salience, identity prominence, and global and role-specific self-efficacy.

Paths are estimated from global self-efficacy to both prominence and salience in the science student identity. The relationship between global self-efficacy and prominence is positive and moderate in size (.14, p < .001). This finding supports Hypothesis 6. Global self-efficacy, however, has no direct effect on salience (.04, p = .88). This large p value well exceeds Frick’s (1995) .50 minimum as a guideline for accepting the null hypothesis, thus supporting Hypothesis 7.

The remaining paths estimated are those from the prior model. The path from role-specific self-efficacy to prominence remains positive and large and is only negligibly attenuated with the addition of global self-efficacy (.30, p < .001). The path from prominence to salience is unchanged with the addition of global self-efficacy (.53, p < .001). However, one of the two paths from the identity constructs to role-specific self-efficacy at Time 4 is altered with the inclusion of global self-efficacy. The path coefficient from salience to role-specific self-efficacy at Time 4 is small but statistically significant (.10, p < .05), unchanged from Model 1. But the path from prominence to role-specific self-efficacy is no longer significant with global self-efficacy in the model. Thus, the linear causal chain previously noted, from role-specific self-efficacy to prominence to salience, might be better described as a loop or circular process: Higher role-specific self-efficacy yields increased prominence which promotes higher levels of salience that, in turn, generate higher levels of role-specific self-efficacy. Finally, the stability coefficient from role-specific self-efficacy at Time 1 to Time 4 remains large and statistically significant (.49, p < .001). Model 2 fits well (RMSEA = .022), with a large negative value of BIC (−1,164).

Discussion

This article began with an explication of two of the key concepts of identity theory. Prominence, the affective valuation of an identity, reflects the aspirational self. Identities high in the prominence hierarchy are those for which the individual cares deeply. Salience, the likelihood of performance of an identity, reflects the actual self. Identities high in the salience hierarchy are those the individual is very likely to enact. Following prior theoretical and empirical work, Brenner et al. (2014) used longitudinal data to demonstrate that prominence influences salience. It is the affective valuation of the identity that increases its likelihood to be enacted. In short, we are more likely to perform the identities that we value. While concordance between prominence and salience is common, these constructs need not agree. Disvalued identities may have a strong propensity for enactment and strongly valued identities may be relatively unlikely to be performed. Thus, although prominence and salience are highly related concepts, they are conceptually distinct and must be treated as such.

Although both salience and prominence attend to a specific focal role, its valuation, and propensity of performance, neither directly assesses consequent role performance nor do they provide an evaluative appraisement of the overall self. Thus, we ask how the individual’s confidence in his or her ability to perform a role affects the value placed on its related identity or the likelihood of identity-relevant role performance. We also examine the reverse relationship: How the affective valuation of the identity and its likelihood of performance affect one’s confidence in its performance. We focus on self-efficacy for two reasons. First, theoretical work integrating identity and self-esteem theories hypothesizes that high role-specific self-efficacy increases the identity prominence (Ervin and Stryker 2001). Second, empirical research links changes in or maintenance of self-efficacy to the verification and evaluation of role-identities specifically, rather than group-identities or person-identities (Burke and Stets 1999; Stets and Burke 2014). Thus, role-specific self-efficacy is hypothesized to be both precedent and product of the identity model.

These hypotheses are tested, fitting structural equation models to four waves of data from a longitudinal study of postsecondary science students. As the key exogenous and outcome variable, role-specific self-efficacy was measured at Time 1 and Time 4. Prominence and salience, as the middle links in the causal chain, were measured at Times 2 and 3, respectively. Findings support these hypotheses. Role-specific self-efficacy emerges as a moderate influence on identity prominence, which in turn has a strong effect on identity salience. Finally, salience influences, albeit modestly, role-specific self-efficacy.

Two other findings are notable. First, role-specific self-efficacy is not a direct cause of identity salience. Rather, their linkage is mediated by identity prominence. Second, with global self-efficacy in the model, the small effect of prominence on the outcome role-specific self-esteem is reduced to nonsignificance; the remaining effect of prominence on role-specific self-esteem is mediated through salience. This finding further informs the relationship between identity theory and the general theory of the self through the functioning of identity prominence: In short, the self-concept influences prominence, which in turn impacts salience. Thus, the connection between these concepts can be described as a circular process or loop: Confidence in one’s ability to perform a role leads the individual to more strongly value that role, which increases the likelihood that the individual will perform it which, returning to the beginning, increases one’s confidence in his or her ability to perform a role.

Our focus has been primarily on the specific level of self-efficacy given the strong link between this level and role-relevant behavior (Rosenberg et al. 1995). However, we have not neglected the global level. These findings include global self-efficacy, a component of global self-esteem, as an influence on the identity constructs and on outcome role-specific self-efficacy. We include global self-efficacy, rather than a measure of the full self-esteem construct, for theoretical consistency with concepts measured at the role-specific level. We allow bidirectional causation between its global and role-specific levels, as posited by Ervin and Stryker (2001), although we do so between waves, given our access to longitudinal data. Paths from global self-efficacy to the identity constructs are also estimated given prior research that suggests that global self-efficacy helps to maintain the identity and can operate as a buffer against potentially disconfirmatory evaluation (Cast and Burke 2002). In line with theory and prior research, the effect of role-specific on global self-efficacy is stronger than the reverse (Rosenberg et al. 1995). Global self-efficacy also provides a small positive influence on prominence. However, the modest size of this effect in comparison with the much larger effect of role-specific self-efficacy also provides support for the link between role-identities and role-specific self-efficacy posited by identity theorists (Burke and Stets 1999; Stets and Burke 2014).

Limitations and Future Directions

This analysis glosses over some of the key concepts involved in the functioning of identities and may use somewhat weak measures of others. Perhaps the most important omission is role performance and the process of identity verification. The individual’s confidence in his or her science student identity at Time 1 is at least partially a product of prior verification. Test scores, positive interactions with primary and secondary school teachers, and supportive words from parents all likely contributed to high self-efficacy at the initiation of data collection. Moreover, over the course of the study period, the science student takes courses; interacts with instructors, mentors, and other science students; and works in science labs. Each of these role performances, if successfully performed, verifies the science student identity (Stets et al., forthcoming). Identity verification potentially increases self-esteem (Cast and Burke 2002), in its self-worth, self-efficacy, and authenticity dimensions and potentially in both the global and role-specific facets. Measures used may not focus centrally on the science student identity, preferring instead to ask students about performing science tasks and identification with this aspirational identity. Future work may choose to more pointedly focus on the student aspect of this identity.

The starting point of the circular process—from role-specific self-esteem to identity prominence to identity salience and back to role-specific self-esteem, if indeed there is a single starting point at all—is not conclusively answered by these analyses. Our data begin measuring the identity process long into the respondents’ educational trajectory, and, presumably, long after these students discover their interests in studying science. Potentially, the process could start with prominence. These students may have decided much early in their education that they strongly value and desire to become a scientist, perhaps given the presence of a role model like a parent in a science career, a prominent scientist in the public eye, or even a fictional character engaged in science. This valuation increases the likelihood the student will enact his or her science identity, which leads to increased role-specific self-efficacy. Alternatively, a starting point at salience would provide an interesting twist on self-perception theory (Bem 1967). Students, while enrolled in a series of science classes as part of a college preparatory program (and thus highly likely to enact the science student identity), gain confidence in their science-related abilities, thus increasing the value placed on the science student identity (Lee 1998, 2002, 2005). Finally, blooming role-specific self-efficacy could be the starting point. For example, a high grade in seventh-grade science class could have initiated the confidence of young science students in their science abilities. While we cannot be certain where the process begins, or if its beginning is the same for each student, such questions are tertiary here. More important for the current project is establishing the relationship between these concepts, although future work with data beginning earlier in the educational trajectory could locate the genesis of the relationship.

Understanding the creation and the maintenance of the science student identity moves this research beyond testing an extension of identity theory. These empirical findings may be of great interest to educational researchers, policymakers, and other stakeholders concerned with encouraging students, especially women and underrepresented racial and ethnic minorities, to pursue STEM fields. In the first “White House Science Fair,” President Obama and Secretary of Education Duncan (The White House 2013) stressed these fields as an economic and strategic need, including “focusing on increasing opportunities for women, girls, and minorities.”

But the promotion of STEM fields is more complicated than encouraging “fun” science activities like a science fair at the White House. Anecdotal reports suggest that the nature of science education at the postsecondary level may discourage even those whose identities as science students have high prominence (see Drew 2011). In short, many students who excelled in high school math and science and who intend to pursue STEM majors become disenchanted by relatively low grades and a lack of engagement in their classes and leave for majors in the social sciences and humanities and professional programs like education and business. Empirical research bears out these anecdotes. Students earn relatively lower grades in STEM courses compared with courses in the social sciences and humanities, and these grades discourage students from choosing or continuing to pursue STEM majors (Rask 2010). In the context of findings of the current research, these lower grades reduce the role-related self-efficacy of a science identity, thereby nudging down both the prominence and salience of that identity, and pushing students toward majors in non-STEM disciplines. For students from underrepresented groups, this effect may be exacerbated.¹⁰

Conclusion

In its original formulation, Stryker’s (1968) identity theory was intentionally minimal, seeking to explain role-related behavior with two key concepts: identity salience and role commitments (Ervin and Stryker 2001). Since its minimalist conception, identity theory has expanded to further delineate its existing concepts (e.g., intensive and extensive commitments), include new concepts from proximate theories (e.g., identity prominence), and draw in the rich theoretical and empirical work on supportive processes (e.g., identity verification and discrepancy outcomes via the perceptual control system). Each development has expanded identity theory from its minimal origins in service of the original central purpose: explaining role-relevant behavior.

Unsurprisingly, these expanded identity constructs are highly related to the original constructs, including those to which attention was paid here: salience and prominence. While these concepts have often been conflated in social-psychological research (Callero 1985; Serpe 1987; Thoits 2012), their conceptual distinctiveness is important. Prominence, as the subjective value of an identity to the individual, is the affective link to the identity. Highly prominent identities are those that reflect the aspirational self. Salience, the likelihood of performance of an identity, is the cognitive link. Highly salient identities reflect the actual, situational self.

Findings from recent research, used as an assumption in the current research, affirm earlier theory and evidence of a strong causal relationship from prominence to salience (Brenner et al. 2014). However, this relationship is not deterministic. While a highly prominent identity is likely to be highly salient, these constructs need not be in concordance. A prominent identity may be unlikely to be enacted, and a salient identity may be disvalued.

In addition to refining and extending its core concepts, research has expanded identity theory into relevant neighboring areas, like emotion, measurement, and, in the present research, self-esteem. Testing and expanding on the theoretical path laid by Ervin and Stryker (2001), self-efficacy at both the global and role-specific levels is integrated as both a force and a product in the structural identity model. Role-specific and global self-efficacies emerged as moderate influences on identity prominence, which in turn has a strong effect on identity salience. Through salience, the role-identity returns its influence to role-specific self-efficacy. In sum, these constructs operate as a circular process: The higher persons’ confidence in their ability to perform a role, the more they will value that role, increasing the likelihood of role performance and, back to the beginning, increasing confidence in their ability to perform the role. As Ervin and Stryker (2001:29) argued, “successfully relating self-esteem (which includes self-efficacy) and identity theory is a step in achieving some degree of theoretical integration over a significant portion of the domain of social psychology.” Moreover, it is a helpful step toward the integration of the multiple domains and theories of the three social psychologies (House 1977; Oishi, Kesebir, and Snyder 2009; Stryker 1977).

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for research, authorship, and/or publication of this article: Funding for this research was provided by a grant from the National Institute of General Medical Science (R01-GM074316).

Notes

Author Biographies

Philip S. Brenner is assistant professor of sociology and senior research fellow in the Center for Survey Research at the University of Massachusetts Boston. His research examines social desirability bias and other measurement errors in interviewer- and self-administered surveys. Brenner is the author of over 20 peer-reviewed journal articles and book chapters, including recent publications in Sociological Methods and Research, Social Psychology Quarterly, Social Forces, and Public Opinion Quarterly.

Richard T. Serpe is professor and chair of sociology at Kent State University. His primary research continues to focus on the relationship between identity/self and social action. He is currently engaged in analysis of a national probability sample of multiple identities to assess the simultaneous relationship relationships of three identities on self, emotions and mental health. He continues engagement of a multi-year panel study funded by the National Institutes of Health assessing identity theory and changes in self and behavior, the link between intention and behavior, and the interrelations of structural identity and social identity.

Sheldon Stryker (born May 26, 1924; died May 4, 2016) was a distinguished professor of sociology emeritus at Indiana University. A career-long contributor to the literature on social psychology focusing in particular on the symbolic interactionist frame and on identity theory, he received the American Sociological Association DuBois Award for Lifetime Distinguished Contributions to Sociology, the ASA Social Psychology Section Cooley-Mead Award for Lifetime Contributions to Social Psychology, the Society for Symbolic Interaction George Herbert Mead Award for Lifetime Contributions, and the International Society for Self and Identity Lifetime Achievement Award.

References

Asencio

Emily K.

2013. “Self-Esteem, Reflected Appraisals, and Self-Views: Examining Criminal and Worker Identities.” Social Psychology Quarterly 76:291-313.

Asencio

Emily K.

Burke

Peter J.

2011. “Does Incarceration Change the Criminal Identity? A Synthesis of Labeling and Identity Theory Perspectives on Identity Change.” Sociological Perspectives 54:163–82.

Bachman

Jerald G.

O’Malley

Patrick M.

1984. “Black and White Differences in Self-esteem: Are They Affected by Response Styles?” American Journal of Sociology 90:624–39.

Bandura

Albert

. 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice Hall.

Bandura

Albert

. 1995. “Exercise of Personal and Collective Efficacy in Changing Societies.” Pp. 1–45 in Self-efficacy in Changing Societies, edited by A. Bandura. New York: Cambridge University Press.

Bem

Daryl J.

1967. “Self-perception: An Alternative Interpretation of Cognitive Dissonance Phenomena.” Psychological Review 74:183–200.

Brenner

Philip S.

2011. “Investigating the Biasing Effect of Identity in Self-reports of Socially Desirable Behavior.” Sociological Focus 44:55–75.

Brenner

Philip S.

2012. “Overreporting of Voting Participation as a Function of Identity.” Social Science Journal 49:421–29.

Brenner

Philip S.

Serpe

Richard T.

Stryker

Sheldon

. 2014. “The Causal Ordering of Prominence and Salience in Identity Theory: An Empirical Examination.” Social Psychology Quarterly 77:231–52.

10.

Browne

Michael W.

Cudeck

Robert

. 1993. “Alternative Ways of Assessing Model Fit.” Pp. 136–62 in Testing Structural Models, edited by Bollen

K. A.

Long.

J. S.

Newbury Park, CA: Sage Publications.

11.

Burke

Peter J.

1991. “Identity Processes and Social Stress.” American Sociological Review 56:836-849.

12.

Burke

Peter J.

Stets

Jan E.

1999. “Trust and Commitment through Self-verification.” Social Psychology Quarterly 62:347–66.

13.

Burke

Peter J.

Stets

Jan E.

2009. Identity Theory. New York: Oxford University Press.

14.

Callero

Peter L.

1985. “Role-Identity Salience.” Social Psychology Quarterly 48:203–14.

15.

Cast

Alicia

Burke

Peter J.

2002. “A Theory of Self-esteem.” Social Forces 80:1041–68.

16.

Chemers

Martin M.

2006. “Science Identity and Self-efficacy.” Unpublished manuscript, University of California, Santa Cruz, CA.

17.

Coopersmith

Stanley

. 1967. The Antecedents of Self-esteem. San Francisco, CA: W.H. Freeman.

18.

Crocker

Jennifer

. 2002. “Contingencies of Self-worth: Implications for Self-regulation and Psychological Vulnerability.” Self and Identity 1:143–49.

19.

Crocker

Jennifer

Knight

Katherine M.

2005. “Contingencies of Self-worth.” Current Directions in Psychological Science 14:200–203.

20.

Crocker

Jennifer

Wolfe

Connie T.

2001. “Contingencies of Self-worth.”Psychological Review 108:593–623.

21.

Drew

Christopher

. 2011. “Why Science Majors Change Their Mind.” The New York Times, November 4, p. ED16.

22.

Ervin

Laurie H.

Stryker

Sheldon

. 2001. “Theorizing the Relationship between Self-esteem and Identity.” Pp. 29–55 in Extending Self-esteem Theory and Research: Sociological and Psychological Currents, edited by Owens

T. J.

Stryker

Goodman

. New York: Cambridge University Press.

23.

Estrada

Mica

Hernandez

Paul R.

Woodcock

Anna

Schultz

P. Wesley

. 2011. “Toward a Model of Social Influence that Explains Minority Student Integration into the Scientific Community.” Journal of Educational Psychology 103:206–22.

24.

Frick

Robert W

. 1995. “Accepting the Null Hypothesis.” Memory and Cognition 23:132-138.

25.

Gecas

Viktor

. 1982. “The Self-Concept.” Annual Review of Sociology 8:1–33.

26.

Gecas

Viktor

. 1991. “The Self-Concept as a Basis for a Theory of Motivation.” Pp. 171–88 in The Self-Society Dynamic: Cognition, Emotion and Action, edited by Howard

J. A.

Callero

P. L.

New York: Cambridge University Press.

27.

Hernandez

Paul R.

Schultz

P.Wesley

Estrada

Mica

Woodcock

Anna

Chance

Randie C.

2013. “Sustaining Optimal Motivation: A Longitudinal Analysis of Interventions to Broaden Participation of Underrepresented Students in STEM.” Journal of Educational Psychology 105:89–107.

28.

Higher Education Research Institute. 2010. “Degrees of Success: Bachelor’s Degree Completion Rates among Initial STEM Majors.” Retrieved February 23, 2017 (http://www.heri.ucla.edu/nih/downloads/2010%20-%20Hurtado,%20Eagan,%20Chang%20-%20Degrees%20of%20Success.pdf).

29.

Hoelter

Jon W.

1986. “The Relationship Between Specific and Global Evaluations of the Self: A Comparison of Several Models.” Social Psychology Quarterly 49:129-141.

30.

Hoge

Dean R.

McCarthy

John D

. 1984. “Influence of Individual and Group Identity Salience in the Global Self-esteem of Youth.” Journal of Personality and Social Psychology 47:403–14.

31.

House

James S.

1977. “The Three Faces of Social Psychology.” Sociometry 40:161–77.

32.

James

William

. [1890] 1950. Principles of Psychology. New York: Dover.

33.

Lazarus

Richard S.

Folkman

Susan

. 1984. Stress, Appraisal, and Coping. New York: Springer.

34.

Lee

James Daniel

. 1998. “Which Kids Can ‘Become’ Scientists? Effects of Gender, Self-concepts and Perceptions of Scientists.” Social Psychology Quarterly 61:199–219.

35.

Lee

James Daniel

. 2002. “More than Ability: Gender and Personal Relationships Influence Science and Technology Involvement.” Sociology of Education 75:349–73.

36.

Lee

James Daniel

. 2005. “Do Girls Change More than Boys? Gender Differences and Similarities in the Impact of New Relationships on Identities and Behaviors.” Self and Identity 4:131–47.

37.

Loehlin

John C.

2004. Latent Variable Models: An Introduction to Factor, Path, and Structural Equation Analysis. New York: Lawrence Erlbaum.

38.

Marcus

Hazel

Kunda

Ziva

. 1986. “Stability and Malleability of Self-concept.” Journal of Personality and Social Psychology 51:858–66.

39.

Marsh

Herbert W.

1986. “Global Self-esteem: Its Relation to Specific Facets of Self-concept and Their Importance.” Journal of Personality and Social Psychology 51:1224–36.

40.

Marsh

Herbert W.

1990. “The Structure of Academic Self-concept: The Marsh/Shavelson Model.” Journal of Educational Psychology 82:623–36.

41.

Marsh

Herbert W.

Shavelson

Richard J.

1985. “Self-concept: Its Multifaceted, Hierarchical Structure.” Educational Psychologist 20:107–25.

42.

McCall

George John

Simmons

Jerry Laird

. 1978. Identities and Interactions: An Examination of Human Associations in Everyday Life. New York: Free Press.

43.

Merolla

David M.

Serpe

Richard T.

Stryker

Sheldon

Schultz

P.Wesley

. 2012. “Structural Precursors to Identity Processes: The Role of Proximate Social Structures.” Social Psychology Quarterly 75:149–72.

44.

National Academy of Sciences. 2011. “Expanding Underrepresented Minority Participation.” Retrieved February 23, 2017 (https://grants.nih.gov/training/minority_participation.pdf).

45.

National Science Foundation. 2015. “Women, Minorities, and Persons with Disabilities in Science and Engineering.” Retrieved February 23, 2017 (http://www.nsf.gov/statistics/2015/nsf15311/digest/nsf15311-digest.pdf).

46.

Nuttbrock

Larry

Freudinger

Patricia

. 1991. “Identity Salience and Mothering: A Test of Stryker’s Theory.” Social Psychology Quarterly 54:146–57.

47.

Oishi

Shigehiro

Kesebir

Selin

Snyder

Benjamin H.

2009. “Sociology: A Lost Connection in Social Psychology.” Personality and Social Psychology Review 13(4):334–53.

48.

Organization for Economic Co-operation and Development. 2012. “Programme for International Student Assessment (PISA) Results from PISA 2012: United States.” Retrieved February 23, 2017 (http://www.oecd.org/pisa/keyfindings/PISA-2012-results-US.pdf).

49.

Organization for Economic Co-operation and Development. 2013. “OECD Science, Technology and Industry Scoreboard 2013: Innovation for Growth.” doi:10.1787/sti_scoreboard-2013-en.

50.

Owens

Timothy J.

Samblanet

Sarah

. 2013. “Self and Self-concept.” Pp. 225–50 in Handbook of Social Psychology, 2nd ed., edited by DeLamater

Ward.

New York: Springer.

51.

Owens

Timothy J.

Serpe

Richard T.

2003. “The Role of Self-esteem in Family Identity Salience and Commitment among Blacks, Latinos, and Whites.” Pp 85–102 in Advances in Identity Theory and Research, edited by Burke

P. J.

Owens

T. J.

Serpe

R. T.

Thoits.

P. A.

New York: Kluwer.

52.

Pearlin

Leonard I.

Menaghan

Elizabeth G.

Lieberman

Morton A.

Mullan

Joseph T.

1981. “The Stress Process.” Journal of Health and Social Behavior 22:337–56.

53.

Pearlin

Leonard I.

Schooler

Carmi

. 1978. “The Structure of Coping.” Journal of Health and Social Behavior 19:2–21.

54.

Pelham

Brett W.

Swann

William B

. 1989. “From Self-conceptions to Self-worth: On the Sources and Structure of Global Self-esteem.” Journal of Personality and Social Psychology 57:672–80.

55.

Raftery

Adrian E.

1995. “Bayesian Model Selection in Social Research.” Sociological Methodology 25:111–63.

56.

Rask

Kevin

. 2010. “Attrition in STEM Fields at a Liberal Arts College: The Importance of Grades and Pre-collegiate Preferences.” Retrieved March 2014 (http://digitalcommons.ilr.cornell.edu/workingpapers/118/).

57.

Rosenberg

Morris

. 1965. Society and the Adolescent Self-image. Princeton, NJ: Princeton University Press.

58.

Rosenberg

Morris

. 1979. Conceiving the Self. New York: Basic Books.

59.

Rosenberg

Morris

Schooler

Carmi

Schoenbach

Carrie

Rosenberg

Florence

. 1995. “Global Self-esteem and Specific Self-esteem: Different Concepts, Different Outcomes.” American Sociological Review 60:141–56.

60.

Schultz

P.Wesley

Hernandez

Paul R.

Woodcock

Anna

Estrada

Mica

Chance

Randie

Aguilar

Maria

Serpe

Richard T.

2011. “Patching the Pipeline: Reducing Educational Disparities in the Sciences through Minority Training Programs.” Journal of Educational Policy Analysis 33:95–114.

61.

Serpe

Richard T.

1987. “Stability and Change in Self: A Structural Symbolic Interactionism Explanation.” Social Psychology Quarterly 50:44–55.

62.

Serpe

Richard T.

Stryker

Sheldon

. 1987. “The Construction of Self and the Reconstruction of Social Relationships.” Pp. 41–66 in Advances in Group Processes.Vol. 4, edited by Lawler

Edward J.

Markovsky

Barry

. Greenwich, CT: JAI.

63.

Serpe

Richard T.

Stryker

Sheldon

. 2011. “The Symbolic Interactionist Perspective and Identity Theory.” Pp. 225–48 in Handbook of Identity Theory and Research, edited by Schwartz

S. J.

Luyckx

Vignoles

V. L.

NY: Springer.

64.

Smith-Lovin

Lynn

. 2003. “Self, Identity, and Interaction in an Ecology of Identities.” Pp 167–78 in Advances in Identity Theory and Research, edited by Burke

P. J.

Owens

T. J.

Serpe

R. T.

Thoits

P. A.

New York: Kluwer Academic/Plenum Publisher.

65.

Sobel

Michael E.

1982. “Asymptotic Confidence Intervals for Indirect Effects in Structural Equation Models.” Sociological Methodology 13:290–312.

66.

Stets

Jan E.

Brenner

Philip S.

Burke

Peter J.

Serpe.

Richard T.

Forthcoming. “The Science Identity and Entering a Science Occupation.” Social Science Research. doi:10.1016/j.ssresearch.2016.10.016

67.

Stets

Jan E.

Burke

Peter J

. 2000. “Identity Theory and Social Identity Theory.” Social Psychology Quarterly 63:224-237.

68.

Stets

Jan E.

Burke

Peter J.

2005. “Identity-Verification, Control, and Aggression in Marriage.” Social Psychology Quarterly 68:160-178.

69.

Stets

Jan E.

Burke

Peter J

. 2014. “Self-esteem and Identities.” Sociological Perspectives 57:409–33. doi:10.1177/0731121414536141.

70.

Stets

Jan E.

Cast

Alicia D.

2008. “Resources and Identity Verification from an Identity Theory Perspective.” Sociological Perspectives 50:517-543.

71.

Stets

Jan E.

Serpe

Richard T

. 2013. “Identity Theory.” Pp. 31–60 in Handbook of Social Psychology, edited by DeLamater

Ward.

New York: Springer.

72.

Stryker

Sheldon

. 1968. “Identity Salience and Role Performance: The Relevance of Symbolic Interaction Theory for Family Research.” Journal of Marriage and Family 30:558–64.

73.

Stryker

Sheldon

. 1977. “Developments in ‘Two Social Psychologies’: Toward an Appreciation of Mutual Relevance.” Sociometry 40(2):145–60.

74.

Stryker

Sheldon.

[1980] 2003. Symbolic Interactionism: A Social Structural Version. Caldwell, NJ: Blackburn.

75.

Stryker

Sheldon

. 1989. “The Two Psychologies: Additional Thoughts.” Social Forces 68(1):45–54.

76.

Stryker

Sheldon

Burke

Peter J.

2000. “The Past, Present, and Future of Identity Theory.” Social Psychology Quarterly 63:284–97.

77.

Stryker

Sheldon

Serpe

Richard T.

1982. “Commitment, Identity Salience, and Role Behavior: Theory and Research Example.” Pp. 199–218 in Personality, Roles, and Social Behavior, edited by Ickes

Knowles.

E. S.

New York: Springer-Verlag.

78.

Stryker

Sheldon

Serpe

Richard T.

1994. “Identity Salience and Psychological Centrality: Equivalent, Overlapping, or Complimentary Concepts?” Social Psychology Quarterly 57:16–35.

79.

Swann

William B.

Jr. Chang-Schneider

Christine

McClarty

Katie Larsen

. 2007. “Do People’s Self-views Matter? Self-concept and Self-esteem in Everyday Life.” American Psychologist 62:84–94.

80.

Tafarodi

Romin W.

Milne

Alan B.

2002. “Decomposing Global Self-esteem.” Journal of Personality 70:443–83.

81.

The White House. 2013. “Third Annual White House Science Fair, April 2013.” Retrieved March 2014 (https://obamawhitehouse.archives.gov/issues/education/k-12/educate-innovate).

82.

Thoits

Peggy A.

2012. “Role-identity Salience, Purpose and Meaning in Life, and Well-being among Volunteers.” Social Psychology Quarterly 75:360–84.

83.

Trzesniewski

Kali H.

Donnellan

M. Brent

Robins

Richard W.

2003. “Stability of Self-esteem across the Life Span.” Journal of Personality and Social Psychology 84:204–20.

84.

Zajonc

Robert B.

Markus

Hazel R.

1984. “Affect and Cognition: The Hard Interface.” Pp. 73–102 in Emotions, Cognition and Behavior, edited by Izard

Kagan

Zajonc

R. B.

Cambridge: Cambridge University Press.