Teaching Tip Sheet: Self-Efficacy

Self-efficacy refers to an individual's belief in his or her capacity to execute behaviors necessary to produce specific performance attainments (Bandura, 1977, 1986, 1997). Self-efficacy reflects confidence in the ability to exert control over one's own motivation, behavior, and social environment. These cognitive self-evaluations influence all manner of human experience, including the goals for which people strive, the amount of energy expended toward goal achievement, and likelihood of attaining particular levels of behavioral performance. Unlike traditional psychological constructs, self-efficacy beliefs are hypothesized to vary depending on the domain of functioning and circumstances surrounding the occurrence of behavior.

Self-Efficacy Theory (SET) has had considerable influence on research, education, and clinical practice. In the field of health psychology, for example, the construct of self-efficacy has been applied to behaviors as diverse as:

• Self-management of chronic disease

• Smoking cessation

• Alcohol use

• Eating

• Pain control

• Exercise

A search of PsycINFO® for the last five years lists more investigations of self-efficacy than of locus of control, sense of coherence, learned helplessness, and other popular constructs. The intuitive appeal of self-efficacy theory in these health-related domains have encouraged its use in research addressing the prevention of HIV.

Self-efficacy is assessed frequently in HIV prevention research but there has been mixed evidence for the relationship between self-efficacy (for safer sex) and sexual risk behavior (Forsyth & Carey, 1998). This pattern of findings might be interpreted to mean that self-efficacy is irrelevant to the study of HIV-related risk behavior, and perhaps other health-related behaviors. However, it is likely that such a conclusion would be premature.

What HIV research has taught us, however, is that reliable and valid measurement of self-efficacy is very challenging. Instruments intended to assess self-efficacy for safer behavior often measure constructs other than self-efficacy. For example, investigators have used measures with content reflecting HIV-related knowledge, behavioral intentions, attitudes toward safer sex behaviors, perceptions of the difficulty of enacting risk reducing behaviors, perceived helplessness, perceived vulnerability to HIV infection, acceptance of sexuality, and other unique operationalizations (Forsyth & Carey, 1998). Imprecise operationalizations of self-efficacy beliefs obscure what is being measured, and attenuate bivariate relationships.

HIV research has also called attention to the limited evidence for the validity of the self-efficacy measures. Brafford and Beck (1991) reported discriminative evidence for the validity of the Condom Use Self-Efficacy Scale (CUSES) by demonstrating that scores distinguish:

(a) consistent, inconsistent, and non-condom users;

(b) sexually experienced and inexperienced participants; and,

(c) participants who did or did not report a history of sexually transmitted disease infection.

In a series of subsequent studies, investigations have corroborated the discriminative validity of CUSES scores (Brien et al., 1994; Mahoney et al., 1995). In each of these studies, self-efficacy ratings distinguished college students on the basis of self-reported consistency of condom use. Considerably less attention has been paid to predictive and construct evidence. A related problem is that attempts to evaluate self-efficacy measures have been limited by validation methods that employ single assessment strategy. Such investigations are unable to demonstrate that observed correlations do not result primarily from shared method variance. This research reminds us that Campbell and Fiske's (1959) recommendations for using multitrait-multimethod matrices for evaluation of convergent and discriminant evidence are needed.

HIV research also reminds us that conceptual clarity about the nature of efficacy beliefs is critical to the development of measures that are consistent with SET. Items intended to assess efficacy beliefs should be operationalized so that they:

(a) assess beliefs in the capacity to

(b) enact domain-specific behaviors in

(c) circumstances that present gradations of challenge.

Studies of HIV prevention frequently do not achieve this level of precision but noteworthy exceptions exist. For example, Basen-Engquist's (1992) multi-item measure of self-efficacy for negotiating safer sex and condom use meets each criterion. The measure assesses students' beliefs in their capacities to enact risk reducing behavior (e.g., initiating a discussion of condom use) across a number of circumstances (e.g., discussing safer sex with a new partner prior to intercourse). This measure also used elicitation-based scenarios to provide details about situational demands that could influence the level and strength of efficacy beliefs. [Use of such elicitation (qualitative) research in advance of quantitative investigations reflects another contribution of HIV research to health-behavior research, in general.]

In addition to these fundamental measurement concerns regarding self-efficacy, HIV research has also demonstrated that methodological issues may attenuate observed efficacy― behavior relationships. Self-efficacy risk reduction associations may be influenced by ceiling effects, response bias, and measurement error associated with self-report measures of risk behavior (Weinhardt et al., 1998). A consistent finding in HIV prevention research is that self-efficacy scores tend to be negatively skewed. In response to inquiries about perceived capabilities, respondents often report being highly efficacious to enact risk-reducing behaviors. This response tendency may lead to censored distributions wherein a considerable proportion of the sample yields maximum self-efficacy scores. One explanation for these ceiling effects is that efficacy measures do not contain sufficient levels of challenge relevant to the target sample (Bandura, 1997). In the absence of contextual cues, responses may reflect performances in "best case" scenarios that yield maximum self-efficacy scores. These responses will obscure real differences between respondents. In addition, scoring protocols that restrict the range of possible responses may also produce truncated data. The resulting lack of sensitivity to differences in self-efficacy limit predictions of behavioral performance. Thus, incorporating sufficient gradations of challenge in items and sufficiently wide response intervals is critical to the development of sensitive self-efficacy measures.

An additional explanation for ceiling effects is that efficacy scores may be influenced by response bias. That is, research participants may respond in ways that reflect well of them. Traditional psychological assessment, which advances a trait conceptualization of social desirability responding, has been adopted in HIV prevention research. Not surprisingly, this approach has revealed no relationship between socially desirability bias and efficacy beliefs (e.g., Forsyth et al., 1997). One limitation of these findings is that investigators attempted to predict dynamic efficacy beliefs from items reflecting stable personality traits, with the latter having no clear relevance to the HIV domain. These traditional measures of socially desirable responding treat assessment items as signs of a larger construct, ignoring the reality that behaviors conferring risk for HIV infection are uniquely stigmatizing. Failure to find significant correlations among social desirability, self-efficacy, and HIV risk behavior may be attributed to incongruencies inherent in the assessment. Participants may present in socially acceptable ways when asked about HIV-risk behaviors, but do so in ways that are not detected by trait measures of presentation bias. Just as risky sexual behaviors may be under-reported, beliefs like self-efficacy for risk reducing behaviors may be over-reported. The assessment of response bias in the context of self-efficacy research warrants increased attention.

Help students to understand the differences among constructs from related social-cognitive theories (e.g., self-efficacy, outcome expectancies, behavioral intentions, behavioral difficulty, self-esteem, optimism, etc.).

Encourage students to develop a measure of self-efficacy for any health-related behavior that avoids the confounding of self-efficacy with these other constructs. If the health behavior is socially stigmatized (e.g., sexual behavior, illegal drug use) or if social norms suggest that one should engage frequently in a behavior (e.g., exercise), discuss how social desirability response biases might inflate self-efficacy scores.

• Discuss measurement (e.g., scale construction) and statistical (e.g., transformation of data) solutions to such problems.

• Encourage students to develop methods to assemble evidence for the validity of their self-efficacy measure.

• Help students to design an intervention program that will enhance self-efficacy, and a research design to measure changes in self-efficacy and whether these changes alter risky behaviors.

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191-215.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.

Bandura, A. (1997). Self-Efficacy: The exercise of control. New York, NY: W. H. Freeman.

Basen-Engquist, K. (1992). Psychosocial predictors of "safer sex" behaviors in young adults. AIDS Education and Prevention, 4(2), 120-134.

Brafford, L. J., & Beck, K. H. (1991). Development and validation of a condom self-efficacy scale for college students. Journal of American College Health, 39(5), 219-225.

Brien, T. M., Thombs, D. L., Mahoney, C. A., & Wallnau, L. (1994). Dimensions of self-efficacy among three distinct groups of condom users. Journal of American College Health, 42(4), 167-174.

Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105.

Forsyth, A. D., & Carey, M. P. (in press). Problems in the measurement of self-efficacy: Review, critique, and recommendations. Health Psychology.

Forsyth, A. D., Carey, M. P., & Fuqua, R. W. (1997). Evaluation of the validity of Condom Use Self-Efficacy Scale (CUSES) in young men using two behavioral simulations. Health Psychology, 16(2), 175-178.

Mahoney, C. A., Thombs, D. L., & Ford, O. J. (1995). Health belief and self-efficacy models: Their utility in explaining college student condom use. AIDS Education and Prevention, 7(1), 32-49.

Weinhardt, L. S., Forsyth, A. D., Carey, M. P., Jaworski, B., & Durant, L. (1998). Reliability and validity of self-report measures of HIV-related sexual behavior: Progress since 1990 and recommendations for research and practice. Archives of Sexual Behavior, 27, 155-180.

Michael P. Carey, PhD and Andrew D. Forsyth
Department of Psychology, Syracuse University