In almost all cancer screening promotion studies, the conceptual framework (if specified) includes some combination of the constructs of intention, self-efficacy, perceived benefit, perceived susceptibility, and/or subjective norms, with intention commonly regarded as the most important determinant of screening behavior (Gochman, 1997
). The theories that contain these constructs are the health belief model (Becker, 1974
), theory of reasoned action, theory of planned behavior (TPB; Ajzen, 1991
), trans-theoretical model (Prochaska & Velicer, 1997
), and health action process approach (Schwarzer, 1992
). A common thread throughout these models is an assumption that human action is guided by
beliefs about likely outcomes of the behavior and the evaluations of these outcomes (behavioral beliefs), beliefs about the normative expectations of others and motivation to comply with these expectations (normative beliefs), and beliefs about the presence of factors that may facilitate or impede performance of the behavior and the perceived power of these factors (control beliefs). (Ajzen, 2005
, p. 135)
Together, these beliefs are said to lead to the formation of a behavioral intention
or motivation to engage in a particular behavior. For example, according to the TPB, one of the dominant health behavior models (Ogden, 2003
), a woman will more likely express the intention to be screened if she holds favorable views about an action (e.g., mammography), perceives her significant others to view mammography positively, and perceives herself to have control over obtaining a mammogram. In most of the theories, in the absence of actual barriers (e.g., lack of insurance) or perceived barriers, having such an intention in itself means that the behavior will almost certainly be carried out, thus treating intention as a more powerful predictor of behavior than attitudes, the positive or negative perceptions one holds about the behavior (Ajzen, 1991
). In fact, intention is regarded as so predictive of screening that many studies have used this construct as the main outcome when actual screening cannot be measured, including studies targeting women from diverse ethnic backgrounds (e.g., Ham, 2005
; Levy-Storms & Wallace, 2003
; Valdez, Banerjee, Ackerson, & Fernandez, 2002
A notable feature of the above theories is an exclusive focus on accessible beliefs, that is, beliefs that individuals are consciously aware of and can report on in answering survey questions. This is as opposed to the core beliefs described in social psychology (i.e., deep-seated automatic thoughts or cognitions) that are held at the level of unconsciousness (Hobbis & Sutton, 2005
). It also differs from social science conceptions whereby people respond both directly and indirectly to situations using culturally derived meanings as a basis for their actions or practices, with some meanings arising from sources outside of individual awareness (Bourdieu, 1990
; Giddens, 1984
). Fishbein and Ajzen (2005)
dismissed such concerns by indicating that these distal or “background factors” (including social status, level of education, and presumably culture) influence intention and behavior only indirectly. These authors asserted that because these influences occur via normative, control, and behavioral beliefs (proximal factors), measurement and intervention restricted to this realm are sufficient and appropriate for understanding and improving behavior.
The normative beliefs component (subjective norms) has been the subject of more debate than the other four constructs under investigation in this study because of overall weak performance in predicting intention (e.g., its theorized function in the TPB) compared with attitudes and control beliefs. In general, there is agreement that the normative component is important but that measurement has been faulty (Armitage & Conner, 2001
). One major aspect of this faulty measurement has been the decision to focus many research studies on a limited set of social roles or people who are presumed to be significant others—such as mothers, husbands, and sisters. Yet others may be strong influences on women’s behavior as demonstrated, for example, by Washington, Burke, Joseph, Guerra, and Pasick (2009)
in this volume and by a study of mammography use among rural women by Steele and Porche (2005)
. The latter, like many others, noted that women can be strongly influenced to obtain mammograms by individuals outside the circle of family and friends such as lay helpers (e.g., Earp, 2002
) or breast cancer survivors (Erwin, Spatz, Stotts, & Hollenberg, 1999
). Yet the standard subjective norms measures fail to capture the influence of these apparently powerful referents. For example, in our own study of mammography among women from five ethnic groups, Stewart, Rakowski, and Pasick (2009)
found cross-sectional associations between recent screening and belief in annual mammography by some referent groups (best friend and the category “most people important to you”) that reached statistical significance for some but not all ethnic groups. Also, recent screening was associated with trying to comply with a belief in annual mammography again for some referent groups (respondent’s sister, doctor, people important to respondent) for the sample overall and for one’s sister and doctor, and only among Latinas for most people important to them.
Studies to test the theorized associations among the above constructs have rarely included ethnically diverse samples, calling into question the extent to which the theories and their constructs can be relied on to alter behaviors associated with health disparities. The findings among studies that have explored construct or theory appropriateness in communities of color or those who are underserved often have not supported the empirical adequacy of the models for behaviors such as Pap smear testing (Jennings-Dozier, 1999
) and exercise (Trost et al., 1999
). In our multiethnic cohort study of mammography mentioned above, Stewart et al. (2009)
found an association between intention at baseline and recent mammography 2 years later that was significant only among White women (odds ratio = 5.0, 95% confidence interval = 2.4, 10), with a statistically significant race/ethnicity interaction (p
= .02). Thus, intention strongly predicted future screening, but only for White women and not for the African Americans, Chinese, Filipinas, or Latinas in this sample. Because this is the only longitudinal cancer screening study we are aware of that involves multiple race/ethnic groups and languages, this finding is important in light of the assumption that intention is an outcome that is treated as nearly equivalent to the practice of screening itself. Yet it may not have the predictive value that has been observed in mainstream populations, suggesting the possibility of ethnic differences in the meaning or relevance of the construct or the validity of the item.
Our analysis of subjective norms found screening to be more strongly associated with normative beliefs than with motivation to comply. Women who reported that their best friend, sister, mother, husband, doctor, or important people (i.e., all the referents measured) believed in annual mammography had about twice the odds of getting regular mammograms than those without such influences. For motivation to comply, associations were found only for some of the referent others. Screening was associated with trying to act on the beliefs of one’s sister or doctor but not of one’s best friend, mother, or husband. Associations with screening did not differ by race/ethnicity for subjective norms regarding some of the specific people (e.g., mother, husband). However, there was a differential effect of subjective norms regarding “most people important to you” by race/ethnicity, which may be because of the classification of different types of people as important.