Definition

Personal control beliefs, also referred to as locus of control and personal mastery beliefs, reflect individuals’ beliefs regarding the extent to which they are able to control or influence outcomes. A wide variety of theorists have emphasized the importance of perceptions of personal control and suggested that the desire to control the world around us (i.e., the desire for behavior-event contingency or personal control) is a fundamental characteristic of human beings (Schultz et al, 1994; see also Haidt & Rodin, 1995; Rothbaum, Weisz and Snyder, 1982 for reviews). Reflecting these varied theoretical perspectives (as well as the extensive research interest in the concept of perceived control), the literature exhibits varying conceptualizations of "perceived control".

Most well known is the concept of "locus of control" which derived originally from Rotter’s social learning theory (Rotter, 1966) and which focuses on "beliefs that individuals hold regarding relationships between actions and outcomes" (Lefcourt, 1991). The earliest instrument developed to measure locus of control beliefs, the Rotter I-E Scale, focused largely on the distinction between belief in internal versus external loci of control. Later instruments, elaborated by Rotter, Lefcourt and others, included more specific assessments of beliefs about personal "internal" control contingencies but also control contingencies manifested by "powerful others" and (similar to the original "external" formulation) perceptions of non-contingency (i.e. "chance") (for review see Lefcourt, 1991). Existing literature on control beliefs in relation to both SES and health largely reflects the "internal vs. external" conceptualization with assessment of individuals in terms of the extent to which they see "control" as residing primarily in themselves versus elsewhere (i.e., in others or chance).

Recently, a "two-process model of perceived control" (Rothbaum et al, 1982) has been elaborated. The model, as presented by Rothbaum et al (1982) and others (Schultz et al, 1994), postulates that "the motivation to feel 'in control' may be expressed not only in behavior that is blatantly controlling (referred to as ‘primary control’) but also, subtly, in behavior that is not (referred to as ‘secondary control’)" (p. 7, Rothbaum et al, 1982). Whereas "primary control" reflects more directly controlling behaviors, "secondary control" reflects behavior that, while not directly controlling, is directed toward promoting a sense of control, not by altering the environment, but by altering oneself (e.g., ones values, priorities, behavior) so as to bring oneself in line with the environment. The central thesis of this formulation of perceived control is that "persons perceive and are motivated to obtain secondary control in many situations previously assumed to be characterized by perceived uncontrollability and an absence of motivation for control" (p.27, Rothbaum et al, 1982).

This two-process model of perceived control may be of particular interest with respect to issues of SES differences in control beliefs as Rothbaum et al (1982) postulate that "secondary control ... is particularly likely in cases of prolonged failure to obtain highly desired and important incentives, or cases in which the inability is perceived as permanent" (p.29, Rothbaum et al, 1982). One might hypothesize that this would be a more frequent scenario in lower SES circumstances. However, to date, there are no reported findings based on this two-process conceptualization of perceived control that explicitly relates primary and/or secondary control to either SES or health.

Other constructs related to personal control include powerlessness (Seeman, 1975), self-efficacy (Bandura, 1977), and sense of coherence (Antonovsky, 1984). Powerlessness is conceptualized in terms of an individual’s general perceptions of a lack of power (vs. "control"), encompassing elements of lack of autonomy, fatalism, and inefficacy (Seeman, 1991). The self-efficacy construct also differs from personal control in that self-efficacy beliefs or expectations focus on evaluations of one’s ability to accomplish certain behaviors or achieve certain outcomes (Bandura, 1977) whereas personal control expectancies related to judgements about whether actions can produce a given outcome (i.e., the extent to which a given outcome is controllable). Bandura differentiates self-efficacy from personal control, suggesting that whereas personal control beliefs focus on the question of whether one can control an outcome, self-efficacy beliefs focus on the evaluation of one’s ability to effectively perform the behaviors necessary to realize that outcome (Bandura, 1977). Sense of coherence has been defined as "a global orientation that expresses the extent to which one has a pervasive, enduring though dynamic feeling of confidence that one’s internal and external environments are predictable and that there is a high probability that things will work out as well as can reasonably be expected" (Antonovsky, 1979, 123) and, differs from personal control in that "the crucial issue is not whether power to determine such outcomes lies in our own hands or elsewhere. What is important is that the location of power is where it is legitimately suppose to be." (Antonovsky, 1979, p. 128).

Measurement

A variety of instruments have been developed to measure personal control beliefs, including (a) global assessments (e.g., the original Rotter I-E scale and Pearlin’s Personal Mastery scale (see Lefcourt, 1991 for details), (b) factorial measures that provide separate measures of beliefs regarding "personal control", "powerful others", and "chance" (e.g., Internality, Powerful Others, and Chance Scales or Spheres of Control; see Lefcourt, 1991 for review of measures); (c) domain specific measures (e.g., "Health Locus of Control" [Wallston et al, 1978 ], Marital Locus of Control [Miller et al, 1983], see Lefcourt, 1991 for more complete review of available scales).

Pearlin & Schooler’s "Personal Mastery Scale" has become perhaps the most widely used measure in health research. It consists of 7 items which are answered on a 4-point (strongly agree, agree, disagree, strongly disagree) scale and has been shown to exhibit reasonable internal reliability (Seeman, 1991) and good construct validity (see Pearlin et al, 1981). More extensive assessments are provided measures such as Levenson’s Internality, Powerful Others, and Chance Scales which includes three subscales, each comprised of eight items in a seven-point Likert format. These measures also exhibit good reliability and validity (Levenson, 1981). [See Lefcourt, 1991 for additional information on other measures of control and their psychometric properties.]

Self-efficacy beliefs have also been a focus of considerable health-related research. Measures of self-efficacy beliefs are largely targeted at beliefs about specific domains of behavior (e.g., exercise efficacy, memory performance: McAuley et al, 1993, Berry et al, 1989). However, a more general 9-item scale was developed by Rodin and colleagues for use in older populations (Rodin & McAvay, 1992). Items ask about perceived self-efficacy in the nine domains deemed most important to older adults and are answered, like the Pearlin scale, in terms of a 4-point (agree – disagree) format. Data from the MacArthur Successful Aging Study indicate reasonable internal reliability and suggest that two subscales can be derived, one measuring perceptions of interpersonal self-efficacy and the other measuring instrumental self-efficacy (Seeman, Rodin, Albert, 1993).

Evidence Linking SES and Perceived Control

Studies have shown a positive association between SES and locus of control beliefs: higher SES (e.g., higher income and/or education) being associated with more internal beliefs (Mirowsky & Ross, 1986; Downey & Moen, 1987; Cicirelli, 1980; Pincus & Callahan, 1994, 1995; Levenson, 1981; Marmot et al, 1997). Similar patterns of association are seen for related constructs such as personal mastery (Pearlin et al, 1981; Gecas, 1989; Lachman & Weaver, in press) and self-efficacy (Gecas, 1989). And, lower SES has been associated with greater powerlessness and anomie (Blauner, 1964; Mirowsky & Ross, 1986).

Research by Kohn, Schooler and colleagues (Kohn & Schooler, 1983), focusing on the impact of working conditions on psychological health, also points to the important and significant ways in which the work environment characteristics affect the development and persistence of personal control beliefs. Work setting characteristics such as environmental complexity and contingency (i.e., control over the process of one’s work) were found to promote the development and persistence of stronger personal agency/control beliefs and to result in enhancing intellectual functioning more generally (Kohn & Schooler, 1983). Studies of the effects of downward mobility with respect to employment status also highlight the negative impact of such experiences on personal control and efficacy beliefs (Pearlin et al, 1981; Gecas 1989). Data such as these suggest that social class differences in personal control beliefs may be importantly influenced by differences in the characteristics of work (and other environmental) settings likely inhabited by those of different social classes.

Control beliefs and health

Evidence linking control beliefs to health is mixed, with evidence for both more positive and more negative health outcomes associated with stronger perceptions of personal control. There is considerable evidence linking sense of control to better psychological health (Rodin, 1986; Rodin, Timko, Harris, 1985; Haidt & Rodin, 1995) as well as evidence of links to better physical health outcomes, including lower incidence of CHD (Karasek et al, 1982, Marmot et al, 1997), better self-rated health and functional status (Seeman & Seeman, 1983; Seeman & Lewis, 1995; Rodin & Langer, 1977), better maintenance of cognitive function (Seeman et al, 1993) and lower mortality risk (Seeman & Lewis, 1995; Rodin & Langer, 1977). However, there is also evidence suggesting that stronger control beliefs can be associated with poorer health outcomes under certain circumstances (Rodin, 1986; Seeman, 1991; Thompson et al, 1988).

Negative health outcomes have been hypothesized to be more likely to occur when there is incongruity between personal control beliefs and actual situational/environmental "control conditions" (Watson & Baumal, 1967; Thompson et al, 1988; Rothbaum et al, 1982; Evans et al , 1993). Stronger control beliefs would thus be predicted to result in poorer outcomes when there is a mis-match between beliefs and environmental contingencies. Support for this prediction can be found in research from both animal and human studies of physiologic reactivity to environmental control conditions where the greatest reactivity (e.g., increase in cardiovascular or neuroendocrine activity or reduction in immune function) is seen when there is incongruity reflecting a combination of general expectancies for control and actual situational "uncontrollablitiy" or difficulty in controlling outcomes (DeGood, 1975; Houston, 1972; Hokanson et al, 1971; Manuck et al, 1978; Sieber et al, 1992). Data from a study undertaken by the MacArthur Successful Aging Study provide further evidence of such effects, showing that men with strong personal control beliefs who perceive that they were NOT "in control" in a driving simulation challenge exhibited the greatest physiological reactivity. By contrast, individuals with similarly strong personal control beliefs who perceive themselves to be " in control" during the challenge situation exhibited the least reactivity (Seeman, unpublished data). Data such as these suggest that having strong internal control beliefs in situations which do not allow for such personal causation will tend to be detrimental in terms of physiologic activation and, if such a "person-environment" mis-match is relatively chronic, may actually result in increased pathophysiology. Known links between the Type A Behavior Pattern and increased risks for heart disease may be an example of such links. Type A’s have been shown to have a strong need for control (Strickland, 1978; Miller et al, 1985), to persist in attempts for control in laboratory situations (Miller et al, 1985; Strube & Werner, 1985) and to exhibit greater physiologic reactivity in the face of uncontrollable situations (Krantz, Glass & Snyder, 1974). Such persistence, in the face of external realities that limit or prevent actual control over outcomes, along with its accompanying physiologic reactivity may contribute to Type A’s increased risk for CHD. Personal control beliefs, however, may also contribute to CHD risk, independent of Type A behavior. The presence of stronger personal mastery beliefs, for example, has been found to be associated with greater coronary atherosclerosis independent of other known risk factors (Seeman, 1991). To the extent that such strong mastery beliefs may tend to promote unrealistic expectations for control, they may be associated with patterns of physiological arousal that promote the development of atherosclerosis.

The complexity of relationships between personal control beliefs and health is also indicated by recent evidence indicating significant SES differences in the patterns of association between such control beliefs and health outcomes. Using data from three national samples, Lachman & Weaver (in press) have demonstrated significant interactions of control beliefs with both education and income in relation to health and well-being. The basic pattern of associations indicated that while control beliefs were associated with more positive health outcomes in all SES groups, the differences in health outcomes associated with stronger versus weaker control beliefs were greater at lower levels of education and income. Indeed, among those with less education or income, those with strong control beliefs reported health outcomes comparable to those seen in higher SES groups for self-rated health, acute physical symptoms, depressive symptoms and life satisfaction.

Limitations

Current limitations relate to the relative lack of research linking personal control to physical health outcomes and the focus, to date, on primarily main effects models for the effects of personal control. As indicated by both laboratory and survey data, there may be important SES differences in the impacts of control beliefs on health status (Lachman & Weaver) as well as possible differences in the patterns of association between personal control and physiological reactivity, depending on environmental contingencies.

Network Usage

Measures of control are included in several databases currently being used for various network projects, including the MacArthur Successful Aging Study, MacArthur Mid-Life Survey, Whitehall, and studies of immune function (Cohen et al).