4 On Who Gets Sick and Why: The Role of Personality and Stress It has long been thought that personality and physical health are related From ancient theories of temperament, through early clinical descriptions of physical disorders, prescientific thinking drew a close association between personality attributes and various somatic disorders Several threads of systematic theory and research on the topic emerged following the birth of psychology and psychosomatic medicine Since the middle of the 20th century, interest in personality and health has intensified considerably It now represents a major focus of psychosocial research concerned with physical disease (Friedman, 1990) Potential points of contact between the personality and physical health domains are numerous Each, by itself, is a large and complex area of inquiry The personality field has undergone considerable expansion and differentiation over the past 50 years During the latter portion of that time period, there has been tension between two major pursuits: construction of a taxonomy of personality descriptors and development of an understanding of personality process (Cervone, 1991; Mischel & Shoda, 1994; Pervin, 1990) This debate reflects an important component of variation in assumptions and approaches within the personality field However, there are also wide differences in the views of investigators within each camp, and many issues in personality research that have implications for understanding physical health not map neatly onto the description/process dichotomy As this handbook will attest, the study of physical health and disease is a vast and diverse enterprise Many physical conditions contribute to morbidity, mortality, and poor quality of life, and any one condition poses several subproblems, including diagnosis, epidemiology, etiology, prevention, treatment, and rehabilitation As a result, the study of physical health and disease is a multidisciplinary endeavor, potentially involving investigators from several health-related fields, including psychologists interested in personality (Schwartz & Weiss, 1977) This chapter is concerned with one portion of the personality- health interface, namely, that involving personality attributes that are thought to have health-damaging consequences because they increase psychological stress or exacerbate its effects Like personality, stress has long been suspected of contributing to physical health problems Moreover, the personality and stress constructs complement one another in that each provides a means of explaining and elaborating the other's role in shaping human adaptation The concept of stress points to social and environmental factors outside the person that influence psychological well- being and physical health, and to psychological and physiological processes that mediate those effects The study of personality points to dispositions within the person that can account for individual differences in responses to a stressor, and to attributes and processes that explain temporal and crosssituational consistency in stress-related response patterns Thus, research that draws from both the personality and stress domains is more likely to provide a comprehensive understanding of psychosocial influences on physical health than does work in which one of these constructs is utilized to the exclusion of the other This chapter provides a discussion of conceptual issues, empirical findings, and methodological concerns that bear on the relations among personality, stress, and health It examines personality as a psychosocial risk factor for disease and as a moderator of psychological stress The review is selective in emphasizing research that supports associations between certain personality dispositions and both measures of physical disorder and markers of diseaserelated processes The focus is primarily on the relation between personality and disease promoting processes that involve direct, psychophysiological effects of stress However, some consideration is also given to behavioral factors that may mediate the health effects of stress-related personality attributes independently of, or in interaction with, psychophysiologic mechanisms Issues and problems that emerge from this discussion are highlighted in a final section that takes stock of available theory and empirical findings and points to some potentially fruitful directions for further study Who Gets Sick? Personality As a Risk Factor “Who gets sick?” is an epidemiological question that can only be answered by programmatic, prospective, multivariate research in which putative risk factors are evaluated with respect to their ability to predict objectively verified disease endpoints independently of potential confounds (Adler & Matthews, 1994) Personality represents but one of several psychosocial domains in which risk factors for physical disease have been sought, with other salient examples including psychological stress, social relationships, and health-related behaviors However, the conceptual and methodological principles that arise from a consideration of the health effects of personality are relevant to a wide range of possible psychosocial risk factors, and there is reason to believe that personality and other psychosocial factors related to health often interact with one another rather than operating independently This section begins by describing major conceptual features that distinguish personality from other psychological constructs, and by discussing the implications of these features for framing the question, “Who gets sick?” An overview is then provided of the numerous personality attributes that have been implicated as possibly influencing vulnerability to physical disease This section concludes with a discussion of those personality attributes for which the epidemiological evidence makes the strongest case for risk factor status Conceptual Elements of Personality The question of how best to define personality and related terms such as personology has received extensive consideration (for classic discussions see Allport, 1937, and Murray, 1938; for more recent treatments, see Mischel, 1968, and Pervin, 1990) These analyses are not reviewed here Instead, the discussion draws on previous work to provide a heuristic overview of some of the major conceptual elements of personality psychology This discussion is necessarily cursory, however, and the reader is urged to consult the sources already cited for more comprehensive coverage of these issues Individual Differences Individual differences refer to between-person variations in behavior In this context, “behavior” may be construed narrowly in terms of a single domain of psychological activity, or it may be defined broadly to include cognition, affect, motivation, overt action, and neurobiological activity Personality psychologists not share a single view of the nature of individual differences per se, or of the importance of any one domain of individual differences in particular Moreover, not all individual differences involve personality Nonetheless, in a general sense, personality and the study of individual differences are intimately related The relevance of individual difference dimensions to the development and course of physical health problems depends on their association with mechanisms involved in the etiology and pathogenesis of disease, or with processes that affect the detection, control, and outcome of physical disorders A rather wide range of individual difference constructs have been implicated as possible risk factors for physical illness The field is narrowed, somewhat, when it is limited to those areas of individual differences that involve personality Patterning in Behavior Much personality research may be distinguished from other areas of psychology by virtue of its focus on two specific forms of patterning in behavior, namely, temporal and cross-situational consistency It is the observed or hypothesized stability of individual differences over time and in different contexts that provides a rationale for inferring drives, motives, traits, cognitive styles, and other dispositional constructs employed in personality psychology Temporal and cross-situational consistency set personality attributes apart from other person factors, such as transient cognitive or emotional states, or highly situation-specific behavioral tendencies Of course, psychological states and individual behaviors can be reflective of enduring personality attributes, and may have significant effects on physical health regardless of such an association However, the nature of those effects and the mechanisms whereby they are mediated may at times differ from those involving personality (Cohen, Doyle, Skoner, Gwaltney, & Newsom, 1995; Scheier & Bridges, 1995) Personality is not the sole source of temporal and cross- situational consistency in behavior Enduring factors that exist outside individuals-such as occupation, economic conditions, and relations between ethnic groupsalso may contribute to regularities in a person's behavior Moreover, as argued from the standpoint of transactional theoretical orientations, the explanation of temporal and cross-situational consistency in behavior may defy a simple, analysis of variance like partitioning of person, situation, and person-by-situation interaction (Lazarus & Folkman, 1984) Instead, person and environment factors may reinforce and sustain one another in ways that make efforts to disentangle their independent contributions difficult or arbitrary Notwithstanding these complexities, the involvement of personality attributes in behavioral patterning has major implications for specifying the role of personality as a risk factor for physical disease The two forms of behavioral patterning associated with personality factors provide a theoretical basis for linkages to health damaging processes Temporal stability in a suspected personality risk factor may indicate a relationship to disease promoting mechanisms that develop gradually over time For example, as an enduring disposition, hostility may be associated with repeated activation of physiologic activity that contributes to slowly progressing disorders such as atherosclerosis (T W Smith, 1992) Crosssituational consistency may operate in a similar manner Consider conscientiousness, a trait that may be related to good health (Friedman et al., 1995) To the degree that conscientiousness involves a pattern of careful, prudent behavior that is displayed in a wide range of situations, the opportunity for the accumulation of risk reducing actions is increased Thus, the two forms of behavioral patterning that define personality attributes as distinct from other psychological factors are also important for their implications regarding associations with disease promoting processes Recent studies involving naturalistic observations have provided evidence of a third form of behavioral patterning that may have interesting implications for the interface between personality and health Mischel and colleagues (Mischel & Shoda, 1995; Shoda, Mischel, & Wright, 1994) demonstrated that individuals show consistent pgtterns of variability in their behavior across different situations For example, children in a residential summer camp reliably displayed higher levels of particular behaviors (e.g., verbal aggression) in some situations (e.g., being teased by a peer, being approached by a peer) than in others (e.g., being warned by an adult, being punished by an adult) These situation behavior profiles consist of stable, meaningful variations in behavior, but are treated as random error in the more traditional focus in personality, where behavior often is aggregated across situations that may not always be psychologically equivalent There may be similar consistencies in patterns of variation in behaviors that individuals display in situations that involve exposure to health risk Organization The term organization is frequently used by personality psychologists, although with more than one meaning In one usage, organization refers to the idea that personality is pervasive, involving the whole person as a unified, although highly complex, system This notion is similar in certain respects to self- regulation perspectives employed in health psychology and behavioral medicine (e.g., Carver & Scheier, 1981; Schwartz, 1979) A systems view of the person is integral to the multilevel, b&psycho-social model of health and disease (Engel, 1977), and also provides a framework within which to conceptualize processes whereby cognitive, affective, and other psychological systems may influence disease promoting mechanisms, a topic discussed later in this chapter In another usage, personality organization refers to the structure of interrelationships of personality descriptors Multivariate methods have generated evidence of hierarchical organization in which relatively specific tendencies (e.g., being talkative, enjoying parties) cluster together to form more general dispositions (e.g., sociability, sensation seeking), which in turn cluster together to form still more general dispositions (e.g., extraversion; Eysenck, 1967) There is growing consensus that at a certain level of abstraction personality organization may be described in terms of a taxonomy of five personality factors that have been labeled extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience (McCrae & Costa, 1985) This five-factor model provides a general framework for characterizing major dimensions of individual differences in personality Some of the traits that form the five-factor model, such as conscientiousness (Friedman et al., 1995) and neuroticism (Bolger & Zuckerman, 1995), have been investigated in relation to stress and health However, many personality variables of interest to health psychologists-such as Type A behavior (Matthews, 1982), hostility (Barefoot, Dodge, Peterson, Dahlstrom, & R Williams, 1989), optimism (Scheier & Carver, 1985), hardiness (Kobasa, 1979), and repressive coping (Weinberger, Schwartz, & Davidson, 1979)involve facets of more than one of the five-factor traits, or are defined in terms of attributes whose location within the five-factor taxonomy has yet to be determined Thus, the five-factor model remains to be more fully explored as a framework for organizing health-related personality attributes (T W Smith & P G Williams, 1992) Personality Structure Structure refers to neurobiological and/or psychological entities that are real and exist beneath the person's skin Personality structures must be distinguished from the individual difference patterns from which they are typically inferred A particular pattern of consistency in behavior across time and context may reflect an underlying personality structure, but the personality structure and the behavior pattern are conceptually distinct, with the former a putative cause of the latter The concept of psychological structure is illustrated by the notion that hostile behavior reflects a set of underlying attitudes characterized by cynicism and distrust (T W Smith, 1992) An example of neurobiological structure may be found in Krantz and Durel's (1983) proposal that the overt display of Type A behavior is, in part, a reflection of activity of the sympathetic nervous system Consideration of the notion of personality structure suggests that, with respect to the role of personality, the question “Who gets sick?” is really asking “What personality structures lead to disease?” The interviews, questionnaires, and other assessment tools used to measure personality necessarily provide only an indirect indication of the presence, content, and form of the underlying psychological structure that presumably gives rise to both the observable manifestations of the personality attribute and to the risk for physical disorders Moreover, the disease promoting structure for which the assessment device provides a marker may operate through mechanisms that not involve all observable manifestations of the personality attribute For example, it may be that cynical, distrusting attitudes need not be expressed in hostile behavior in order to increase coronary risk; it may suffice for those attitudes to operate through more subtle behavioral expressions to undermine the person's ability to develop and to maintain a supportive social network (T W Smith, 1992) Similarly, an underlying tendency toward hyperactivity of the sympathetic nervous system may be toxic to the coronary arteries regardless of whether it promotes the overt display of Type A behavior (Contrada, Krantz, & Hill, 1988) Although this problem is but a specific instance of the usual, third variable alternative to causal hypotheses, it is often overlooked in research concerning personality and health Context Corztext refers to factors outside the skin that may influence behavior Context is a multilevel concept Revenson (1990) referred to four broad contextual dimensions: situational (immediate stimulus configuration), interpersonal (social relationships, group affiliations), sociocultural (socioeconomic status, reference group), and temporal (life stage) Of particular relevance to health problems are situational factors whose interaction with personality gives rise to stress and influences the coping process (Lazarus, 1966) These interactions must be viewed within the framework of interpersonal relationships from which stressful situations may respect to heart disease, as a consequence of stress-related physiological mechanisms (Camey, Freedland, Rich, & Jaffe, 1995) There is also reason to believe that disengagement can influence health through effects on the immune system Kamen-Siegel, Rodin, Seligman, and Dwyer (1991) reported an association between the pessimistic attributional style and lower cell-mediated immunity More recently, Herbert and Cohen (1993) conducted a meta-analysis that provides evidence of a strong and consistent relation between clinically significant depression and three parameters of immune competence, namely, lower proliferative response of lymphocytes to mitogens, lower natural killer cell activity, and alterations in numbers of several types of white blood cells There was also evidence that depressed mood in nonclinical populations is associated with lower functional immunity The meta-analysis did not address the mechanisms that account for the linkage between depression and immunity One set of plausible mechanisms discussed by Herbert and Cohen (1993) involves the link between depression and neuroendocrine mechanisms mentioned earlier Behavioral Responses to Stressors Attributes falling in the disengagement cluster may produce health damaging effects as a consequence of behavioral responses to psychological stress For example, the pessimistic explanatory style has been associated with poor health habits, such as cigarette smoking, excessive alcohol use, and low exercise levels (Peterson, 1988), and with behaviors reflecting poor management of illness, such as failure to increase fluid intake and to get adequate bed rest (Peterson, Colvin, & Lin, 1992) Herbert and Cohen (1993) noted that altered immune function in depressives might reflect behavioral correlates of depression, such as inadequate sleep, low levels of exercise, poor diet, cigarette smoking, and alcohol and drug use Depression also has been associated with low levels of compliance with various medical regimens (e.g., Blumenthal, R S Williams, Wallace, R B Williams, & Needles, 1982; Eisen, Hanpeter, Kreuger, & Gard, 1987; Guiry, Conroy, Hickey, & Mulcahy, 1987) As with anger/hostility and emotional suppression/ repression, it is not clear whether these behaviors are best conceptualized as a reflection of stress-related processes Conceptual and Methodological Issues Research reviewed in this chapter suggests possible relationships between personality attributes falling in three domains -anger/hostility, emotional suppression/repression, and disengagement -and both physiological and behavioral pathways to physical disease that may involve responses to psychological stress However, support for this hypothesis is not entirely consistent Even with the focus in this chapter on research reporting positive associations, methodological problems and the small numbers of available studies provide reason to view the findings with caution Perhaps the most appropriate overall conclusion to draw at the present time is that the evidence supports the plausibility of the personality- stress-disease hypothesis.Research on personality, stress, or health poses many conceptual and methodological difficulties The additional challenges that emerge from efforts to establish linkages across these three domains can be daunting This section discusses some of the more salient issues and attempts to identify theoretical perspectives and methodological strategies that may prove fruitful in this endeavor These issues are grouped into two general categories, namely, those concerned with units of analysis in personality, and those concerned with establishing linkages with health-related processes Units of Analysis in Personality and Their Measurement Research suggests a number of observations regarding personality units that appear to warrant further examination as possible risk factors for physical disease The following are three of the more salient observations: Content: The tendency to experience negative emotions is strongly implicated as a personality domain that is relevant to stress and health Structure: The breadth and complexity of the behavior patterns associated with promising personality attributes remain unresolved Context relevance: Relations between health-related personality attributes and relevant factors in the social and physical environment have yet to be adequately explored Content Perhaps the clearest theme to emerge from this chapter is that all three promising personality risk factors involve negative emotion Thus, as pointed out previously (e.g., Contrada et al., 1990), descriptions of health damaging personality attributes have changed little since the theory of temperaments was put forth by Hippocrates and Galen This, observation agrees with conclusions based on previous reviews (e.g., Friedman & BoothKewley, 1987), and it is, perhaps, unsurprising, given long-standing and well-documented linkages between negative emotions and biological responses to stressors that have been implicated in the production of disease (Cannon, 1925; Mason, 1975) Several implications may be drawn from this observation One implication of the central role of emotion in health- related personality attributes is that it underscores the need for greater clarification of the statistical structure of affective trait measures This issue, and the related issue of statistical structure of state affect measures, have received considerable attention, but a consensus has yet to emerge The matter is often framed as a choice between two alternatives, one pointing to a set of five to seven qualitatively distinct affects (e.g., Ekman, 1992), and another to a two-factor space involving separable dimensions of positive and negative affect (e.g., Watson, Clark, & Tellegen, 1984) A structural model of affect also has been described that integrates these two perspectives (e.g., Diener & Fujita, 1995) Many of the studies reviewed here involved a single measure of a particular emotional trait; less frequently, a nonspecific measure (e.g., “neuroticism”) was used Both approaches permit only limited inferences regarding the contribution of emotion-related personality to health outcomes It would seem useful for future studies examining the relation between emotional traits and physical health to include measures representing both dimensional and discrete emotions perspectives Another suggestion that might be taken from the salience of emotion in health-related personality attributes, and one that appears to be promising, is the potential utility of working backward from experienced and overtly manifested emotion to the personality structures and processes that might constitute its determinants As noted earlier, many theories of emotion (e.g., Lazarus, 1991) emphasize cognitive processes, such as appraisal and problem representation, in the production of emotional states Of the personality attributes discussed in this chapter, the concept of pessimistic explanatory style is arguably most explicit in attending to cognitive process, given its focus on the attributions that individuals generate to account for negative experiences Cognitive factors also have received extensive examination in accounts of depression (e.g., Coyne & Gotlib, 1983), and Scheier and Bridges (1995) discussed the potential for processes involved in setting and evaluating progress toward goals to account for behavioral and emotional responses in individuals characterized by depression, fatalism, and pessimism Similar analyses of anger/hostility (Contrada et al., 1990) and emotional suppression (Newton & Contrada, 1992) also have been described A better understanding of the cognitive structures and processes underlying emotion-related dispositions should lead to improvements in the measurement of those constructs and to more detailed models of the interface between personality and stress Whatever the structure of emotional &aits and the cognitive determinants of personality-related emotional reactions, there is ample evidence to indicate that expressive behaviors play a major role in the subjective experience (Izard, 1990), physiologic concomitants (Siegman, 1994), and interpersonal consequences (Gottman & Levenson, 1986) of emotion However, although the items on conventional trait measures of negative emotions tap into expressive and interpersonal aspects of emotion, they not necessarily provide optimal assessments of those factors Conceptual analyses and multivariate methods have only recently begun to identify coherent facets of emotional traits such as anger/hostility (Barefoot et al., 1989; A H Buss & Perry, 1992; Contrada & Jussim, 1992; T Q Miller et al., 1995), a step that is prerequisite to efforts to establish the relative contributions of experiential versus expressive components It remains to be seen whether paper-andpencil measures will be adequate to the task, or whether it will be necessary to supplement these with direct observations of vocal, facial, and other behavioral responses to actual, emotion provoking situations (e.g., Dembroski et al., 1978; Seigman, 1994) Still another implication of the central role of emotion in health-related personality characteristics is the potential for confounding that arises in research that involves subjective measures of physical health (Watson & Pennebaker, 1989) Negative affectivity, or neuroticism-a tendency to experience and to express distressing emotions and other negative statesmay contribute both to scores on emotion-related personality attributes, and to responses on symptom checklists, use of health care, and medical diagnoses that reflect patient complaints The result may be a correlation that is spurious in the sense that it does not reflect a direct, physiologically mediated association between personality and the etiology or progression of a disease process On the other hand, such an association may be of considerable substantive interest as an instance in which personality contributes to illness or illness management through behaviorally mediated processes involving symptom detection and control Thus, further examination of negative affectivity in producing associations between personality and health should be directed toward elucidating the mechanisms, both artifactual and substantive, that may bring about such linkages (T W Smith & P G Williams, 1992) Structure The study of personality and health has at different times and for different purposes focused on both broadly and narrowly defined personality attributes This is well illustrated by the history of research in the Type A area What began as a multifaceted but coherent pattern consisting of competitiveness, achievement striving, time urgency, hostility, and vigorous speech and motor mannerisms was dissected into its constituent elements (Dembroski et al., 1978) One of those elements- hostility-was then identified as the most promising predictor of coronary disease Subsequently there have been efforts to resolve hostility into still more circumscribed facets Several componential schemes of hostility have been proposed, including one distinguishing between antagonistic interpersonal style, hostile content, and anger-in (Suls & Wan, 1993), and one involving cynicism, hostile affect, and hostile attributional style (Barefoot et al., 1989) Conversely, attempts also have been made to weave these components together into broader dimensions, such as anger experience and anger expression (e.g., Musante et al., 1989) At the same time, meta-analysis has been used to infer a generic, “disease- prone” personality that is broad and nonspecific in subsuming tendencies to experience other negative emotions in addition to anger (Friedman & Booth-Kewley, 1987) Although factor analysis, component analysis, and meta-analysis of existing measures will likely continue to prove useful tools, efforts to determine the appropriate level of generality for describing health-related attributes might also benefit from alternative approaches derived from social- cognitive approaches to personality (Cantor & Zirkel, 1990; Mischel & Shoda, 1995) Rather than seeking personality units solely within the structure of items and scales for measuring traits, various “middle-level” personality constructs have been defined in terms of their relevance to the individual's identity, goals, and life concerns Ewart and Kolodner (1991) followed such an approach in demonstrating the utility of a social competence interview for psychophysiological research An interview protocol that required subjects to describe and to re-experience a recurring life goal or challenge provided a means of eliciting cardiovascular activity that may reflect disease promoting processes Another example of an alternative to traditional structured personality tests is the assessment of self-discrepancies, or negative selfevaluations, the activation of which may provoke negative emotion states and associated reductions in immune competence (e.g., Strauman, Lemieux, & Coe, 1993) Personality structures reflected in life goals and standards for self- evaluation may lie closer to the personality-stress interface than more traditional, traitlike dispositions because their contents are defined by the individual, rather than by the investigator, and may have more explicit links to the social-environmental context from which stressors often emanate At the same time, it should be possible to establish associations between these cognitive-motivational units and more traditional personality constructs, especially where the latter are well- established, major dimensions of personality, or are firmly tied to health or to health-related processes (T W Smith & P G Williams, 1992) An issue related to the question of breadth concerns the structural complexity of health-related personality attributes Type A, Type C, and hardiness are explicitly multifaceted but, at least in some formulations, the facets were thought to be highly interrelated, yielding single, albeit broad constructs By contrast, the study of emotional suppression/ repression suggests that it may be profitable to conceptualize health-related personality attributes as combinations of two or more relatively distinct dispositions Both the repressive coping construct (Weinberger et al., 1979), and the concept of defensive hostility (Helmers et al., 1995; Jamner et al., 1991) are defined in terms of two, substantially independent personality factors, an emotional trait (anxiety or hostility) and a defensive or avoidant orientation toward negative emotion (as reflected in Marlowe-Crowne social desirability scores) Another example is the inhibited power motivation construct, which involves high power motivation, low affiliation motivation, and high activity inhibition (Jemmott, 1987) Beyond these few examples, very little is known about the health implications of particular combinations of distinctive personality attributes, and multicomponent models of personality dispositions raise conceptual and analytic issues that have yet to be resolved (Carver, 1989) Context Relevance Personality measures that have proved useful in identifying health damaging attributes differ in the way they represent contextual factors Instruments such as the Cook-Medley Hostility scale, the Marlowe-Crowne Social Desirability scale, and the Life Orientation Test (a measure of trait optimism) (Scheier & Carver, 1985) contain items that refer to or imply general types of situations, for example, occasions that might stimulate anger or aggression, situations that raise questions about the respondents' or others' pro-social motives, or possible future scenarios involving positive or negative outcomes By contrast, the attributional style questionnaire (a measure of pessimistic explanatory style) (Peterson, 1988) presents the subject with specific but hypothetical situations involving various negative events The Structured Interview (used to assess global Type A and various anger-related attributes) (Dembroski et al., 1978) also contains questions that refer to hypothetical situations, but differs from questionnaire approaches in that it primarily is used to create a real situation in which to observe directly the respondent's behavioral response to social challenge To the degree that the display of personality dispositions is contingent on the presence of conceptually congruent situation factors (Mischel & Shoda, 1995), the measurement of health-related personality attributes might benefit from more systematic attention to situational parameters contained in assessment devices The issue of context extends beyond the content of questionnaires and interviews Very little is known about the manifestations of many healthrelated personality dispositions in naturalistic settings, although technological developments may stimulate efforts to address this gap in knowledge (Stone & Shiffman, 1994) Naturalistic research efforts guided by Mischel and Shoda's (1995) notion of situation-behavior relations might reveal individual consistencies in the use of certain coping strategies to manage some types of types of stressors, but not others, or to engage in health damaging behaviors under certain psychological conditions, but not others The generalizability of personality-health associations across social and cultural categories also warrants much more attention Findings involving even the fairly well-studied anger/hostility constructs have raised questions about gender differences in the pathogenicity of attributes falling in this domain (Stoney & Engebretson, 1994) Much remains to be explored concerning the interplay between health-related personality attributes and gender, ethnicity, and socioeconomic status (Anderson & Armstead, 1995) Personality and Health-Damaging Processes Research reviewed above also suggests a number of observations regarding the causal analysis of personality-disease linkages Two of these are discussed below: (I) Psychophysiological advances: Technological advances permit more penetrating examinations of mechanisms linking personality to stress-related physiological activity than are available in the published literature; (2) Health-related behavior: Although direct, psychophysiological pathways to illness may be thought of as more closely associated with the health effects of personality and stress, behavioral pathways may warrant greater consideration Psychophysiological Advances The application of more sophisticated psychophysiological methods would advance knowledge concerning personality and health in a number of ways Within the laboratory, the availability of techniques such as spectral analysis, which distinguishes between sympathetic and parasympathetic nervous system influences on cardiac activity, makes possible a more fine-grained analysis of the autonomic control of cardiac responses to stressors (e.g., Kamada, Miyake, Kumashiro, Monou, & Inoue, 1992) Similarly, impedance cardiography can be used to determine the relative contributions of myocardial and vascular processes to stress-related blood pressure elevations (Kamarck, Jennings, Pogue-Geile, & Manuck, 1994) Refinements in these and other psychophysiologic methods permit the examination of neurogenic and hemodynamic mechanisms operating in healthy individuals Still other techniques have become available to measure stress-induced physiological changes in coronary patients, such as myocardial ischemia and abnormalities in myocardial contraction (e.g., Helmers et al., 1995) Thus, in addition to focusing on psychophysiologic processes thought to be involved in the development of cardiovascular disorders, it is now possible to examine personality and stress in relation to psychophysiologic processes that may be responsible for the precipitation of clinical disease states at later stages of those disorders Outside the laboratory, the availability of instruments for the ambulatory monitoring of physiologic activity can generate evidence bearing on the relationship between personality and stress-related cardiovascular responses in the naturalistic setting (e.g., Jamner et al., 1991) As with laboratory methods, ambulatory methods also permit the assessment of clinically relevant events in patients with disease (Helmers et al., 1995) Evidence of this sort-involving associations linking personality, stress, and disease in naturalistic settings-is critical to arguments that psychophysiologic correlates of personality dispositions are relevant to processes that culminate in physical disease (Krantz & Manuck, 1984) In addition to the foregoing examples from cardiovascular psychophysiology, it is possible to point to increasing sophistication in methods for assessing a range of parameters that reflect biological responses to psychological stress, including various measures of immunity, neuroendocrine activity, neurotransmitter and peptide receptor function, and metabolism (e.g., Herbert & Cohen, 1993; Schneiderman, Weiss, & Kaufman, 1989) Future research examining personality in relation to physiologic aspects of the stress response must exploit these developments if it is to add detail to current models concerning the psychosocial causation of physical disease and generate data that increase confidence in the validity of those models If so, it will be possible to hold the next generation of studies in this area to a criterion of persuasiveness, rather than mere biological plausibility Personality and Health-Related Behavior Although the focus in this chapter has been on personality factors thought to influence health directly, through psychophysiological pathways, there appears to be a sufficient basis for examining these attributes in relation to behavioral pathways as well In some instances, evidence for associations between suspected personality risk factors and pathogenic behaviors was generated through efforts to introduce statistical controls for factors that might confound personality-disease relations deriving from physiological mechanisms In other cases, the examination of associations between personality and health damaging behaviors appears to have been exploratory, or based on the general negative or maladaptive connotation of a personality descriptor Research based on explicit theory linking personality attributes to health behavior or reactions to illness (e.g., S M Miller et al., 1996) might be expected to yield more consistent and robust findings In addition to the benefits of a more theoretical approach to work in this area, it may be useful to give greater consideration to the proper role of personality in this research Some research appears to reflect a premature focus on one particular personality disposition, to the exclusion of others, and to rely too heavily on it alone in attempting to predict measures of health-related behavioral processes Other research is based on a normative health behavior model in which personality is included in a list of relevant background variables in the absence of a systematic examination of potential points of connection between personality and the key elements of the behavioral model If research guided by psychophysiological models has provided any lesson at all for more behavioral approaches, it is that personality is neither an overriding factor in physical illness nor an inconsequential one Conclusions The major constructs and research approaches used in the study of personality, stress, and disease have evolved only to a modest degree in recent years Current concerns about lack of theory, measurement problems, and other issues not differ dramatically from those raised in reviews published decades ago Nonetheless, although the robustness of beliefs in the importance of stress-related personality attributes is not always matched by effect sizes or theoretical sophistication, research in this area has progressed Assertions that once were based on a single, prospective epidemiologic investigation, or on a handful of psychophysiologic experiments, are now based on consistent patterns of findings from large numbers of studies There is a greater tendency to select for study types of stressors that appear psychologically relevant to the particular personality attribute under investigation Personality factors that were once vaguely conceptualized as capable of exacerbating stress responses are now embedded in detailed theoretical frameworks positing specific stress promoting processes In short, it would seem that the study of personality, stress, and disease is alive and reasonably well