James (1904) proposes that our self-feeling depends entirely on what we back ourselves to be and do Our self feeling, he says, "is determined by the ratio of our actualities to our supposed potentialities" (p. 187). His formula, a fraction, the denominator of which is our pretensions and the numerator our successes, is perhaps the most widely quoted formula in psychology.

    This study observed self-esteem, one's self feeling, at several discrete points in time. We consider that each of these self-feelings is at once related to and appropriated with all the "self-feelings" that lay, as it were, in the consciousness of the self. Global self-esteem is not deemed to be a single snapshot of self or "self-image. Rather, the term self-esteem is consistent with the idea expressed by Demo (1985),

    Self-esteem is also consistent with the idea expressed by Rosenberg (1965), "global self-esteem is based not solely on an assessment of [one's] constituent qualities but on an assessment of the qualities that count," (p. 18). Self-esteem is not considered to be a "trait" or a "static state." It is more akin to an attitude brought about as a result of a dynamic process.

    The proposed study approaches the individual experience of "self-esteem" from the point of view of social behaviorism as explained by Mead (1934). Behaviorism in the broader Meadian sense involves a study of the inner experience of the individual. In this wider sense Mead (1934) says it involves "the study of the individual from the point of view of his conduct . . . but not exclusively the conduct as it is observable by others." (p. 2). Mead suggests that an "attitude" is recorded in some fashion in the nervous system of the individual (p. 11). Whether or how the attitude toward the self is recorded is not pertinent to this study. Attitudes, whether considered structures within the organism, or a flow of thoughts that contribute to the interaction of the individual to the situation, evoke responses by the individual to social situations and. as such, are called out of the individual. As the individual participates in the response, the response changes the environment. The "Me" perceived by the individual is a part of that environment and thus the perception of the "Me" is always subject to change. The "attitude" of the individual toward the self," is assumed to be reflected in the individual's on-going evaluation of self, one's self-esteem.

    The process of self-evaluation suggests a duality. The reality of the self, however, is a unity. The self is no better, and no worse than it is. For the depressed, however, the empirical ego and the object of "unworthiness" is the perceived "Me." As successive evaluations separate the "I" and the "Me" spatially, perhaps the "I, pure ego," eventually succumbs to the view that it is no better than the perceived "Me," the "empirical ego," and ultimately identifies with it thereby entering a new reality, the reality of the depressed. It is this dynamic process that this study attempts to record.

    The literature on depression may well reach back to the Book of Job, and include a 3900 year old Egyptian manuscript (Kovacs and Beck 1978). Beck (1973) concludes that the historical accounts indicate depressive behavior is "manifested in… the traditional psychological divisions of affection, cognition and conation," and comments that few clinical descriptions are as constant through successive periods of history (p.5). Klerman and Weissman (1980) also point to the similarity among ancient descriptions of depressive states. Beck (1973) defined depression as a clinical entity, marking a well-defined onset with a progression in the severity of the symptoms until the condition bottoms out. He noted a steady regression (improvement) of the symptoms until the episode was over.

    The Diagnostic and Statistical Manual of Mental Disorders DSM III-R, (rev. 3rd Ed. 1987) describes depression:

    Researchers generally recognize depression disorders as "affective disorders. However, wide differences have existed concerning the classification of these disorders. Researchers use different terms to distinguish among depressive conditions. For example, endogenous, (autonomous, psychotic), stands for disorders that seemed to come from within. Exogenous, (reactive, neurotic), signifies those caused by some external psychosocial factor (Beck 1973). These distinctions appeared in the diagnostic and statistical manual of the American Psychiatric Association as late as 1968. The 1968 edition of the diagnostic manual classified depression into two broad categories. The first category, neurotic depression and psychotic depressive reaction, characterize depression by antecedent psychosocial conditions and is, presumably, reactive to life events, and, therefor more or less psychogenic. The second category includes envolutional melancholia and manic depressive illness. These occur when the mood does not seem related directly to a precipitating life experience, (hence endogenous or as coming from within) (DSM 1968). The implication is that the latter are biological in origin.

    Beck objected to this kind of classification. Beck argues that the difference between psychotic and neurotic depressive reactions is quantitative rather than qualitative (Beck 1973). The APA modified its approach in 1980 with the adoption of DSM III. The manual included a new reference, affective disorders. The essential feature of this group is described as a disturbance of mood not due to any other physical or mental disorder. The DSM III R (1987) classifies depressive disorders as Mood Disorders, under Major Affective Disorders. This classification includes Manic Episode and Major Depressive Episode. The debate continues because the American Psychological Association's National Task Force on Women and Depression Final Report, (1990) comments that depression is heterogeneous. The report also says that depression is not a continuum from the blues to major depression and, eventually, suicide. Rather, it argues, depression varies in kind and in the contribution of different risk factors.

    Adolph Meyer, Sigmund Freud, and Emil Kraepelin begin the modern era in the study of depression (Friedman and Katz, 1974). During the first half of the twentieth century researchers viewed depression as either biological in origin or (Kraepalian), "ego" centered or (Freudian) or originating in an interaction of biological, psychological and environmental forces, (Meyerian). (Friedman and Katz, 1974).

    Kraepelin (1913), working with manic-depressives conceptualized depression as a disease entity with a physical pathology. Kraepelin worked to develop a systematic classification of depressive illness, and maintained that the origins of depression were biological. On the other hand, Meyer (1951) advocated a dynamic psychobiological approach. Meyer viewed depression as a form of reaction, the result of biological, psychological and environmental forces. He reasoned that abnormal mental developments are "attitudes and reactions of the person as a whole" (p. 601). Meyer acknowledged the importance of constitutional genetic factors, but would extend the study of depression to psychosocial forces. While considering constitutional make-up to count for a great deal in dagnosing and discovering the origins of depression, Meyer believed studying the deterioration of habits and the undermining of instincts and their somatic components would yield much more (p. 152).
 

    Freud's (1968) approached depression etiologically, from a clinical data basis. Freud, the psychoanalyst, focused on the ego and the superego. Friedman and Katz (1974) observe that the Freudian influence on depression theory and research was the dominant force during the first half of the twentieth century and continues to be profound. In the development of the psychoanalytic explanation of depression, Freud (1968) distinguished melancholia from the normal affect of mourning. Each results from loss. Freud (1968) described mourning as an uncomplicated grief reaction, a conscious process. Melancholia on the other hand, he argues, works at the unconscious level. Both he claims involve work. Both exhibit "a profoundly painful dejection, cessation of interest in the outside world, loss of the capacity to love, and an inhibition of all activity. . . [but] the disturbance of self-regard is absent in mourning" (p. 51). Depression follows from a loss of the love object withdrawn from consciousness. Fenichel (1968) observes that the object may be a person or an object, and that after the loss a narcissistic identification with the object is established carried out by the instinctual process of oral incorporation. Hostility, initially directed toward the lost object, is turned against the subject's own self, "abusing it, defaming it, making it suffer" (Freud, 1968, p.60). The attack results in a fall in self-esteem. Gaylin (1968) reasons, if a man is depressed over the loss of a job it is not because the job symbolizes a loved object but, like a loved object, "it can symbolize one's self-esteem" (p.17).

    Fenichel (1968) hints that diminishing self-esteem may be an operative component of an adaptive process, or even a warning of eventual melancholia (p. 111). He also suggests the whole depressive process appears as an attempt at reparation, intended to restore the self-esteem that has been damaged (p. 128).

    Bibring (1953) viewed depression "as the emotional correlate of a partial or complete collapse of the self-esteem of the ego since it feels unable to live up to its aspirations, (ego ideal, superego), while they are strongly maintained" (p. 26). For Bibring, self hatred is secondary to a breakdown of self-esteem. Bibring also injected the term "helplessness," subjective powerlessness, into psychoanalytic literature to describe the ego's inability to cope (Bibring 1953, p. 24). Gaylin (1968) suggests that for Bibring diminished self-esteem is the cause of depresson. (p. 388).

    Psychoanalytic theory. while holding that depression follows upon a loss, real or symbolic, holds firmly to two additinal notions. The first is that with major depression, diminished self-esteem, whether treated as an effect, Freud, Rado, and Fenichel, or as a cause, Bibring, culminates in the total collapse of the ego. The second is that the depressed person is predisposed to depression, by reason of a traumatic experience usually occurring in early childhood (Bibring 1953 p. 39). Under psychoanalytic theory, the depressed have a depressive personality. These were described as the "pathogenic, phenotype" or the class of "moodcyclic phenotypes" (Rado, 1968) - the orally fixated (Freud, 1968) - and the "person who is fixated where his self-esteem is regulated by external supplies" (Fenichel, 1968). Bibring (1953) alone hints that perhaps depression extends beyond personality, "a human way of reacting to frustration and misery" whenever the ego finds itself in a state of (real or imaginary) helplessness against "overwhelming odds" (p.36).
 

    Behaviorists in the classic tradition approach depression differently. Ferster (1974) observes that the most obvious aspect of depression is a marked reduction in the frequency of certain kinds of activity and in the increase in the frequency of others, usually avoidance and mistake. Lewinsohn (1974) outlined the behavioral theory of depression:
The occurrence of behavior followed by positive reinforcement is vital if depression is to be avoided. We predict depression when the probability is low that the individual's behavior will be followed by reinforcement, and also when the probability is high that the individual will be "reinforced" when he does not emit the behavior (e.g., the retired person receiving his paycheck regardless of what he does). Under both conditions the probability of the individual emitting behavior is reduced . . . . Statements like, (a) "I am weak or otherwise inadequate," or (b) "I am not likable" are alternative explanations available to the depressed individual to label an experience of an unpleasant feeling state (dysphoria) Such statements are looked upon as secondary elaborations of the feeling of dysphoria, which in turn is presumed to be the consequence of a low rate of response contingent positive reinforcement, (resconposre).(p. 160).

    The study conducted with this dissertation measured internal behavior (Mead 1934) using a subjective self-report measure. As such, it is not within the classical behavior model. However, the frequency with which an individual makes the kinds of statements of the feeling of dysphoria before the onset of depression quoted by Lewinsohn, could be indications of a change in the level of self esteem in a negative direction. Put in behavioral terms, a change in the rate of resconposre in a negative direction resulting in secondary elaborations of the feeling of dysphoria before the onset of a depression episode may be prodromal signs of the onset. It is pertinent to observe that behavioral methods used to treat depressions are those which add to, modify or change an individual's behavioral repertoire from a passive style to an active style (Ferster, 1971). Viewed from a self-perspective, the result of behavior therapy is to change self-esteem in a positive direction.
 

    Recognizing that the behaviorial helplessness model did not explain lowered self-esteem as a symptom of depression, Abramsom, Seligman & Teasdale (1978), offered the reformulated helplessness model to remedy this cognitive shortcoming. They state:

1. Depression consists of four classes of deficits: motivational, cognitive, self-esteem, and affective.
2. When highly desired outcomes are believed improbable or highly aversive outcomes are believed probable, and the individual expects that no response in his repertoire will change their likelihood, (helplessness) depression results.
3. The generality of the depressive deficits will depend on the globality of the attribution for helplessness, the chronicity of the depression deficits will depend on the stability of the attribution for helplessness, and whether self esteem is lowered will depend on the internality of the attribution for helplessness.
4. The intensity of the deficits depends on the strength, or certainty, of the expectation of uncontrollability and, in the case of the affective and self-esteem deficits, on the importance of the outcome (p. 68).

    Beck (1974) argues for a cognitive model to explain depression. Beck's formulation came as a result of clinical studies and observation of depressed patients. Beck and his colleagues found that depressed patients persisted in making indiscriminate negative predictions of the future, and developed the Beck Depression Inventory to measure the behavioral manifestations of depression and to provide an instrument for use for research purposes (Beck , Ward, Mendelson, Mock, & Erbaugh 1961). Beck and his colleagues conducted a series of correlation studies and found significant correlation between the inventory and measures of pessimism and negative self-concept, and between the inventory and measures of negative view of the future and of the self. Beck (1974) describes a "negative triad" upon which depression revolves, a "distortion regarding the patient's. evaluation of his world, of himself, and of his future" (p. 6). Beck reports that the clue to understanding the origin of depression is understanding the depressed individual's sense of loss,

Depressed patients [say] "I have no future, I've lost everything, My family is gone. . . " He regards himself as lacking some element or attribute he considers essential for his happiness: competence in attaining his goals, attractiveness to other people, closeness to family or friends. . . . he either misinterprets or exaggerates the loss (p.6).

    Beck, Steer, Epstein and Brown, (1990) developed a self-concept test to measure aspects of     Beck's cognitive model of psychopathology. They describe a negative self-concept that is developmental, schematic, and progressive. Beck, Steer, Epstein and Brown (1990) argue that the self-concept is the product of input of self-relevant data and that relatively stable structures (self-schemata) serve as information processors leading to data supporting a depressed person's negative self—concept, data more readily acceptable to the individual than data fostering positive self-evaluations. Kovaks and Beck (1978) do suggest, however, that the maladaptive cognitive schemata described by Beck can be modified.
 

Some Commonalities and Differences among the Models

Diminished self-steem follows upon depression.

Loss — real - imagined — symbolic.

    Both the reformulated helplessness model and the Beck model are reminiscent of Bibring's (1953) analysis:

The feelings of helplessness are not the only characteristic of depression. On further analysis…one invariably finds the condition that certain narcissistically significant, that is, for the self-esteem, pertinent, goals and objects are strongly maintained. Irrespective of their unconscious implications, one may roughly distinguish between three groups of such persisting aspirations of the person: (1) the wish to be worthy, to be loved, to be appreciated, not to be inferior or unworthy; (2) the wish to be strong, superior, great, secure, not to be weak and insecure; and (3) the wish to be good, to be loving, not to be aggressive, hateful and destructive…In the first group, depression sets in whenever the fear of being inferior or defective seems to come true. . . he is definitely doomed to be a "failure." In the second group…depression is due to the shocklike (actual or imaginary or symbolic) evidence that this goal will never be achieved due to the ego's weakness' . . .In the third. . .the narcissitic shock . . . is due to the unexpected awareness of the existence of latent agressive tendencies within the self with all the consequences involved. (p. 24).

Personality trait or vulnerability

    The psychoanalytic school suggests the origin of this affective state lays in a premorbid personality, grounded in a fixation in development. Bibring (1953) describes the process:
It is further assumed on the basis of clinical material that such traumatic experiences usually occur in early childhood and establish a fixation of the ego to the state of helplessness. This state is later on repressively reactivated whenever situations arise which resemble the primary shock condition, i.e., when for external or internal reasons those particular functions which serve the fulfillment of the important aspiration, prove to be inadequate. (p. 39)

    The reformulated helplessness model (Abramsom, Seligman & Teasdale (1978), with its emphasis on "attributional style," a cognitive set, offers a pre-morbid personality explanation of depression, and the model predicts that the attributional style will produce depression proneness, perhaps the "depressive personality. (p.68)"

    Beck (1974) uses the term "depression prone" to signify the development of a negative schema (Beck 1974, 1978; Beck, Steer, Epstein, & Brown, 1990). Kovacs and Beck (1978), arguing for the efficacy of cognitive therapy, state that cognitive structures, "probably acquired early in development, and if uncritically carried into adulthood, serve to predispose the individual to depression" (p. 525)

    The kind of reasoning that arrives at the notion that depression follows from being "depression prone" or because one has a "premorbid personality" or a personality trait or schematic structure is not new (e.g. Chodoff 1973). The tendency of modern researchers to conclude that depressives are depression prone brings to mind the debate between the Kraepelin school and the Meyerian school. Kraepelin (1913) insisted on an organic explanation. Meyer (1951) included a psychological factor as well as an organic factor in attempting to explain mental disorders. Meyer sought to expand medicine from its insistence on physical disease. Meyer wrote,

We cannot afford any longer to ignore the chains of conduct and behavior or mental reactions . . . As soon as we put ourselves on a dynamic psychobiological foundation, we make unnecessary the continual yearning for something (physical) back of the events, at the expense of the plain facts in evidence (p. 602)

    Dohr, Rush and Bernstein (1989) compared symptomatically depressed, clinically remitted, and normal controls in cross sectional and longitudinal designs using cognitive measures to investigate predepression bias. They found that depressive episodes affect cognition, but the cognitions measured were more statelike than traitlike. This is not inconsistent with the theory that self-esteem may change prior to onset.

    A longitudinal prospective study to determine whether cognitions known to be correlated with depression precede, accompany, or follow an episode of depression has been reported. The study was an attempt to contribute to the resolution of the question of whether negative cognitions precede depression and in some way contribute to its occurrence (Lewinsohn, Steinmetz, Larson and Franklin, 1981). The study failed to support a hypothesis that depression-related cognitions are antecedents of depression. From the hypothesis it was predicted that individuals who became depressed during the course of the study would differ in their cognition patterns from those who remained free of depression. The results did not support the prediction. Self-esteem was included as a general variable but the findings were not indicative of a diminished self-esteem prior to depression.
 

Depression: A Multicausal Psychosocial Approach

    Some researchers have expanded the search for the origin of depression to psychosocial factors. Klerman (1974) argues this kind of an approach. He writes, the "Meyerian psychological approach was an explicit reaction against Kraepelin's synthesis of the ninetieth century continental tradition Kraepelin (1921) which emphasized biological causation" (p.129). He suggests the view that depression represents an attempt at adaptation that has failed. Using adaptation in a Darwinian sense, he compares depression to infant distress. In the infant, depression serves a "signal" function. (p. 132). Infant distress, he claims, represents an important signal of distress that suggests the infant is seeking such resources as nurturance, support and protection necessary to his/her biological survival. Klerman believes a depressive episode may be initiated as a response to helplessness and fallen self-esteem, and thus may serve as the signal for the individual that there has been a discrepancy within the self-system between ideal expectations and practical reality. He argues that the depressive episode per se, while initiated as an attempt at adaptation in response to environmental change, must be regarded as an index of failure having maladaptive consequences (p.139). Klerman advances the notion that a significant factor accounting for the occurrence of depression is the vulnerability of the individual.

    Assessing the relative roles of predisposition and of environmental events in the precipitation of the overt clinical depressive state Klerman (1974) concluded that clinical states vary according to the balance between precipitating external stress and the vulnerability of the individual. Environmental stress seems to play a role in the timing and precipitation of acute events, but these events are not universal or specific to depression. "It seems that the most significant factor accounting for the occurrence of depression is the predisposition or vulnerability of the individual" (p. 142).

    Klerman and Weissman (1980) addressing depression among women concluded that a simple cause for depression "is unlikely to emerge, and that more complex, multicausal explanations are necessary to understand this complex human experience," (p. 72) They argue that clinical depression is a maladaptive outcome of partially successful attempts of adaptation and observe:
Internalized and subjective affects also play important functions in psychic reflection. They permit goal-setting and self-evaluation, which are again representative of evolutionary attainments. . . . The depressive episode is often initiated as a response to helplessness and fallen self-esteem, but this attempted adaptation fails. . . According to the sociopsychological explanation for depression, we could expect to find the highest rates of depression among that group of people who experience the greatest number of stressful events life demands, and at the same time the fewest actual possibilities for mastery of them. (pp. 88-89).

    The multicasual psychosocial approach gives support to the underlying theory of this dissertation. The multicasual psychosocial approach recognizes a dynamic process involving an attempted adaptation of the self-system to the ever-changing environment that includes, among other factors, the attitude toward the self .
 

Environmental Factors Associated With Depression.

    The role of a number of psychological and social factors in depression, including life events, have been investigated. Paykel (1978) reviewed past epidemiological research about the onset of psychiatric illness and concluded, "a multifactorial [chain is indicated] in which events are of importance but interact with a host of other factors" (p. 252). Paykel suggests a shift toward prospective studies of events in the general population.
In a study involving 40 depressed patients and 41 normal controls Roy, Breier, Doran and Pickar (1985) found that depressed patients had experienced significantly more life events and significantly more undesirable life events than controls in the six months before the onset of depression. Life events included changes in work, education, finances, health, bereavement, residence, legal issues, the family and marital situation and other selected areas. This dissertation theorizes that life events may interact with a change in self-esteem prior to depression onset.

    Aneshensel and Stone (1982) tested the hypothesis that social support networks lessen the adverse psychological consequences of stress. They found life-event losses and perceived strain were positively related to depressive symptomatology, while close relationships and perceived support were negatively related to those symptoms. The researchers presumed that stress leads to depression. They also recognized that an alternative model would suggest that existing psychological impairment may result in the occurrence of stressful events like losing one's job. A change in the level of self esteem could occur before or after the loss of a close relationship that provided support or after the loss of a job making one vulnerable to the onset of depression.

    In a longitudinal study Brown, Bifulco, Harris, and Bridge (1986) found negative evaluation of self, at first contact, associated with an increased risk of depression one year later. Negative evaluation of self, NES, consisted of a measure of three components. The three components were (a) attributed self, physical attractiveness, intelligence, ability to get along with people and so on, (b) self definition, perception of her competence in roles such as wife, mother and worker, and (c) self acceptance, more generalized feelings about the way she sees herself. The study also examined the role of chronic subclinical symptoms, CSCs, among women not clinically depressed at first contact with a subsequent onset of depression. The study included 215 "normal" women, that is, those without subclinical symptoms. Of the normal group, 15 of those followed up experienced an onset of depression. Of these 67% (10/15) had NES at first contact, and met with a provoking agent before the onset of depression. The study did not examine NES between the first contact and depression onset. A provoking agent was operationally defined as a severe event or a major difficulty. The focus of the study was etiologic, but Brown, Bifulco, Harris, & Bridge, (1986) were unable to offer a definitive conclusion about the role of CSCs in the genesis of depressive disorders at a caseness level (p. 17). They did conclude, however, that both NES and CSC are required in the best model predicting depression. The authors suggest that NES and CSC may be indicators of a common depression-prone disposition (p. 15).

    In a prospective study, Murrell and Norris (1984) found a significant interaction between resources and undesirable events on depression in a sample of 1,166 adults, age 55 and older. Resources included health, self-esteem, social support, education, and urbanicity. Events consisted of the person's perception of the nature of the impact on him or her of quite specific environmental changes. They found that older adults with relatively stronger resources were much better off psychologically to begin with than those with weaker resources and maintained this psychological advantage regardless of their subsequent life event experiences. They also found that within both strong and weak resource groups that persons who subsequently experienced high levels of undesirable events were more depressed that those who later experienced low levels of undesirable events. They did not look at the effect of a change in resources over time.

    Phifer and Murrell (1986) examined the additive and interactive roles of (a) six sociodemographic factors, (b) resources and (c) life events, in the development of depressive symptoms in a sample of 1,233 person, 55 years of age and older. They found that "health and social support played both additive and interactive roles, life events had weak effects, and sociodemographic factors did not contribute to depressive onset" (Phifer and Murrell 1986, p. 282). Self-esteem was not included as a variable.

    A study by Dura, Stukenberg and Kiecolt—Glaser, (1990) compared current and lifetime rates of DSM—III—R disorders in 86 older adults caring for a spouse with a progressive dementia and sociodemographically matched control subjects. Dura et al, found that Dementia caregivers were significantly more dysphoric than non—care givers. They concluded that chronic strains of care giving appear to be linked to the onset of depressive disorders in older adults with no prior evidence of vulnerability. Dura et al, (1990) did not examine self-esteem prior to depression onset.
The psychosocial approach admits of the possibility that a deterioration in the level of self-esteem occurs prior to the onset of a depression episode, but that possibility has not been examined. The study proposed by this dissertation does just that.
 

    A review of the literature reflects the etiology of depression is as much an enigma as ever. A common factor involved in all psychological models offered to explain depression is self-esteem. The unresolved issue has been whether loss of self-esteem is antecedent to depression or whether diminished self esteem is a result of depression. Each of the models reviewed suggest that depression follows a "loss." Other than the psychosocial, all models suggest a depressive personality. Only one study, Lewinsohn, Steinmetz, Larson, & Franklin (1981) attempted to assess self-esteem in non-depressed subjects who later become depressed. The authors reported that self-esteem measures did not predict depression. However, Ingham, Kreitman, McMiller, Sashidharan, & Surtees, (1986) suggest that the design of the investigation perhaps limits the force of the conclusion.

    Fenichel (1968) hints at an adaptive process. Bibring (1953) suggests that the ego becomes aware of its incapacity to live up to ideals. Beck (1974) suggests the development of a negative view of self. Abramsom, Seligman, & Teasdale (1978) posit the acquisition of an attributional style. The classical behaviorist view suggests a low rate of response contingent positive reinforcement. The psychosocial approach admits of an interaction of self-esteem with life events or life situations.
A deteriorating self-system, manifested in a progressive loss of self-esteem, prior to the onset of depression is not inconsistent with these explanations of depression.
 

    Notwithstanding the extensive research examining depression and self esteem, and the knowledge that low self-esteem and depression are linked, research examining the possible diminishment in self esteem prior to depression onset is virtually non-existent. Low self esteem has been suggested as establishing a propensity for depression (Brown, Bifulco, Harris, & Bridge 1986). Established theories suggest that loss of self-esteem follows upon depression ( Lewinsohn, Steinmetz, Larson, & Franklin 1981; Beck, 1974; Abraham, Seligman, & Teasdale, 1978; Gaylin, 1968). Researchers investigating cognitive factors related to depression conducting prospective designs have reported measures of self-esteem at Time 1 and Depression at Time 2 (Brown, Bifulco, Harris, & Bridge 1986; Lewinsohn, Steinmetz, Larson, & Franklin 1981; Lakey 1988, Essex & Klein, 1989), but these researchers did not examine changes in self-esteem during the time interval. Researchers have called for longitudinal studies that examine the dynamic and reciprocal relation between self-esteem and depression Essex & Klein (1989). Some have suggested an attempt to asses self-esteem in non-depressed subjects who later become depressed and to compare them with those who remain free of depression would be a worthwhile strategy Ingham, Kreitman, McMiller, Sashidharan, & Surtees (1986). The present study was designed to do that. Possible antecedent conditions of depression were controlled by eliminating from consideration scores of subjects with present or antecedent levels of psycopathology.

    The theory underlying this research postulates that the self-system is a dynamic unity engaged in a process of adjusting to its environment which includes the self. The study observed the attitude toward self through self report, and looked for a change in attitude that occurred prior to depression onset.

    Research has linked life events to depression, and the mediating role of resources including self-esteem prior to the occasion of undesirable life events. (Paykel, 1978; Lakey, 1988; Murrell & Norris, 1984; Roy, Breier, Doran and Pickar, (1985). The present study was an effort to locate a prodromal sign of depression, a change in the level of self esteem prior to the onset of a depression episode, to which undesirable life events were linked.
 
 

Hypotheses

    The study also hypothesized that those who exhibit depressive symptoms during the study will report more undesirable life events than Controls.
 

Design:

    The study utilized a prospective design, exploratory in nature, conducted in a natural setting. The study employed the strategy of assessing self-esteem in subjects at three month intervals over a period of fifteen months and then comparing the degree of change in the level of self-esteem scores of those subjects who reported depression symptomatology with the degree of change in the level of self-esteem the scores of subjects who remained depression free.

Subjects:

    Subjects were recruited from the residents of South Whidbey Island, the southern portion of Whidbey Island. Whidbey Island is located in the Puget Sound, North of Seattle Washington. South Whidbey is a unique area known for its exceptional quality of life. The terrain is generally rolling with elevations from sea level to 580 feet. Over 80% of the land area is undeveloped and it features trees, prairie areas of agricultural lands, marshes and open bays. The environment is rural. There are four village settlements on South Whidbey. Each has a grocery store and a few shops. The largest, the incorporated town of Langley has fewer than 1200 residents. Langley is a tourist attraction featuring a small movie theater, (that doubles as a playhouse), a library, marina, beach and shops. South Whidbey is served by a State ferry that runs between Clinton, an unincorporated village at the southern tip of the Island, and Mulkiteo, a small town on the mainland.
The sheltering Olympic and Cascade mountains together with the moderating effects of the surrounding water provide a mild temperate climate. The climate and rural character of the area has served to attract a large number of retirees who appreciate the mild short winters and cool comfortable summers. They likewise appreciate the scenic views and the freedom from the hustle, bustle and tensions of metropolitan life.

    Recruiting was completed through Senior Services of Island County, Washington a county wide non-profit service organization, congregations of various Church denominations located on South Whidbey, notices posted on public bulletin boards, and talks given at service organizations. The notice is a basic outline of the talks.

    Each volunteer received a letter of invitation that explained the project. The letter of invitation served as the Consent Form. The letter invited the reader to take part in a study concerned with the relationship between mood changes in an individual and what that individual thinks and feels about herself or himself. Each volunteer was asked to sign and return the Consent Form and to complete and return the Confidential Information Questionnaire to indicate their willingness and consent to participate in the study. The Consent Form advised the prospective participant that his or her participation was voluntary and that all responses would be held in confidence. The Confidential Information Questionnaire included the demographic items, age, gender, marital status, education (in years), annual family income and employment status.
Fifteen women and seventeen men agreed to participate. All but seven participants were over fifty years of age and of the total number fifteen were retired. All but five reported membership in an organized religion. Twenty eight reported annual income greater than the median income of their peers on the Island. From the standpoint of education, all had graduated from high school and twenty seven had experienced a college education.

Self-esteem :

    Self-esteem was measured by the Rosenberg Self-esteem Scale, RSE (Rosenberg, 1965). This scale is widely used in social science research (Lakey, 1988). The scale is a measure of global self-esteem. A high score indicates a feeling of unworthiness, while a low score suggests the person feels himself or herself to be a person of worth, but not necessarily superior to others. Responses are made to ten items on a 4-point scale ranging from strongly agree to strongly disagree. The scale exhibits good convergent validity with other self-esteem measures (Silber & Tippett, 1965) and strong predictive validity. (Rosenberg, 1965). Baker & Gallant, (1985) in a comparison study of self-esteem measures found that the Rosenberg Self-esteem Scale was superior to other scales in terms of both convergent and predictive validity, and that a simple additive index proved superior to the Guttman version. A simple additive index, was employed in this study in lieu of the Guttman version. A higher score at a later interval from an earlier interval indicates a negative change in the level of self esteem. Scores can range from 10 - 40. Scores _ 20 indicate positive healthy levels of self-esteem.

Depression:

    The Beck Depression Inventory, BDI, (Beck, 1978: Beck, Rush, Shaw, & Emery, 1979) was utilized to detect depressive symptomatology. The BDI has become one of the most widely accepted instruments for detecting depression in normal populations and has been used as a screening instrument in research and practice (Steer, Beck, & Garrison, 1986; Ponterotto, Pace, and Kavan, 1989). In a twenty-five year, (1961-1986), review of research studies focusing on the psychometric properties of the BDI, Beck Steer and Garbin, (1988) report high internal consistency estimates and high concurrent validities. Beck et al, report that with respect to factor analytic studies, the BDI measures three highly inter-correlated factors, negative attitudes, performance difficulties, and somatic complaints. All reflect the DSM III (R) criteria. The BDI is quick and easy to administer, and when self-administered takes from 5-10 minutes to complete. The respondent is asked to describe how he or she has been feeling the past week, including today. The BDI consist of 21 items, rated on a 4-point scale (0-3) of intensity. The BDI is scored by summing the ratings given to each of the 21 items of depression. The proposed study employed the cut-off scores recommended by the Center for Cognitive Therapy to classify cases having depressive symptomatology, that is, none or minimal depression < 10, mild to moderate 10-18, moderate to severe 19-29; and severe 30-63, Beck, Steer, & . Scores range from 0-63, with higher scores indicating greater severityGarbin (1988).
Life Events :

    Life events were measured by a scale adopted from the Louisville Older Persons Event Scale. The events, their frequency and their desirability ratings with respect to the project sample are reported in Murrell, Norris, and Hutchins (1984). To minimize any bias that may be caused by depression, life events were inquired of at each time segment and the inquiry was limited to the preceding three month period and dates were requested. A loss event category consisted of the following events: separation, divorce, lost home, lost job, less money to live on, friend moved, child moved, and lost a pet. A Bereavement events category includes: spouse died, child died, parent died, friend died, sibling died, and grandchild died. The final category consists of health problems: new personal injury or illness of the subject or of his or her spouse. The event measure score consisted of a simple count of the number of events experienced in a given category for the measurement period.

    Volunteers were enrolled by letter of invitation. It took approximately three months to enroll all the participants. Each participant had a different beginning date. A booklet containing all the measures was mailed to each participant after he or she returned the completed Consent Form and the Confidential Information Questionnaire. Each participant's beginning date, T0, was the date the initial booklet was dated. No adjustments were deemed necessary or made with respect to different times of beginning. Each participant was mailed a new booklet approximately three months after he or she completed and returned the earlier measures. The dates of mailing and completion varied with each participant depending on the date the earlier measures were returned. Not all who started completed the several self report measures from T1…5.

    A booklet containing the measures was mailed to each subject at the beginning of each three month interval relating to that subject. The booklet instructed the subject to complete each measure and to return it as soon as possible. Subjects were instructed to date their returns. Subsequent booklets were mailed three months after the return of the earlier booklet. Some participants did not date their returns and, in such case, the date of the return envelope was used to segregate the data as to time interval. Return of the booklet required minimum first class postage. Return envelopes and postage were not furnished. One subject commented on the fact that return envelopes and postage were not furnished. This subject completed all six rounds. The scores of all participants were recorded and sorted according to the scores on the BDI

[I have omitted the table of results]

    This study examined self-esteem as it relates to depression. Initial symptom level was controlled as was overall health level.  Assessing prior psychological state, initial experienced self-esteem, and health status minimized the potential of confounding these measures with later psychological state. The community sample used was relatively homogenous in environmental  influences, age, income, and education level, and balanced as to gender.

    Overall, the data generated do not reveal that  individuals suffer a decline in self-esteem prior to the onset of depression symptoms, nor do the data demonstrate that a loss of self-esteem is simply an epiphenomenon of depression.  The question is still an open one.

    The study produced interesting data that deserves comment.

    Because all respondents were volunteers, highly educated, financially secure, and all enjoyed the relatively tranquil environment of South Whidbey Island, we conclude that the sample may have been composed of psychologically strong and capable individuals in a uniquely supportive environment who were able to draw upon their personal resources to respond and adapt their respective self systems to the stress occasioned by the undesirable life events they experienced over the course of the study.

    Recall that depression is explained by the theoretical approaches reviewed in our review of the literature, as occurring in the "depression prone, " those with (a) a "premorbid personality," (Bibring 1953), (b) a developed attribution style (helplessness) that produces depression proneness, (Abramson, Seligman & Teasdale, 1978), or (c) an early developed cognitive structures that serve to predispose the individual to depression, Kovacs & Beck (1978). From the point of view of these theorists it would be argued that this sample did not include vulnerable individuals predisposed to depression. However, individuals in this study with high, positive, stable self-esteem levels did report depression symptoms. This suggests that individuals with healthy levels of self-esteem may not be invulnerable to experiencing depression symptoms and possibly depression. The nature and character of individuals who experience more severe depression is still shrouded in mystery, but we conclude that the data generated in this study, if confirmed in a study of a larger sample, would lend support to the theory that depression may be better explained as following a dynamic process resulting from the inability of a particular individual to adapt his or her self-system to  cope with internal and external stress. We suggest that even the poverty stricken data  of this study indicates that state-like vulnerability as contrasted with trait-like vulnerability is as likely an explanation of the depression mechanism. We believe a larger broader study would disclose individuals who begin free from depression symptoms and progress to severe depression symptoms. The approach of this study is compatible with the Multicasual Psychosocial Approach advanced by Klermann (1974).

    We recognize the limitations of the present study particularly because of the small number of persons reporting depressive symptoms and the fact that only mild to moderate symptoms were reported.  We point out, nevertheless,  that we did observe slight changes in BDI scores and changes in self-esteem levels. We believe that we can tentatively conclude that the depression process is dynamic over time,  and that the process involves an effort on the part of the individual to adapt the self-system to the ever changing environment which includes the respondent's attitude of the self carried forward in time. The data suggests that the process will be influenced by the kind of ecosystem in which the person finds herself or himself. Individuals free of depressive symptoms at T0 reported changes in levels of depressive symptoms, albeit in the mild to moderate range. These  individuals can be said to have adapted successfully, perhaps aided by the ecological system in which they found themselves. These observations are consistent with the findings by Brown, Bifulco, Harris, & Bridge (1986). The question remains, what happens with those who are located in more stressful environments, physically and sociodemographically and who do not adapt?

    In any case, of significance is the fact that the study did demonstrate that a prospective design could be utilized to produce relevant data. One lesson learned from conducting this study is that researchers must be willing to take a vow of poverty that includes not only the willingness to accept poverty of finances, but also a poverty of experimental results. A larger, broader, community sample will be necessary to flush out persons who eventually exhibit  severe depression symptoms. The demonstrated correlation between self-esteem scores and subsequent BDI scores in this study is encouraging for future research that implements the prospective design employed in this study with the recommended adjustments.

        The ecosystem peculiar to South Whidby Island, reputedly attractive to the retired elderly, might, as mentioned, in itself, have contributed to the relatively healthy stable self-esteem levels reported by the respondents and the mild levels of the reported depressive symptoms. On the other hand it could be concluded that healthy self-esteem alone provides a valuable resource to individuals in any setting to cope with external and internal stresses on mood, with high positive stable self-esteem levels acting as a coping resource to deal with circumstances that might otherwise lead to depressed mood. In the study sample, for example, even the individuals who reported depressive symptoms at the outset and who continued to report depressive symptoms reported only minimal  negative levels of self-esteem at each T.  We cannot discern from the scores reported whether the mild levels of depression symptoms are a function of high self-esteem or whether stable self-esteem levels are simply unaffected by minimal levels of depression symptoms. Acknowledging the limitations associated with volunteerism and the environmental setting, the results reported in this study would be valuable as a basis for comparing results in future prospective studies that include individuals whose initial self-esteem levels are similar at outset to those in this study but who are located in environmental settings containing stress factors normally associated with a metropolitan setting and who later experience depression symptoms in both the mild, moderate and severe ranges.

    The study's reliance on self-report measures is another obvious limitation. Self report measures tend to contain self-report bias. However, the scores reported demonstrate that the Beck Depression Inventory is well adapted to reveal subtle changes in mood, See Table 1. Additionally, when it is understood that the principal objective of the research is to measure changes in self-esteem prior to the onset of depressive symptoms as contrasted with determining the onset of a clinically defined depressive disorder, the Beck Depression Inventory is well suited to the design utilized. On the other hand, the results of this study  suggest that the Rosenberg Self-esteem Scale, while widely utilized in social science research and an accepted measure of global self-esteem, may not be sufficiently sensitive to periodic fluctuations in self-esteem and thus inadequate to observe more subtle fluctuations in experienced self-esteem. In future studies a more sensitive measure of self-esteem is indicated, e.g. the Coopersmith Self-Esteem Inventory, SEI, (SEI; Coopersmith, 1967). The Coopersmith measure assesses several areas of self-esteem including self-derogation and self-worth and may serve to detect shifts not differentiated by the global assessment offered by the Rosenberg Self-Esteem Scale. In a study designed to test the validity of methods to measure experienced self-esteem, (Demo, 1985), the SEI correlated significantly with every other measure tested providing strong evidence of convergent validity with respect to personal feelings of competence and effectiveness.

    The results do show that some respondents who exhibited mild depression symptoms did not report antecedent deterioration of self-esteem or report negative self-esteem concurrent with depression symptoms. It is tempting to conclude that it is probable that self-esteem deficits do not precede or accompany mild to moderate depression symptoms. However the nature of the sample and its ecology precludes generalization and as previously observed both may have contributed to the poverty of the results. Future research of a much larger sample extending over different environmental settings that employs the prospective design utilized here with the adjustments noted could well develop data from which to generalize with respect to these phenomena.

    The data generated by this study were insufficient to make the proposed comparisons concerning the impact of loss events on those experiencing depressive symptoms with those who did not.  Indeed, a potential limitation exists concerning the use of the event measure at any T to assess the impact of events on that T  depressive status. The life event measure is designed to elicit a factual response, but a respondent's  subjective recollection of prior events may be confounded with his or her current status with respect to reported depressive symptoms. The measure adopted does call for events occurring prior to the date of the measure and asked for a date. This may have reduced the limiting effect. However, some respondents appear  to have reported the same event at successive time intervals. Also, the change in depressive symptoms may have preceded the event reported. Thus the use of the T event measure may overestimate the causal effects of life events. Only speculative conclusions can be drawn from the observed effect for loss events pending further research utilizing more accurate dating of time of onset and event occurrence. A well designed structured interview used in any future study could assist in determining whether the loss event preceded the change in mood and the use of such a structured interview is indicated for future research. Additionally, the inclusion of a wider age range of respondents is indicated to assess whether the impact of events on younger persons is different from the impact of events on  older persons.

    Subject to the limitations noted above, when the scores of those reporting depressive symptoms from the outset of the study are included the study did reveal that, overall, respondents reporting depression symptoms at any T, however slight, reported more undesirable life events than those who were free of depression symptoms at that T.  This suggests that loss events had modest effects on mood. These findings also support the conclusion that a dynamic process of adaptation is at work.

    Because of the lack of data no statements concerning self-esteem associated with severe depression episodes can be postulated. That question remains for future research. As a practical matter whether a sufficient number of respondents who experience more severe symptoms will appear in a prospective study is a question upon which this effort sheds no reliable information. If sufficient financial resources are available it could be expected that a large sample could be obtained in a major metropolitan setting that offers different ecological systems from which respondents could be obtained. The character of the individuals attracted in this study and the environmental context in which they were found coupled with the fact that some of these individuals reported changes in mood is encouraging; and, given that large samples in earlier cross sectional studies involved individuals who experienced depressive symptoms, the probabilities are that with a larger sample sufficient data for analysis will emerge.

    The results also indicate that a study extending over twelve to fifteen months can be maintained without serious attrition, Table 1. In future studies it is recommended that self-addressed stamped envelops be furnished with the booklet. Doing so might encourage continued participation. We believe a study of eighteen months would be difficult to maintain. Given that mood changes were reported by respondents initially free of depressive symptoms during a  twelve month period in this study suggests that a twelve to fifteen month study should be long enough to develop sufficient data to test the hypotheses.

    This study shows that a prospective design extending over a period of twelve or fifteen months can be implemented. The BDI, together with a more sensitive self-esteem measure, and a shorter more definitive life events measure is indicated. A much larger sample that includes individuals from a similar environment and individuals from a metropolitan environment is also indicated. One would expect to obtain a wider range of demographic characteristics in the larger sample. A carefully prepared structured interview to assess the validity of the life events reported is also indicated. It is recommended that the interviewers be ignorant of the hypothesis of the study.

    All measures employed were self-report measures and as previously mentioned contain self-report bias. Both the BDI, the RSE and the SEI are based upon the assumption that individuals are able, and willing, to state their feelings honestly and accurately. The fact that in the present study scores between successive T's reflect variability consistently suggests that the measures are reliable, and the assumption reasonable. See Table 1. While scoring the BDI we concluded that in future studies random presentation of the statements offered at the time intervals, without identification of the score associated with the answer, would create a more reliable instrument. This conclusion was based on intuition and not on noting any test-retest similarity of scores.

    Self-evaluation is a continual process. Mood changes do occur. The measures utilized in this study and the SEI proposed for future research are generally accepted as an indicator of experienced self-esteem and the BDI of the existence of depression symptoms. All self-reports are nevertheless only single frames in an ongoing process. The process of self-evaluation itself changes the environment because each antecedent evaluation becomes part of the environmental setting for the subsequent evaluation. While the report of a decline in self-esteem based on a single measurement is a snapshot as opposed to a change in the base level of the subject's self-esteem, a decline continuing over more than one time period coupled with an increase in depressive symptoms would tend to suggest  that  the way to depression is more progressive than precipitous.
Of course, the BDI cannot be substituted for a clinical diagnosis of depression. When the findings are limited to the expression of depressive symptoms the condition of  compared groups in future research may differ quantitatively and qualitatively from clinical depression. However, our experience in this study  suggests that future research proposed could produce data that is subject to manipulation as proposed.

    The research design adopted for this dissertation study proved workable. The attrition rate was acceptable and could probably be improved. The prompt responses by the participants who continued and their willingness to participate coupled with the ease of participation also testify to the design's utility. The participants' willingness to reveal depressive symptoms  is encouraging. Larger samples and improved measures of experienced self-esteem and life events are indicated.  The proposed change in the presentation of the BDI should enhance its reliability. One important lesson was learned in the conduct of this study. The environmental setting, the ecosystem in which the respondents reside, could affect the ability of the individual to adapt to internal and external stresses. This fact alone makes this study useful in evaluating outcomes in studies emanating from different ecosystems. The demographic questionnaire could be adapted to define the ecosystem of the respondent. More than individual resources, however,  will be required to develop sufficient data for analysis.

    Future research of the kind employed here, adjusted to reflect the lessons learned in this study, can lead to a better understanding about the nature of the association of self-esteem with depression.