Contribution of the MMPI—2 Content Scales to the Differential Diagnosis of Schizophrenia and Major Depression

The contribution of the MMPI—2 content scales to the differential diagnosis of psychopathology was evaluated by investigating their ability to aid in the differentiation between schizophrenia and major depression in inpatient psychiatric settings. A total of 160 inpatients (76 with schizophrenia and 84 with major depression) were administered the MMPI—2 in 3 psychiatric facilities. Results indicated that both the original clinical scales and the new MMPI—2 content scales contain information that is useful in the differential diagnosis of these 2 conditions. Further analyses indicated that the content scales possess incremental validity with respect to this diagnostic question in that they contain information beyond that which is available from the clinical scales.

The Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1940 ) has had a long history of use in the task of differential diagnosis of psychopathology. In fact, the initial goal of its developers was to provide a screening instrument to differentiate various forms of psychopathology ( Hathaway, 1964 ). Soon after its publication, however, it became evident that the test's primary strength lay in the empirical extra-test correlates associated with scale scores and configural patterns, such as the ones identified by Marks, Seeman, and Haller (1974) .

In conjunction with research designed to identify correlates of MMPI scale scores, investigators continued their efforts to demonstrate the MMPI's utility in the differential diagnosis of psychopathology. Following Meehl's (1954) illustration of the advantages of statistical prediction, numerous efforts, varying in their statistical sophistication, were undertaken to develop MMPI indexes to be used for differentiating various forms of psychopathology (e.g., Goldberg, 1965 ; Meehl & Dahlstrom, 1960 ). Although some of these efforts were moderately successful, primarily in differentiating between neurotic and psychotic conditions, in recent years the number of attempts to produce additional data pertaining to the MMPI's utility in differential diagnosis of psychopathology has declined (e.g., McFall, Moore, Kivlahan, & Capestany, 1988 ; Mezzich, Damarin, & Erickson, 1974 ; Roy, 1984 ; Walters, 1984 ; Walters & Greene, 1988 ).

The relative dearth of recent efforts to study the use of the MMPI in differentiating forms of psychopathology may, in part, reflect the kind of problems that led to its recent revision. The basic clinical scales of the MMPI mirror the psychiatric nomenclature of the 1930s. The items making up these scales have not been updated since that time, yet considerable changes have taken place in psychiatric nosology–changes to which the MMPI has been ill-equipped to respond until recently. In addition, the norms on which MMPI scores were based had not been updated in over 50 years. It has been noted since as early as the 1950s that these norms may yield elevations on the clinical scales when there is no clinical evidence of psychopathology.

For these and other reasons, the MMPI was recently revised and updated, culminating in the publication of the MMPI—2 ( Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989 ). Two features of the revision may potentially enhance the effectiveness of the MMPI—2 in the task of differential diagnosis. First, a new, nationally based normative sample was collected, resulting in the recalibration of the original scales to represent more accurately the population of the United States. Second, new item content was incorporated in the revision, with the aim of updating the assessment of existing constructs (e.g., the incorporation of statements reflecting depressive cognitions to enhance the assessment of depression) and expanding the pool to areas not previously covered in the MMPI (e.g., Type A behavior). This new item content was incorporated in the development of a new set of scales, the MMPI—2 content scales ( Butcher, Graham, Williams, & Ben-Porath, 1990 ), whereas the original clinical scales have been left almost entirely intact ( Butcher et al., 1989 ).

Armed with new norms and scales, users and researchers of the MMPI may wish to reexplore the instrument's potential for the differential diagnosis of psychopathology. To demonstrate the utility of the MMPI—2 content scales in the differential diagnosis of psychopathology, it is not enough to show that they are equal to the clinical scales in their capacity to differentiate forms of psychopathology. To do so would mean that we have advanced no further than where we were before the revision. What is necessary is a demonstration that the content scales possess incremental validity, that they go beyond what can be accomplished with the traditional clinical scales, that they contain diagnostic information not available in the clinical scales.

In the present study, the potential contribution of the MMPI—2 content scales to differentiating between schizophrenia and major depression is investigated. New content incorporated in these scales includes items dealing with frequently found positive symptoms of schizophrenia (e.g., thought broadcasting) and various cognitions known to be associated with depression. The content scales also incorporate many of the original MMPI items that related specifically to these disorders. Thus, they possess the potential to meet the above-noted criterion of incremental validity by providing information that goes beyond what is available in the original clinical scales.

Diagnoses, using the criteria of the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM—III; American Psychiatric Association, 1980 ), which were in place at the time of data collection, were based on the clinical examination of patients by clinicians on staff at the three facilities where data were collected. At the acute care facility in Minnesota, diagnoses were based on examination of the patients by staff psychiatrists and psychiatric residents who, after consultation with a team consisting of the psychiatrist, a clinical psychologist, a psychiatric resident, psychology trainees, a psychiatric social worker, and psychiatric nurses, rendered a primary Axis I diagnosis (schizophrenia or major depression for subjects included in this study) and, if warranted, secondary Axis I or Axis II diagnoses. MMPI data were not available to the clinicians when the initial diagnosis (used in the analyses) was made. All of the subjects included in the sample from this facility had the same diagnosis at the time of discharge.

At the second acute care facility, in Ohio, diagnoses were provided by staff psychiatrists following a clinical evaluation and consultation with treatment staff. Again, a primary Axis I diagnosis was provided along with secondary Axis I or Axis II diagnoses, and MMPI data were not available to the clinicians at the time when these diagnoses were rendered. Over 90% of the subjects included in the study from this facility had the same diagnosis at discharge as they did at intake. (Intake diagnoses were used in the analyses because in some cases MMPI data were available by the time the discharge diagnosis was formulated.) For the small number of cases for which there was a change in the primary diagnosis, there were no changes that might imply major affective symptomatology for subjects in the schizophrenic group or psychotic symptomatology related to nonaffective disorders for those in the major depression group.

The third facility was a state psychiatric hospital in Minnesota. At this setting the initial diagnosis was made following a clinical examination by a staff psychiatrist. The patient's diagnosis at this facility is periodically updated and changed (when indicated) in consultation with a treatment team consisting of the staff psychiatrist, a clinical psychologist, psychiatric residents, psychology trainees, psychiatric social workers, and psychiatric nurses. The most recent diagnosis at the time the MMPI—AX was administered was used in this study. Data from the MMPI—AX were not available to the clinicians when the diagnoses used in this study were rendered.

Due to limitations on staff time, no interrater reliability data were available. However, the procedures used, including consultation with other clinicians, should ensure an acceptable degree of diagnostic reliability. In addition to their primary diagnosis, patients in the schizophrenic group had the following secondary diagnoses: 9 patients, chemical dependency/abuse diagnosis; 1, sexual disorder; 1, antisocial personality disorder; 1, schizotypal personality disorder; and 1, dependent personality disorder. Of the patients in this group, 63 had no secondary diagnosis. Of the 84 patients with a primary diagnosis of major depression, 15 had a secondary diagnosis of chemical dependence/abuse; 2, secondary diagnoses of dysthymia; 6, mixed personality disorder; 2, antisocial personality disorder; 2, borderline personality disorder; and 1, histrionic personality disorder. Of the patients in this group, 53 had no secondary diagnosis.

Subjects in the two groups did not differ significantly in age (schizophrenic men: M = 30.4, SD = 7.6; depressed men: M = 31.2, SD = 9.4; schizophrenic women: M = 38.2, SD = 10.0; depressed women: M = 35.3, SD = 11.1), education (schizophrenic men: M = 12.4, SD 2.4; depressed men: M = 12.2, SD = 2.4; schizophrenic women: M = 12.9, SD = 1.6; depressed women: M = 12.0, SD 2.3), or racial composition (percentage of Whites in each subsample was 72.7, 87.8, 87.1, and 90.2, respectively).

All subjects completed the MMPI—AX, the experimental form of the MMPI used in the restandardization project. This form contained all 550 original MMPI items plus 154 new, experimental items. The MMPI—2 contains 567 items, all of which were included in form AX. In the present study scores on the MMPI—2 were derived from the MMPI—AX. T scores presented in the tables that follow are uniform (see Butcher et al., 1989 ) for the clinical (omitting Masculinity—Femininity [ Mf ] and Social Introversion [ Si ]) and the content scales, and K corrected for scales Hypochondriasis ( Hs ), Psychopathic Deviate ( Pd ), Psychasthenia ( Pt ), Schizophrenia ( Sc ), and Hypomania ( Ma ). Because the content scales investigated in this study are new to the MMPI—2, a brief description of each scale can be found in the Appendix .

The MMPI—2 content scales' internal consistency in clinical samples ranges from .71 to .91, with most of the scales' coefficients alpha falling in the .80s range ( Butcher et al., 1990 ); test—retest reliability in a normal sample ranges from .78 to .91, with most coefficients falling in the .80s ( Butcher et al., 1990 ). Initial validity analyses in a normal sample provided good evidence of the scales' validity ( Butcher et al., 1990 ). Item overlap between the content scales and the clinical scales varies from scale to scale. Thus, for example, 9 of the 33 items scored on the DEP content scales appear on Clinical Scale 2, and 8 of the 23 items scored on BIZ appear on Clinical Scale 8.

Mean scores for the schizophrenic and major depression groups were computed on the 13 MMPI—2 basic validity and clinical scales and on the 15 MMPI—2 content scales. Table 1 presents the mean T scores and standard deviations for the two groups on the MMPI—2 basic scales, along with the results of statistical tests of the significance of differences between scores found for the two groups.

As seen in Table 1 , a multivariate analysis of variance (MANOVA) indicated, for both genders, that the two groups differed significantly across the 13 scales that make up the basic profile. Univariate analyses of variance (ANOVAs) were then conducted within each gender to determine which specific scales were contributing to the overall difference between the groups. Because these analyses involved 13 concurrent statistical tests, the Bonferroni correction was applied, and alpha was set at p < .004. Among the men in this study, only one scale was found to differentiate the groups significantly: Men diagnosed with major depression scored significantly higher on Scale D (Depression) than did the schizophrenic cohort. Among the women, the two groups differed significantly on only one scale, Pt, with the depressed group scoring significantly higher than did the schizophrenics.

Turning to scale elevation (i.e., scales on which the group mean exceeded a T score of 65), the results presented in Table 1 indicate that both groups' mean basic profiles present a rather heterogeneous clinical picture. As a group, schizophrenic men had elevated scores on Scales F, D, Pd, Paranoia ( Pa ), Pt, and Sc, whereas depressed men had clinical elevations on Scales F, Hs, D, Hysteria ( Hy ), Pd, Pa, Pt, and Sc . Schizophrenic women produced elevations on Scales F, Pd, Pa, and Sc, whereas depressed women had elevated scores on Scales F, D, Pd, Pa, Pt, and Sc .

Table 2 presents group means and standard deviations and the results of statistical tests of the significance of differences found between the two groups on the MMPI—2 content scales. For both genders, MANOVAs indicated significant group differences across the content-scales profile. Once more, univariate ANOVAs were conducted, with alpha set at .004, to identify specific content scales contributing to these differences. These analyses revealed that schizophrenic men scored significantly higher on BIZ than did men in the depressed group, and the latter scored significantly higher on Depression ( DEP ) and Social Discomfort ( SOD ) than did the schizophrenics. Among the women, the depressed group scored significantly higher on Anxiety ( ANX ) and Depression ( DEP ) than did the schizophrenics.

With respect to scale elevation, the content scales produced a far less heterogeneous picture than did the clinical scales. As a group, the schizophrenics had elevated scores on DEP, Bizarre Mentation ( BIZ ), Low Self-Esteem ( LSE ), Family Problems ( FAM ), and Work Interference ( WRK ), whereas the depressed men had elevated scores on DEP, LSE, SOD, WRK, and Negative Treatment Indicators ( TRT ). Schizophrenic women's scores were elevated only on BIZ, whereas depressed women had elevated scores on ANX, DEP, LSE, SOD, WRK, and TRT .

Results presented to this point indicate that both the traditional clinical scales and the new MMPI—2 content scales may contribute to the differential diagnosis of schizophrenia and major depression. What has yet to be addressed is whether the content scales contribute incrementally to this task or all the information available through the MMPI—2 could be obtained from the clinical scales alone. To answer this question, we performed hierarchical regression analyses to test statistically the incremental contribution of the content scales to differentiating between the two groups. These analyses were performed in the following manner. The dependent variable was a dichotomous dummy variable representing a diagnosis of schizophrenia or major depression. The independent variables were entered into the analysis in two blocks. The first block consisted of the 13 validity and clinical scales on the basic profile; the second block contained the 15 MMPI—2 content scales. Within each block, variables were entered in steps up until the point at which they no longer contributed significantly to the prediction of diagnosis. The first variable entered was the scale correlating most highly with the dependent variable. The variable entered in the next step was the scale with the highest partial correlation with the dependent variable, with the correlation between all previously entered variables and the dependent variable partialed out. Thus, all scales that entered into the regression equation in the second block contributed incrementally to the prediction of diagnosis.

Table 3 presents the results of these analyses. Focusing first on the findings for the men, two of the basic scales examined in the first block, D and F, contributed significantly to the prediction of diagnosis. In the second block, two content scales, DEP and BIZ, contributed incrementally to the differential diagnosis of schizophrenia and major depression, raising the total amount of variance in diagnosis accounted for by the MMPI—2 from 22% to 37%. For the women, three basic scales, D, F, and Pt, contributed significantly to the prediction of diagnosis, and one content scale, BIZ, added significantly to this prediction. Here, the increase in the prediction of diagnosis went from 19% to 27% of the variance.

A second set of analyses was conducted to ascertain the extent to which the content scales alone could account for variance in diagnosis. Here, the order of blocks entered into the regression procedure was switched so that the content scales were entered first, followed by the validity and clinical scales. It is interesting that, for both men and women, none of the basic scales contributed significantly after the content scales had been entered into the equation. Thus, for men, Scales DEP and BIZ alone accounted reliably for 36% of the variance in diagnosis, and for women, scales DEP, BIZ and ANX accounted reliably for 25% of the variance.

The most informative test of the potential contribution of the content scales to the task of differential diagnosis involved the multiple regression analyses presented in Table 3 . These analyses demonstrated clearly that the MMPI—2 content scales contain information pertaining to the differentiation between schizophrenia and major depression beyond what is found in the clinical scales. In other words, the content scales contributed incrementally to this differential diagnostic task. Subsequent analyses indicated that, with respect to differentiating between schizophrenia and major depression, the MMPI—2 content scales appear to carry almost all of the predictive weight. In fact, these analyses demonstrated that, for both genders, the clinical scales do not possess incremental information pertinent to differentiating between the two groups beyond what is found in the MMPI—2 content scales.

What are the implications of these findings for clinical applications of the MMPI—2? On the basis of this study's results, we conclude that the MMPI—2 content scales are a viable source of information pertaining to the task of differential diagnosis of schizophrenia and major depression. They convey information that can be useful in refining the clinical picture suggested by the basic profile. Thus, we suggest that the MMPI—2 content scales be used as adjuncts to the interpretation of scores on the basic profile of validity and clinical scales.

The finding that the content scales in fact outperformed the clinical scales in the particular diagnostic task involved in this study should not be interpreted to mean that they might replace the clinical scales. These findings do suggest, however, that, with the enhancement of the MMPI—2 item pool to include more modern formulations of these disorders, the MMPI—2 content scales may be better suited for this specific task. Nevertheless, personality assessment in clinical settings goes far beyond the proposal of a diagnosis, and the clinical scales, based on a strong foundation of research and clinical experience that has accumulated over the years, will continue to be at the forefront of MMPI—2 interpretation. In addition, content scales, by the obvious nature of their items, are susceptible to faking, both good and bad, which may limit their utility with some patients and in certain settings.

The current study has a number of limitations that should be addressed in future investigations. First, the diagnoses used in this research were not derived by structured psychiatric interview, and we were not able to conduct interrater reliability analyses. Second, the ratio of subjects to variables studied was somewhat low. Because the focus of this investigation was more on whether the MMPI—2 content scales possess incremental validity, and less on the identification of specific diagnostic cutoff scores or regression weights, the major findings of this study were unaffected by these limitations. However, to provide more precise information on the degree of incremental information contained in the content scales and specific psychodiagnostic recommendations, these findings should be replicated with larger samples and more rigorous diagnostic procedures. Further studies, focusing on additional diagnostic groups, are also necessary to test the full scope of the content scales' ability to add to the clinical picture conveyed by the traditional MMPI—2 scales. On the basis of this study's findings, the MMPI—2 content scales appear to possess considerable potential for supplementing the clinical scales in differentiating various forms of psychological disorder.

Table 4

Table 4b

Table 1.

Table 2.

Table 3.