Skip to main content

Empirical correlates of naturally occurring MMPI subgroups in an inpatient dual diagnosis population

17 April 1992


The purpose of this study was to investigate the presence of discrete childhood variables within subgroups of the dually diagnosed population. The Minnesota Multiphasic Personality Inventory (MMPI) profiles of 212 dually diagnosed patients were sorted into subgroups based on code types identified in a review of the literature. This sort resulted in 28% of the population being classified into six MMPI subgroups. In order to increase the number of MMPI subgroups, the sample was re-sorted using Friedman, Webb, and Lewak's (1989) rules. This resulted in 80% of the population being sorted into 19 subgroups. There were five subgroups which occurred in both sorting procedures resulting in a total of 20 unique MMPI code types. Patient charts were then reviewed for the presence of specific childhood variables. A list of these childhood variables was generated using information from the literature review and from a preliminary chart review. using the chi-square statistic (n=212), significant differences were found between subgroups in the presence or absence of one or more childhood variables. Among the six subgroups with primary scale 4 elevations, the 4-9 elevation subgroup had the largest number of significant childhood variables. These included a history of childhood behavior problems, abuse by parents, and parents getting divorced. The 4-9-8 subgroup, unlike the 4-9 subgroup, was more destructive and assaultive in childhood and had more familial alcoholism. The spike 4, 4-6, and 4-2 subgroups had a significant absence of childhood behavior problems and family pathology. Unlike other scale 4 subgroups, the 4-6 subgroup had a significant incidence of familial drug use and the 4-8 subgroup had a significant incidence of the death of their fathers. six subgroups had primary elevations on scale 2. The 2-4-8 subgroup was the most heterogenous with more significant childhood variables than any other subgroup in the sample. These variables included physical deformity, depression, oppositionality, parental abuse and neglect, and mental illness in relatives. The 2-4-6-8 subgroup also had a significantly high incidence of parental abuse and neglect, along with childhood sexual abuse and death of patients' fathers. The 2-7-8 subgroup was all male and had a significant incidence of parental neglect, along with mental illness and alcoholism in their sisters. Familial mental illness was also significant in the 2-4-7 subgroup. The 2-8 subgroup had a significant incidence of arrests and out of home placements. The spike 2 code type had paternal depression as its sole significant childhood variable. There were five subgroups with primary elevations on scale 8. The 8-9 subgroup was unique in reports of school behavior problems; physical illnesses, and physical and mental illnesses in relatives. The 8-6 and 8-7-6 subgroups had significantly more emotional problems than other subgroups. The 8-7-6 subgroup had a significant incidence of paternal verbal abuse and the 8-6 subgroup had a significant incidence of paternal sexual abuse. The 8-7 subgroup was nondistinctive from other scale 8 subgroups. Among the three remaining subgroups, the scale 6 subgroup had a significant absence of childhood behavior problems, and a significant presence of paternal depression and alcoholism. The scale 3 subgroup was predominantly female, and was significant for physical illness in the patient and mother. The scale 1 subgroup reported a confusing set of variables that were not indicative -of a particular pattern. The results of the birth order variable were unique and worth special note. The 4-9-8 and 8-9 subgroups were predominantly oldest children, while the spike 4 and 8-7 subgroups were youngest children. Based on the above results, support was demonstrated for the hypothesis of this project that there are unique childhood variables in the subgroups of the dually diagnosed population.


Files are restricted to Pacific University. Sign in to view.