Psychological Testing

Misuse of the MMPI and MMPI-2

Scale 5 0verinterpretations

Overinterpretation of the K Scale in Court or Custody Settings

Failure to Recognize the Situational Factors in a Scale 6 Elevation

Departing from Standard Administration Procedures

Overinterpretation of the MMPI Supplementary Scales

Ignoring a Within Normal Limits Profile and Finding Pathology with Projective Tests

Millon Clinical Multiaxial Inventory (MCMI and MCMI-II)

Multiphasic Sex Inventory

The Penile Plethysmograph

Misuse of the MMPI and MMPI-2

Ziskin (1981) notes that the MMPI better fits the forensic requirements for evidence to be believable and understandable than do other assessment methods. The MMPI has years of validation research and the data obtained from it are objective and quantifiable. The goal of the MMPI-2 revision committee was to develop the MMPI-2 so that the research on the original MMPI was still relevant and usable. There is dispute, however, as to whether this goal was realized (see Chapter 12 in this volume).

The major problem with the MMPI and MMPI-2 is that mental health professionals may give testimony that is far beyond what the test can assess. In their reports, depositions, and testimony these professionals make interpretations and draw conclusions about how an individual's MMPI is in some fashion typical or not typical of sex offenders.

The MMPI and the MMPI-2 have no scales that determine whether or not an individual is a pedophile or a sex offender. There has been research on MMPI scale elevations in sex offenders, but there is no typical sex offender MMPI profile. Although mean profiles often involve scales 4 and 8, with 9 and 2 also sometimes elevated, these elevations were also found in murderers, arsonists, and property offenders in a forensic psychiatric facility (Quinsey et al., 1980). Elevations on scale 4 are common in prison populations (Murphy and Peters, 1992). The MMPI cannot establish whether an individual is a sex offender.

The MMPI and MMPI-2 can provide information about personality characteristics that can be useful in the overall analysis of a case. It is most useful when there are allegations of highly deviant, low base rate, or sadistic abuse which the individual denies, and a valid, within normal limits MMPI suggests the absence of psychopathology. In such cases, the clinician must pay attention to the discrepancy.

We have observed several specific errors in interpretation made with the MMPI in child sexual abuse cases. MMPIs are often overinterpreted and misinterpreted. Such erroneous interpretations are not simply a matter of a difference of opinion; they are wrong and cannot be justified by the literature. Psychologists making such interpretations should be confronted and required to produce the research supporting their claims.
 

Scale 5 0verinterpretations

A scale 5 (masculinity-femininity) elevation may be interpreted as reflecting sexual conflict and sexual dissatisfaction and thus making it likely that the person committed a sexual offense. We have seen a psychologist testify that a scale 5 elevation meant the person had a tendency to act out sexually with a child.

It is mistake to interpret an elevation on scale 5 as reflecting sexual conflicts or as meaning it is likely that the person is homosexual or a child molester, since there are many factors behind such an elevation. The MMPI-2 norms have resulted in much lower scale 5 elevations in males, so perhaps the frequency of this particular misinterpretation will be less in the future. Scale 5 is the least well defined and understood of the MMPI clinical scales (Butcher, 1990).

An elevation on 5 in males is believed to reflect an intelligent, tolerant, imaginative, creative, sensitive, and empathic individual with a wide range of interests which do not fit the masculine stereotype. Scale 5 is highly correlated with education, intelligence, and social class and interpretations must take these factors into account (Butcher, 1990). A very high elevation (76 and above) is believed to be found in males who do not identify with the traditional masculine role and such elevations may indicate passivity and conflicts over sexual identity. However, there is no indication in the MMPI literature that child molesters or other sex offenders are more likely to score high on scale 5. Any testimony that a scale 5 elevation is typical of pedophiles or child sexual abusers should be countered by the lack of empirical support for such an assertion.
 

Overinterpretation of the K Scale in Court or Custody Settings

An overinterpretation of a high K (defensiveness) scale in a court or custody setting is a common error. Any conclusions about defensiveness on the MMPI must be qualified in terms of the testing situation. Elevations on the K scale in persons taking the MMPI in custody and court situations are common and must not be interpreted as signifying defensiveness as a personality characteristic. It is a normal and adaptive response to the situation. Graham (1988) notes that, if he doesn't see an elevation on K in a custody evaluation, he wonders what is the matter-doesn't the person want the child?

We have seen numerous forensic cases where a K elevation in an otherwise within normal limits MMPI was interpreted by the psychologist as "clinically significant." In one case, the psychologist claimed the K elevation meant that the client was defensive and was trying to "present himself in the best light psychologically and emotionally" and was "trying to answer the questions in the direction of looking good." He further claimed that "Sexually, this kind of thing (an elevation on the K scale) is expected." There were no qualifications in terms of the setting in which the MMPI was taken. In addition, this was a professionally and occupationally successful man with college education. The person's social class and educational level must be considered in interpreting K since persons from higher social classes typically produce K scores on the MMPI-2 between 55 and 65 (Butcher, 1990).
 

Failure to Recognize the Situational Factors in a Scale 6 Elevation

An elevation in scale 6 (paranoia) is a common response in persons who have been accused of sexual abuse and who deny the allegations. This is due to the affirmation of such items as he knows who is responsible for most of his troubles, someone has it in for him, he believes he is being plotted against, and he is sure he is being talked about. Rather than reflecting anger, hostility, suspiciousness, and paranoia as a pathological personality trait, the endorsement of these persecutory items reflects the individual's current reality and is a normal response to the situation. We have done research on this (Wakefield and Underwager, 1988a and 1988b), and Ziskin (1981) also discusses such situational effects on scale 6 and recommends caution in interpreting scale 6 elevations in such circumstances.

It is an error to interpret a scale 6 elevation in such a situation as indicating high defensiveness, anger, distrust, sexual conflict, poor behavioral controls, and tendencies toward acting out conflicts and impulses. In one case, a scale 6 elevation in a person accused of sexual abuse was labeled "seriously abnormal," a "very pathological profile," "scary" and the conclusion was made that the person was very likely to be a sexual abuser.
 

Departing from Standard Administration Procedures

Occasionally, a psychologist will send MMPIs home to be finished, or deviate from the standardized administration in other ways. In one case, the client left several items unanswered and the psychologist called him up and read the questions and recorded the answers over the telephone.

Whereas psychologists may sometimes deviate from standardized administrations with therapy clients, it is never acceptable for a forensic evaluation where the results of the evaluation are to be presented in the justice system and are to be used in making decisions about people's lives. Ziskin (1981) warns against this practice:

The "take home" MMPI should be avoided in the forensic situation. . . . This practice can lead to questions as to whether the individual took the test in the standard way and whether all of the responses are purely his own, as highlighted by Graham's amusing anecdote about the mental hospital patient who had his ward colleagues assist him by voting on the appropriate answers. (p. 7)
 

Overinterpretation of the MMPI Supplementary Scales

The supplementary scales must be interpreted cautiously when the basic clinical scales are within normal limits and the interpretations must be on the basis of rules based on research. For example, in one case in a custody evaluation, the clinical scales for the father were all well within normal limits but the dominance scale was elevated. The MMPI was interpreted as indicating that the father had a "highly assertive and domineering style," whose leadership is "characterized by determination, inflexibility, and an almost autocratic control." In his trial testimony, the psychologist said that the father was "a very willful man" who has "not played the game right" and added that "All the time, I suspect what I saw in my tests undercuts that quite a bit, because assertiveness, being aggressive, dominance, can become autocraticness, and I think that's what has happened."

This is a misinterpretation of a dominance scale elevation in an otherwise within normal limits profile. Caldwell (1988) says the following about the appropriate interpretation of the Do (dominance) supplementary scale:

Although based on peer nominations of subjects as strong, confident, influential, unintimidated in face-to-face situations, and showing initiative and leadership. . .the title "dominance" may be partially misleading. That is, the scale reflects taking charge of one's own life-or not taking charge-considerably more than bossiness or being overbearing. Do should be interpreted as taking charge of one's life. . .e.g. as self-organizing, making workable plans, and meeting deadlines. (p. 56)

This description, was, in fact, quite accurate for this man.
 

Ignoring a Within Normal Limits Profile and Finding Pathology with Projective Tests

We see this frequently. The MMPI or MMPI-2 is valid and within normal limits, but a Rorschach or TAT, or even a Bender, often administered and scored idiosyncratically, forms the basis for a diagnosis of serious psychopathology. Two examples:

· The MMPI-2 was within normal limits and not defensive (K = 56). But the evaluator, who was very sympathetic to the woman who had accused her former husband of sexually abusing their child, said that this was because the man, a physician, was "in a sophisticated way, understating concerns in his life." On the basis of his clinical impressions, a few TAT stories, and a Rorschach interpreted with no scoring system, he diagnosed the man as Paranoid Schizophrenic and said that he was threatening and potentially dangerous. The man, a successful physician, had no history of mental illness nor dangerous or violent behavior, but expressed his anger at being falsely accused of sexually abusing his child.

· The MMPI was moderately defensive and within normal limits. But, on the basis of a Rorschach and the House-Tree-Person test, the man was said to have tied up his son with a blue bicycle chain and sodomized him. The Rorschach (which had no unusual responses) was interpreted as: ". . .highly defensive stance which is accompanied with blocking, censoring, and inhibition of his underlying affect. . . .an undercurrent of anxiety, unrequited love, and cloaked sexuality...difficulty with relating appropriately to others...latent polymorphous perverse orientation to the environment. . .fantasies (that may include) homosexual, bisexual, and exhibitionist feelings. . .hostility toward women. . ."

In such cases, the attorney can have the psychologist read the interpretation out loud, ask for the scientific literature supporting the assertions and the scientific literature supporting any contrary interpretations and make it clear to the finder-of-fact that this is meaningless jargon.
 

Millon Clinical Multiaxial Inventory (MCMI and MCMI-II)

When the Millon Clinical Multiaxial Inventory, along with the computerized interpretation is used, the psychologist is apt to report significant psychopathology. The computerized interpretation of the MCMI-II may be lifted verbatim and without qualification from the computerized printout which accompanies the test scoring.
This practice is a particular problem with the MCMI-II, which is normed on and intended to be used for a clinical population. When used for other assessment purposes, the MCMI-II must be interpreted extremely cautiously because of its tendency to overpathologize. The result of using these computerized interpretations greatly exaggerates psychopathology.

The problem is not in the test, but in its misuse. The test is normed entirely on clinical samples and is only intended for persons who have psychological symptoms and are being assessed for treatment and evaluation. The manual (Millon, 1987) clearly states that this test is "not a general personality instrument to be used for 'normal' populations or for purposes other than diagnostic screening or clinical assessments." (p. 7) Millon has repeatedly warned against using the inventory with people who are not psychiatric patients because the test norms may not be valid if the subject does not fit the standardizing (psychiatric) group (Choca et al., 1992) .

The MCMI can provide useful information when interpreted cautiously and conservatively. Choca et al. (1992) state that there is nothing intrinsically wrong with using the MCMI to test "normal" people as long as the evaluator is aware that the test was designed for and standardized with a psychiatric population. The user will have to make the appropriate adjustments. But this is seldom done. The computerized narrative must never be lifted verbatim into the report since it may find serious psychopathology and personality disorders in just about everyone. The attorney should vigorously cross-examine a psychologist who does this.
 

Multiphasic Sex Inventory

The Multiphasic Sex Inventory (MSI) (Nichols and Molinder, 1984) is a self-report questionnaire which consists of statements about sexual activities, problems, and experiences. It has scales which assess the level of openness about the deviant sexual behaviors. The authors state that it has been used by over 1400 clinicians, clinics, universities, and institutions. Although the authors report on the use of the MSI in studies of sex offenders, it has not been reviewed in Buros. It is intended to be used in assessing sex offenders to develop treatment plans and to be used during treatment to assess progress. However, it is also sometimes used to assess an individual who denies sexual abuse to determine whether the individual actually is an abuser.

This test is not intended for this purpose. It must never be used when the defendant is denying the offense. The manual accompanying the MSI states, "[I]t is important to remember that the MSI is not appropriate for use in the legal pursuit of guilt or innocence. The alleged offender must acknowledge culpability in order for the inventory to be used" (Nichols and Molinder, 1984, p. 39). It must never be used on an individual who denies being a sex offender or as part of an assessment to determine whether someone who denies an alleged sex offense is likely to have actually done it.
 

The Penile Plethysmograph

The penile plethysmograph is a technique which attempts to measure sexual arousal by recording the penile responses during the presentation of sexual stimuli. The stimuli consist of slides of nude male and female adults and children and the audiotapes portray a variety of sexual activities. During the presentation of the stimuli, the penile responses are recorded with a volumetric or a circumferential device. Supporters claim that this technique permits assessment of sexual arousal and hence, sexual preferences and deviancy.

This technique is controversial and should never be used with someone who denies sexual abuse in order to assess the veracity of the denial. Plethysmograph researchers claim that plethysmography can be useful in treatment, but is of limited use with known sex offenders in predicting future behavior, and is of no use in screening a normal population. It cannot be used to determine whether a person who has been accused of sexual abuse and is denying it is telling the truth. There are virtually no data related to the use of the plethysmograph with adolescents (Murphy et al., 1991). Despite these limitations, the plethysmograph is often used in evaluations of both adults and adolescents in sexual offense cases.

Problems with the penile plethysmograph include:

· There is a lack of standardization for training in the use of the plethysmograph (Murphy and Peters, 1992).

· There are no standards controlling the type of erotic stimuli used and the method of presentation (Barker and Howell, 1992; Murphy and Peters, 1992; Schouten and Simon, 1992; Simon and Schouten, 1991).

· There are no generally agreed-upon guidelines as to normal and deviant phallometric response ranges (Simon and Schouten, 1991).

· There is a lack of adequate normative data in which the sexually deviant population is compared to a normal population. Without standardized norms, interpretation is impossible (Barker and Howell, 1992).

· Studies with normal controls indicate that a high percentage of control subjects respond with deviant arousal patterns (Annon, 1993; Freund and Watson, 1991; Simon and Schouten, 1991). The high percentage of normal controls who show arousal to deviant stimuli on the plethysmograph means that arousal to deviant themes does not confirm sexual deviance.

· Subjects are readily able to manipulate their erectile responses. There is no completely adequate way or generally accepted procedures for detecting, preventing, or correcting for faking on the plethysmograph (Barker and Howell, 1992; Hall et al., 1988; Langevin, 1988; Murphy and Peters, 1992; Proulx et al., 1993; Quinsey and Laws, 1990; Schouten and Simon, 1992; Simon and Schouten, 1991; Travin et al., 1988).

· Although some research with adults has been able to separate offenders from nonoffenders on the group level, statistically significant differences between groups does not automatically translate into functionally significant differences for interpreting an individual's pattern of erectile responding (Marshall and Eccles,1991; Murphy and Peters, 1992).

· Incestuous offenders tend to show arousal patterns that are similar to nonoffenders (Murphy and Peters, 1992).

· Efforts to calculate a "pedophile index" and use a cutoff score or the use of discriminate analyses results in many misclassifications and produces a high rate of both false negatives and false positives (Murphy and Peters, 1992; Simon and Schouten, 1991).

· Although reliability is necessary for the plethysmograph to be valid, the reliability in studies is influenced by variables such as the length of the test-retest interval, selection bias, stimulus content, and scoring methods. The research shows reliability coefficients ranging from .38 to .94 (Simon and Schouten, 1991).

· The research on the relationship between changes in arousal patterns after treatment and recidivism is limited and the evidence is that changes in erectile responding with treatment do not predict outcome (Blader and Marshall, 1989).

· Although the rationale for using the plethysmograph is that psychophysiological assessment is necessary because sex offenders cannot be taken at their word, one study (Day et al., 1989) found that the self-report measures (MSI scales) were superior to psychophysiological measures in discriminating between groups classified on the basis of their offenses.

· Although the rationale for using the plethysmograph is that it can detect deviant arousal in offenders who are not truthful concerning their erotic likes and dislikes, the plethysmograph is not very sensitive for offenders who do not admit to a corresponding erotic preference (Freund and Watson, 1991).

· The evidence does not provide adequate support for the hypothesized relationship between sexual arousal in the laboratory and overt sexual acts (Barker and Howell, 1992; Hall et al., 1988; Simon and Schouten, 1991).

· Not all sex offenders have deviant arousal patterns that correspond to their criminal sexual behavior (Hall et al., 1988; Marshall and Eccles, 1991).

Murphy and Peters (1992) conclude about the forensic use of the penile plethysmograph:

The results of the studies using erection data suggest that, although group differences are reliably found, the ability to classify an individual would produce error rates that would not be appropriate for the trial situation. In addition, in cases of incest or when patients deny charges, one would even expect to find either no responding in the laboratory or a normal response pattern. Further, it is clear that individuals can fake their responses and the absence of significant responding is basically meaningless in terms of a clinical interpretation. Like the MMPI literature, we find the conditions under which the test has been validated do not meet legal requirements. (pp. 32-33)

Simon and Schouten (1991) argue:

The use of phallometric findings for important clinical and legal decisions and scientific inquiry should reflect a full appreciation of the measurement technique and the assumptions underlying its use. This becomes possible given adequate empirical support and clear explication of general principles. The validity and clinical utility of plethysmography in the assessment and treatment of sexual deviance remain to be established. (p. 87)

Barker and Howell (1992) state:

Misuse of the plethysmograph is a major concern. Using the plethysmograph to predict innocence, guilt, or likelihood of reoffending is beyond the scope of the test's validity. In this application the plethysmograph has not "gained the general acceptance" required by Frye vs. United States to be acceptable in a court of law. (p. 22)

McConaghy (1989) observes:

Though never validated as a measure of individuals' sexual arousal, PVR measures of erection are currently widely recommended for assessment and determining treatment of individual sex offenders. If these assessments could affect or are believed by the offenders to affect the outcome of the legal processes in which they are involved, the procedure is not only scientifically unsupported, it is unethical. (p. 357)

Pithers (quoted in Annon, 1993), in a deposition, states:

I know of no psychometric procedure or psychophysiological procedures that can be used to demonstrate with psychological certainty that a person has committed a legal offense or engaged in child sexual abuse or is likely to do so in the future. That is the province of sorcerers and witches, not of a psychologist. It clearly asserts that the practitioner has special powers beyond which most psychologists would assert themselves to have; and therefore, I believe it is a highly inappropriate response and potentially one for consideration by an ethical board. (p. 40)

In conclusion, research does not support the use of the plethysmograph as a technique to determine whether an individual who denies abuse is, in fact, sexually deviant, to make sentencing recommendations, or to predict recidivism. It is not generally accepted in the scientific community and meets neither the Frye test nor Daubert.