• No results found

Interpreting the MMPI-2 Validity Scales

James N. Butcher

F- Back Scale (FB)

4 Interpreting the MMPI-2 Validity Scales

Several methods and a number of basic and special scales are available to assess complex validity issues on the MMPI-2. Nine basic MMPI-2 validity measures provide the clinician with the essential frame of reference for interpreting the clinical and supplementary scales. These nine basic validity measures (?, VRIN, TRIN, F, FB, F(p), L, K, and S) give some indications of (a) whether the client understood and answered all the items in a consistent fashion; (b) the extent to which the client is endorsing items claiming the experience of psychological pain and unusual, bizarre experiences, behaviors, or ideas; and (c) the extent to which the client might be trying, intentionally or unintentionally, to give a favorable or unfavorable self-portrayal. In other words, the basic validity scales are useful in determining the consistency and accuracy of reported symptoms, complaints, and traits. In addition to measuring validity, some of the validity scales also have personality correlates. The interpretations that follow, then, focus both on the validity aspects of the validity scales and, where applicable, the personality aspects.

Additional special validity scale guideline ranges are presented for: Ds, Mp, Sd, Ss, F-FB Index, Dissimulation Index (F – K), and the True Percent index.

As previously described in Chapter 3, the original MMPI was constructed with three basic validity scales: L, F, and K. The VRIN, TRIN, FB, F(p), and S scales were later developed for the MMPI-2 to assist in evaluating the consistency and accuracy of a given protocol. Except for some slight differences in cut scores, minor rewordings, and the loss of four items on F, the three original validity scales (L, F, and K) are the same for the MMPI-2 as they were for the MMPI. The contemporary normative sample did slightly alter the T-score values for the L, F, and K scales.

Chapter 3 provides a detailed description of the construction methodology, psychometric operating features, and reliability and validity research for all of the scales described in this chapter. This chapter also describes various response sets (e.g. All-True, All-False, Random) and illustrates profiles with common validity patterns.

Cannot Say Score (?)

The ? scale is misnamed in that it is not a scale, but simply a report of the number of unscorable items (i.e. double-marked or omitted items). Consequently, we refer to it as the Cannot Say score. On the MMPI-2, the number of unscorable items is recorded on the bottom left of the profile form, and no T-score value is assigned. The original MMPI test instructions were less stringent than those adopted for the MMPI-2, with test takers now asked to answer all of the items. The mean number of omitted items

100 Interpreting the MMPI-2 Validity Scales

on the MMPI-2 thereby reduced and their significance is thereby increased over that for the MMPI.

Normal Range: 5 or Fewer Unscorable Items

Little, if any, validity significance is associated with the omission of five or fewer items, although examining omitted items may yield clinical hypotheses. It is unlikely that omissions in this range will distort the profile unless they are narrowly confined to a particular content area. In forensic settings, every omitted item may be significant and should be noted by the expert witness, as the opposing attorney might question the omission(s) during cross-examination. In such settings, a protocol could be returned for completion even if only four or five items are left unanswered. The psychologist should be aware that a cross-examining attorney may ask, “Did the defendant respond to all the items?” and, if the answer is no, “Which items were not answered?” Thus, if feasible, the client should be encouraged to complete the omitted items on the test, and to provide an explanation for them.

Moderate to High Range: 6 to 29 Unscorable Items

Raw scores of 6 to 29, if reading or comprehension difficulties can be ruled out, generally indicate idiosyncratic and contentious interpretation of items and/or obsessional indecision, often with extreme intellectualization. This process may be seen in the legalistic over-cautiousness of some paranoid clients or the indecision of anxious, obsessive–compulsive ones. In other individuals, this range may reflect a reticence to respond openly to test items that probe personal experiences or private feelings and experiences in particularly sensitive areas for that person. It is recommended that the clinician examine the content and scale membership(s) of omitted items.

Marked Elevation (Possibly Invalid Profile): 30 or More Unscorable Items

If more than 30 items are omitted, the elevation and the code pattern of the clinical profile is likely to be significantly lowered (except for Scale 5 for women, which could be elevated), or otherwise altered, reducing the probability of obtaining an accurate picture of the person’s unique personality characteristics. Extremely high raw scores can occur among severely emotionally impaired persons, who are unable to perform the decision-making task because of profound depression, intrusive thoughts, excessive preoccupation with the exact meaning of the items, or apprehension that the items’

dichotomous nature will unfairly put them in some negative or inaccurate category. It is probable, however, that such a high score indicates a resistive, uncooperative person or a person with substantial reading or comprehension problems.

The clinician should attempt to discover the reasons for the individual’s Cannot Say propensity when so many items are omitted before asking the individual to retake the test. The clinician should sensitively probe the reasons for items being unanswered or double marked and reassure the individual that the assessment process is collaborative in nature—unless, of course, the client is taking the test in a litigation, personnel selection, or employment fitness-for-duty context. Even so, it is possible to interpret profiles with

limited accuracy when as many as 30 items are omitted—although the attenuating effect on the profile must be considered because clinical scales can drop significantly and the code pattern can also change. Note that if the omitted items all or mostly occur after Item 370, the basic validity and clinical scales will not be affected.

If up to 30 items are omitted, Greene (2011) suggested augmenting the profile by re-plotting it after scoring the omitted or double-marked items in the keyed direction.

See Chapter 2 under “Test Scoring and Profile Plotting” for a fuller description of this procedure. In general, if more than 30 items have been omitted, the profile is likely invalid.

Variable Response Inconsistency Scale (VRIN) (67-item pairs)

This MMPI-2 scale was designed to detect inconsistent responding. It is particularly useful when F and/or FB (or both) is elevated and there is uncertainty as to whether the elevations reflect severe psychopathology, random, or careless responding, or symptom exaggeration. In such cases, an elevated VRIN would suggest that the F and FB elevations are due to confusion or careless responding, whereas a low VRIN score would suggest that the high F or FB score (or both) is due to severe psychopathology or deliberate exaggeration of psychopathology. Cut scores for invalidity are inexact, but the following ranges provide interpretive guidelines.

Consistent Responding: Raw Scores Between 0 and 9, T-Scores Between 30 and 66 Because the determination of validity is a multivariate process, it is important to examine other indicators of response consistency when assessing the validity of the profile. Raw scores between 0 and 9 on VRIN typically reflect consistent responding, with a raw score of 9 equaling a T-score of 65 for males and 66 for females. Scores in this range suggest the test taker understood and responded to the items in a consistent manner.

Profile Indeterminate: Raw Scores Between 10 and 14, T-Scores Between 69 and 86 In this range, the profile could reflect either inconsistent or consistent responding. The MMPI-2 Manual (Butcher et al., 2001) suggests that a raw score of 13 or higher (T-score of 80 for men, 82 for women) should be the cut score for determining inconsistent responding. However, because the Mean VRIN score for random responding is only 16.75, a cutting score of 13 may be too liberal, and raw scores above 11 should be viewed with suspicion. Examine the TRIN, F, FB, and F(p) scores to further determine the validity of the profiles. Also consider where the inconsistently-endorsed VRIN item pairs are concentrated, and especially on which side of item 370 the endorsements occur. An imbalance in the proportion of True to False responses could also account for scores in this range; so, the T percent value should be checked (see Chapter 3 and

“T Percent Index” in this chapter). Severely disturbed individuals (whose confusion and uncertainty is part of their symptom picture), or uncooperative clients, often score in this range on the VRIN scale. However, if the VRIN score is in the upper end of this large range, it is unlikely that the MMPI-2 profile validly reflects the individual’s stable personality, unless TRIN, F, FB, and F(p) appear valid.

102 Interpreting the MMPI-2 Validity Scales

Inconsistent Responding: Raw Scores Greater Than 14, Τ-Scores of 87 or Greater Raw scores over 14 (T-84 [males]; T-86 [females]) on the VRIN scale strongly suggest that the profile is invalid due to inconsistent or random responding. At this level, it can be difficult to discern whether such random elements are caused by carelessness or by a compromised mental state characterized by confusion and uncertainty. Even if the profile is valid by F, FB, F(p), L, and K criteria, the high VRIN score suggests so much inconsistency that the profile’s validity is compromised.

True Response Inconsistency Scale (TRIN) (23-item pairs)

This 23-item scale developed for the MMPI-2 measures a participant’s tendency to answer True or False to the MMPI-2 items without careful consideration of the meaning of the items. Some individuals may answer the items with a tendency to acquiesce, responding True (yea-saying) to the items even if their answers are inconsistent. Others respond with resistance or defensiveness, responding False (nay-saying) without full regard for the meaning of an item.

A majority of MMPI-2 items are scored in the pathological direction when marked True; thus, severely disturbed, non-defensive individuals will produce high F scores, high scores on Scales 6 and 8, and a higher percentage of True responses. The exception, however, are somatizing profiles (e.g. 123/321 and 13), where more of the pathological responses are False. These individuals, like others who are approaching the test with a highly positive self-presentation, will produce a higher False than True response percentage.

The TRIN scale can be useful in determining a yea- or nay-saying response set, especially when True or False percentages are high. The TRIN score is computed by adding a point for every True TRIN item pair endorsed and subtracting a point for every False TRIN pair. Nine is added to the equation to ensure there are no negative numbers.

A profile is showing no evidence of yea- or nay-saying with a TRIN score of 9 (0 True items endorsed + 0 False items endorsed + 9 = 9). A raw score of 13 or greater suggests considerable yea-saying, and a score of five or less suggests nay-saying.

Raw Scores Between 8 and 10

If other validity indicators suggest the profile is valid, then TRIN scores in this range confirm it. The mean raw score for the re-standardization sample was about nine, which corresponds to a T-score of 50.

Raw Scores of 6 and 7, and 11 and 12

Raw scores of 6 and 7, and 11 and 12 are of uncertain significance. When TRIN exceeds VRIN and True percent is low (≤ 25 percent), raw scores of 6–7 suggest an indiscriminate preference for False-marking or nay-saying, and retesting may be warranted. Similarly, when TRIN exceeds VRIN and True percent is high (≥ 60 percent), raw scores of 11–12 suggest an acquiescent or yea-saying response style, and may also warrant retesting.

In all other cases, such moderate elevations on TRIN may be allowed, unless VRIN

exceeds TRIN, in which case the interpretation of the VRIN score should supersede the interpretation of TRIN (Nichols, 2011).

Raw Scores of 5 or Less and 13 or Greater

The profile is invalid because of excessive yea- or nay-saying and the test should be retaken with appropriate test-taking instructions and/or proctoring.

Infrequency Scale (F) (60 items)

The F items reflect the degree to which thoughts, attitudes, and experiences reported by a person differ from the normative population. Persons who randomly mark items, who cannot comprehend item content, or who consciously attempt to portray a negative image of themselves endorse F items to a higher degree than the general population, and usually to a higher degree than genuinely disturbed populations. However, individuals who are authentically disturbed and unhappy also elevate the F scale, as would persons with unusual attitudes (e.g. sociopaths) or alienating experiences (e.g. minority populations). Because about one third of the F items overlap with Scales 6, 7, 8, and 9, elevations on the F scale are often accompanied by elevations on these clinical scales, particularly Scales 6 and 8. The following descriptions deal primarily with validity issues and, to some extent, address clinical correlates. The clinical information, however, is minor compared with the validity information, and should be ignored if contradicted by the information from the basic clinical scales. As described earlier in Chapter 3, it is imperative that other validity scales be examined when F is elevated; particularly VRIN, TRIN, FB, Ds, F(p), F-FB, and F-K indices, and the True percent index.

Low Scores: Raw Scores of 2 or Less, T-Scores Between 42 (Males) and 44 (Females) Raw scores of two or less on F suggest conventionality, if not over-conventionality, in that the person is reporting few, if any, unusual thoughts, beliefs, or behaviors, and is in no particular distress at the time of testing. Individuals without elevations on the clinical scales, then, may be described as calm, dependable, sincere, unpretentious, and honest, attributes one would expect from a person who is basically psychologically healthy. For some individuals, however, the possibility exists that they are denying or minimizing problems and that they have systematically avoided admitting to distressing or unusual feelings or thoughts. The other validity scales, however, would provide a clue to the source of the low F scores; thus a low L score and a normal K score would suggest that the low F score is, in fact, reflecting mental health rather than positive malingering.

Sometimes an individual obtains a very low F score but also obtains a moderately elevated clinical profile. This would suggest that the individual’s disturbance is not acute but more longstanding, and perhaps not psychologically disorganizing.

Moderate Scores: Raw Scores Between 3 and 7, T-Scores Between 45 and 61

Raw scores between three and seven indicate acknowledgment of a number of unusual experiences, attitudes, feelings, and behaviors that differ somewhat from the norm.

104 Interpreting the MMPI-2 Validity Scales

With adolescents, college populations, creative persons, or moderately disturbed people, such F elevations are not atypical. Similarly, when persons become intensely involved in religious, political, or social causes that are different from the mainstream, the scores on the F scale tend to be slightly elevated. Occasionally, however, particularly in mental health settings, such scores will occur in persons with some psychopathology, but who have adjusted to chronic problems, and therefore are not in great immediate distress.

Their problems are “contained” and perhaps ego-syntonic.

High Scores: Raw Scores Between 8 and 16, T-Scores Between 61 and 92

Valid raw scores between 8 and 16 suggest distress, increasing significantly toward the middle and high end of this range. Scores in this range indicate unusual or markedly unconventional thinking and attitudes, as well as psychological distress. Identity and self-esteem issues, as well as psychopathology, are indicated, and the person is likely to be described as moody, changeable, restless, unstable, dissatisfied, talkative, and opinionated but self-deprecating, all descriptions associated with psychological disturbance. The clinical scales are almost always elevated when the F scale is in this range, so the basic personality picture is described by the code pattern. The personality descriptors associated with F elevations in this range reflect general disturbance and distress. Young people struggling with identity problems and a need to define themselves through nonconformity frequently score in this range on F. Generally, as raw scores approach 16, the profile almost always reflects a serious, often panicked disturbance. F raw scores between 8 and 12 are typically seen in individuals with a longstanding and serious disturbance.

Borderline Valid Scores: Raw Scores Between 17 and 25, T-Scores Between 88 and 113 (for Men) and Between 88 and 120 (for Women)

Raw scores between 17 and 25 are in an indeterminate range and suggest the following possibilities: (a) failure of the client to understand the items; (b) resistance or lack of cooperation; (c) random or haphazard reporting of numerous F items because of delusional thinking, confusion, or other psychotic processes; (d) conscious distortion and exaggeration of psychological symptom for secondary gain; (e) panicked over-endorsement of F items, perhaps to draw attention to current fear, pain, and distress;

or (f) seriously disturbed individuals, perhaps in inpatient or prison settings. As the F raw score approaches 17, the importance of all the other validity indicators to rule out symptom magnification becomes even more critical. With scores this high, the examiner should suspect that the profile is invalid and should verify this by considering the characteristics of the remainder of the profile, the actual test-taking and interview behavior, and possible motivations for exaggeration that might derive from the client’s current situation. Even profiles with F scores of 25 may be occasionally interpreted in a psychiatric setting. An interpretable F score in this range is almost always associated with a profound disturbance, disorganized thinking and psychotic symptoms, although sometimes it may be a “cry for help.” A clinical interview usually reveals very atypical thought processes, poor judgment, and distortions of reality, often accompanied by confusion, withdrawal, and lack of cooperation.

Interpreting a profile with an F score between 22 and 25 is hazardous. Certainly, in a forensic setting, the clinician would have to be armed with solid supportive evidence to withstand a cross-examination to show that such a profile is validly interpretable.

Information would have to be presented from other tests, history, and interviews to paint an integrated picture of the individual so that the elevated profile has some context. For example, an expert witness psychologist might say, “Yes, by standard criteria this F score of 25 suggests invalidity. However the other tests suggest a serious psychotic disturbance and the patient’s history confirms it. In light of such data this MMPI-2 may reflect his current psychosis.” Although F scores in the very high range can be interpreted in some situations, they should not be interpreted if the scales showing confusion, inconsistency, or a biased response set (i.e. VR1N, TRIN, F(p), and Ds) are also significantly elevated.

Increasingly, it has been our experience that attorneys send, through a psychologist, their clients’ or opponent clients’ MMPI-2 to one of the computer-generated report services. These services tend to be conservative in validity cut scores so that an F score over 18 is ruled invalid. This may be too conservative in some cases, but the psychologist would need to be prepared to defend a claim of validity as the F score reaches 18 and above. It can be difficult to explain under cross-examination why a respected computer-generated report is inaccurate in suggesting that a profile is malingered or invalid and why the psychologist is accurate in suggesting it is valid. Consequently, the expert witness needs to be well prepared and have the ability to integrate the MMPI-2 data with other persuasive data if claiming an elevated F score is valid.

Very High/Invalid Scores: Raw Scores of 26 or Greater, T-Scores of 116 or Greater (for Men) and 120 or Greater (for Women)

Although F scores in the very high range (raw scores of 26 and above) sometimes can be very cautiously interpreted in situations in which there appears no motivation for

Although F scores in the very high range (raw scores of 26 and above) sometimes can be very cautiously interpreted in situations in which there appears no motivation for