interpreting validity.pdf

Interpreting the MMPI-3 Validity Scales 8 T his chapter provides interpretive guidelines for the MMPI-3 Valid- ity Scales. Develop...

Interpreting the MMPI-3 Validity Scales 8 T his chapter provides interpretive guidelines for the MMPI-3 Valid- ity Scales. Development of the scales is described in chapter 3. The literature available to help guide and support their interpretation is reviewed in chapter 4. We begin by discussing a conceptual framework for un- derstanding and using Validity Scales as measures of threats to protocol validity. Next, we map the MMPI-3 Validity Scales onto this framework and identify potential confounds that complicate their interpretation. In the subsequent section we provide interpretive guidelines for each of the MMPI-3 validity in- dicators, which include consideration of the confounds just mentioned. The chapter concludes with a review of cases that illustrate MMPI-3 Validity Scale interpretation. THREATS TO PROTOCOL VALIDITY To provide useful information in response to the statements that make up a self- report inventory, a test taker must read, comprehend, and respond accurately to the test items. Failure to do so, intentionally or unintentionally, compromises the utility of the resulting test scores, and, in extreme cases, renders them uninter- pretable. Therefore, prior to drawing any substantive inferences from self-report measures, careful consideration must be given to the quality of information pro- vided by the test taker, that is, to the validity of the individual test protocol. Ben-Porath (2013) described a conceptual framework for understanding and using Validity Scales as measures of protocol validity. This approach dis- tinguishes between instrument validity, which refers to the extent that scale 333 334 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S scores predict relevant criteria, canvas a relevant content domain, or reflect an underlying construct (i.e., criterion, content, or construct validity), and pro- tocol validity, which refers to the interpretability of a test taker’s scores. The premise underlying this approach is that even if a hypothetically 100% valid self-report measure were available, any given administration of the instrument could nonetheless yield invalid results if the test taker was unable or chose not to respond accurately to the test items. Threats to protocol validity fall broadly into two categories that reflect the role of item content in invalid responding: non-content-based and content-based. Important distinctions can be made within each of these categories as well. Non-Content-Based Invalid Responding Non-content-based invalid responding occurs when the test taker’s responses are not based on an accurate reading and comprehension of the test items. Its deleterious effects on protocol validity are obvious: to the extent that a test taker’s responses do not reflect their reactions to the actual items, the responses cannot gauge the individual’s standing on the constructs of interest. This in- valid test-taking approach can be divided further into three subtypes: non- responding, random responding, and fixed responding. Nonresponding Nonresponding occurs when the test taker fails to provide a scorable response to an item. Typically, this is the absence of a response, but if the test taker answers both True and False to an item (which is only possible with paper and pencil ad- ministration), that is also considered a nonresponse. Nonresponding may occur for a variety of reasons. Test takers who are uncooperative or defensive may fail to respond to items, but excessive nonresponding may also reflect a test taker’s lack of reading and language comprehension, cognitive functioning deficits lead- ing to confusion or obsessing over responses, or limited introspection and insight. The effect of nonresponding on protocol validity depends partly on the re- sponse format of the instrument. In tests that use a True/False response format, a nonresponse is typically considered a response in the nonkeyed direction be- cause raw scores are derived by counting the number of responses given in the keyed direction. However, it cannot be assumed that these de facto responses approximate how the test taker would have actually responded. Therefore, to the extent that nonresponding occurs in a protocol, it will distort the result- ing test scores by lowering them artificially. If not identified and considered, nonresponding can lead to underestimation of the individual’s standing on the constructs measured by the affected scales.  INTERPRETING THE MMPI-3 VALIDITY SCALES | 335 Random Responding Random responding is characterized by an unsystematic response pattern that is not based on an accurate reading and comprehension of test items. It is not a dichotomous phenomenon, meaning that random responding may be present to varying degrees in a test protocol. Two types of random responding can be distinguished. Intentional random responding occurs when the individual has the capacity to respond relevantly to test items but chooses to respond irrelevantly in an unsystematic manner. An uncooperative test taker may engage in intentional random responding instead of becoming involved in a confrontation with the examiner over their refusal to participate. In this example, the test taker provides answers to items without reading or considering their content. They may do this intermittently or consistently throughout the test protocol. This form of responding is also most common among uncooperative research participants. Unintentional random responding occurs when the individual is unable to respond relevantly to test items but responds anyway without understanding their content. Test takers are often not aware that they are responding in this way. Several factors may contribute to unintentional random responding. Read- ing difficulties could compromise the test taker’s ability to respond relevantly. Most current self-report measures require a fourth- to sixth-grade reading level for the test taker to be able to read, comprehend, and respond relevantly to test items. This is not synonymous with having completed 4 to 6 years of educa- tion. Comprehension deficits can also lead to random responding. The individ- ual may be able to read the test items but does not have the necessary language comprehension skills to process and understand them. This could be a product of low verbal ability or, for nonnative speakers, a lack of facility with the lan- guage in which the test is administered. Reading and comprehension difficulties tend to be relatively stable test taker characteristics that will likely compromise protocol validity regardless of when a test is administered. Other factors, such as confusion and thought disorganization, may be transitory. Finally, uninten- tional random responding may result from response recording errors. If the test taker mismarks responses on the answer sheet, they are essentially providing random responses. This could occur if an individual accidentally skips an item on the answer sheet or generally has a careless approach to response recording. Fixed Responding Fixed responding is an invalidating test-taking approach characterized by a sys- tematic response pattern that is not based on an accurate reading and com- prehension of test items. In contrast to random responding, the test taker provides the same non-content-based responses (e.g., True) to various items 336 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S without considering their content. If the test taker provides both True and False responses indiscriminately, then they are engaging in random responding. In fixed responding, the indiscriminate responses are stereotypic, either True or False, or in the case of a Likert scale response format, the test taker marks items indiscriminately at the same level without considering content. Like nonresponding and random responding, fixed responding is a matter of degree rather than a dichotomous all-or-none phenomenon. Unlike nonre- sponding and random responding, fixed responding has received a great deal of attention in the assessment literature. Jackson and Messick (1962) sparked this discussion when they proposed that much (if not all) of the variance in MMPI scale scores was attributable to two response styles, termed acquiescence and social desirability. Acquiescence was defined as a tendency to respond True to MMPI items without consideration of their content (i.e., fixed True respond- ing). These authors factor analyzed MMPI scale scores in a broad range of samples and recurrently found that two factors accounted for much of the vari- ance in scores. They attributed variance on these factors to two response styles, acquiescence and social desirability, and cautioned that MMPI scale scores appeared primarily to reflect individual differences on these nonsubstantive dimensions. Furthermore, they suggested that MMPI scales were particularly vulnerable to the effects of acquiescence and its counterpart, counteracquies- cence (defined as a tendency to respond False to MMPI items without consid- eration of their content), because the scoring keys were unbalanced; that is, for some MMPI scales, many, if not most, of the items were keyed True, whereas on other scales, most of the items were keyed False. Contrasting with Jackson and Messick’s findings, Block (1965) demonstrated in an extensive and sophis- ticated series of analyses that the two primary MMPI factors reflected substan- tive personality dimensions rather than stylistic response tendencies. Although fixed responding does not pose as broad a threat to protocol valid- ity as Jackson and Messick argued, in the infrequent cases when a test taker uses this response style excessively, the resulting scale scores will be invalid and un- interpretable. Contrary to Jackson and Messick’s assertion, constructing scales with balanced keys or Likert scale response formats does not make self-report measures less susceptible to this threat to protocol validity. An indiscriminate set of True responses is invalid regardless of whether the scoring key is balanced, and Likert scales provide even more opportunity for stereotypic responses. Content-Based Invalid Responding Content-based invalid responding occurs when the test taker skews their re- sponses to items and creates a misleading impression as a result. This test- taking approach falls broadly into two classes discussed under various labels in the literature. The first class has been termed overreporting, feigning, fak-  INTERPRETING THE MMPI-3 VALIDITY SCALES | 337 ing bad, negative response bias, and malingering; the second has been termed underreporting, faking good, positive response bias, denial, defensiveness, and positive malingering. Because both types of content-based invalid responding can be generated intentionally and unintentionally, the more neutral descriptive terms overreporting and underreporting are preferred. Overreporting Overreporting occurs when a test taker reports problems they do not actually have or exaggerates the significance of difficulties they do have. In a hypo- thetical situation, if a completely objective assessment of the individual’s func- tioning were available, the overreporter’s subjective self-report would indicate greater dysfunction than the objective assessment. Intentional overreporting occurs when the individual knowingly slants their self-report to appear dysfunctional. Such a test taker may be motivated by some external gain and thus fit the DSM-5 definition of malingering (American Psy- chiatric Association [APA], 2013). However, intentional overreporting is not synonymous with malingering because, for example, in the absence of an exter- nal incentive, it may correspond to the DSM-5 definition of factitious disorder, or merely uncooperativeness. Moreover, intentional overreporting is not in itself an indication that psychopathology is absent. An individual with genuine psy- chological difficulties may amplify their extent or significance or may fabricate others but may nonetheless be experiencing significant dysfunction. In such a scenario it is important for the interpreter to be aware of the likelihood that the test taker’s substantive scale scores reflect a level of dysfunction greater than they are experiencing. Unintentional overreporting occurs when a test taker is unaware that they are describing themself in an unrealistically negative manner. It is the test taker’s self-concept rather than the self-report that is skewed. Individuals who engage in this test-taking approach mistakenly believe that their responses are accurate when in fact they are overreporting. Individuals with somatic symp- tom disorders, for example, report significant somatic problems that cannot be explained by objective medical findings (Lamberty, 2008). They believe their symptoms to be the result of some heretofore undiagnosed condition. Test tak- ers who tend to catastrophize and see things as worse than they actually are may also unintentionally overreport in response to self-report measures. Underreporting Underreporting occurs when a test taker describes themself as having less seri- ous or a smaller number of difficulties (or both) than they have. Referring back to the hypothetical objective benchmark just mentioned, an underreporting 338 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S test taker would paint a picture of better functioning than would be indicated by an objective assessment. Here, too, a distinction may be drawn between in- tentional and unintentional underreporting. In intentional underreporting, the individual knowingly denies or minimizes the extent of the psychological difficulties or negative characteristics they know they have. As a result, the test scores underestimate their level of dysfunction. Differentiating denial from minimization is important but complex. In the for- mer, an individual blatantly denies problems that they know exist; in the latter, the individual acknowledges some difficulties or negative characteristics but minimizes their extent or impact. Unintentional underreporting occurs when the individual unknowingly denies or minimizes difficulties or negative charac- teristics. Here, too, objective and subjective indicators of psychological func- tioning would be at odds; however, in unintentional underreporting, this dis- crepancy results from the individual’s distorted self-concept rather than from an intentional effort to produce misleading test results. Assessing Protocol Validity Threats With the MMPI-3 and Consideration of Confounds Table 8.1 lists the possible threats to protocol validity assessed by the MMPI-3 Validity Scales as well as other potential influences that can confound Valid- ity Scale interpretation if not considered. The “Non-Content-Based Threats” listed in the first column of Table 8.1 reflect the approach used to detect incon- sistent responding with the MMPI-3. As described in chapter 3, the Combined Response Inconsistency (CRIN), Variable Response Inconsistency (VRIN), and True Response Inconsistency (TRIN) scales assess for inconsistent responding by considering a test taker’s response to item pairs. The item pairs (called com- posites) scored on VRIN are keyed such that a True response to one and a False response to the other, or vice versa, adds a point to the scale’s raw score. The scale label reflects that these are variable (one True and the other False) inconsistent responses. The composites scored on TRIN are keyed such that a combination of inconsistent True responses or a combination of inconsistent False responses contributes to the scale’s raw score. The inconsistent True re- sponses and inconsistent False responses are counted separately, with the latter then subtracted from the former followed by the addition of a constant to avoid negative raw score values. The resulting raw score indicates whether a test taker has engaged in a predominant pattern of fixed True responding or fixed False responding (reflected by a letter T or F following the TRIN T score). CRIN is scored by adding the count of variable inconsistent responses, the count of fixed True responses, and the count of fixed False responses to provide an over- all indication of the level of combined inconsistent responding in a protocol.  INTERPRETING THE MMPI-3 VALIDITY SCALES | 339 TABLE 8.1. MMPI-3 Validity Scales: Threats to Protocol Validity and Confounds CNS CRIN VRIN TRIN F Fp Fs FBS RBS L K Threats NON-CONTENT-BASED Nonresponding X ¬ ¬ ¬ ¬ ¬ ¬ ¬ ¬ ¬ ¬ Random responding X X + + + + + + + Variable inconsistent X + + + + + + + responding Fixed True responding X + + + + + ¬ ¬ Fixed False responding X + + + + + + + CONTENT-BASED Overreporting X X X X X Underreporting X X Extratest Confounds Reading/comprehension + + + + problems Psychopathology/ + + + + + psychological distress Medical conditions + + Traditional upbringing + Good adjustment + Note. x = Scale designed to assess this threat; + = Confound artifactually increases score; − = Confound artifactually lowers score; CNS = Cannot Say; CRIN = Combined Response Inconsistency; VRIN = Variable Response Inconsistency; TRIN = True Response Incon- sistency; F = Infrequent Responses; Fp = Infrequent Psychopathology Responses; Fs = Infrequent Somatic Responses; FBS = Symptom Validity Scale; RBS = Response Bias Scale; L = Uncommon Virtues; K = Adjustment Validity Shaded area identifies confounds that can invalidate scores on the corresponding validity scales. An X in the grid of Table 8.1 identifies the primary MMPI-3 scale(s) for as- sessing each threat. Interpretation of scores on the validity indicators marked with an X can be confounded by invalid responding other than the type assessed by a scale as well as by the extratest confounds listed in Table 8.1. A plus sign (+) in the grid indicates a confound that can artifactually increase scores on a valid- ity indicator; a minus sign (−) indicates a confound that can artifactually lower scores. For example, nonresponding can confound interpretation of scores on all the MMPI-3 validity indicators (except CNS) by lowering the resulting score (as indicated by the minus sign [−] in the Nonresponding row under each of these scales). Confounds listed in the shaded cells can reach levels that invalidate scores on a validity indicator, rendering them uninterpretable. For example, under CRIN, nonresponding is identified as a confound that can artifactually lower scores on this scale. In an extreme example, if a test taker does not provide 340 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S scorable responses to any of the CRIN items, the resulting raw score of zero is uninterpretable insofar as assessment of random responding is concerned. Although extratest confounds do not render validity scale scores uninterpre- table, they must be considered carefully to avoid misinterpreting their effects as indications of intentional invalid responding. INTERPRETIVE GUIDELINES Tables 8.2 through 8.12 provide interpretive guidelines for the MMPI-3 Validity Scales. For each Validity Scale, these tables list the protocol validity concerns, possible reasons why a test taker may score in a designated range, and the inter- pretive implications of scores within those ranges. The lists of possible reasons for a score and interpretive implications include guidance for identifying the protocol validity threats and possible confounds just discussed. For example, the tables describing scales that assess for content-based invalid responding in- dicate the need to rule out inconsistent responding before reaching inferences about overreporting or underreporting. Three subsets of the MMPI-3 Valid- ity Scales address the threats to protocol validity—content nonresponsiveness, overreporting, and underreporting. Content Nonresponsiveness Scales Cannot Say (CNS) The Cannot Say (CNS) score is a count of unscorable responses to the 335 MMPI-3 items and is a raw score. The most common type of unscorable re- sponse is no response; however, if the test taker marked both True and False for an item, that response would also be unscorable. As discussed earlier, unscorable responses artifactually lower scores on the MMPI-3 because the scoring method used (counting the number of items on a scale answered in the keyed direction) treats these responses as though the test taker answered in the nonkeyed direction. Scores on scales with at least 90% of the items answered are not likely to be af- fected by unscorable responses to a degree that would compromise interpretability (Dragon et al., 2012). However, the absence of elevation (including a low score) on a given scale becomes increasingly uninterpretable as the proportion of unscorable item responses on that scale goes beyond 10%. Elevated scores on a scale are still interpretable as such when the test taker has responded to less than 90% of the items, but the scores may underestimate the problems assessed by that scale. Table 8.2 provides interpretive recommendations for the CNS score. A score of 15 or greater raises concerns about the possibility of compromised validity owing to unscorable responses. Which scales are affected can be deter-  INTERPRETING THE MMPI-3 VALIDITY SCALES | 341 TABLE 8.2. Cannot Say (CNS) Score Interpretation Raw Protocol validity Possible reasons score concerns for score Interpretive implications ≥ 15 Scores on some Reading or language Examine the content of unscorable items to detect scales may be limitations possible themes. The impact is scale dependent. For invalid. Severe psychopathology scales on which less than 90% of the items were Obsessiveness scorable, the absence of elevation is uninterpretable. Lack of insight Elevated scores on such scales may underestimate the Lack of cooperation significance or severity of associated problems. 1–14 Scores on some of Selective Examine the content of unscorable items to detect the shorter scales nonresponsiveness possible themes. The impact is scale dependent. For may be invalid. scales on which less than 90% of the items were scorable, the absence of elevation is uninterpretable. Elevated scores on such scales may underestimate the significance or severity of associated problems. mined only by calculating the percentage of items answered on each scale—a process greatly facilitated by computerized scoring (see chapter 7). As indi- cated in Table 8.2, an excessive number of unscorable responses may reflect lack of cooperation by the test taker but may also occur if the test taker lacks adequate reading or language comprehension skills, is seriously disturbed, is overly obsessive, or lacks the necessary insight or self-awareness to respond to some test items. The interpreter will need to consider which of these factors may be involved based on extratest information (e.g., background, interview, results of other testing). Combined Response Inconsistency (CRIN) The Combined Response Inconsistency (CRIN) scale is a global measure of response inconsistency. CRIN scores indicate the overall level of inconsistent responding in an MMPI-3 protocol, combining information about both vari- able and fixed inconsistent responding. It includes the 86 item pairs of both the Variable Response Inconsistency (VRIN) and True Response Inconsistency (TRIN) scales (described in the next sections). The CRIN scale can be par- ticularly helpful in identifying test takers who intermittently engage in invalid quasi-random and fixed responding and/or alternating patterns of fixed True and fixed False responding; such test takers may not have elevated scores on VRIN or TRIN. Table 8.3 provides interpretive recommendations for CRIN. A score of 80T or higher indicates that the protocol is invalid because of ex- cessive response inconsistency. This score does not necessarily mean that the test taker was intentionally uncooperative. Other possible reasons for eleva- tion need to be considered based on extratest information. However, this score does indicate that scores on the remaining content-based Validity Scales (the 342 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S TABLE 8.3. CRIN (Combined Response Inconsistency) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 80 This protocol is invalid Reading or language limitations The protocol is because of excessive re- Cognitive impairment uninterpretable. sponse inconsistency. Errors in recording responses Intentional random responding An uncooperative test-taking approach 70–79 There is some evidence of Reading or language limitations Scores on the content-based response inconsistency. Cognitive impairment invalid responding indicators Errors in recording responses and the Substantive Scales Carelessness should be interpreted with An uncooperative test-taking approach some caution. 39–69 None The test taker was able to comprehend No concerns and respond relevantly to the test items. 30–38 There is evidence of remark- The test taker was deliberate in their No concerns ably consistent responding. approach to the assessment. overreporting and underreporting indicators) and the Substantive Scales can- not be interpreted. Variable Response Inconsistency (VRIN) The Variable Response Inconsistency (VRIN) score is based on the test taker’s responses to 53 item pairs selected so that members of each pair are similar in content. The raw VRIN score equals the number of pairs answered inconsis- tently (i.e., True-False or False-True). The T score derived from the raw score is used to identify protocols marked by excessive variable inconsistent responding. Table 8.4 provides interpretive recommendations for VRIN. A score of 80T or higher indicates that the protocol is invalid because of excessive variable response inconsistency. As noted, this score does not necessarily mean that the test taker was intentionally uncooperative. Other possible reasons for elevation need to be considered based on extratest information. However, this score does indicate that scores on the remaining content-based Validity Scales (the overreporting and un- derreporting indicators) and on the Substantive Scales cannot be interpreted. True Response Inconsistency (TRIN) The True Response Inconsistency (TRIN) score is based on the test taker’s responses to 33 item pairs. The members of each pair are quasi reversals in content and keyed so that the raw TRIN score equals the number of pairs (in- consistently) answered True-True minus the number of pairs (inconsistently) answered False-False. Thus, high raw TRIN scores indicate fixed (semantically inconsistent, indiscriminate) True responding, whereas low scores indicate  INTERPRETING THE MMPI-3 VALIDITY SCALES | 343 TABLE 8.4. VRIN (Variable Response Inconsistency) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 80 The protocol is invalid Reading or language limitations The protocol is because of excessive variable Cognitive impairment uninterpretable. response inconsistency. Errors in recording responses Intentional random responding An uncooperative test-taking approach 70–79 There is some evidence Reading or language limitations Scores on the content-based of variable response Cognitive impairment invalid responding indicators inconsistency. Errors in recording responses and the Substantive Scales Carelessness should be interpreted with some caution. 39–69 There is evidence of consis- The test taker was able to comprehend No concerns tent responding. and respond relevantly to the test items. TABLE 8.5. TRIN (True Response Inconsistency) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 80T The protocol is invalid because An uncooperative test-taking approach The protocol is of excessive fixed, content- Difficulties with double negatives uninterpretable. inconsistent True responding. 70T–79T There is some evidence of An uncooperative test-taking approach Scores on the content-based fixed, content-inconsistent Difficulties with double negatives invalid responding indicators True responding. and the Substantive Scales should be interpreted with some caution. 50–69 There is no evidence of No concerns fixed, content-inconsistent responding. 70F–79F There is some evidence of An uncooperative test-taking approach Scores on the content-based fixed, content-inconsistent Difficulties with double negatives invalid responding indicators False responding. and the Substantive Scales should be interpreted with some caution. ≥ 80F The protocol is invalid An uncooperative test-taking approach The protocol is because of excessive fixed, Difficulties with double negatives uninterpretable. content-inconsistent False responding. fixed False responding. The TRIN T scores were derived by first transform- ing the raw scores into linear T scores and then reflecting all T-score values below 50 (those deviating from the mean in the counteracquiescent direction). For example, an initial T score of 80, indicative of acquiescence, is left un- changed, but a T score of 20, indicating an equally large deviation in the coun- teracquiescent direction, is reflected and consequently also becomes 80. To dis- tinguish acquiescent from counteracquiescent scores, the former are displayed with the letter T (e.g., 80T) and the latter with the letter F (80F). Table 8.5 344 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S provides interpretive recommendations for TRIN. A T score of 80 or higher on TRIN (in either direction, True or False) indicates that the protocol is invalid because of excessive inconsistent fixed responding and that scores on the re- maining content-based Validity Scales (the overreporting and underreporting indicators) and on the Substantive Scales cannot be interpreted. Overreporting Scales Overreporting is defined as occurring when the test taker’s self-presentation portrays a degree of dysfunction that is “noncredible” (i.e., more extreme than would be indicated by a hypothetical objective assessment of the individual). The general term overreporting is preferred over expressions that imply inten- tionality, which typically require consideration of extratest data. “Faking bad” is an example of such an expression. A diagnosis of malingering is another example because it requires evidence of an incentive to exaggerate or fabricate symptoms. In other words, neither faking bad nor malingering can be inferred from the test data alone; however, elevations on the MMPI-3 overreporting scales raise and support the possibility that the test taker is faking bad or malingering. Particularly in regard to intentionality, it is necessary to consider the pos- sible impact of various psychological disorders characterized by unintentional overreporting of symptoms (e.g., somatic symptom disorders) or mispercep- tion of reality (e.g., thought disorders). Moreover, even when there is extratest evidence of intentionality, the current diagnostic system calls for a differential diagnosis of malingering, requiring an external motive (e.g., avoidance of legal responsibility), rather than a factitious disorder that would be inferred to have an internal motive (i.e., a psychogenic need to assume the “sick role”). Scores on the MMPI-3 Validity Scales do not provide specific indications of intentional- ity or motivation when there is evidence of overreporting. However, test results indicative of overreporting (or lack thereof) can be used to support (or refute) inferences about feigning or malingering, just as scores on the Substantive Scales can suggest diagnostic possibilities to be considered and evaluated with the aid of extratest data (e.g., historical records, other test results, behavioral observa- tions, interviews). In addition, as outlined next, indications of overreporting on the MMPI-3 Validity Scales have implications for the interpretability of scores on the Substantive Scales. As reflected in Table 8.1 and discussed earlier, scores on the MMPI-3 over- reporting indicators are artifactually elevated by random and fixed responding. Therefore, as indicated in the following interpretive guidelines, overreporting can be inferred only if excessive inconsistent responding has been ruled out. As defined here, overreporting is the inaccurate reporting of dysfunction. Broadly  INTERPRETING THE MMPI-3 VALIDITY SCALES | 345 speaking, test takers may overreport psychopathology symptoms (e.g., depression, psychosis), cognitive symptoms (e.g., attention difficulties, memory impairment), or somatic symptoms (e.g., pain, poor health). The MMPI-3 Validity Scales are differentially associated with these various areas of noncredible responding. These differences are highlighted in the interpretive recommendations that follow. As discussed earlier, it is important to note that a finding of overreporting or malingering is not evidence that a test taker is free of dysfunction. Psychopa- thology and overreporting are not mutually exclusive. Individuals with genuine disorders may overreport their symptoms or fabricate others for a variety of reasons. Therefore, positive findings on the MMPI-3 overreporting indicators do not, in themselves, rule out the possibility that the test taker is psychologi- cally disordered. Infrequent Responses (F) The Infrequent Responses (F) scale consists of 35 items rarely answered in the keyed direction by members of the MMPI-3 normative sample. Elevated scores on F are associated with overreporting a broad range of psychological, cogni- tive, and somatic symptoms. However, as indicated in Table 8.1, elevated scores up to a certain level can also be generated by individuals experiencing genuine difficulties manifesting in psychopathology or psychological distress. There- fore, extratest considerations (e.g., whether the individual has a documented history of significant dysfunction) are critical to proper interpretation of scores on this scale. Table 8.6 provides interpretive recommendations for F. As outlined, scores in the 75T–99T range raise concerns about possible overreporting of psycho- logical dysfunction that need to be considered in the context of a history or current extratest findings of dysfunction. Significant psychopathology and/or pronounced emotional distress can also result in deviant scores on this scale. As the score on F rises, evidence of a greater degree of genuine dysfunction is needed to rule out overreporting. A T score of 100 or greater on F is uncom- mon even in individuals with genuine, severe dysfunction and indicates that the protocol is invalid and uninterpretable. Infrequent Psychopathology Responses (Fp) The Infrequent Psychopathology Responses (Fp) scale consists of 21 items rarely answered in the keyed direction by individuals with genuine, severe psy- chopathology. As a result, in contrast with F, scores on Fp are less likely to be confounded with severe disorder or distress. Therefore, this scale is particularly helpful in assessing overreporting in settings and with populations characterized 346 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S TABLE 8.6. F (Infrequent Responses) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 100 The protocol is invalid. Inconsistent responding Inconsistent responding should be Overreporting is indicated Overreporting considered by examining the CRIN, by an excessive number of VRIN, and TRIN scores. If it is ruled infrequent responses. out, note that this level of infrequent responding is uncommon even in indi- viduals with genuine, severe psycho- logical difficulties who report credible symptoms. Scores on the Substantive Scales should not be interpreted. 90–99 The protocol may be invalid. Inconsistent responding Inconsistent responding should be Overreporting of psychologi- Severe psychopathology considered by examining the CRIN, cal dysfunction is indicated Severe emotional distress VRIN, and TRIN scores. If it is ruled by a considerably larger than Overreporting out, note that this level of infrequent average number of infrequent responding may occur in individuals responses. with genuine, severe psychological difficulties who report credible symptoms. However, for individuals with no history or current corroborat- ing evidence of dysfunction, it most likely indicates overreporting. 80–89 Possible overreporting of Inconsistent responding Inconsistent responding should be psychological dysfunction is Significant psychopathology considered by examining the CRIN, indicated by a much larger Significant emotional distress VRIN, and TRIN scores. If it is ruled than average number of Overreporting out, note that this level of infrequent infrequent responses. responding may occur in individuals with genuine, substantial psychologi- cal difficulties who report credible symptoms. However, for individuals with no history or current corroborat- ing evidence of dysfunction, it very likely indicates overreporting. 75–79 Possible overreporting of Inconsistent responding Inconsistent responding should be psychological dysfunction Significant psychopathology considered by examining the CRIN, is indicated by a larger than Significant emotional distress VRIN, and TRIN scores. If it is ruled average number of infrequent Overreporting out, note that this level of infrequent responses. responding may occur in individuals with genuine psychological difficulties who report credible symptoms. However, for individuals with no history or current corroborat- ing evidence of dysfunction, it likely indicates overreporting. < 75 There is no evidence of No concerns overreporting. by high base rates of significant psychological disorders, most notably those marked by psychotic symptoms or severe emotional dysfunction. Table 8.7 provides interpretive recommendations for Fp. As indicated, scores in the 80T–99T range raise substantial concerns about the possibility of symp- tom exaggeration, even for test takers with a significant history of mental health problems. Scores that reach or exceed 100T indicate that the protocol is invalid owing to the strong likelihood of substantial overreporting.  INTERPRETING THE MMPI-3 VALIDITY SCALES | 347 TABLE 8.7. Fp (Infrequent Psychopathology Responses) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 100 The protocol is invalid. Inconsistent responding Inconsistent responding should be Overreporting is indicated by Overreporting considered by examining the CRIN, assertion of a considerably VRIN, and TRIN scores. If it is ruled larger than average number out, note that this level of infrequent of symptoms rarely described responding is very uncommon even by individuals with genuine, in individuals with genuine, severe severe psychopathology. psychopathology who report credible symptoms. Scores on the Substantive Scales should not be interpreted. 80–99 Possible overreporting is indi- Inconsistent responding Inconsistent responding should be cated by assertion of a much Severe psychopathology considered by examining the CRIN, larger than average number Overreporting VRIN, and TRIN scores. If it is ruled of symptoms rarely described out, note that this level of infrequent by individuals with genuine, responding may occur in individuals severe psychopathology. with genuine, severe psychopathology who report credible symptoms, but it could also reflect exaggeration. For individuals with no history or current corroborating evidence of psychopa- thology, scores in this range very likely indicate overreporting. 70–79 Possible overreporting is Inconsistent responding Inconsistent responding should be indicated by assertion of a Severe psychopathology considered by examining the CRIN, larger than average number Overreporting VRIN, and TRIN scores. If it is ruled of symptoms rarely described out, note that this level of infrequent by individuals with genuine, responding may occur in individuals severe psychopathology. with genuine, severe psychopathology who report credible symptoms. How- ever, for individuals with no history or current corroborating evidence of psychopathology, scores in this range likely indicate overreporting. Scores on Fp can be particularly helpful when a test taker produces a mark- edly elevated score on F and the interpreter needs to determine whether this reflects overreporting or genuine dysfunction. The lower the score on Fp, the less likely it is that an elevation on F reflects overreporting and the more likely it is that an F elevation reflects accurate reporting of experiences that are un- common in the general population but not in individuals with significant psy- chological difficulties. This is particularly true if the individual presents with symptoms of thought dysfunction. However, test takers who overreport prob- lems associated with emotional (e.g., depression, anxiety) rather than thought dysfunction may also produce substantially higher scores on F than on Fp. F scores may also be higher than Fp scores when test takers report noncredible somatic or cognitive symptoms. These individuals would also be expected to generate elevated scores on Fs, FBS, and/or RBS as described next. 348 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S TABLE 8.8. Fs (Infrequent Somatic Responses) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 100 Scores on the somatic scales Inconsistent responding Inconsistent responding should be may be invalid. Overreport- Overreporting of somatic considered by examining the CRIN, ing of somatic symptoms is complaints VRIN, and TRIN scores. If it is ruled reflected in the assertion of out, note that this level of infrequent a considerably larger than responding is very uncommon even in average number of somatic individuals with substantial medical symptoms rarely described problems who report credible symp- by individuals with genuine toms. Scores on the somatic scales medical problems. should be interpreted cautiously. 80–99 Possible overreporting of so- Inconsistent responding Inconsistent responding should be matic symptoms is reflected Significant and/or multiple considered by examining the CRIN, in the assertion of a much medical conditions VRIN, and TRIN scores. If it is ruled larger than average number Overreporting of somatic out, note that this level and type of of somatic symptoms rarely complaints infrequent responding may occur in described by individuals with individuals with substantial medi- genuine medical problems. cal conditions who report credible symptoms, but it could also reflect exaggeration. In individuals with no history or corroborating evidence of physical health problems, this prob- ably indicates noncredible reporting of somatic symptoms. Scores on the somatic scales should be interpreted cautiously. < 80 There is no evidence of No concerns overreporting. Infrequent Somatic Responses (Fs) The Infrequent Somatic Responses (Fs) scale consists of 16 items with somatic content uncommonly endorsed by medical patients receiving treatment for various physical diseases. Based on the same rare-symptoms rationale as the other two MMPI-3 infrequent response indicators (F and Fp), Fs is designed to identify test takers who overreport somatic symptoms by endorsing many somatic complaints rarely reported by medical patients. Table 8.8 provides in- terpretive recommendations for scores on Fs. As shown, scores in the 80T–99T range raise concerns about possible overreporting of somatic symptoms. Scores of 100T or higher indicate that overreporting of somatic complaints has likely occurred, limiting the interpretability of scores on Somatic Complaints (RC1), Malaise (MLS), and Neurological Complaints (NUC) scales. To reiterate an important point made earlier, overreporting is not synony- mous with faking or malingering. Noncredible somatic complaints may stem from a variety of sources, including an external incentive but also the internal psychological factors underlying a somatic symptom disorder or somatic delu- sions. Extratest data (in this case, a detailed medical and psychological history) are needed to make these distinctions.  INTERPRETING THE MMPI-3 VALIDITY SCALES | 349 TABLE 8.9. FBS (Symptom Validity Scale) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 90 Scores on the Somatic/ Inconsistent responding Inconsistent responding should be Cognitive Scales may be Overreporting of somatic considered by examining the CRIN, invalid. Overreporting is indi- and/or cognitive symptoms VRIN, and TRIN scores. If it is ruled cated by a very unusual com- out, note that this combination bination of responses that is of responses is very uncommon associated with noncredible even in individuals with substantial reporting of somatic and/or medical problems who report credible cognitive symptoms. symptoms. Scores on the Somatic/ Cognitive Scales should be interpreted cautiously. 78–89 Possible overreporting is Inconsistent responding Inconsistent responding should be indicated by an unusual com- Significant and/or multiple considered by examining the CRIN, bination of responses that is medical conditions VRIN, and TRIN scores. If it is ruled associated with noncredible Overreporting of somatic out, note that this combination of reporting of somatic and/or and/or cognitive complaints responses may occur in individuals cognitive symptoms. with substantial medical problems who report credible symptoms, but it could also reflect exaggeration. Scores on the Somatic/Cognitive Scales should be interpreted cautiously. < 78 There is no evidence of No concerns overreporting. Symptom Validity Scale (FBS) The Symptom Validity Scale (FBS) consists of 30 items. A longer version of the scale was originally developed to complement the MMPI-2’s F scale by identifying individuals presenting with noncredible symptoms in the context of civil litigation. Table 8.9 provides interpretive recommendations for FBS. As indicated, scores in the 78T–89T range identify possible overreporting as reflected in an unusual combination of responses that is associated with noncredible presentation of somatic and/or cognitive symptoms. Scores of 90T or higher indicate likely overreporting of such symptoms, limiting the interpretability of the scales RC1, MLS, NUC, and COG. The extratest data needed to make inferences about possible motives for noncredible symptom reporting for Fs should be considered in the interpretation of scores on FBS as well. As reflected in the respective interpretive recommendations, both Fs and FBS provide information about possible noncredible somatic symptom reporting. Scores on the two scales are only moderately correlated, indicating relatedness but by no means redundancy. Therefore, scores on both scales need to be con- sidered with appropriate attention to the setting. 350 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S TABLE 8.10. RBS (Response Bias Scale) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 90 Scores on the Cognitive Com- Inconsistent responding Inconsistent responding should be plaints scale may be invalid. Overreporting of memory considered by examining the CRIN, Overreporting is indicated by complaints VRIN, and TRIN scores. If it is ruled a very unusual combination out, note that this combination of of responses that is strongly responses is very uncommon even in associated with noncredible individuals with substantial emotional memory complaints. dysfunction who report credible symptoms. Scores on the Cognitive Complaints scale should be inter- preted cautiously. 75–89 Possible overreporting is Inconsistent responding Inconsistent responding should be indicated by an unusual com- Significant emotional considered by examining the CRIN, bination of responses that is dysfunction VRIN, and TRIN scores. If it is ruled associated with noncredible Overreporting of memory out, note that this combination of memory complaints. complaints responses may occur in individuals with substantial emotional dysfunc- tion who report credible symptoms, but it could also reflect exaggeration. Scores on the Cognitive Com- plaints scale should be interpreted cautiously. < 75 There is no evidence of No concerns overreporting. Response Bias Scale (RBS) The Response Bias Scale (RBS) is composed of 28 items associated empiri- cally with scoring below established cutoffs on psychological tests designed to detect inadequate effort on cognitive testing, known as performance validity tests (PVTs). Table 8.10 provides interpretive recommendations for RBS. As in- dicated, scores in the 75T–89T range indicate possible overreporting reflected as an unusual combination of responses associated with noncredible memory complaints. However, scores in this range may also indicate that individuals are experiencing substantial emotional dysfunction. Scores of 90T or higher indicate likely overreporting of memory problems, limiting the interpretabil- ity of scores on the COG scale. Although PVTs were used in the development of RBS, the scale is not intended or recommended for use as a measure of ef- fort on cognitive tests. Underreporting Scales Underreporting is defined as occurring when the test taker’s self-presentation suggests a level of functioning that is better than would be indicated by a hy- pothetical objective assessment of the individual. As with the term overreport- ing, the term underreporting is preferred over terms such as “faking good” or  INTERPRETING THE MMPI-3 VALIDITY SCALES | 351 “positive malingering,” which connote an intentionality that cannot be inferred from test data alone. Self-report measures of personality and psychopathology are inherently susceptible to intentional underreporting, which is most likely to occur when, given the assessment context, good adjustment is a highly de- sirable quality and the individual has a great deal at stake (e.g., child custody evaluations, preemployment assessments, or release from involuntary commit- ment). However, underreporting can also result from lack of awareness of or insight into psychological dysfunction. As with overreporting, differentiating intentional from unintentional under- reporting requires consideration of extratest data. For example, if a test taker presents as extraordinarily well adjusted but is experiencing considerable psy- chosocial difficulties, this increases the likelihood that the individual is know- ingly underreporting. At the same time, as will be explained shortly and is reflected in Table 8.1, elevated scores on the two MMPI-3 underreporting indi- cators can to some extent reflect factors (confounds) other than underreport- ing (e.g., being raised in a very traditional environment or being considerably better adjusted than average). Regardless of whether the underreporting is intentional, unintentional, or can be explained by other factors, elevated scores on the MMPI-3 underreport- ing scales indicate a need for caution in the interpretation of scores on the Sub- stantive Scales. Specifically, nonelevated and, in particular, low scores (i.e., 38T or lower) on the Substantive Scales are uninterpretable and cannot be relied upon to rule out problems when there are indications of possible underreport- ing. Elevated scores on one or more of these scales are interpretable but may underestimate the problems associated with the elevation(s). As is the case with the overreporting indicators, scores on the underreport- ing scales can be significantly distorted if a protocol is marked by considerable inconsistent responding. This is particularly true of fixed False responding, which artifactually elevates scores on both underreporting indicators because most of the items on these measures are keyed in that direction. Uncommon Virtues (L) The Uncommon Virtues (L) scale consists of 14 items. Elevated L scores indi- cate that the test taker presented themself in a favorable light by denying minor faults and shortcomings that most individuals acknowledge. An important consideration in understanding the significance of an elevated L score in each case is whether the test taker was raised in an environment stressing traditional values, an important cultural factor discussed in chapter 4. If so, inferences about underreporting need to be tempered, particularly for moderate levels of elevation on this scale. 352 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S TABLE 8.11. L (Uncommon Virtues) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 80 The protocol may be invalid. Inconsistent responding Inconsistent responding should be Underreporting is indicated Underreporting considered by examining the CRIN, by the test taker presenting VRIN, and TRIN scores. If it is ruled themself in an extremely out, note that this level of virtuous positive light by denying self-presentation is very uncommon many minor faults and short- even in individuals with a background comings that most people stressing traditional values. Any acknowledge. absence of elevation on the Substan- tive Scales is uninterpretable. Elevated scores on the Substantive Scales may underestimate the problems assessed by those scales. 70–79 Possible underreporting is Inconsistent responding Inconsistent responding should be indicated by the test taker Traditional upbringing considered by examining the CRIN, presenting themself in a very Underreporting VRIN, and TRIN scores. If it is ruled positive light by denying out, note that this level of virtu- several minor faults and ous self-presentation is uncommon shortcomings that most but may, to some extent, reflect a people acknowledge. background stressing traditional values. Any absence of elevation on the Substantive Scales should be interpreted with caution. Elevated scores on the Substantive Scales may underestimate the problems assessed by those scales. 65–69 Possible underreporting is Inconsistent responding Inconsistent responding should be indicated by the test taker Traditional upbringing considered by examining the CRIN, presenting themself in a very Underreporting VRIN, and TRIN scores. If it is ruled positive light by denying out, note that this level of virtu- some minor faults and short- ous self-presentation may reflect comings that most people a background stressing traditional acknowledge. values. Any absence of elevation on the Substantive Scales should be interpreted with caution. Elevated scores on the Substantive Scales may underestimate the problems assessed by those scales. Table 8.11 provides interpretive recommendations for L. Scores in the 65T–69T and 70T–79T ranges reflect possible underreporting, with higher scores indicating an increased likelihood of this being the case (and a reduced possibility that a traditional upbringing can account fully for the elevation). Scores of 80T or higher indicate that substantial underreporting very likely oc- curred and raise the possibility that the protocol is consequently of limited util- ity or invalid.1 When L scores reach this level, elevated scores on the Substantive Scales are interpretable but may underestimate the problems associated with these elevations. Absence of elevation on the Substantive Scales is uninterpre- table when L scores are this high.  INTERPRETING THE MMPI-3 VALIDITY SCALES | 353 TABLE 8.12. K (Adjustment Validity) Interpretation T score Protocol validity concerns Possible reasons for score Interpretive implications ≥ 70 Underreporting is indicated Inconsistent responding Inconsistent responding should be by the test taker presenting Underreporting considered by examining the CRIN, themself as remarkably well VRIN, and TRIN scores. If it is ruled out, adjusted. note that this level of psychological adjustment is rare in the general popula- tion. Any absence of elevation on the Substantive Scales should be interpreted with caution. Elevated scores on the Sub- stantive Scales may underestimate the problems assessed by those scales. 66–69 Possible underreporting is Inconsistent responding Inconsistent responding should be indicated by the test taker Very good psychological considered by examining the CRIN, VRIN, presenting themself as very adjustment and TRIN scores. If it is ruled out, note well adjusted. Underreporting that this level of psychological adjust- ment is rare in the general population. For individuals who are not especially well adjusted, any absence of elevation on the Substantive Scales should be interpreted with caution. Elevated scores on the Sub- stantive Scales may underestimate the problems assessed by those scales. 60–65 Possible underreporting is Inconsistent responding Inconsistent responding should be con- indicated by the test taker Good psychological sidered by examining the CRIN, VRIN, presenting themself as very adjustment and TRIN scores. If it is ruled out, for in- well adjusted. Underreporting dividuals who are not well adjusted, any absence of elevation on the Substantive Scales should be interpreted with cau- tion. Elevated scores on the Substantive Scales may underestimate the problems assessed by those scales. < 60 There is no evidence of No concerns underreporting. Adjustment Validity (K) The Adjustment Validity (K) scale consists of 14 of items. Elevated K scores indicate that the test taker presented themself as well adjusted, with higher scores representing a higher level of adjustment. This type of self-presentation is associated with underreporting. However, the possibility that the test taker is in fact better adjusted than average also needs to be considered in the inter- pretation of an elevated K score. Extratest indications that the individual is not well adjusted would support a conclusion that an elevated K score indicates underreporting, whereas evidence that they are well adjusted would temper this interpretation. Table 8.12 provides interpretive recommendations for K. Scores in the 60T–65T and 66T–69T ranges indicate possible underreporting, with higher scores suggesting a greater likelihood of underreporting and requiring evidence 354 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S of better adjustment to rule out this interpretation. T scores above 70 indicate that the test taker presented themself as remarkably well adjusted. Without ex- tratest indications to the contrary, this level of elevation on K indicates in most settings that nonelevated scores on the Substantive Scales represent favorable self-portrayal reflecting an underreporting tendency and cannot be interpreted as indicating the absence of the problems they are intended to assess. Elevated substantive scale scores may not fully reflect the magnitude or severity of as- sociated problems when the K score reaches this level. CASE ILLUSTRATIONS We turn next to a series of cases selected to illustrate MMPI-3 Validity Scale interpretation. Figure 8.1 presents the validity scale scores for a test taker who provided 32 unscorable responses (the Cannot Say raw score appears in the fourth row under the profile). The third row provides the percent scorable responses (Response %) for each of the MMPI-3 Validity Scales. As noted earlier, the score on any scale for which this percentage falls below 90 cannot be interpreted following the standard guidelines in Tables 8.2 through 8.12. Specifically, the absence of elevation on these scales cannot be interpreted as indicating the absence of the protocol validity threat assessed by that scale. In Figure 8.1, this applies to the scores on CRIN, VRIN, TRIN, FBS, and K, for which the percent scorable responses are all lower than 90% and hence are printed in bold. In the case of the CRIN scale, which has 83% scorable responses, this translates into 15 unscorable item pairs. Had the test taker re- sponded to at least 10 of those item pairs inconsistently, their CRIN T score would reach 80, invalidating the protocol. Therefore, it is not possible to rule out that this is an invalid MMPI-3 protocol. As seen in Figure 8.2, this individual’s scores on the MMPI-3 Higher-Order (H-O) and Restructured Clinical (RC) Scales are all well within normal limits. However, the percent scorable responses on more than half of these scales falls below 90. The ab- sence of elevation on these scales cannot be interpreted as indicating the ab- sence of the problems they assess. The Validity Scales profile depicted in Figure 8.3 illustrates a case in which neither the VRIN nor the TRIN score reaches a level that would invalidate this MMPI-3 protocol. A T score of 83 on CRIN indicates nonetheless that this is an invalid protocol. This case illustrates one way in which the CRIN scale provides incremental information beyond VRIN and TRIN. As described ear- lier, the CRIN raw score equals the sum of the inconsistent variable (one True the other False), inconsistent True, and inconsistent False responses, which is 17 in this case. This random pattern, reflected in intermittent variable and fixed responding, invalidates the protocol. As indicated under the Possible rea- sons for score heading in Table 8.3, this level of inconsistent responding can  MMPI-3 Validity Scales 120 --- --- --- --- --- --- --- 110 --- 100 --- 90 80 --- 70 60 50 --- --- --- 40 --- --- --- --- 30 --- --- 20 CRIN VRIN TRIN F Fp Fs FBS RBS L K Raw Score: 6 4 13 2 0 1 4 9 1 6 T Score: 51 51 50 47 41 47 37 61 40 50 Response %: 83 82 84 97 95 100 87 100 100 79 Cannot Say (Raw): 32 The highest and lowest T scores possible on each scale are indicated by a "---"; MMPI-3 T scores are non-gendered. CRIN Combined Response Inconsistency F Infrequent Responses L Uncommon Virtues VRIN Variable Response Inconsistency Fp Infrequent Psychopathology Responses K Adjustment Validity TRIN True Response Inconsistency Fs Infrequent Somatic Responses FBS Symptom Validity Scale RBS Response Bias Scale FIGURE 8.1. MMPI-3 Score Report Validity Scales Profile Showing Nonresponding FIGURE 8.2. MMPI-3 Score Report Validity Scales Profile Showing Nonresponding  MMPI-3 Validity Scales 120 --- --- --- --- --- --- --- 110 --- 100 --- 90 80 F --- 70 60 50 --- --- --- 40 --- --- --- --- 30 --- --- 20 CRIN VRIN TRIN F Fp Fs FBS RBS L K Raw Score: 17 9 9 6 3 4 14 14 8 5 T Score: 83 72 73 F 60 67 64 64 80 69 47 Response %: 100 100 100 100 100 100 100 100 100 100 Cannot Say (Raw): 0 The highest and lowest T scores possible on each scale are indicated by a "---"; MMPI-3 T scores are non-gendered. CRIN Combined Response Inconsistency F Infrequent Responses L Uncommon Virtues VRIN Variable Response Inconsistency Fp Infrequent Psychopathology Responses K Adjustment Validity TRIN True Response Inconsistency Fs Infrequent Somatic Responses FBS Symptom Validity Scale RBS Response Bias Scale FIGURE 8.3. MMPI-3 Score Report Validity Scales Profile Showing Combined Inconsistent Responding 358 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S result from reading or language limitations, cognitive impairment, errors in recording responses (unlikely in this case because responses were recorded for all 335 items), intentional random responding, or an uncooperative test taker. Determining which of these factors play a role and to what extent (they are not mutually exclusive) requires consideration of extratest information relevant to each of the possible contributing factors just listed. Regardless of why this indi- vidual engaged in a random pattern of inconsistent responding, the protocol is invalid. An important clarification is that in cases in which one or more of the inconsistency scales indicate that the protocol is invalid, scores on the content- based invalid response indicators and the Substantive Scales (but not the in- consistency scales) are invalid. In the case depicted in Figure 8.3, this includes elevations on RBS and L, which are invalid and cannot be interpreted. Figure 8.4 shows a protocol with the same CRIN score as the case depicted in Figure 8.3; however the T score of 93 on VRIN indicates that in this case the predominant pattern of inconsistent responding was variable. This can be discerned by subtracting the raw score on VRIN (14) from the CRIN raw score (17), indicating that the test taker provided 14 variable inconsistent response and only 3 fixed inconsistent responses. In contrast, in the case depicted in Figure 8.3, the test taker provided 9 variable inconsistent responses and 8 fixed inconsistent responses (determined by subtracting the VRIN raw score from the CRIN raw score). The same possible reasons for elevation listed for the case in Figure 8.3 apply to the elevated CRIN and VRIN scores in Figure 8.4. Here too, regardless of why the test taker engaged in a substantial pattern of variable inconsistent responding, the protocol is invalid, and the score on Fp cannot be interpreted as indicating the possibility of overreporting. This is reflected in the interpretive implications listed for Fp scores in the 70T–79T range in Table 8.7, which indicate that inconsistent responding must be ruled out before consider- ing other possible reasons (including severe psychopathology or overreporting) for the test taker’s score of 76T on Fp. In the case depicted in Figure 8.5, the test taker produced an invalidating VRIN T score of 85, whereas neither the CRIN nor the TRIN score reached a level that would invalidate the test protocol. The CRIN raw score of 14 includes the 12 VRIN item pairs that were answered inconsistently and two additional TRIN item pairs answered inconsistently. A TRIN T score of 50 indicates that the two inconsistent pairs of responses offset each other, with one pair keyed True and the other False. Thus, nearly all the inconsistent responding in this protocol was variable (one item keyed True the other False). The result- ing scores on the content-based invalid response indicators and the Substantive Scales are invalid and uninterpretable. Figure 8.6 shows a Validity Scales profile marked by a very high level of fixed, content-inconsistent True responding that invalidates the test protocol. Exami- nation of the interpretive guidelines for TRIN in Table 8.5, indicates that this  MMPI-3 Validity Scales 120 --- --- --- --- --- --- --- 110 --- 100 --- 90 80 --- 70 60 F 50 --- --- --- 40 --- --- --- --- 30 --- --- 20 CRIN VRIN TRIN F Fp Fs FBS RBS L K Raw Score: 17 14 12 6 4 6 9 9 4 9 T Score: 83 93 54 F 60 76 75 51 61 52 59 Response %: 100 100 100 100 100 100 100 100 100 100 Cannot Say (Raw): 0 The highest and lowest T scores possible on each scale are indicated by a "---"; MMPI-3 T scores are non-gendered. CRIN Combined Response Inconsistency F Infrequent Responses L Uncommon Virtues VRIN Variable Response Inconsistency Fp Infrequent Psychopathology Responses K Adjustment Validity TRIN True Response Inconsistency Fs Infrequent Somatic Responses FBS Symptom Validity Scale RBS Response Bias Scale FIGURE 8.4. MMPI-3 Score Report Validity Scales Profile Showing Variable Inconsistent Responding FIGURE 8.5. MMPI-3 Score Report Validity Scales Profile Showing Variable Inconsistent Responding  MMPI-3 Validity Scales 120 --- --- --- --- --- --- --- 110 --- T 100 --- 90 80 --- 70 60 50 --- --- --- 40 --- --- --- --- 30 --- --- 20 CRIN VRIN TRIN F Fp Fs FBS RBS L K Raw Score: 12 3 22 23 6 11 18 13 2 1 T Score: 69 47 104 T 113 93 103 75 76 44 35 Response %: 100 100 100 100 100 100 100 100 100 100 Cannot Say (Raw): 0 The highest and lowest T scores possible on each scale are indicated by a "---"; MMPI-3 T scores are non-gendered. CRIN Combined Response Inconsistency F Infrequent Responses L Uncommon Virtues VRIN Variable Response Inconsistency Fp Infrequent Psychopathology Responses K Adjustment Validity TRIN True Response Inconsistency Fs Infrequent Somatic Responses FBS Symptom Validity Scale RBS Response Bias Scale FIGURE 8.6. MMPI-3 Score Report Validity Scales Profile Showing Fixed, Content-Inconsistent True Responding 362 | I N T E R P R E T I N G T H E M M P I - 3 V A L I D I T Y S C A L E S test taker’s T score of 104T falls well above the cutoff for invalidity. The pos- sible reasons listed for this elevated score include an uncooperative test-taking approach and/or difficulties with double negatives. The other possible reasons listed for CRIN and VRIN elevations (e.g., reading and language limitations) do not apply to elevated TRIN scores because there is no reason to assume that they would result only in indiscriminant True responding. Note that the VRIN T score (47) indicates no significant pattern of variable inconsistent re- sponding. Examination of the row labeled Fixed True responding in Table 8.1, identifies this response pattern as a potential confound that can artifactually elevate scores on the overreporting scales. In this case, the highly elevated F, Fp, and Fs scores are therefore invalid and cannot be interpreted as indicating pos- sible overreporting. The same row of Table 8.1 also indicates that fixed True responding can artifactually lower scores on the underreporting scales, which likely accounts for the low scores on L and K in this case. The Validity Scales profile depicted in Figure 8.7 reflects a very high level of fixed, content-inconsistent False responding that invalidates the protocol. Here too, the possible reasons listed for this score in Table 8.5 include an uncoopera- tive test taker or difficulties with double negatives. As seen in the row labeled Fixed false responding in Table 8.1, in contrast with fixed True responding, this response pattern can artifactually elevate scores on the underreporting scales. Consequently, in the case depicted in Figure 8.7, fixed False responding invali- dates the elevated L and K scores. Figure 8.8 depicts a case in which a highly elevated T score on F (116) indi- cates that the protocol is invalid owing to overreporting. As seen in Table 8.6, possible reasons for an F score at or above 100 include inconsistent respond- ing and overreporting. Nonelevated scores on CRIN, VRIN, and TRIN in this case rule out inconsistent responding. The interpretive implications listed in Table 8.6 indicate that this level of infrequent responding is uncommon even in individuals with genuine severe psychological difficulties who report credible symptoms. Therefore, scores on the Substantive Scales are invalid and should not be interpreted. Note that the possibility of malingering is not listed for this