6 Personality and Broadband Self-Report Scales

Perhaps there is head scratching at the sight of a separate chapter devoted entirely to self-report broadband scales. After all, the titles of many self-report scales are identical to the titles of teacher- and parent-completed scales seen in the preceding chapter (e.g., BASC-3, Conners CBRS). Despite familiar titles, self-report scales possess different capabilities and perspectives than their informant-completed counterparts. Some differences derive from an entirely self-evident fact. Thinking is a private experience. The same holds true regarding the texture of emotional life, perceptions of the social world, the nature and flow of thoughts, personal preferences, and predispositions that may never appear in one’s overt actions. Consider for a moment, ratings of anxiety. It’s true that anxiety often includes overt aspects (e.g., reluctance in the face of novelty). But covert aspects are also overwhelmingly common. For example, many adolescents suffer troubling private worries as they confront age-related biological, cognitive and identity-related changes (Dugas, Laugesen & Bukowski, 2012). And such worries may be fully appreciated only by each teen herself. What’s more, informants may miss important aspects of anxiety not only because of its covert nature but also because it is not readily witnessed in  certain  settings. For example, a teen, uniquely, may register strong anxiety felt during peer interactions that transpire with no parent-informants around to see it (Keeley et al., 2018). Equally important, a high school teacher can easily overlook strong overt expressions of anxiety because her class contains too many students to permit close vigilance of each one. Well-devised self-report scales might fill the gaps inherent in “other-completed” scales.

That said, in at least a few practice venues, with a least a few practitioners, the subjective world is apt to get short shrift. Self-report scales (as well as student interviews) may be downplayed or skipped. This is arguably true because school psychology’s dominant theoretical orientation during the last 30 years has been behavioral (see Bergen, 1990), characterized by a preference for behavioral techniques in both assessment (e.g., Kilgus, Kazmerski, Taylor & von der Embse, 2017) and intervention (e.g., Steege, Pratt, Wickerd, Guare & Watson, 2019). As further evidence of a preference for overt behavior, informant-completed rating scales top the list of the most-used assessment instruments, whereas self-completed instruments are far down that same list (Benson, Floyd, Kranzler, Ekert, Fefer & Morgan, 2019). In fact, the top-ranked parent-completed and teacher-completed scales were reportedly used between two and five times more often than self-report scales. In a contemporary world where behaviorism predominates many professional practices, self-report and the internal world may need advocacy. In this chapter it is argued that self-report scales help tap that world and prevent school psychologists from missing important aspects of students’ personality and adjustment.

There is an additional reason for a separate self-report chapter. That is because we need to talk briefly about one monumentally important self-report scale, the Minnesota Multiphasic Personality Inventory (MMPI). You will hear about it twice. Once concerning  self-report scales generally; this is because of its role as a trailblazer. You will hear about it again (as the MMPI’s adolescent version) because of its occasional use in middle school and high schools.

Things to Keep in Mind for All (Most) Self-report Scales and Personality Inventories

There are some general considerations when using self-report scales in schools. These considerations are covered before we turn to individual self-report scales.

Some Youth, Especially the Very Young, Are Poor Self-informants

Research suggests that by age five years children typically demonstrate an understanding of their own psychological characteristics and possess a formulated conception of self (Jia, Lang & Schoppe-Sullivan, 2016). This might imply, on its surface, that any child with a five-year-old’s developmental level would prove a good candidate to rate herself. Not so. In fact, the BASC-3 “Self-report of Personality” and Conners CCBRS self-report version are suited to those only 8-years and older. There seem to be two considerations that might constrain valid self-reporting: general developmental maturity and literacy level. Consider development first. We know as psychologists that there is certainly more to responding to self-report questions than a recognition of one’s own emotions. For example, consider all of the psychological processes needed to respond to the simple true/false question: “I feel sad.”

1. After this item is read, working memory sufficient to hold the item (and its meaning) online is needed while long-term memory is accessed.
2. Memory search strategies are invoked to scan prior experiences for occurrences of sadness, perhaps considering their frequency, intensity, and recentness.
3. Analytical capabilities are tapped to permit a summary judgment about “sad feelings.”
4. Judgment is applied to consider any off-setting feelings of happiness that might change a sadness rating.
5. Judgment is used to convert the just-reviewed level of sadness against response options (true/false).

These steps, for just a single item, require cognitive maturity. What’s more, all of these considerations are magnified when many items, not just one, are confronted. No wonder that self-report scales rarely reach down to children younger than age eight years. An immature brain may fail to approach these requirements with the ease of a parent or a teacher. Although a manual-specified age level might be listed as eight years, it seems plausible that children with developmental delays (e.g., low full-scale IQ) might fail to respond validly despite their age. The same is true for delayed language development. Indeed, delayed language development, is known to constrain young children’s emotional understanding and their ability to talk about first-hand emotional experiences (Rieffe & Wiefferink, 2017).

Next consider the issue of literacy. A child has to read fluently enough to complete a 100+ item scale. She also has to possess sufficient reading comprehension to grasp the meaning of each item. Of course, school psychologists routinely encounter students with poor literacy. What’s more, poor literacy is rife among students placed in emotional disorders programs (Hurry, Flouri, & Sylva, 2018), with average reading scores < 25th percentile in one study (Lane, Barton-Arwood, Nelson & Webby, 2008). This would imply that despite their reliance on items with simplified reading requirements, that self-report scales, at least on some occasions, will represent a challenge to read. It’s tempting to opt for simply reading the items for such a student. But for many self-report scales, each student reading items for herself was the method used during standardization. Consequently, the published norms are not applicable when standard administration procedures are violated.

Self-ratings: Advantageous for Internalizing Problems?

It’s intuitive that self-report scales hold certain advantages, even if selective ones, over scales completed by a teacher or parent. The most obvious of these relate to internalizing symptoms. After all, might not a teenager judge the degree of his subjective worry or the intensity of his own unhappiness better than a parent or a teacher? Consequently, shouldn’t a parent or teacher rating prove less valid? Accordingly, shouldn’t self-ratings miss fewer problems (produce fewer false negatives) than ratings coming from a teacher or a parent? The answer appears to be perhaps yes or perhaps no. There are hints from the published literature. For example, Youngstrom, Findling and Calabrese (2004) compared adolescents’ own ratings of symptoms of elevated mood (mania) and depressed mood compared to their parents. “Youth self-report of mania symptoms showed lower correlations with clinician ratings than did parent ratings” (Youngstrom et al, 2004, p. S5). The same was true regarding depressive symptoms. What about tools that school psychologists might use and more common concerns? Let’s look at the of Conners CBRS and anxiety. Table 6.1 indicates the classification accuracy of four Conners CBRS scales concerning their ability to predict membership in an anxiety clinical group compared to control youth. If correct classifications arising from self-report exceed those derived from informant reports, then the first column should have the highest values. This is does not seem to be the case. In fact, overall accuracy arising from self-reports rarely exceeds accuracy produced via informant reports. This is just one summary from one aspect of internalizing problems using just one scale, but it gives rise for caution. The adage that self-reports are uniformly best for internalizing problems may not always prove correct.

Same Title, Different Details

Beth Kazan is an experienced school psychologist just assigned to a high school after three years of elementary-school practice. At her elementary site, Beth was a consistent and confident user of both the BASC-3 Teacher Rating Scale (TRS) and Parent Rating Scale (PRS). She had, however, never touched a BASC-3 Self-Report of Personality (SRP). Beth recently conducted her first high school evaluation. Cognizant of the ability for self-appraisal among high school students lacking among her former elementary students, Beth opted to include in her assessment battery the BASC-3 SRP. But examining her first BASC-3 SPR printout proved shocking. Although she certainly found some well-known composite and clinical scale names (e.g., Internalizing Problems, Hyperactivity), other familiar scores were nowhere to be found (e.g., BSI, Externalizing Problems). More bewildering was what to make of the printout’s group of completely unfamiliar SPR scales. For example, the SRP reports a dimension entitled Sense of Adequacy and another entitled Ego Strength. Perhaps one can intuit the meaning of “Sense of Adequacy” and of “Ego Strength.” And, perhaps, those intuitions would prove to be correct. But guessing at the nature of a score’s meaning represents a poor way to practice. This means that novel users of self-report scales, like Beth, need to do due diligence and pre-acquaint themselves with their tools. You are advised against assuming that a BASC-3, is a BASC-3, is a BASC-3. Chagrined, Beth eventually grasp this reality as she opened and commenced  reading the BASC-3 manual.

Personality Inventories Same as, and Different from, Self-report Behavior Rating Scales

Self-report measures of personality are sometimes called personality inventories, which are common in adult practice settings. This title is applied because these tools can take an inventory of various aspects of personality via informants’ self-rating. These inventories, in turn, are applied to diverse aspects of practice. This is exemplified by adult-specific instruments such as the Revised NEO Personality Inventory (Costa & McCrae, 1992). This tool is designed to measure five factor-analytically-derived personality traits, the so-called “Big Five.” These are as follows:

1. Openness to experience
2. Conscientiousness
3. Extraversion
4. Agreeableness
5. Neuroticism

The five traits then can be used in research and practice. For example, marriage satisfaction across life stages can be examined by looking at marriage partners’ similar versus complementary personality structures as characterized by the Big Five (Shiota & Levenson, 2017). Similarly, the Big Five personality traits are found to be differentially related to vocational interests, such as in science, technology, engineering and math–STEM (Perera & McIlveen, 2018). Closer to the topic of this book, the Big Five also predicts adult psychopathology. For example, in a study of 4,927 adults (college students) using the Revised NEO Personality Inventory it was found that “some predicted associations between self-evaluation and psychopathology were not found, and unanticipated associations emerged” (Durrett & Trull, 2005, p. 359). Although marriage and family therapist may employ them, and the same for guidance counselors aiding their clients in making career choices, the Big Five are generally of secondary importance for clinical diagnostic purposes. It turns out that personality inventories tapping the Big Five (including the 16 PF, see Cattell & Schuerger, 2003) look a lot like self-report behavioral rating scales that you will soon hear about (e.g., they have multiple dimensions, are norm-referenced). But they are primarily interested in generic personality, not clinical diagnosis, such as might be used in career and vocational planning. Thus, these instruments use nomothetic elements (norm referencing) but they are also idiographic. They depart from many purposes of self-report rating scales, such as assistance in reaching categorical, not dimensional, conclusions. Once again, assessment tools need to match the purpose at hand, as is often reflected in the referral question.

Contextualizing Self-report Tools: Start with the MMPI

Many of the features of self-report scales can be traced to a singular innovative endeavor. This was the creation and refinement of the original Minnesota Multiphasic Personality Inventory (MMPI). Note the use of the term personality inventory, even though the MMPI seems nearly identical to now-common self-report rating scales. A portion of the MMPI’s importance is foundational and trail blazing. Another portion is exemplified by revealing just what self-report scales can do with sufficient refinement. Yet another portion concerns watershed research conducted on the MMPI that exposed clinicians’ limitations of human judgment. In light of these reasons, a brief digression into the general version (adult) MMPI is in order before we turn to consideration of individual tests suited for school use, including the adolescent version(s) of the MMPI.

Way back in 1937, a psychologist, Starke Hathaway, and a psychiatrist, Charnley McKinley, at the University of Minnesota, formulated an instrument that became the MMPI. They described the first MMPI as “an objective aid in the routine psychiatric case work-up of adult patients and as a method of determining the severity of the conditions” (Dahlstrom, 1972, p. 4). The authors devised a vast array of one-line questions that adults read and confirmed or disconfirmed (i.e. answered “yes” or “no”). You can already see similarities to contemporary instruments. The MMPI’s items covered multiple dimensions of potential psychopathology, in someways paralleling a detailed diagnostic interview. But item wording was often subtle, avoiding the transparency that might allow respondents to discern the authors’ purpose and thus outsmart the test. The MMPI soon achieved widespread adoption during World War II (Hathaway & McKinley, 1942), a period short of trained psychologists and psychiatrists. As an alternative to time-consuming and professionally-taxing diagnostic interviews, it helped to usher in the era of objective assessment techniques comprised of many individual items (i.e., personality inventories). The early MMPI foreshadowed the objectivity of today’s tools: every administration consisted of identical items, all respondents were provided identical directions, raw scores were always generated in the same manner, plus derived scores (in this case T-scores) were invariably calculated and plotted. Standardization, in turn, facilitated quantification, soon followed by vast empirical efforts at scale interpretation.

Seminal MMPI research, conducted during an era when clinicians’ judgment was often viewed as sacrosanct, included one particularly eye-opening finding. It concerned the radical notion that mere scores (treated in algorithmic manner) are more accurate than clinicians’ non-numeric, expert judgments. The classic work of University of Minnesota professor Paul Meehl (1954), Clinical versus statistical prediction: A theoretical analysis and review of the evidence, proved to be a pivotal publication. This is because it documented this very finding—cold, hard numbers typically outperform the conclusions of trained mental health specialists. In other words, for diagnostic purposes, nomothetic procedures nearly universally outperform idiographic procedures. (Incidentally, an meta-analysis conducted 50 years after Meehl’s original work largely confirmed his original conclusion; Ægisdottir et al., 2006).

But what might an interpretative algorithm look like and why should we care about it today? In 1956, Paul Meehl published yet another important (and provocative) article. This one was entitled, oddly, “Wanted-A good cookbook.” Echoing his earlier finding that clinicians’ descriptions of clients’ characteristics routinely over-valued their own insights and conclusions, Meehl proposed his radical alternative, which was dubbed the “cookbook method.” This method was envisioned to assume various forms depending on the availability of research associated with an instrument’s cookbook. “Whatever its form, the essential point is that the transition from psychometric pattern to personality description is an automatic, mechanical, ‘clerical’ kind of task, proceeding by the use of explicit rules set forth in the cookbook.”

For the MMPI, the cookbook would revolve around a fixed interpretative process using a “two-point code.” Thus, various MMPI profiles would not be interpreted subjectively based on overall configuration or magnitude of several scale elevations. Rather, the basic 10 MMPI scales would always be examined and if two scales were elevated, then those two elevations (regardless of the order and magnitude) would be listed to anchor interpretation. For example, if an MMPI resulted in the high scales being scale 1 (Hypochondriasis, with a T-score of 74) and scale 3 (Depression, with a T-score of 77), then a 1-3 code would be documented.

According to his approach, the clinicians need not surmise what the admixture of hypochondriacal characteristics and depressive symptoms might portend for their case. Instead—and this is the key point—they next consult a code book that list empirically-established characteristics associated with a 1-3 code. Practitioners skip (or minimize) clinical judgment; they merely consult the empirical correlates of each case’s objective code type. The upshot of all of these factors helped to make the MMPI wildly popular among psychologists. In fact, it has now been translated into more than 50 languages. Over the past 75 years no topic in psychology has been the subject of more publications than the MMPI. For example, in 2017 alone the number of peer-reviewed, English language publications devoted to the MMPI was at least 88! For context, this one-year total is far more than can be found  for the BASC-3 (and BASC-3) over an entire decade.

You can probably envision the importance of a methodology like a user-friendly and quickly-determined two-point code. You can also probably envision push back from clinicians. Meehl, more than 60 years ago, saw this coming: “I am quite aware that the mere prospect of such a method will horrify some of you [speaking to his clinical psychology colleagues]; in my weaker moments it horrifies me” (Meehl, 1956, p. 264). Because Meehl’s “statistical prediction” and his cookbook proposal is all about rules and numbers, not clinical wisdom. You can sense that he was arguing for something akin to the Bayesian approach (probability nomogram) you saw in Chapter 2. And, of course, he was a champion of the nomothetic approach and harsh detractor of the idiographic approach (at least for most referral concerns).

There is more to the historical legacy of the MMPI; this concerns validity scales. Even for school psychologists who never touch an MMPI-A, its validity scales are still worth contemplating. This is the genesis of most validity scales you see in your current practice (e.g., as found on the BASC-3 or Conners CBRS). In the 1940s, validity scales were a novel feature and an insightful one. Adults seen in clinical settings sometimes wanted to make themselves look favorably. Perhaps a family member was advocating for unwanted treatment or (especially in the old days) institutionalization. Consequently, a set of MMPI items that could flag flagrant problem minimization might alert clinicians to a potentially invalid profile. This effort produced the L (Lie) scale. A related validity scale, the K scale, also concerned defensiveness. The K scale, however, was used to make statistical corrections to certain clinical scales to account for defensiveness. The K scale’s application allowed for fewer false negative profile interpretations (i.e., instances in which a genuine problem was missed). Research shows, however, that K-corrections do not enhance classification decisions with youth on the MMPI-A (Alperin, Archer & Coates, 1995), even though they may work with adults.

Of course, respondents may not only wish to present themselves positively, sometimes they sought the opposite. In other words, they attempted to feign severity. This might concern, for example, someone seeking a Workmen’s Compensation claim after a trauma who would like to appear as impaired as possible. It might also be someone attempting to avoid conscription in the Armed Forces (a real consideration for baby boomers during the Vietnam War). Accordingly, a scale comprised of items that are rarely confirmed by individuals was created (F scale). Critically, these items are infrequently endorsed even among adults with true psychopathology (cleverly, however, some of the F scale items possessed face validity suggesting psychopathology). Table 6.2 provides more information.

Some Self-report Scales and Personality Inventories

Minnesota Multiphasic Personality Inventory-Adolescent (MMPI-A)

All of the popularity enjoyed by the MMPI in the adult clinical world notwithstanding, it is a much more modest player in the adolescent clinical world, and apparently even less when the venue is a school campus. School psychologists’ place it far down a list of potentially-used self-report scales (Benson et al., 2019). The MMPI-A was created using items from the original MMPI and the updated MMPI-2 in 1992. It was devised expressly for use with adolescents. The MMPI-A comprises 478 true/false items (Butcher et al., 1992) with an updated manual (Butcher et al., 2006) appearing some years ago. A new MMPI-RF (revised form) has been published based on the MMPI-A (Archer, Handel, Ben-Porath & Tellegen, 2016), although the prior MMPI-A remains in use and is indeed marketed concurrently with the newer MMPI-A-RF. The MMPI-A-RF is discussed a few paragraphs below. The MMPI-A is suited for those 14 to 18 years of age with reading levels of mid-to-high fourth grade; it is supported by norms derived from a national sample of approximately 1,600 teens.

MMPI-A Organization and Characteristics

The MMPI-A uses the 10 traditional clinical scales and associated single-digit numbers to indicate them. These are:

• 1 Hs – Hypochondriasis
• 2 D – Depression
• 3 Hy – Hysteria
• 4 Pd – Psychopathic Deviate
• 5 Mf – Masculinity/Femininity
• 6 Pa – Paranoia
• 7 Pt  – Psychasthenia
• 8 Sc – Schizophrenia
• 9 Ma – Hypomania
• 0 Si  – Social Introversion

Many of these terms sound archaic to our 21st century ears. Words like hysteria and psychasthenia have largely vanished from today’s lexicon (although terms like depression and paranoia persist). Scale names, however, may matter relatively little in MMPI interpretation. This is because there is a MMPI-A-specific system used to aid interpretation that often frees diagnosticians from the need to make inferences based on individual scales, as you saw with use of the two-point code earlier.

MMPI-A: Exemplifying Specialized Interpretation

Fortunately, use of the MMPI-A is supported by a text for psychologists, which comprises 459 pages (Archer, 2017). The usefulness of this vast MMPI information storehouse is exemplified by the case of 15-year-old Jennifer, a student referred for plummeting report card grades and waning school interest. Indeed, review of her cumulative file suggested two semesters of falling marks as well as mounting conflict with siblings. More troubling was a report of two episodes of loud disagreement coupled with agitation during second period English class. There was no evidence of health problems, and an interview with her parents contraindicated the presence of changes in family/living circumstances or noteworthy environmental stressors. With background information like this, the school psychologist’s preliminary hypotheses concerned (1.) a mood problem, perhaps involving rapid mood swings, (2.) emerging anti-social tendencies or (3.) perhaps the occurrence of trauma not detected by simply interviewing Jennifer’s parents. The school psychologist estimated that the base rates for the first and third hypotheses were fairly high in her practice setting (a middle-class high school where school psychologists embrace a mental health role). In contrast, she estimated the base rate for the second hypothesis as rare, but not unheard of, among adolescents seen at her site. Before conducting an interview, the school psychologist had Jennifer complete an MMPI-A. A 2-0 code type resulted. Relying on the two-point-code rubric, she need not surmise what scale 2 (depression) actually tapped or just what was implied by an elevation corresponding to the one Jennifer was expressing on this scale. Similarly, she did not need to conjecture about her scale 0 (social introversion) elevation and try to amalgamate elevations on scales 2 and 0 into a portrait of Jennifer’s social-emotional status. Even more important, there was no need to attend to and synthesize several other clinical elevations found in Jennifer’s MMPI-A profile. Instead, she wrote down the 2-0 code and turned to a published codebook. Consider the bullet points associated with this code:

• quite uncommon in clinic settings (occurs among 1.8% of males; 3.8% of females)
• often referred for shyness, lethargy, interpersonal sensitivity, symptoms of depression, anxiety, withdrawal, social introversion and feelings of inferiority
• conforming, passive
• very unlikely to engage in antisocial behavior
• unlikely to use drugs or drink alcohol
• most are compliant
• there may be an elevated risk of bulimia or anorexia
• assertiveness training and cognitive behavior therapy may help to decrease depressive thoughts

Although stand-alone MMPI use is ill advised, it nonetheless remains clear that much might have been learned about Jennifer by simply considering her two-point code. Speculation about a mood problem seems to have been substantially strengthened, whereas hypotheses about anti-social behavior and trauma seem to have been weakened. There is obviously more assessment work to be done, but Jennifer is arguably much better understood by virtue of the bullet point information listed above.

Reliance on two-point codes presupposes that adolescents designated by each two-point code share important characteristics (i.e., they are at least somewhat homogeneous regarding personality characteristics and mental health problems). Crucially, it also depends on research correlating code types with clinicians’ descriptions. Use of two-point codes also fuels applied research on success during treatment, gold standard diagnoses, common comorbidities, and the risk of various life events. Research suggests that relying on codes and concomitantly diminishing the importance of clinicians’ judgment fosters accuracy. This is Paul Meehl’s famous argument from nearly 70 years ago.

In part enhanced accuracy occurs because clinicians, like all humans, can be swayed by  transient factors, as well as longstanding biases, unavailable to conscious awareness (Kahnemann, 2011). And associated reliance on the Automatic System can stifle good judgment. Influences as trivial as time of day, level of comfort in the office where one is working, and even insignificant aspects of a client’s presentation (whether he selected a blue shirt, which the clinician likes, rather than a gray shirt, which the clinician detests). For clinicians working in mental health settings, where caseloads may consist of one adolescent after another, it’s easy to envision frequent use of the MMPI-A. In a practice like this, unique aspects of the MMPI-A and familiarity with various code types are possible. But for the infrequent user (e.g., a school psychologist who partitions her time among elementary and high school venues or one for whom there is limited mental health involvement), effective use of the MMPI-A may prove quite difficult.

MMPI-A Supplemental Scales

The adult MMPI offers many research and practice-relevant supplemental scales, an important advent. The same is true of the MMPI-A. The formulation of validity scales is just one of the MMPI’s historically important contributions. Introduction of content and supplemental scales is another. Here’s the logic. The initial MMPI item pool might be reconfigured to formulate alternative, novel scales whose items tapped considerations of interest to the clinicians. Indeed, this very proposal has garnered progressive popularity. Table 6.3, for example, indicates that at least 87 such scales exist for diagnosticians who work with teens. Many of these, obviously, might be useful for school psychologists seeking to confirm or disconfirm hypotheses (or even for formulating hypotheses when no obvious ones are yet apparent). For example, a scale called Alcohol/drug Problem Proneness (PRO) and another one called Low Achievement Orientation (A-las-1) might help clarify students’ problems. The same is true for two other scales, Immaturity (IMM) and Adolescent-low self-esteem (A-lse), which turn out to afford value in determination of high school students with emotional disturbance (as seen a few paragraphs below). Incidentally, as a review of the BASC-3 and Conners CBRS manuals will confirm, these instruments also use this methodology (i.e., creating novel supplemental scales via reconfiguring existing items).

MMPI-A and Classification Validity

It may now be obvious that complex algorithms are possible because of the MMPI-A’s objective scores and its many clinical, content, and supplemental scales. In fact, inventive researchers have demonstrated the capability of MMPI-A scores to predict an interesting array of outcomes. In one example, researchers concerned with predicting delinquency (Morton, Farris & Brenowitz, 2002) examined several scales working in concert. Theory were blended with straightforward empiricism. Excitatory tendencies (as measured by the psychopathic deviation and schizophrenia scales) contrasted with inhibitory tendencies (as measured by masculinity/femininity and social introversion) were envisioned to predict tendencies toward delinquency. This was confirmed. The score best combinations produced a DLR^positive of 3.5 and a DLR^negative of .28. Moreover, the rate of correct prediction remained as strong with cross-validation. In a second study of delinquency, researchers (Edner, Glaser, Calhoun, Dukes-Murray, Khan & Donaldson, 2020) used a specially crafted MMPI-A scale entitled the Trauma Scale for Juvenile Offenders (TSJO). The goal was to distinguish adjudicated youth with a history of trauma from those lacking such a history. High levels of TSJO endorsement resulted in a  DLR^positive of 1.83; low levels with a DLR^negative of .59. It’s not these values per se that matter but what else the researchers included in their article. Perhaps foreshadowing a happy future were diagnostic utility statistics abound, the authors proved vast information. This included: sensitivity, specificity, total accuracy, positive predictive value, negative predictive value, positive DLR, negative DLR, indication of statistical significance, effect size, all calibrated relative to changing cut-points.

Turning to more practical examples of classification validity, consider the research conducted in Texas (Miller, Li & Kabell, 2015) a state where school psychologists’ gatekeeping duties concern emotional disturbance (ED) but not specific learning disability (an exceptionality managed largely by educational diagnosticians). These researchers addressed, in part, the MMPI-A’s ability to distinguish students with emotional disturbance from controls as well as another group of children who were experiencing a current psychiatric hospitalization or treatment in a residential treatment center. Favorably, the researchers calculated diagnostic utility statistics—just the kind of analysis needed to inform classification decisions. As seen in Table 6.4, MMPI-A scores were predictive of ED status (e.g., three scales are associated with increased risk of ED), although effect sizes were unimpressive. Perhaps there will be more practice-friendly studies like this in the future.

MMPI-Adolescent-Restructured Form (MMPI-A-RF)

The MMPI-A keeps evolving, included are some fairly recent changes now marketed as the MMPI-A-RF (Archer, Handel, Ben-Porath & Tellegen, 2016). The MMPI-A-RF is briefer than its MMPI-A counterpart, comprised of 241 true/false items, all of which are shared with the MMPI-A. Interestingly, the standard MMPI-A remains available at the same time and from the same publisher (i.e., Pearson). Revisions were hoped to generate better accounting for variance associated with demoralization, an unwanted influence on traditional MMPI scales among adolescents (Stokes, Pogge & Archer, 2018). One obvious effect of the restructuring, however, was the advent of a hierarchical organization of score reporting (i.e., Higher-Order Scales, Restructured Clinical Scales and Specific Problem Scales), as reported by Handel 2016 and seen in Table 6.5. There were also updates in terminology. As seen in Table 6.5, the MMPI-A-RF’s terms seem modern (e.g., somatic complaints and dysfunction scale resonate to the contemporary practitioner more than archaic and arcane terms like hypochondriasis and psychasthenia as found in the MMPI-A). Regarding school-based use, scores for several comparison groups (e.g., youth in residential treatment, those in outpatient psychiatric settings, those in school settings) can be accessed from a test manual.

Regarding applications of both the MMPI-A and MMPI-A-RF, per-use scoring online (i.e., the “score report” option) may discourage some cash strapped school districts because of relatively high cost. Some school psychologists, however, may choose to use a hand score option. The complexity of the MMPI-A as well as the MMPI-A-RF is such that occasional use is probably contraindicated. MMPI-A dabbling risks interpretative errors and poor use of professional time. For these same reasons, some school psychologists are frustrated to find MMPI-A scores accompanying a teen who returns from a residential treatment stay or one who has undergone a clinic-based assessment. The effort to make sense of scores, especially when only a brief narrative report exists, can be dismaying. The concise volume prepared by Archer and Krishnamurthy (2002), although now dated, may still help in situations like this.