As we pointed out at the beginning of this chapter, personality traits are internally based, relatively stable characteristics that define an individual’s personality. Traits are continuous dimensions, and people vary from each other along the dimensions for each of the various traits that make up personality. But what are the specific traits that make up human personality? Think about all of the adjectives that can describe a person. There are thousands, but this doesn’t mean that thousands of traits are necessary to describe human personality. Trait theories attempt to discover exactly how many traits are necessary. You will be surprised to find out that the major trait theories propose that this number is rather small. How are such numbers derived?

Trait theorists use factor analysis (and other statistical techniques) to tell them how many basic personality factors (or traits) are needed to describe human personality, as well as what these factors are. Remember, factor analysis is the statistical tool from Chapter 6 that we used to discuss the question of whether intelligence was one ability or many. Factor analysis identifies clusters of test items (in this case, on a personality test) that measure the same factor (trait). Allport and Odbert (1936) did the initial analyses, starting with 18,000 words that could be used to describe humans. Their analyses left them with around 200 clusters. Later trait researchers have reduced this number considerably; the consensus number of necessary traits is only five.

In this section, we’ll consider two major trait theories that were developed using factor analysis, Hans Eysenck’s three-factor theory and the Five Factor Model of personality. We’ll also discuss personality assessment, because personality tests are often developed to assess a theorist’s proposed basic traits. The major uses for personality tests, however, are to diagnose personality problems, to aid in counseling, and to aid in making personnel decisions; therefore, we’ll focus our discussion on the two main types of personality tests used, personality inventories and projective tests.

According to trait theorists, basic traits are the building blocks of personality. Each trait is a dimension, a continuum ranging from one extreme of the dimension to the other. A person can fall at either extreme or anywhere in between on the continuum. An individual’s personality stems from the amount of each trait that he has. This is analogous to the trichromatic theory of color vision that we discussed in Chapter 3. Remember that according to the trichromatic theory, all of the thousands of different colors that we can perceive stem from different proportions of the three primaries (red, green, and blue). In trait theories, all personalities stem from different proportions of the basic traits.

355

The number and kind of personality traits. Different theorists have come up with different answers to the question of how many basic personality traits there are and even what they are. For example, using factor analysis, early trait researcher Raymond Cattell found that 16 traits were necessary (Cattell, 1950, 1965). More recently, Hans Eysenck, also using factor analysis, has argued for only three trait dimensions (Eysenck, 1982; Eysenck & Eysenck, 1985). There are two major reasons for the different numbers—the level of abstraction and the type of data. First, the number of traits depends upon the level of abstraction in the factor analysis. For example, if you factored Cattell’s 16 traits further, you would probably get something closer to Eysenck’s three higher-order factors (Digman, 1990). This means that Eysenck’s theory is at a more general and inclusive level of abstraction than Cattell’s theory. More important, the number of traits depends upon which databases and what types of data different theorists use in their factor analysis. Different data could, of course, lead to different traits, as well as a different number of traits. Surprisingly, contemporary trait researchers have reached something of a consensus about the number and the nature of the basic personality traits. The three traits proposed by Eysenck are insufficient, and the 16-factor theory of Cattell is needlessly complex (McCrae, 2011). The consensus number is five, and these five traits (or factors) are referred to as the “Big Five.” First, we’ll discuss Eysenck’s three-factor theory, because he (unlike most trait theorists) proposes actual causal explanations for his traits, and then we’ll discuss the Five Factor Model of personality.

Eysenck’s three-factor theory. Hans Eysenck’s three trait dimensions are extraversion–introversion, neuroticism–emotional stability, and psychoticism–impulse control. These are continuous dimensions, with extraversion, neuroticism, and psychoticism, respectively, at the high ends. Eysenck argues that these three traits are determined by heredity and proposes biological bases for each one of them (Eysenck, 1990, 1997). We will briefly discuss some of the proposed biological mechanisms for each of Eysenck’s three trait dimensions so you can see why this is a biological trait theory of personality.

Extraversion and introversion have their normal meanings, indicating differences in sociability. Extraverts are gregarious, outgoing people with lots of friends; introverts are those who are quiet, are more introspective, and tend to avoid social interaction. For example, in a study of students going off to college, the extraverted ones struck up new friendships more quickly than the introverted ones (Asendorpf & Wilpers, 1998). The main proposed biological mechanism for the extraversion–introversion trait is variance in the normal level of cortical arousal (neuronal activity) in a person. According to Eysenck, an introvert has a higher resting level of arousal in the brain than an extravert because of differences in their reticular formations. Remember, as you learned in Chapter 2, the reticular formation is responsible for our different levels of arousal. According to Eysenck, the brains of introverts are sufficiently active normally, but an extravert has to seek out external stimulation in order to raise the level of arousal in the brain to a more optimal level. Brain-scan studies have generally provided support for this contention. Introverts are typically found to show more cortical arousal compared with extraverts (Hagemann et al., 2009).

356

People who are high on the neuroticism–emotional stability dimension tend to be overly anxious, emotionally unstable, and easily upset; people who are low on this dimension tend to be calm and emotionally stable. People high on this dimension experience more lingering negative emotional states in their daily lives and suffer from higher rates of anxiety and depression (Widiger, 2009). Eysenck proposes differences in the limbic system, which plays a major role in regulating our emotions, and the sympathetic nervous system, which controls the body when it is in a stressful situation, as the main biological bases for this trait. People high on the neuroticism trait tend to have greater activity in their limbic system and a more reactive sympathetic nervous system, causing them to be on average more anxious and to become even more emotionally aroused and upset when they are threatened or find themselves in stressful situations.

The psychoticism–impulse control trait dimension is concerned with aggressiveness, impulsiveness, empathy (seeing situations from other people’s perspectives), and antisocial behavior. According to Krueger and Walton (2008), psychoticism seems to have been a poor choice of labels given its exaggerated image of pathology, so disinhibition might be a better label for the high end of this dimension. Eysenck assumes that the relevant biological mechanisms for this trait include a person’s levels of testosterone and MAO, a neurotransmitter inhibitor. For example, a high level of testosterone and a low level of MAO lead to behavior that tends to be aggressive, impulsive, antisocial, or lacking in empathy (disinhibited behavior reflective of the high end of this dimension).

Eysenck’s specific biological assumptions make his theory rather easy to test experimentally. Thus, it has generated much research, literally thousands of studies (Geen, 1997), and the findings have been fairly supportive. However, most research on personality traits seems to indicate that five factors, rather than three, are necessary to describe personality (Funder, 2001; Goldberg, 1990; John, 1990; McCrae & Costa, 2003; Wiggins, 1996). Unlike Eysenck, five-factor theorists use only the high end of each trait dimension to identify the factor. So how do you get to five from Eysenck’s three factors? Eysenck’s extraversion and neuroticism factors are retained, but his psychoticism factor is broken into two parts, conscientiousness and agreeableness, and an openness factor is added (Nettle, 2007). Now let’s take a closer look at the Five Factor Model of personality.

357

The Five Factor Model of personality. The Five Factor Model of personality was formulated by Robert McCrae and Paul Costa (McCrae & Costa, 1999). The descriptions of the high and low ends of the five trait dimensions (called the Big Five) given in Table 8.3 will provide you with a brief overview of the five proposed factors. To remember the five factors, try forming acronyms using their first letters. Order is not important. You could use OCEAN (Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism) or CANOE (Conscientiousness, Agreeableness, Neuroticism, Openness, and Extraversion). Similarly, you might use PEN to remember Eysenck’s Big Three traits (Psychoticism, Extraversion, and Neuroticism).

We already briefly discussed extraversion and neuroticism when we described Eysenck’s trait theory, so let’s see what conscientiousness, openness, and agreeableness entail and what real-world behaviors they are related to. Conscientiousness is a tendency to show self-discipline, dependability, and organization and to prefer planned versus spontaneous behavior. Conscientiousness is positively correlated with high school and college GPAs (Noftle & Robins, 2007) and is linked to living a healthier and longer life (Bogg & Roberts, 2004; Kern & Friedman, 2008; Roberts, Smith, Jackson, & Edmonds, 2009). Openness (often referred to as openness to experience) reflects a person’s degree of intellectual curiosity, creativity, and preference for novelty and variety. A person who scores high on the openness factor is willing to try new things, is imaginative, and tends to have liberal values and be open-minded (McCrae & Sutin, 2009). High scorers on openness also tend to score high on intelligence tests and the SAT Verbal test (Noftle & Robins, 2007; Sharp, Reynolds, Pedersen, & Gatz, 2010). Agreeableness is the tendency to be compassionate and cooperative rather than suspicious and antagonistic toward others. People who score high on the agreeableness factor tend to be friendly, altruistic, empathetic, and to have an optimistic view of human nature (Graziano, Habashi, Sheese, & Tobin, 2007). Variation in agreeableness also correlates well with the variation in social-cognitive theory of mind functioning in adults, with high scorers performing better on theory of mind tasks (Nettle & Liddle, 2008).

358

According to Costa and McCrae (1993, p. 302), the Five Factor Model “is the Christmas tree on which findings of stability, heritability, consensual validation, cross-cultural invariance, and predictive validity are hung like ornaments.” Indeed, recent research indicates that these five factors may be universal. They have been observed across gender, and in many diverse languages (such as English, Korean, and Turkish) and cultures (including American, Hispanic, European, African, and Asian cultures), and not just in Western industrialized societies (McCrae & Costa, 1997, 2001; McCrae et al., 2004; Schmidt, Allik, McCrae, & Benet-Martinez, 2007). In addition, research has found that the Big Five factors have about a 50% heritability rate across several cultures (Bouchard & McGue, 2003; Jang et al; 2006; Loehlin, 1992; Loehlin, McCrae, Costa, & John, 1998), indicating a strong genetic basis. The Big Five factors also seem consistent from about age 30 to late adulthood (Costa & McCrae, 1988) and are generally consistent across situations (Buss, 2001). Given their general consistency over time and situations, some psychologists think that each trait must be associated with a specific pattern of brain behavior or structure (e.g., Canli, 2006). In fact, DeYoung et al. (2010) have proposed a theory of the biological bases of the Big Five personality factors and found supporting evidence using magnetic resonance imaging for four of the five factors.

Given the importance of the Big Five factors, Robert McCrae and Paul Costa have developed personality inventory assessment instruments, the NEO-PI-R, the NEO-PI-3, and the NEO-FFI, which provide personality profiles of these five factors (Costa & McCrae, 1985, 1992, 2008; McCrae, Costa, & Martin, 2005). Rammstedt and John (2007) have developed a 10-item, 1-minute-or-less measure of the NEO-PI-R that retains significant levels of reliability and validity. An adapted version of their test is given in Table 8.4. Go ahead and take the test. It will help you to get a better understanding of not only the Big Five personality traits but also personality inventories in general and how they differ from the better-known projective personality tests. After taking the test and before going on to read about personality assessment, review Table 8.5, which summarizes all of the theoretical approaches to personality that we have discussed. Make sure that you know and understand the various theorists and their key concepts.

359

360

Person–situation controversy. In 1968, personality theorist Walter Mischel argued that situations and not personality traits determine a person’s behavior. He pointed out that the correlation between a person’s assessed traits and their behavior in personality laboratory studies was very weak and that a person’s behavior was situation-dependent—a person’s behavior varies in accordance with the situation and will vary across situations. Mischel’s work led to what is known as the person–situation controversy in personality theory. Do traits or situations best predict behavior? Seymour Epstein (1979) answered Mischel’s criticism by pointing out that Mischel’s conclusions were based on looking at the relationship between traits and small slices of a person’s behavior in laboratory research studies. Epstein argued that accurate assessment of the relationship between a person’s traits and behavior requires assessment of a person’s behavior over time and many different situations and not just single instances of behavior; a person’s average behavior across time and situations will reflect his traits. Epstein argued that traits are general behavioral tendencies and not invariable predictors of behavior. Subsequent research has supported Epstein’s argument (Funder, 2001). Research has also found that traits are relatively stable across time from childhood on (Roberts & DelVecchio, 2000). Please note, though, that this does not mean that traits are immutable. Although traits are generally consistent across time, there is evidence of trait change during the period of childhood through young adulthood, especially in the latter stage of development (Caspi, Roberts, & Shiner, 2005), and to a lesser degree from middle adulthood on (Helson, Jones, & Kwan, 2002).

361

Personality tests are a major tool in personality trait research, but the main uses of personality tests are to aid in diagnosing people with problems, in counseling, and in making personnel decisions. There are many types of personality tests, but we will focus on two, personality inventories and projective tests. Most of the other types of personality tests have more limited use. For example, Carl Rogers and humanistic psychologists have developed questionnaires that only address the self-concept, and Julian Rotter has developed a test that specifically assesses locus of control.

Personality inventories. A personality inventory is designed to measure multiple traits or, in some cases, disorders. It is a series of questions or statements for which the test taker must indicate whether they apply to him or not. The typical response format is True–False or Yes–No. Often a third response category of Can’t Say or Uncertain is included. The assumption is that the person both can and will provide an accurate self-report. These tests include items dealing with specific behaviors, attitudes, interests, and values. We will discuss the MMPI, the Minnesota Multiphasic Personality Inventory (Hathaway & McKinley, 1943), revised to become the MMPI-2 in 1989 (Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989). It is the most used personality test in the world (Weiner & Greene, 2008), and it has been translated into more than 100 languages.

The MMPI-2 uses a True–False–Cannot Say format with 567 simple statements (such as “I like to cook”) and takes about 60 to 90 minutes to complete. Most of the statements are from the original MMPI. Some were reworded to update the item content and to eliminate sexist language, and most of the questions pertaining to religion and sexual practices were eliminated (Ben-Porath & Butcher, 1989). The MMPI was developed to be a measure of abnormal personality, with 10 clinical scales. Many of the test statements have obvious connections to disorders, such as “I believe I am being plotted against,” but many are just simple ordinary statements, such as “I am neither gaining nor losing weight.”

The MMPI test statements were chosen in the same way that Binet and Simon chose test items for their first valid test of intelligence (discussed in Chapter 6). Binet and Simon developed a large pool of possible test items and then chose only those items that differentiated fast and slow learners. Similarly, a large pool of possible test items (simple statements of all types) was developed, and then only those items that were answered differently (by a representative sample of people suffering from a specific disorder versus a group of normal people) were chosen for the test. It didn’t matter what each group responded or what the content of the statement happened to be; what mattered was that the two groups generally responded to the item in opposite ways. Thus, the test developers were only concerned with whether, and not why, the items were answered differently by the groups. This test construction method allowed the test developers to choose items that clearly identified the response patterns of people with 10 different disorders (for example, depression or schizophrenia).

The MMPI and MMPI-2 are scored by a computer, which then generates a profile of the test taker on 10 clinical (disorder) scales. There are also test statements that comprise the basis for three validity scales, which attempt to detect test takers who are trying to cover up problems and fake profiles or who were careless in their responding. For example, there is a lie detection scale that assesses the extent to which the test taker is lying to create a particular test profile. The lie scale is based on responses to a set of test items such as “I get angry sometimes,” to which most people would respond “True.” The profile for a person consistently answering “False” to these questions would be invalidated.

362

Personality inventories like the MMPI are objective tests and are objectively scored by a computer. The computer converts the test taker’s responses into a personality profile for the trait or disorder dimensions that the test is measuring. This is why the MMPI has been so popular in clinical settings for making diagnoses of personality problems. Its test construction method leads to good predictive validity for its clinical scales (Garb, Florio, & Grove, 1998), and its objective scoring procedure leads to reliability in interpretation. Projective personality tests, which we will discuss next, do not have good validity or reliability, but are widely used.

Projective tests. The chances are that when personality tests are mentioned, you think of the Rorschach Inkblot Test, a projective personality test. In contrast with personality inventories, a projective test is a series of ambiguous stimuli, such as inkblots, to which the test taker must respond about his perceptions of the stimuli. An objective response format is not used. Test takers are asked to describe the stimulus or tell a story about it. We will discuss the two most used projective tests, the Rorschach Inkblot Test and the Thematic Apperception Test (TAT).

Projective tests aren’t administered, scored, or interpreted in the same way as the MMPI, MMPI-2, and other objective personality inventories. They are not as objective, especially with respect to interpretation. The most popular and widely used projective test is the Rorschach Inkblot Test (often referred to as The Rorschach), which Swiss psychiatrist Hermann Rorschach included as part of a monograph in 1921 (Rorschach, 1921⁄1942). Actually Rorschach never intended this to be a test of personality. He created it for use as a diagnostic tool for schizophrenia. Sadly, in 1922, the year after his monograph was published, he died from appendicitis; he was only 37 years old. There are only 10 symmetric inkblots used in the test. Five are in black ink, two in black and red ink, and three are multicolored. All have a white background. The test taker is asked what he sees in each inkblot, though the inkblots are ambiguous and have no inherent meaning. The examiner then goes through the cards and asks the test taker to clarify his responses by identifying the various parts of the inkblot that led to the response.

363

The assumption is that the test taker’s responses are projections of his personal conflicts and personality dynamics. This means that the test taker’s responses have to be interpreted, which makes the test scoring subjective and creates great variability among test scorers. In fact, one could view the test scorer’s interpretations as projections themselves. The Rorschach Inkblot Test has been widely used (Watkins, Campbell, Nieberding, & Hallmark, 1995), but many researchers have questioned its scoring system and hence its reliability and validity (e.g., Dawes, 1994; Lilienfeld, Wood, & Garb, 2000), leading to a call for a moratorium on its use (Garb, 1999). However, recently Mihura, Meyer, Dumitrascu, and Bombel (2013) reported their findings for meta-analyses of the validity studies using externally assessed criteria (e.g., psychiatric diagnoses) for the 65 main variables assessed by the Rorschach. The resulting estimated validity coefficients were very supportive for the majority of these variables. The strongest validity was observed for those variables that target cognitive processes, such as perception.

Some leading Rorschach critics promptly responded to Mihura et al.’s meta-analytic findings and validity claims. Wood, Garb, Nezworski, Lilienfeld, & Duke (2015; cf. Mihura, Meyer, Bombel, & Dumitrascu, 2015) concluded that Mihura et al.’s meta-analyses for the cognitive variables (e.g., Conventional Form, which assesses the tendency to perceive the world as others do) were trustworthy and thus lifted their moratorium on the use of these variables in research and applied settings. However, given the results of some new meta-analyses that they conducted, Wood et al. did not lift their moratorium on the use of the noncognitive variables (e.g., Suicide Constellation, which assesses suicide risk) for which Mihura et al. had found varying levels of support. They thought that although there was some support for these variables, it was still insufficient. Hence, the validity of these noncognitive variables remains debatable.

364

In addition, a new Rorschach scoring system, Rorschach Performance Assessment System, R-PAS, has been published (Meyer, Viglione, Mihura, Erard, & Erdberg, 2011). This scoring system also utilizes normative data so the results are reported in a manner similar to using standard deviation scores like those used for intelligence tests that we discussed in Chapter 6. Given that this new assessment system only selected Rorschach variables that have been empirically supported in the research literature, it should help to increase the test’s reliability and validity in the future.

Now let’s consider another popular projective test, the Thematic Apperception Test (TAT), developed by Henry Murray and his colleagues in the 1930s (Morgan & Murray, 1935). The TAT consists of 31 cards (the original version had only 20), 30 with black and white pictures of ambiguous settings and one blank card (the ultimate in ambiguity). Only 10 or so of the 31 cards are used in an individual testing session. The test taker is told that this is a storytelling test and that for each picture he has to make up a story. He is to tell what has happened before, what is happening now, what the people are feeling and thinking, and how things will turn out. In scoring these responses, the scorer looks for recurring themes in the feelings, relationships, and motives of the characters in the created stories. However, it has yet to be demonstrated that the TAT scoring (interpretation) is either reliable or valid (Lilienfeld, Wood, & Garb, 2000). You are probably wondering why these two projective tests have been used so much and are so popular if the TAT has no demonstrated reliability or validity and if these essential test characteristics have only recently been demonstrated for some of the variables assessed by the Rorschach Inkblot Test. Therapists who use them and other projective personality tests tend to believe that such tests allow them to explore aspects of the test taker’s personality that the reliable and valid objective personality tests do not. These therapists tend to be psychoanalysts who believe projections are meaningful; therefore, the tests’ subjective nature is not problematic for them.