PERSONALITY THEORIES : PSYC-322 LEARNING JOURNAL 2

HOW IS PERSONALITY ASSESSED?

137

to generate detailed images of the brain (e.g., DeYoung et al., in press). Conversely, studies concerned with brain activity may use fMRI (e.g., Canli, 2004) or PET (e.g., Johnson et al., 1999), whereas researchers interested in measuring brain processes as they occur (Wacker et al., 2006) may prefer to use EEG. Not all physiological techniques measure brain activity, however. Other measures employed in research include salivary cortisol, which relates to hypothalamic pituitary axis stress response, body temperature, blood pressure, heart rate, skin conductance, and eye-blink startle response.

These techniques can, in addition, be used in two differ- ent experimental designs: recording biological signals under pure resting conditions or in response to tasks or challenges presented in the laboratory. The fi rst approach involves the measurement of base levels of neural activity that may underlie personality differences. For instance, as discussed in the previ- ous chapter, Eysenck’s personality model of Extraversion posits that introverts and extraverts differ in their “natural” levels of cortical arousal. These differences should therefore be detect- able with measures of resting brain blood fl ow using fMRI or electrical activity in hypothesized brain regions using EEG (as was done by Gale, 1983).

The second approach involves the measurement of brain activity during laboratory tasks. For instance, one of Cattell’s measures included in the objective analytic test battery assessed startle response (such as pupil dilation) to a gunshot. Other examples include measuring differences in brain activity while participants are exposed to words with different connotations, such as strong negative (e.g., vomit, rape, murder, etc.) ver- sus positive (e.g., win, happy, love, party, etc.). Indeed, much recent work has employed this latter approach, looking at vari- ous indexes, such as the magnitude, timing, and topography of changes in brain activity of participants, while they engage in experimental tasks.

So, what evidence is there for the reliability and validity of psychophysiological measures of personality? First, research has

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 137 8/24/2012 6:36:31 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

138

shown that signifi cant and reliable correlations exist between physiological measures and personality traits measured by self- reports. For instance, people higher in Neuroticism have been shown to have larger startle modulations to fearful scenes than emotionally stable participants (Caseras et al., 2006). These fi ndings have also been confi rmed with fMRI responses (e.g., Canli et al., 2001). Similarly, extraverts have been found to have faster habituation of the startle refl ex as measured by vari- ous physiological measures (Blumenthal, 2001). Furthermore, extraverts compared to introverts also tend to show higher reactivity in the brain when presented with stimuli associated with rewards (Cohen, 2005). A number of additional studies such as these show that there are indeed signifi cant correla- tions between physiological measures of brain activity and self- reported personality traits.

On the other hand, the magnitude of these correlations is not always very strong. Thus, a large part of the variance between physiological measures and self-reports is not shared, meaning that the different measuring paradigms will tap into substan- tially different constructs. A question one might ask, therefore, is which method of assessment better captures the personality constructs in question. To address this, we need to fi rst examine the reliability and validity indices of physiological measures. In terms of reliability, the results of research on psychophysiologi- cal measures have been mixed. For some measures, test–retest reliabilities have been found to be high (e.g., Smit, Posthuma, Boomsma, & De Geus, 2007). Other studies, however, have shown poor test–retest correlations of physiological responses (e.g., Anokhin et al., 2007). These discrepancies are often due to the difference in methodology and the question a researcher is trying to address in the study. Regardless, it is clear that the overall reliability of psychophysiological measures remains to be addressed.

In regard to the validity of psychophysiological measures, the literature is much less developed. Indeed, surprisingly little is known about the predictive validity of these measures. While

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 138 8/24/2012 6:36:31 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

139

some studies have provided preliminary validity evidence in regard to psychopathology, very few have actually examined the ability of psychophysiological measures to predict more common behavioral outcomes, such as career or relationship success. Thus, while there is clearly an increased enthusiasm among researchers about the use of psychophysiological mea- sures in personality assessment, a lot of work remains to be done before any fi rm conclusions can be drawn in regard to their usefulness. A fi nal point, which physiological research will need to address, is whether there actually is a causal rela- tionship between brain areas detected by these measures and trait manifestations. In essence, there is nothing to say that a brain area that is activated during a particular task is not merely a “by-product,” or “side effect,” occurring concurrently while the actual causal, yet distinct, brain areas are doing all the work.

Taken together, while recent years have seen a renaissance of objective measures of personality, questions still remain in regard to their reliability, validity (usefulness), and, on some occasions, theoretical rationale. Nevertheless, with the rate of technological improvements in software and computer devel- opments, objective personality measurement is certainly not an unimaginable ordeal. Furthermore, given the scientifi c aim of objectivity in measurement, the quest for objective assessment of personality traits is likely to continue. It remains to be seen, however, when and whether these measures will be adequate alternatives to more established methods of assessment (e.g., self-reports). Until reliability and validity analysis of objective measures are established, they can, at best, be seen as comple- mentary to these methods.

Self-Reports

We have alluded to self-report measurement of personality throughout this chapter. We noted that self-reports are the most commonly used methods for assessing personality today.

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 139 8/24/2012 6:36:31 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

140

The interesting thing about self-reports, and particularly per- sonality questionnaires, is that most people seem to fi nd them poor alternatives for assessing personality. They do not have an intuitive appeal. Questionnaires seem easy to fake (and it is not diffi cult to think of situations in which people would want to fake). Furthermore, even when people don’t want to fake, one could question their self-knowledge. You would probably not fi nd it hard to think of someone you know who thinks they are someone they are not.

Despite these apparent limitations, personality psycholo- gists extensively employ these tests. Clearly, then, they must have merit for assessing personality; or else psychologists must be doing something wrong. We will evaluate the merit of person- ality inventories below. We will start by describing the various features and types of self-report inventories before critically eval- uating their usefulness, or accuracy, for assessing personality.

The majority of personality psychologists see personal- ity scales as scientifi c and technical measuring instruments. A main reason for this is that these scales can be used to gather structured information from people concerning personality and other psychological characteristics. That is, the key feature of self-report inventories is that they are standardized. They con- sist of a standard list of questions or statements, a fi xed set of response options, and standardized scoring systems. Everyone taking the particular personality questionnaire in question gets the same questions, the same number of questions, and the same response options. Everyone’s score is calculated through the same principles, or formula, often called the “scoring key.”

Today, there are a vast number of personality inventories available, both within academia and outside it. Although most share some common features, they can generally be differenti- ated in a few basic ways. First, personality inventories may dif- fer in the number of attributes they measure; some assess single characteristics, others assess multiple attributes. For instance, a measure such as the Core-Self Evaluations (Judge & Bono, 2001) measures a person’s self-perceptions on a single scale. On the

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 140 8/24/2012 6:36:31 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

141

other hand, the NEO-PI-R (Costa & McCrae, 1990) measures fi ve different traits.

A second way in which personality inventories differ is in their response format. Some inventories include a Yes/No response format. For instance, the item “Must you be in plenty of time if you have to catch a train?” (Eysenck & Eysenck, 1975) can only be answered as either a “yes” or a “no.” Others include True/False (e.g., “I regularly feel sick before exams”; MMPI; Hathaway & McKinley, 1951), Like/Dislike, and Forced Choice. The most common response format, however, is a rating scale, which may differ on a continuum from, say, “strongly agree” to “strongly disagree.” Thus, the measure of entrepreneurial tendencies and abilities (Ahmetoglu, Leutner, & Chamorro- Premuzic, 2011) item “I am quick to spot profi table opportuni- ties” may be answered as “strongly agree,” “agree,” “uncertain,” “disagree,” or “strongly disagree.”

In addition to differences in number of attributes measured and type of response format, personality inventories can dif- fer in the approach taken to their development. Most common test construction involves both theoretical considerations and empirical analysis. Sometimes, however, they are simply based on the validity of items, without any theoretical conceptions. For instance, items in the MMPI were chosen not because their content seemed relevant to the trait in question, but because the item could discriminate between different psychiatric popula- tions. That is, an item was kept in the MMPI if it could differen- tiate between people with depression and people with anxiety (and eliminated if it couldn’t). Indeed, hundreds of such scales have been developed using this approach (Kline, 2000).

Most commonly, researchers fi rst have a theoretical concep- tion of the trait(s) in question before constructing a personal- ity inventory. For instance, a researcher in leadership may have a theoretical conception of charismatic leadership—what this construct constitutes, and what sort of statements a person has to agree with to be considered a charismatic leader. Based on this concept, the researcher then usually assembles (writes) a

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 141 8/24/2012 6:36:31 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

142

large number of items for “piloting.” The items are thus admin- istered to participants (usually in the form of questionnaires) and analyzed for their effi cacy.

Effi cacy of items can be judged in several ways: First, they need to be able to discriminate between people’s responses. Is there enough variability between responses or does pretty much everyone agree (or disagree) with this item? For instance, 99% of people will agree with the item “Do you have a good sense of humor?” (Hogan, 2007). Conversely, very few would disagree with the item “I prefer to have autonomy in my life.” Such items are pretty much useless as they give no information about indi- vidual differences.

Second, individual items need to correlate well with the total score. An item that does not correlate with the total score usually measures a different variable. For instance, in a test of say, self- confi dence, each item should correlate with the person’s overall level (or score) of confi dence. If it doesn’t, it probably doesn’t measure self-confi dence. This relates, as mentioned above, to the internal consistency of a scale. In addition, an item should not measure more than one variable, or the researcher could have a mixed test. An item that correlates with, for example, both self- confi dence and achievement motivation (two related but dis- tinct concepts), would render the test less reliable.

If items are discriminating and internally consistent (i.e., they all correlate well with the total score), and are univariate (measure only one variable), usually the test will be reliable. Of course, as mentioned earlier in addition to reliability, the validity of the test needs to be established. In this respect, the construct validity of the questionnaire should be demonstrated by meth- ods already mentioned above (i.e., predictive validity, criterion validity, and concurrent and discriminant validity).

What Makes Personality Inventories So Popular?

One reason for the widespread use of personality inventories is undoubtedly the “convenience factor.” Questionnaires are

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 142 8/24/2012 6:36:31 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

143

easy to administer and score. Data can be collected from a large number of people at the same time, even without an examiner present. Today, web-based surveys make it easy to reach large, diverse, and arguably more representative, samples, quickly and without much effort. Researchers can, in addition, give instant feedback to test-takers about their personality, which is often an effective way of enticing participants. Thus, compared to other assessment methods reviewed above, which can be cumbersome (and very costly), personality measures are highly practical.

Nevertheless, ease of data collection is hardly the main reason for the popularity of this method. Ultimately, the key advantage of personality inventories is that they are highly suit- able for reliability and validity analyses. That is, simply by fol- lowing the principles outlined above, of the test-construction process, researchers are able to increase a test’s reliability, deter- mine its validity, and establish its norms. This can be done with self-reports far more readily than with other available methods. Good personality scales often take months, if not years, to devise, pilot, and standardize. During this process, revisions are made in order to increase reliability and validity. Items are eliminated until a test enables researchers to compare individuals fairly and with confi dence. Large samples can be, and usually are, tested, which enables standard scores, means, and variability around the means to be established. Thus, self-reports are considered scientifi c and technical measuring instruments because they go through a rigorous scientifi c construction process. The rationale is simple. If criteria in each step of the process are scientifi cally validated, eventually one will have a scientifi cally validated tool for measuring personality. Thus, self-report questionnaires per se may not be good tests. Indeed, there are a vast number of self- report tests that are very poor indicators of personality traits. This simply means, however, that they have not met the required standards during test construction. The bottom line is that good self-report inventories are good because they have been and can readily be made so. Compared to other methods, this process is relatively easy.

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 143 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

144

Why Are They So Unpopular?

Despite benefi ts of self-reports, however, there is a prevailing skepticism to the use of such tests, especially in applied settings (but also in general). The reason is not very complicated and has been mentioned before: Self-reports are vulnerable to what are known as response biases. Several forms of response biases exist. Some are much more serious than others.

For instance, one common response bias is known as “yea- saying”—some people tend to agree with most responses, regard- less of their content. One reason for this can be that the items are ambiguous. Another is that people are simply acquiescent by nature. Nevertheless, there are also people who tend to dis- agree with most items; this is known as “nay-saying.” Another response bias is the tendency to endorse extremes. Some indi- viduals tend to either strongly disagree or strongly agree with most items. Conversely, some individuals have a greater ten- dency to avoid extremes. They seem to be uncertain with regard to most items. Yet, others simply respond in a random man- ner. They simply don’t care, or are otherwise not motivated to respond in a true fashion.

These are systemic biases. They can make interpreta- tions about people’s “true personalities” diffi cult and weaken test validity. However, test-developers have usually been able to deal with these response biases by taking steps in the test construction. For instance, “yea” and “nay-saying” can be nul- lifi ed by having both positive and negative statements about the attitude in question (also known as “reverse scoring”). For instance, “I get stressed out easily” and “I am relaxed most of the time” are two items from the International Personality Item Pool (IPIP; Goldberg, 1990), which require the participant to disagree and agree, respectively, to be responding in an emo- tionally stable direction. Avoiding ambiguous items is another way of reducing “yea” and “nay-saying.” Although it is diffi cult to avoid people’s tendencies to endorse or avoid extremes, care- ful item construction can minimize these effects. Similarly,

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 144 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

145

while carelessness can be diffi cult to eliminate, it can easily be detected. If internal consistency of a scale has previously been established, participants who respond inconsistently or in con- tradiction to themselves on items (thus lowering internal con- sistency) can be eliminated from the analysis.

Response biases, such as the ones outlined above, can gener- ally be cancelled out by test-developers. However, as mentioned before, there is one bias that seems particularly serious when it comes to evaluating the robustness and actual value of person- ality questionnaires: faking. Faking is the tendency of people to dishonestly present themselves in an overly favorable (or overly unfavorable) light. It is no doubt the most obvious criticism of self-report inventories. The criticism is there because it is usu- ally easy to “see the point” of items in questionnaires and to identify good from bad responses. So, can self-reports be valid given that people can distort their results? And, how much can we trust self-report scores when people’s necks are on the line, for instance, in job applications or clinical interviews?

The question of faking has been investigated by person- ality psychologists for over 60 years. We will review in detail the literature on faking in Chapter 5; however, suffi ce to say at this point that although faking is a valid issue in personal- ity research, it is not a decisive problem. The data from a vast amount of research show that people do not fake as much as one would imagine, and that being able to fake in turn may be a desirable personality trait in itself. Nevertheless, there is still an ongoing debate between advocates and critics of self-reports about the issue of faking in personality assessment, and we will critically evaluate both sides of the argument in Chapter 5.

Other Ratings

It has been suggested that one of the ways of overcoming fak- ing is to obtain ratings from others. As mentioned above, these ratings may be done by peers, bosses, subordinates, and so on. Thus, if there is high agreement between self and other reports,

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 145 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

146

one may conclude that the person isn’t lying. Other reports are popular methods for assessing personality. Indeed there are those who argue that other reports are even more valuable than self-reports (e.g., Hogan, 2007). There are two reasons why this may be the case. First, others may actually know you better than you know yourself. Secondly, what you think about your- self (your identity) is arguably less important than what others think of you (your reputation). You may “deep inside” believe that you are a great leader or a very warm person but unless others (e.g., your peers or the person you fancy) agree, what you think will matter very little for outcomes.

An important question, of course, concerns the conver- gence between other ratings and self-ratings. Do other ratings converge with self-ratings? Research has shown a remark- able consistency in identifying the structure of personality, with the Big Five dimensions of personality emerging also with other ratings (Digman, 1990). Studies have also found clear convergence when comparing self-reports with ratings of knowledgeable others, suggesting that people are not com- pletely oblivious to others’ views and are honest. Nevertheless, the effect size of these correlations is often moderate, which means that signifi cant differences between self and other rat- ings do exist. This discrepancy could be interpreted in several ways. The most obvious is that while people are not entirely dishonest, they do self-enhance to a degree. Differences may also refl ect some limitations in introspection. However, increasingly, researchers suggest that ratings from different sources are complementary rather than indicative of measure- ment error. That is, reliable differences in ratings may simply refl ect distinct information that each source receives by the ratee. Each source, in turn, would provide unique informa- tion through their ratings. As a result, a more complete pic- ture of the person can be obtained. Indeed, research shows that other reports and self-reports explain unique variance in the prediction of relevant criteria (Chamorro-Premuzic & Furnham, 2010a, 2010b).

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 146 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

147

Whether other-reports can remedy faking is diffi cult to judge. However, it is clear that obtaining other reports, in addi- tion to self-reports, may provide researchers with more informa- tion than each method may by itself. Of course, one limitation with other reports is that they are often not easy to collect. Moreover, it has been pointed out that the accuracy of other reports seems to depend on the level of familiarity between the rater and ratee. The level of familiarity may, in addition, itself be related to distortion in ratings.

Interview Methods

Another self-report method is the interview. Interviews are one of the oldest methods for assessing the personality of other peo- ple. In applied settings, it is still one of the most widely used (Chamorro-Premuzic & Furnham, 2010a, 2010b). In fact, it is estimated that 90% of employment selection decisions involve interviews (Cook, 2004). Interviews are also highly popular in clinical settings. Advocates of the interview argue that this method provides an opportunity like no other to obtain rich and detailed information about a person. First, interviews allow people to express themselves freely and answer a ques- tion in an elaborate way. This is not possible with self-report questionnaires, where responses are fi xed. Second, interviews provide information not only about the content of the response, but also allow the interviewer to assess nonverbal behavior, appearance, mannerisms, and so on, which are impossible to capture in questionnaires. Finally, this method allows greater scope for maneuvering; for instance, an interviewer can, at will, explore further a particular response, focus on an issue in greater depth, or prompt the interviewee to provide additional information about a particular event.

Of course, how much fl exibility there is in an interview may vary quite substantially. It has long been common practice to differentiate between what has been called structured and unstructured interviews (though, strictly speaking, they are

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 147 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

148

really on a continuum from completely unstructured to rigidly and infl exibly structured). The ultimately unstructured inter- view is a little like an informal discussion, where interviewers ask whatever questions come to mind and interviewees follow up with answers in an intuitive and whimsical way. Crucially, questions are open-ended and attempt to avoid “leading” the interviewee’s answers into any specifi c direction. Given its theoretical orientation, psychoanalysis often takes the form of unstructured interviews. This allows the interviewer to dig deeper into the recesses of the patient’s mind to uncover uncon- scious confl icts, motives, and drives.

The structured interview, on the other hand, is preplanned to ensure every candidate receives exactly the same questions, in the same order, and at the same pace. Structured interviews also employ rating scales, use checklists for judgment, allow for few or no follow-up questions (to limit interviewees’ response time and standardize it), and leave little autonomy for the inter- viewer. In that sense, totally structured interviews are little more than standardized personality questionnaires. Researchers tend to favor structured interviews, in order to obtain more standard- ized and less biased information about people. It is probably true to say, however, that most interviews are unstructured or semi- structured (Chamorro-Premuzic & Furnham, 2010a, 2010b).

Interviews undoubtedly provide a richer source of infor- mation than most other assessment methods. This could explain its popularity and widespread use in applied settings. Indeed, research has shown, for instance, that people applying for jobs not only approve of interviews, but often expect them (Chamorro-Premuzic & Furnham, 2010a, 2010b). They may even be surprised and disappointed if they are not asked to give them. Given that many people see interviews as fair and good methods of assessment, the question of how reliable and accurate inter- views are in assessing personality is of utmost importance.

A large amount of research has been carried out to address this question. Results in the fi eld have led to a num- ber of conclusions. The fi rst is that interviews, and particularly

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 148 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

149

unstructured ones, show poor reliability. For instance, in a meta-analysis, Conway et al. (1995) found reliabilities of .53 when observers watched different interviews of the same can- didate. This is substantially below the accepted standard of .7 for test reliability. Furthermore, different interviewers watching the same interview frequently show low levels of agreement. That is, interviews often also have poor interrater reliability. Given the fl exibility of interviews in general, this is perhaps to be expected. There are several sources for low reliability. One is individual differences between interviewers. Interviewers ask different questions, record and weigh answers differently, and may have radically different understandings of the whole pur- pose of the interview. Furthermore, their personal backgrounds and life experiences affect the questions asked, how answers are interpreted, and the format or process of the interview. In addition, interviewees may react and respond differently to dif- ferent interviewers, who differ in appearance, age, personality, interpersonal style, and so on.

Apart from individual differences between interviewers, sources of unreliability may derive from common cognitive biases. For instance, studies looking at the process of selec- tion interviews have shown all too often how prior knowledge about the interviewee (based on the application form or cur- riculum vitae [CV] of the candidate) can strongly infl uence the interviewer, before the interview even occurs (Harris, 1989). Research also shows that interviewers make up their minds too quickly based on fi rst impressions (superfi cial data, Cook, 2004). Equally, they overweigh or overemphasize negative information or bias information not in line with the algorithm. A fi nal source of unreliability is that interviews, which are often face to face, make people even more concerned about managing impressions. Thus, socially desirable responses (i.e., faking) may be even more likely in interviews than in self-report measures.

Given the numerous sources of error in interviews, it is unsurprising that research fi nds mixed and often unsatisfactory results also with regard to the validity of interviews. A suffi cient

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 149 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

150

number of meta-analyses have been done to examine the valid- ity of interviews. Studies looking at vocational outcomes, for instance, show that interviews often provide very little unique information about a candidate and show little incremental validity over established personality tests in the prediction of future job performance (Schmidt & Hunter, 1998). Similar con- clusions are also reached in clinical settings.

Despite their limitations, interviews can still provide valu- able information. This is particularly true in clinical settings where patients may have little awareness of their symptoms and are, therefore, not able to report these in questionnaires. Furthermore, in organizational settings, interviews can be designed to be more reliable and accurate. Most reviewers have seen that the single, simplest way to improve reliability is to introduce consistency and structure to the interview. Studies also show that it is possible to increase an interviewer’s reli- ability by different, but important, steps, including doing a job analysis, training the interviewer, having structured interviews, and having behaviorally based and anchored rating scales (Chamorro-Premuzic & Furnham, 2010a, 2010b). If used in conjunction with other methods, interviews can indeed work well. However, it is probably fair to say that laypeople and prac- titioners have overrated the usefulness of interviews compared to other methods for assessing personality, such as the ques- tionnaire methods.

Other Methods for Assessing Personality

Although we have reviewed the most commonly used methods for assessing personality, several other methods exist. These include the diary methods (where people keep a diary of their everyday behavior over a period of time), case studies (where a psychologist gets in-depth information about a person through lengthy conversations and tests), life history records (objective information obtained through records and reports about a per- son’s past behaviors), behavioral observations (where people

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 150 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

151

are observed in experimental settings or in their everyday life), narratives, and more. Although these methods are used in research, they are often done so for very specifi c purposes, by researchers with particular theoretical orientation and in a minority of studies. A major reason for this is that they are often not very practical, can be cumbersome to employ, and, in gen- eral, show lower levels of reliability and validity than standard- ized and empirically validated self-report questionnaires.

What About Astrology, Graphology, and So On?

Before we conclude this chapter, we want to say a few words about some assessment methods that seem to have won the hearts and minds of the general population but are rarely subject to scien- tifi c discussions. The most popular of these no doubt is astrology. Astrology is any of several traditions or systems in which knowl- edge of the apparent positions of celestial bodies is held to be useful in understanding, interpreting, and organizing knowledge about reality and human existence on earth. It is probably useful to talk about methods such as astrology, simply because people often use them to make inferences about other people’s character and personality. Indeed, chances are that you will occasionally be asked about your zodiac sign by someone you have recently met at a party or a social event (we can certainly attest to this). It seems that people are prepared to make instant judgments about your personality based on the month you were born, or zodiac sign, sometimes regardless of your actual behavior.

Making judgments based on zodiac signs may often be a fun exercise. This is also undoubtedly the most handy person- ality assessment tool out there; it takes virtually no time and there is no test that needs completing. However, there is evi- dence to show that people (i.e., aside from astrologers) actually believe in astrology (Hamilton, 1995) and do accept the person- ality descriptions it offers (Glick, Gottesman, & Jolton, 1989). For example, Hamilton (1995) found that undergraduates pre- sented with one-paragraph descriptions of the characteristics

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 151 8/24/2012 6:36:32 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

152

of their own astrological sun sign and an alternative sun sign, chose their own sun sign paragraph as a better representation of their personality than the alternative sun sign description. Van Rooij (1999) found that participants presented with indi- vidual trait words associated with the personality descriptions of each of the 12 sun signs chose the traits of their own sun sign as more personally descriptive than the traits associated with the other 11 signs. Thus, more than being mere amuse- ment tools, it is likely that astrological readings have real-life implications. This could involve simple (but important) things such as people deciding their compatibility as partners based on their zodiac signs. However, it could also involve important decisions made in applied settings, such as an employer’s evalu- ation of an employee’s personality or a clinician’s evaluation of a patient’s symptoms. The signifi cant implication is that, even if the inferences made are invalid, they may bias judgments or have self-fulfi lling effects. It is of practical interest therefore to examine the empirical grounds of astrological readings.

There have been several scientifi c investigations of the reliability and validity of astrology as a tool to assess per- sonality. For instance, Eysenck and his colleagues examined relationships between astrological and “Giant Three” person- ality factors of Extraversion, Neuroticism, and Psychoticism (Gauquelin, Gauquelin, & Eysenck, 1979; Mayo, White, & Eysenck, 1978). A couple of studies have even reported associ- ations between these established traits and astrological factors (e.g., Gauquelin, Gauquelin, & Eysenck, 1979). Indeed, more recently, Sachs (1999) attempted to put astrology on a scien- tifi c footing by using statistical methods to explore associa- tions between the zodiac and human behavior (in particular, criminal behavior).

While some studies have found signifi cant correlations, however, many other results have failed to support the role of astrology in personality (van Rooij, 1994). For instance, Clarke, Gabriels, and Barnes (1996) explored the effect of positions of the sun, moon, and planets in the zodiac at the moment of birth

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 152 8/24/2012 6:36:33 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

153

and found no evidence that tendencies toward Extraversion and emotionality are explained by such signs. Even when whole charts are used in “matching tests,” astrology seems unreliable and invalid. Furthermore, astrologers seem no better at iden- tifying different personality profi les (e.g., highly extraverted versus highly introverted people) than random guessing (Dean, 1987). They also fail comprehensive tests when they themselves provide the required information (Nanninga, 1996). In a com- prehensive review, Kelly (1997) concluded that:

The majority of empirical studies undertaken to test astrological tenets did not confi rm astrological claims, and “The few studies that are positive need additional clarifi cation.” (p. 1231)

Various authors have similarly dismantled Sachs’ (1999) “scientifi c” claims, leading at least one group of authors (von Eye, Lösel, & Mayzer, 2003, p. 89) to claim that:

If there is a scientifi c basis to astrology, this basis remains to 1. be shown, and; If there exists a link between the signs of the zodiac and 2. human behavior, this link remains to be shown too.

Finally, it is worth noting that very few, if any, psychomet- ric studies on the structure of personality have found 12 fac- tors of personality indicated by the zodiac signs. Most studies converge (as discussed in Chapter 1) on fi ve or three factors. In addition, astrology is proposing a typological perspective of personality, which contrasts with the dispositional view. According to the dispositional theory of personality, people do not differ in terms of which category they fall in (e.g., extravert or introvert) or which trait they possess (e.g., introversion or Extraversion), but rather in terms of how much they possess of each trait (e.g., how extraverted; Chamorro-Premuzic, 2011). In this (nomothetic) paradigm, which has represented the state- of-the-art approach to the study of individual differences for

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 153 8/24/2012 6:36:33 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

154

the past 50 years, traits cannot form psychological categories or types, such as those proposed in astrology.

It is worth noting that several methods, in addition to astrology, such as graphology, phrenology, physiognomy, and, certainly, projective tests, are still in use in applied set- tings, despite the lack of any scientifi c merit for their effective- ness. This is perhaps a testament to the naivety, ignorance, and perhaps desperation of laypeople and many practitioners. However, it may simply be because people are more accepting of information that is ambiguous, general, and positive (known as the Barnum effect). These features are inherent in many of the tests mentioned. What is clear is that while astrology and other methods have stood the test of time, they have consis- tently been shown to be both unreliable and invalid methods for assessing personality.

CONCLUSIONS—WHICH IS BEST?

HOW TO CHOOSE?

We started by asking: “Can personality be assessed?” The answer is “yes.” The question then becomes how well? This, in essence, asks how reliably and how accurately can we measure person- ality. Clearly, there is more than one method one can employ. Some seem more intuitively appealing and compelling than others. It is clear that many people want objectivity. However, objective tests with acceptable reliability and validity levels are yet to be found. Projective tests may be compelling and inter- esting, but they are virtually useless. Even the interview, which is perhaps the most common method for assessing personality, is shown by research to have poor reliability and relatively low validity. Self-report inventories have little intuitive appeal. They seem easy to fake and work out. They also seem subjective in nature. Yet, they are also the most amendable to improvements

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 154 8/24/2012 6:36:33 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

HOW IS PERSONALITY ASSESSED?

155

in reliability and validity. Thus, psychometric tests are suitable for meeting the demands of good scientifi c work. They are con- cerned with sound quantifi cation and place great emphasis on sampling, research design, and statistical analysis. For instance, tests such as the NEO-PI-R (McCrae & Costa, 1999) or the EPQ (Eysenck & Eysenck, 1975) have been validated extensively throughout the years. They show high internal consistency, test– retest reliability, cross-cultural consistency, and high construct validity. They are, in addition, able to predict a large amount of real-life outcomes better than any other available methods. We will review these in the next chapter; however, suffi ce it to say that the rationale behind the use of self-reports is far beyond the “convenience” factor. Good self-report instruments have been found to be highly reliable and valid for the assessment of per- sonality. Thus, they are the most accurate instruments currently possessed by researchers for measuring personality, and they are scientifi cally proven to be fi t for the job.

Despite the validity of self-reports, they are by no means the only methods worth using. Several other methods have shown to have acceptable reliabilities and incremental validity over self-reports. Thus, it is advisable that researchers try to use multiple methods. Further advances in measurement should also soon make it possible to assess personality objectively. This is not necessarily an either-or scenario. We may well need each type of measure, self-report, and objective to assess personality accurately. What is clear, however, is that currently personality psychologists are able to measure personality and are able to do so fairly accurately. As Meyer et al. (2001) conclude:

Data from more than 125 meta-analyses on test validity and 800 samples examining multimethod assessment suggest four general conclusions: (a) Psychological test validity is strong and compel- ling, (b) psychological test validity is comparable to medical test validity, (c) distinct assessment methods provide unique sources of information, and (d) clinicians who rely exclusively on inter- views are prone to incomplete understandings. (p. 128)

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 155 8/24/2012 6:36:33 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .

CHAPTER 3

156

This statement is a testimony to the incredible progress that has been made in the measurement and assessment of person- ality. The sophistication of psychological measurement will no doubt continue to improve and contribute incrementally to our understanding of personality. Their informed use will hope- fully also increase in applied settings to make more informed and rigorous decisions about people and their likely behavior.

Ahmetoglu_PTR_Ch-03_24-08-12_101-156.indd 156 8/24/2012 6:36:33 PM

Chamorro-Premuzic, T., & Ahmetoglu, G. (2012). Personality 101. Retrieved from http://ebookcentral.proquest.com Created from claflin on 2018-05-29 12:03:40.

C op

yr ig

ht ©

2 01

2. S

pr in

ge r P

ub lis

hi ng

C om

pa ny

. A ll

rig ht

s re

se rv

ed .