The ELEVATOR SPEECH
There are two problems with current personality tests. One is that psychologists don’t agree on the number or types of traits that make up someone’s personality. The second is that, for all their importance, personality traits, as they’re currently measured, are generally not great predictors of people’s behavior. For example, the best trait at predicting behavior, Conscientiousness, typically only predicts about 10% of someone’s job performance.
Our test tries to address both problems. First, we took a practical, “shotgun,” approach to choosing traits to ask about. Instead of limiting ourselves to a particular theory of personality, we included as many traits as we could practically fit into a personality test. And we chose those traits based on our training and experience in work psychology. What traits would be most important on the job or in the classroom?
Second, rather than the broad generic questions you often get on personality tests, we asked the respondents to put themselves in particular situations and tell us how they would react or what they would think IN THOSE PARTICULAR SITUATIONS.
In addition to addressing those problems, we think this will have two additional benefits:
Asking questions about thoughts and actions in a specific situation can likely give answers that better characterize what a person will typically say or think in that situation—so more connection between the test answers and the real world.
People can answer a specific question more accurately than a general question. By making our personality test questions about their thoughts and actions in a specific setting, we can likely improve overall accuracy.
—AN IMPORTANT REQUEST:
If any of the above piqued your interest, then PLEASE keep reading. Elevator speeches aren’t exactly lies, but they’re not the whole truth, either. All elevator speeches (or “executive summaries,” if you prefer) are dumbed-down versions of what’s really going on. Actually, executives should always remember that (but often they don’t seem to). So again, PLEASE read further when you have a minute…
——But first, a really important WARNING:
Be VERY careful about words. In psychology a lot of terms mean something VERY different to psychologists than those same words mean in everyday speech. For example, commonly used words like “personality” and “intelligence” mean something very different to psychologists. To name a major example, both of those words, in psychology, are understood to be generic, umbrella terms for a host of different cognitive structures and processes. For psychologists, there really isn’t a singular thing called “personality” or a singular thing called “intelligence” (and that even includes those psychologists who lean toward a general factor in intelligence). It’s understood that those simple terms stand for more complex, multi-variate phenomena.
Some “psychology” terms commonly used don’t even refer to the thing that people think they refer to (shades of that famous line from the Princess Bride!). “IQ” is commonly used to refer to intelligence itself but it’s actually a specific way of scaling intelligence scores (specifically, Intelligence Quotient = 100 + ((a score on an intelligence test – the mean of that test)/the standard deviation of that test)*15). And there are lots of other ways to score “intelligence tests.” So, it helps a lot if you put aside any pre-conceived notions about what “personality” is.
Some quick side notes:
All disciplines do this, not just psychology. Ask a physicist what they mean by terms like “charm,” “color,” “information,” or “observation” (to name a few) and you’ll get an answer very different from what non-physicists mean by those terms.
Those last two benefits listed above are benefits in accuracy and consistency, respectively; in “validity” and “reliability” in technical terms. Since reliability puts an upper bound on validity (the equation is: observed validity/square root of reliability), a psychometrician would put the second benefit first, but in an elevator speech I thought talking about accuracy first would be better.
The 10% number for Conscientiousness is the usual reported validity. I believe I’ve seen it as high as 16% but that’s only when job performance was measured very narrowly. For most real-world jobs Conscientiousness is important, but it’s far from the only important trait. So, the 10% is just about right. I should point out that Jo disagrees with me about that—she believes it’s higher. But it’s important to remember that Conscientiousness is by far the BEST of the traits for prediction. The other ones you find on most personality tests absolutely suck at real-world prediction. (For some reason the “intuitive” characteristic comes to mind…)
What IS personality? And do you really know what it is?
What’s personality? Well first let’s talk a little about what it is not. Most of us think of personality as a permanent, built-in part of our brain that determines how we act. Like a lot of misconceptions, there are some grains of truth in that definition. But as some famous architect apparently said, “God is in the details” (I find that version more optimistic than the widely quoted “The devil is in the details”. Either way…)
Without going into a complete literature review of the decades and decades of research* into personality, let me summarize a couple of the main problems in that definition:
If personality is a set of built-in parts, then how many parts are there and what are they? If they’re built-in, then we should be able to find, count, and describe them, right? And do we all agree on what kinds of human characteristics should even be called “personality”? What’s the difference between “temperament” and “personality”? Or are they just two different words for the same thing? (By the way, the quick and dirty answer is NO);
Are they really permanent? And what does that mean? Are we born with them? Do we develop them in our childhood? (If you’re a Freud fan – which you shouldn’t be—you’ll say yes here). Either way, do they never change in adulthood—do you have the same personality at 20 that you have at 60? How about 40 and 80?;
Does personality really determine how we act? Is our behavior in a given situation driven entirely or even mostly by our “personality”? Isn’t that what personality is supposed to be, a predictor of how we’re going to act?
You can probably guess that the answers to all of those questions is: “Sorry, but it’s not nearly that simple.” And you’d be correct. And one note about that… Personality is part of the human brain-mind (sorry about the hyphenated term but the brain and mind duality is WAY too complicated to get into here; for our purposes just think of the brain as the hardware and the mind as the software—that’s not right either, but close enough). And it’s been said, probably correctly, that the human brain is the most complicated structure in the known universe (look at the number of neurons, the number of neural connections, and the way neurons operate if you don’t believe that). Moreover, I once heard a great talk by Nobel Prize–winner Herbert Simon in which he made a point that even simple brains (if I recall correctly, he used ants as an example) can exhibit complex behaviors if they’re in a complex environment. And humans live in a VERY complex social, cognitive, and physical environment. So DON’T expect personality to be simple. It isn’t.
By the way, the same thing is true of other human “individual differences,” as well. For example, the nature and operation of “intelligence” has many of those same problems as personality. Not to mention there are LOTS of other individual differences that don’t get nearly as much attention—all the way from “temperament” to relatively obscure ones like “field dependence/independence” (which I was told is correlated with how many traffic accidents you have—but that was a very long time ago.) And note that term “individual differences.” When I was in graduate school at the University of Illinois Urbana-Champaign (and that was a very long time ago also) I was in a program called “Social, Organizational, and Individual Differences” or “SOID” for short. That’s how important these human characteristics like personality, intelligence, etc. are; they’re up there with social psychology and industrial-organizational psychology (and are actually a lot older, as a field of inquiry).
But ironically, one of the biggest misconceptions humans have about themselves is how much of their actual behavior is driven by their personality. We like to think that, as unique individuals (which we are), our behavior is driven from the inside. What we think, feel, and decide determines what we do. But to a very large extent it doesn’t. And this misconception shows up in everything from the Barnum Effect to the Fundamental Attribution Error—more on those later… Actually, this is not as bad as it may sound. What it means is that humans are marvelously adaptive to their environments and that includes all the social, cognitive/meaning, and physical parts of that environment. Sometimes people refer to that adaptability as “plasticity” (a term which also has a related but quite different meaning to physiological psychologists). Humans live in all kinds of societies, cultures, climates, you name it (just ask any handy anthropologist for the full list—you’ll be amazed). But DON’T assume that your personality is the primary driver of your behavior. Yeah, I know that’s a funny thing for a purveyor of a personality test to say, but it’s true nonetheless.
Types—Why Personality Doesn’t Come in “Types”
A large part of the personality problem is the constant and persistent public misunderstanding of the trait part of personality. Your personality and the factors that make up your personality are supposed to be personal characteristics, what are referred to as “traits.” But the most widely used and widely publicized personality tests are type theories, not trait theories. But wait, aren’t “traits” and “types” basically the same thing? Emphatically, NO. Here’s why…
Type theories make a basic mistake about human characteristics. Personality is a set of continuous, not categorical, variables. Type theories artificially dichotomize this continuum. Example: Introversion-Extraversion is what statisticians call “normally distributed” (i.e., Gaussian distributed). In other words, most people are in the middle, with a few very introverted and a few very extraverted individuals out at the extremes. One popular personality test artificially divides that range of human characteristics into two boxes: “I” or “E.” But people come in ranges of characteristics, not in boxes (aka “types”). Humans are naturally prone to categorize (things and people). It saves us mental energy and it makes the world appear to be more predictable (which makes us feel more secure). Unfortunately, that is also what drives incorrect categorization—stereotypes, prejudices, biases, etc. Neil deGrasse Tyson has multiple chapters in his book Starry Messenger: Cosmic Perspectives on Civilization, in which he points out the dangers of categorization.
And be wary: While some publishers of type-theory personality tests state that they specifically warn people against categorization, this statement is disingenuous at best: You’re warning people not to use a personality test to stereotype others, then you give them a set of labels to slap on other people (and themselves). You can even buy T-shirts and caps with your type, or your spouse’s type, or a friend’s type (and a cute little cartoon illustrating the type). If that’s not stereotyping, then what is?
Type theories get the number and content of basic distinguishable personality characteristics wrong. For example, a typical type theory may have four general dimensions: Introversion-Extraversion, Intuition-Sensing, Feeling-Thinking, and Perception-Judging. To be fair, test publishers sometimes postulate some interesting interactions among these dimensions, but unfortunately those postulated interactions are not supported by good empirical research in legitimate journals. But decades of independent research into the actual dimensions of personality so far indicate that there are at least five dimensions of personality (the “Big 5”). And those dimensions apparently aren’t nailed down—for example, Conscientiousness may actually consist of two independent dimensions: Orderliness and Dutifulness. Later, we’re going to make the argument that this inability to pin down the actual dimensions of personality is evidence that we’re not looking at personality correctly. More on that later… But even if the Big 5 is exactly right, that list doesn’t really include the dimensions above except for Introversion-Extraversion. The I-S, F-T, and P-J dimensions seem to be a mish-mash of temperament, preferred judgment strategy, and some other cognitive characteristics (but defined very poorly and ambiguously; again, NOT based on the empirical research that has been out there for decades).
Type theories are based on poor theory. For example, some popular type theories are based on Carl Jung’s theory of human mind/personality. Jungian theory is not at all supported by modern psychological research. Jung was essentially a neo-Freudian and Freudian theory has been debunked for almost 50 years now (famous personality researcher Hans Eysenck once said that Freud has been a “wholesale disaster for psychology”). Here’s a glaring example: the Jungian concept of synchronicity—that apparently coincidental events have an underlying meaning… Your Aunt Tilly thinks she’s psychic. She lives in Los Angeles (that could account for why she thinks she’s psychic, but that’s not it…). Last month Aunt Tilly dreamed of earthquakes, and sure enough, the very next day Los Angeles had a noticeable temblor. Now she’s convinced she is psychic. The problem is that Aunt Tilly doesn’t realize that in Los Angeles it’s very likely that tens of thousands of people dream about earthquakes every night. And since LA does have quakes, occasionally the dreams and the quakes coincide. And then thousands of people are convinced that their dreams predicted an earthquake.
Theories that don’t take into account human cognitive processes make for poor measures of personality.
One prominent type theory was studied by the National Research Council’s Committee on Techniques for the Enhancement of Human Performance (in Phase 2: “In the Mind’s Eye: Enhancing Human Performance.” 1991). Note that the National Research Council is part of the National Academy of Sciences—possibly the most prestigious scientific organization in the U.S. and charged with advising the United States government on scientific matters.
The Association for Psychological Science Observer (September 1994) summarized that report: “Committee took strong position against typing people as if they fit into a box and never change, rather than treating them as dynamic, learning people. At best, (type theory) gives a picture of an individual at a given point in time. It should not be used to assemble teams or pick people for jobs” (emphasis mine).
By the way, note the date on this report. As of this writing that was 30 years ago. One philosopher of science (many years ago and certainly pre-social media) estimated that it took 50 years for findings at the leading edge of science to percolate into the consciousness of the general public. Given that only just now are people in the public starting to realize the many issues with the type theories, that philosopher was probably pretty close.
A very real danger with type theories (and to a lesser extent with trait theories) is that they tend to stereotype and oversimplify people and their behavior. There’s a real danger of self-fulfilling prophecies: e.g., “My wife and I are incompatible types so marital counseling won’t do us any good.” These tendencies also explain a lot about why these incorrect theories are so widely accepted (at least by the general public)…
Why “Type” Personality Tests Are So Popular: Shortcuts and Illusions
Why these tests are popular (and how a misperception of personality has persisted for 3000 years)
Type theories of personality are at least 3000 years old (and probably older). The ancient Greeks famously formulated one of the first known theories of personality—the “temperaments.” The Greeks thought that there were four personality types, each caused by an excess of a certain category of bodily fluids. Too much blood, and you’re sanguine—energetic, optimistic, outgoing. Too much phlegm, and you’re calm, cool, and collected (and yes, that’s where “phlegmatic” comes from). The other two types involved black bile and yellow bile. Enough said.
Sheldon’s Body-Type Theory as an Example
In answer to the question, wait a minute, if type theories of personality are so off base, then why have they been around so long (and believed by so many)? Sheldon’s Somatotype theory is a perfect illustration of why. I’ve observed that no theory, no matter how bad, ever really dies (no matter the weight or quantity of evidence against it). I call them “zombie theories.” Personally, I think Freudian theory is a classic example of a zombie theory. But in regard to personality, to this day I still see popular press articles citing Sheldon’s Somatotype/Constitutional theory of personality. In case you’re not familiar with it, in the 1940’s William Sheldon proposed that your personality was based on your body type. He identified three body types (humans love to categorize into 3–5 categories, probably because that’s the capacity of our working memory; it’s likely not a coincidence that the MBTI has four factors, the Big 5 has, well, five, etc. …). According to Sheldon you’re an ectomorph, a mesomorph, or an endomorph. In plain English, you’re thin, medium, or fat (respectively). And according to Sheldon your personality is a function of your body type: thin people are loners, secretive, have little sense of humor, etc.; meso people are (according to Wikipedia) “courageous, assertive, direct, and dominant,” while endomorphs are basically jolly and lazy. Clearly Sheldon must have read Shakespeare… “yon Cassius has a lean and hungry look.” And I’ve never seen a picture of Sheldon but I strongly suspect he saw himself as a mesomorph.
Of course, all of this is complete bunk. But I recently saw it cited in a national magazine—in the context of giving advice to athletic and personal trainers about their clients. Good luck with that…
Why do humans love categories (types) when most human characteristics are “continuous” not categorical?
There’s another, very good cognitive reason why type theories are so popular. Types are easy and simple. And they’re categories. If you can pigeon-hole somebody into a category you can also believe that you can predict their behavior and reduce your uncertainty about it. And reducing uncertainty makes people feel better…
One of several very real dangers with the type approach is self-stereotyping. And recall what stereotypes actually are: They are predictions of behavior based on a categorization. Years ago, I did some consulting with a large textile company. They had some plants in which the illiteracy rate among the workers was disturbingly high. There had been several incidents in which the processing and inventory control tags had been misread by workers and this had led to major (and costly) manufacturing mistakes, e.g., producing 10,000 shirts with no buyer. As part of a program (for which the company deserves a LOT of credit for doing the right thing, in addition to the obvious financial benefits) to reduce illiteracy in the workforce, a previous consultant had administered a type personality test to many of the workers. In between my project-related conversations with people I asked, just out of curiosity, what they thought of the test results. One woman I talked to was practically bubbling over with enthusiasm about it—”I think it may have saved my marriage! I made my husband take it too and we both realized that he’s a Badger and I’m a Wolf. I think we understand each other much better.” Happy ending, right? And perhaps that couple went to have a long and happy marriage. But when I went home and thought about what the woman had said, a chill went down my spine. What if they had come to the opposite conclusion? “Well, you’re a Badger and I’m a Wolf, clearly we’ll never be compatible. It’s time to call the divorce lawyer.” It would have been all too easy for them to draw the conclusion that their personalities, and therefore their behaviors, are fixed and immutable, and drive all of their behaviors. And then consider the vast possibility of self-fulfilling prophecies. You act like a “Badger” because you’re convinced you’re a Badger and therefore can’t act any differently. Divorce court, here, we come…
There are two lessons here. One is that even an ill-conceived, ill-supported instrument can have some positive benefits in getting people to think about their own characteristics—especially in the context of a relationship. But the other lesson is much more ominous: That instrument can lead people to make unwarranted conclusions about their own psyche and their own behavior—stereotype themselves and their partners—with a huge potential to do harm.
Ironically, this idea that people love to put themselves (much less others) into categories is very well known in psychology. So well-known, in fact, that it has a name: The Barnum Effect. Yes, named after P.T. Barnum the showman (notorious for, among other things, labeling an exit from his traveling tent show and exhibition with the sign “This way to the Egress –>”, thus luring the vocabulary-challenged to exit the show prematurely and have to pay another dime to get back in and see the rest).
When I was a young psychologist (okay, a “rookie” psychologist—I wasn’t that young) I had been taught about the Barnum Effect but I really didn’t believe it. Or at least I didn’t believe that it could be as powerful as the textbooks said. So, I conducted a little exercise with my Research Methods class. I collected a list of general personality test–style statements (“I like spending time by myself”). In the interest of full disclosure, I must admit that I stole a few of them from my local newspaper’s astrology column. Then I wrote a computer program that had a slick interface (or what passed for a slick interface back then) and asked a “battery” (in this case, ten) questions about your personality. Then I had the computer appear to be processing your answers and then it displayed “your personality profile according to the Switzer All-Purpose Personality Test.” In reality, the computer just randomly selected five of the personality statements and displayed them to the respondent. The kicker was that after the respondent had a chance to read their personality “profile,” they were asked how accurate that profile was: “On a scale of 1 to 100, where 100 is perfectly accurate, does this profile accurately characterize your personality?” The average answer to this question was 92%. My intelligent, college-educated, psychology-major (!) students thought that a randomly selected personality description was 92% accurate! By the way, this phenomenon is in no way limited to Psych students—it’s a general finding—people love to self-categorize and to find ways to fit themselves into a nice, safe, comfortable, predictable category.
There’s nothing special about the human need to categorize personality. Humans are, in general, high-capacity, high-efficiency, categorizing machines. Categorization is often a situation, perceptual, memory, and decision-simplifying process that is useful across diverse situations. The famous psychologist Paul Meehl (one of my academic heroes) has made a very persuasive argument that we often under-use categorization (specifically taxonomies) in psychological theory and research. It is VERY likely that our ancestors found categorization to be an adaptive trait and it is likely that we may have a general-purpose categorization subroutine built into our brains. In psychological terms there may be a System 1 (or “The Autonomous Set of Systems” to use Keith Stanovich’s excellent term) module for categorizing. Your System 2 (the “Analytical Processing System”—which does exactly what it sounds like—and is a general-purpose, reprogrammable simulation and forecasting machine) can pick a domain and a few salient characteristics and bam!, your categorizing module is off and running. But while this sounds great (hey, our ancestors did it! It must be “natural”!), we don’t live in the same “ecological niche” that our hunter-gatherer ancestors did. And that can lead to some serious misjudgments (and that’s a euphemism for total disasters—see the “availability” example below).
Wait, categorization can lead you seriously astray? Yep. Combine this with another system 1 module—your Ingroup-Outgroup module—and you can relatively easily and neatly explain racism, both implicit and explicit (and a whole bunch of other “-isms,” as well). Again, see Tyson’s Starry Messenger for multiple excellent chapters about the dangers of categorization. Here’s a practical example of misusing categorization: Imagine that you own a trucking company. Of course, you want to hire the best drivers. So, who drives better, men or women? When I asked this question (many times) in my classes the vast majority of the students said “men!” And note that the majority of the women said that too. But ask any car insurance company—the accident record shows that women have fewer wrecks than men. In other words, they’re safer drivers. So, should you hire women drivers for your trucking company? Well wait a minute, aren’t there safe men drivers and dangerous women drivers? Of course there are, driving ability DOESN’T come in categories—it’s another one of those human continuous variables. So, should you hire drivers based on gender (another over-categorized variable)? Of course not, you should hire drivers based on their safety record, regardless of gender. And note that it is literally true that “women drive better than men” (because you categorized everybody). But what that statement means, if you’re more precise, is that “the mean number of accidents by women is lower than the mean number of accidents by men” (and presumably that difference is “statistically significant—but that’s a technical can of worms we don’t want to get into, trust me).} But when you hire truck drivers you don’t hire means, you hire people from continuous distributions (standard deviations are important too!).
Another side note (that is potentially relevant to the nature of personality and its measurement): As I said above, our ancestors’ System 1 modules evolved in THEIR ecological niche, not in ours. Two of the most influential psychologists of the modern era, Amos Tversky and (Nobel Prize winner) Daniel Kahneman noted (almost forty years ago now) that transplanting cognitive processes (that probably worked well for hunter-gatherer clans out on the savannah) can often shoot us in the foot when they are used in our modern environment. Our System 1–based automatic responses, in the form of general-purpose judgment heuristics, can lead us to make tragically wrong responses. The saddest instance of this may be the unsung victims of 9/11. One of the heuristics that Kahneman and Tversky described was the “availability” heuristic. This an automatic tendency (again, probably based in a System 1 module) to judge the likelihood of an event based on how easy it is to retrieve an instance of that event from memory (how “available” it is). Note that this is generally a pretty good (i.e., adaptive) idea: When things happen a lot (and therefore are generally likely to happen in the future), they’re easy to remember. If they’re easy to remember, they’re likely to happen in the future. The problem is that, especially in the modern world (think media and instant world-wide communications), your brain is exposed to events that are VERY easy to remember. “Flashbulb memory” and PTSD-induced memories are extreme examples of this. Shortly after 9/11 it was very easy to remember anincident of a plane crashing. Many people had plane tickets (or were going to get plane tickets) for trips they were anticipating. When those people thought about taking a commercial airliner, a dramatic instance of a plane crash was very “available” in memory and so their estimates of likelihood of dying in a plane crash during their anticipated trip went up dramatically. That, in turn, led them to drive to their destinations instead of flying. The sad conclusion to this story of course is that, by multiple estimates, thousands of people were injured or died in automobile wrecks. People who would have arrived at their destinations totally unharmed if they had stuck to their original plans to take the substantially safer option of commercial airline flights.
Small side note: I’ll share what Daniel Kahneman said about personality (in his book Noise) a little later…
Bottom line
So, we come by categorization naturally and honestly. And categorization can be an excellent tool to simplify very complex knowledge structures (ask any biologist about taxonomies!). The problem is, of course, that categories aren’t real (at least until you get down to the quantum level). What I mean by that is that there are many, many human characteristics that are actually continuous variables, and the distributions of those variables can have different shapes, fine details, and other important characteristics that one must ignore in order to categorize a target stimulus (whether person, flower, or personality “type”). And recall, in the words of the architect, “God is in the details” (often misquoted as “the devil is in the details–but same basic idea…)
So, here’s the way to think about types: If you need a taxonomy (in biology, psychology, or wherever) to simplify a complex idea enough to make it manageable, enough to give you a way to wrap your head around it, then categories are a great way to do it. But, as any good biologist will tell you, you should NEVER forget that the categories are human inventions and don’t really tell the full story about nature. And often there are better ways than categorization to get your head around something.
TRAITS
So, if types are incorrect, what’s so great about traits?
In the scientific community the predominant view is traits, not types. What’s a “trait”? Basically, a trait is a fairly stable (more on that later) cognitive (thinking) characteristic that influences your behavior and cognition (acting and thinking). Traits are always shown as a RANGE or AMOUNT of that characteristic (not a category). For example, you’re not Introverted or Extraverted. You are somewhere on the range between the two extremes of introversion and extroversion. And just as important, personality traits are generally not distributed uniformly across the introversion-extraversion range. They are typically distributed in a shape that statisticians refer to as “normal” or “Gaussian” (also known as a “bell-shaped” curve). The important part of that is that most people in such a distribution AREN’T at the introversion or extraversion end of the range, but rather in the MIDDLE or somewhere around the middle. The bottom line here is that you, dear reader, are probably NOT an “introvert” or an “extravert,” but rather a little bit of both (a lot more on that later).
So how many traits are there in someone’s personality? The general view is that there are about five (but maybe six, maybe more) fairly stable traits that help drive behavior. Trait theories have lots of advantages, not the least of which is that they fix the continuous variable/distribution problem (discussed above) inherent in type theories. And that fix alone has much deeper implications than most people realize. Conversely, trait theories don’t offer the emotional reassurance that type theories do, so they’re not as popular outside the scientific community. But trait theories have MUCH better empirical support, i.e., research support (and I’m referring to real research—methodologically sound, peer-reviewed, published in respectable journals, etc. —not the bogus self-promotion that often claims to be research).
But the trait approaches also have their own issues. And the scientific community is starting to realize the extent and implications of those issues…
Most people overestimate the effects of personality in general (i.e., no matter which theory of personality you subscribe to). Even the strong proponents of personality (e.g., Barrett, Judge, et al.) estimate that personality characteristics such as conscientiousness (possibly the best-studied single personality factor except for maybe introversion-extraversion) account for only 10%-20% of the variability in job performance (for example). Here’s another notable example. In a recent meta-analysis (meta-analyses, by their nature, are better indicators than single studies; meta-analyses tend to provide more stable, insightful information than the individual studies that go into them because they reduce noise), well-known and respected researchers Sam McAbee and Fred Oswald (current President of the Society for Industrial-Organizational Psychology) looked at how well personality predicted grade point average. Only conscientiousness had any substantial relationship with GPA. And even that was only r=0.23—mathematically, this means that even knowing people’s level of conscientiousness, you could only predict about 5% of the variance in their GPAs. Pretty lame… Note that they refer to this result as “robust” but they mean that relatively; it’s robust compared to the other dimensions of personality—which performed abysmally (r’s less than 0.10).
We STILL can’t agree even on the number of dimensions of personality. The most widely-accepted trait theory is the Big 5, but conscientiousness is now split into “subfactors”: orderliness and dutifulness. And now all of the Big 5 factors have “facets”—more factors that have varying relationships to the Big 5 characteristics they are ostensibly a part of. Even the Big 5 has a competitor, the HEXACO model (this adds humility/honesty to the Big 5). These are red flags.
For 3000+ years smart people have been trying to pin down the “dimensions” or “factors” of personality—something psychologists call “factor structure.” The fact that we STILL cannot definitely agree on the factors, the most basic components of a trait personality, is a serious issue. It’s a bit of a truism in science that if the exact answer keeps eluding you, you may be asking the wrong question.
Even the stability of traits themselves is being called into question. It’s a widely accepted (and typically unquestioned) view that whatever the actual dimensions of personality are, at least those traits are stable over time. Even Walter Mischel shares this view. But there is increasing evidence that traits are NOT stable over periods longer than probably 5–10 years. A recent longitudinal study of the stability of personality traits (the longest to date, at 63 years) at the University of Edinburgh found NO significant correlations between personality at age 14 and age 77.
Again, there’s that truism I mentioned above: A repeated failure to find a consistent answer means you’re asking the wrong question…
Here’s a personality example
Take one of the best-known (popularly and scientifically), most-researched traits, extraversion (and the one that appears in virtually every theory of personality going back to the ancient Greeks). While the term has somewhat changed in meaning over the decades (and millennia), people who are outgoing, enthusiastic, and upbeat are considered to be extraverted. Extraversion, sometimes called positive emotionality, is at the opposite end of the scale from introversion, which characterizes those who are “solitary, withdrawn, and reserved in emotion.” Cross-cultural studies (using the crude binary distinction) have found that in the United States, more people can be described as extraverts (about 35%) than introverts (25%); this is probably the result of living in an individualistic culture (per Hofsteder). At this point please read on, but DON’T forget what I said before about how personality traits are actually distributed in the population.
Obvious differences in behavior have been demonstrated in experiences. In one study, paired extraverts eagerly asked each other questions to learn about the other person; a follow-up study of introverts had some pairs of introverts sitting together without saying a word to each other. So aspects of this laboratory study fit the usual categorical view of personality. But more extensive research on extraversion found that the broad category of extraversion can be broken down into the following six(!) facets:
- Friendliness—increased tendency to form close relationships, express positive feelings towards others, and to genuinely like others.
- Gregariousness—enjoy being with others.
- Assertiveness—more likely to take charge and direct others.
- Activity level—increased levels of energy and number of activities involved in.
- Excitement-Seeking—increased risk-taking and thrill-seeking.
- Cheerfulness—increased levels of happiness and optimism.
These facets tend to be modestly correlated—those who are “extraverts” tend to be higher on all of them. However, note that even though extraversion is supposed to be a unitary construct, some extraverts may score high on four or five them, while other extraverts may score high on a different set of four or five facets. And note the inherent situational differences even among the facets: “assertiveness” and “gregariousness” have very different connotations, especially across situations. In lots of social situations (for example), gregariousness is generally good while assertiveness is probably rude. A smart “extrovert” probably knows that (and remember that how smart you and the other people are is actually part of the situation in this context). Likewise, assertiveness is appropriate in many business situations in which high gregariousness would be considered unprofessional. A smart extravert probably knows that, too.
What Nobel Prize–winner Daniel Kahneman observed about personality:
Note that last sentence on page 207: “behaviors are a function of personalities and of situations.” More importantly for us, note the sentence on page 208: “Andrew and Brad may share the same score on a test of aggression, but they are unique in their pattern of response to triggers and contexts” (emphasis mine).
And that’s the whole point of our personality test. We can’t put Andrew and Brad in a situation that might trigger their aggression response, but we can activate their schemata for specific work-/school-related contexts. And that alone should improve validity (per Kahneman et al.’s comment on page 207).
And one more time: It’s blindingly apparent from the literature that there are more dimensions than the Big 5 (and it was a huge scientific leap beyond the Type theories and their ilk…). So that’s the other thing we did: We cast the “personality” dimension net a LOT wider.
*Noise: A Flaw in Human Judgment is by Daniel Kahneman, Olivier Sibony, and Cass Sunstein.
Kahneman and Tversky’s work formed the basis of my dissertation (long before Danny won the Nobel Prize; Amos would have been a co-winner but tragically, he died just before the nominations. (By the way, I use their first names here not out of disrespect–even though I didn’t know them personally I’ve spent so much of professional life immersed in their work that I feel as though I did know them. And as a result of that immersion, I have huge respect and admiration for both of them.) Cass Sunstein is co-author of Nudge with Richard Thaler (another Nobel Prize–winner and Kahneman and& Tversky’s protegé; the whole fascinating story is in Michael Lewis’s book The Undoing Project. Lewis also wrote The Big Short (and Moneyball) and that explains why Thaler appears with Selena Gomez in the movie version of The Big Short).
An interesting side note: Researchers in positive psychology have taken note of the facet of cheerfulness and its relationship to extraversion. Research into what makes people happy has exploded in the past 20 years, and the relationship between happiness and extraversion was one of the earlier ones noted. In fact, the relationship between the two is so strong that some researchers wonder if they aren’t measuring the same thing. One other interesting finding in the extraversion/introversion-happiness research is that even introverts are happier when they are with others, thanks to the social nature of humans. Also, extraversion is positively correlated with socioeconomic class. The effect is very slight, however, with conscientiousness having a much stronger correlation (remember, as noted above, “stronger” is still a pretty wimpy correlation in the +0.20’s).
Like other traits, extraversion has both biological and environmental origins. Twin studies have demonstrated a genetic influence on extraversion, with identical twins being significantly more similar in extraversion than fraternal twins (.5 vs. .2). The question has arisen of whether the environment was the main cause (I don’t think we need to comment on the fact that “environment” is often another word for situation). Follow-up studies showed that identical twins reared apart were more alike than fraternal twins reared together. Biological differences are also found between extraverts and introverts, with introverts having a higher level of physiological arousal. This higher arousability translates into more conditioned inhibitions, leading them to be less comfortable in social situations. MRI studies revealed a correlation between the volume of brain regions that process reward with extraversion. And therein lies a big clue to the bases of the our approach. Imagine that you are born with a relatively high-functioning perceptual system but also a fairly reactive limbic system—that’s the higher level of physiological arousal. In certain social situations (think party) there is a high level of incoming stimuli (lots of conversations, loud music, even flashing colored lights if it’s a really good party…). In order to adapt your systems to the situation/environment, what strategies will you evolve? You’ll probably learn to have fewer conversations with fewer people; you may even evolve the strategy of standing near a corner rather than in the center of the room in order to cut down the range of sources of incoming stimulation. But what about the same person in a relatively quiet back room of a restaurant with six of his or her closest friends? That same person, with that same genetic endowment, may be the loudest person in the room, may sit right in the middle of the group, and have the most conversations. But we are compelled, as humans, to try to simply our view of this variability by force-fitting that person into a category. (And by the way, are you cheerful across all situations? I didn’t think so.)
The Power of Situation
and maybe a different way of looking at personality
A major influence on how (and why) we wrote our personality test in this way was a paper by the noted psychologist Walter Mischel at Columbia University. Specifically, Mischel’s paper, “Toward an integrative Science of the Person,” Annual Review of Psychology, 2004, 55: 1–22, was a large part of the motivation for our approach to personality. I think it was an extraordinary look at the fundamental issues in personality and other individual differences. BUT…
I do want to put in a disclaimer here: This is MY interpretation of Mischel and may not be exactly what he meant. AND this is our attempt to “operationalize” some of his points in a real-world personality test. By the way, just as I discussed way back at the beginning of this white paper, DON’T assume that you know what “operationalize” means exactly. The best explanation of what it takes to operationalize a concept in psychology came from a biologist, Jared Diamond (yes, the same person who wrote Guns, Germs, and Steel). See his article, Soft Sciences Are Often Harder Than Hard Sciences, Discover magazine, August, 1987, 8(8): 34ff.
—Another reminder: Don’t forget Mischel’s part about how eliminating context comes by aggregation across situations (and as somebody said, “There’s no text without context”). Basically, this is a self-fulfilling prophecy (or is it a tautology?): If you eliminate situations, then you won’t see any effect of situations.
—Emphasize how the dynamic aspect of personality can only be captured by trying to measure across the dimension in which the dynamics occur. In this case, it’s situation (and a little bit across time but not enough to be important in our situation).
Our approach—What We’re Doing, Why, and How It Addresses the Problems
—Directly and practically addressing the two major problems noted in the elevator speech.
First, the WHY… That’s easy. (See above.)
What We’re Doing—So, what’s the fix?
To understand the essence of Mischel’s different approach to “personality,” you have to understand something that frequently causes confusion, even among psychology graduate students. Here we’re talking about a concept that is frequently referred to by the phrase “Interpret the interaction first!” Correctly, that phrase should be: “Interpret the interaction first, then MOVE ON!” This concept often falls under the aegis of “moderator variables” (a closely related concept that also often causes confusion). Here’s an example of an interaction/moderator (and why it’s foolish to interpret the “main effects” first and sometimes foolish to interpret them at all…).
Imagine you’re doing an experiment on reading speed. You have a theory that reading speed is partly a function of the color of the paper (assuming you’re reading on paper – kind of old-fashioned but bear with me…) and the font size. You randomly assign subjects to take a reading speed test, half on green paper and half on white paper. Half of the green paper tests and half of the white paper tests use a 14-point font, the other two halves use a 10-point font. Some terminology before we continue: paper color is an “independent variable” in research methods parlance. Font size is another independent variable (and reading speed, the thing we’re measuring, is the “dependent variable”). When we conduct our experiment we may find that paper color does, indeed, have an effect. The people who read on green paper had a higher average reading speed than the folks who read on white paper. That’s called a “main effect” for paper color. We might find that the folks who got the 14-point font read faster than the 10-point font people. That’s a main effect for font size. BUT, we might find an “interaction”. That is, green paper had a positive effect on reading speed but only for the 14-point font – people actually read SLOWER on green paper with a 10-point font. And here’s the central point: IF you have an interaction like this, then it makes no logical sense to talk about main effects. If someone asked you “Hey, you study reading speed, does green paper help you read faster or not?,” then you can’t answer the question without saying, “It depends! Are you talking about a 10-point font or a 14-point font?”
And that’s precisely the current situation in personality. People keep asking “What’s the effect of extraversion on behavior?” And Mischel (and we) have to say “It depends! Exactly what situation are we talking about?” Asking questions about traits, without a context, without a situation, is basically meaningless. And that explains a lot about some of the long-standing mysteries about personality inventories, including why many well-researched personality tests do a mediocre job, at best, of predicting real-world behaviors.
An Important Side Note
There are other problematic aspects of personality tests that neither we (nor anybody else at this point) have a perfect solution for. The fact that most of these tests are self-report measures, and rely on a person to correctly observe themselves and remember accurately both the situation and their behavior, contributes to lack of predictive validity. But worse, “Deceptive responses” and cheating by trying to answer in a socially acceptable way rather than in an honest way is a long-standing problem in personality measurement—and despite some modest progress—is still an unsolved problem, despite what some people claim. We don’t have and don’t claim to have any fixes for those issues. But given those limitations, inherent in almost any self-response situation, we still think personality can be better measured…
How We See It
Here’s how this concept plays out in our situation (no pun intended). Personality traits interact with the situation – in exactly the same sense as in the paper color/font example above. You cannot interpret one in the absence of the other. All the recent hubbub about “introverted extraverts” (or sometimes “extraverted introverts”) is a perfect example. People are realizing that the traditional view of personality characteristics is missing something fundamental. They’re starting to realize (without putting it in these terms) that they’re trying to interpret/describe an interaction but referencing only one of the variables. And you just can’t sensibly do that.
And note that it’s not just a problem of interpretation—Mischel’s whole point is that personality is not “a stable trait that gets expressed differently depending on the situation,” personality IS the interaction. And your personality is the specific set of interactions that you carry around in your head. Another way to think about it, and a way I often describe it to my graduate students, is that “personality” is a collective noun that we use to describe a set of specific social strategies that you can pull out as needed. Personality is a toolbox.
By the way, there’s a fascinating potential research question here: Are there (and how big are) the differences in the size and diversity of the tools in the toolbox? If we see someone who has a “stable” personality across situations (e.g., somebody who is ALWAYS extroverted), are we just looking at somebody who only has one tool in the toolbox? Is this a “When you have a hammer everything looks like a nail” phenomenon?
And giving credit where it’s due: Both Mischel and we point out that the basic concept that you MUST talk about traits and situations simultaneously probably originated way back in 1936 with an acknowledged pioneer of Psychology, Kurt Lewin (A Dynamic Theory of Personality: Selected Papers, New York: McGraw-Hill Book Company, 1935).
How do you make a Mischel-compatible personality test?
Let me draw your attention to a quote from the Mischel paper I referenced at the beginning: “In short, the route to finding the invariance in personality requires taking account of the situation and its meaning for the individual, and may be seen in the stable interactions and interplay between them…”
In English: if you want to find a stable human characteristic that can predict cognitions and behaviors, then you have to categorize the personality trait and the situation and its personal interpretation by the individual simultaneously. Mischel gives a great example in the next paragraph: You don’t have “conscientiousness,” you have “college conscientiousness.” I would argue that it goes further… You have “courses in your major conscientiousness,” “courses in your minor conscientiousness,” and “elective courses conscientiousness.” And if you’re going to measure THOSE kinds of conscientiousness, you’ll need a new approach to constructing personality test questions.
Unfortunately, as yet there’s no really good term describing that simultaneous intersection of trait and situation. I hate neologisms but one might be appropriate here. For brevity we’ll refer to a specific combination of trait and situation as a “TxS.” Not very catchy but it communicates the idea.
Note that the vector of characteristics (we gave you something like 19) and the vector of situations can lead to a pretty large matrix. And that leads directly to some tough psychometrics issues (especially given the practical limitations on how many questions you can ask somebody before their head explodes). But we have some potential solutions. In the interest of full disclosure, let me point out that these are NOT well-documented, “tried and true” approaches. Our approach is riskier (and substantially more mathematically “interesting”), without going too far down the Big Data rabbit hole (yet).
This is why we delivered such a long list of characteristics (and note that even those are crude labels/categorizations for the nuanced measures we have in mind). We’ll need to tap into characteristics that include how people encode situations, the interactions between their self-regulatory competencies and the situations (that’s where “grit” comes in), and their “beliefs, expectancies, and goals” (exactly the stuff that I studied way back when in my dissertation).
The Psychometric implications of all this (and there are many)…
The traditional view of personality implies the kind of measurement approach that you see in traditional personality tests (type or trait). You get asked several questions, presumably about behaviors characteristic of the target trait. For example, here are some Extraversion items (from the International Personality Item Pool):
I talk to a lot of different people at parties.
I don’t mind being the center of attention.
I start conversations.
I keep in the background.
I have little to say.
I don’t like to draw attention to myself.
I am quiet around strangers.
These are very typical of personality tests (and even occasionally reference situations, e.g., the “parties” in the first item).
After collecting responses from an individual, those responses are typically averaged across the items to yield a score on the extraversion “scale.” Ideally, one’s responses to those items are almost identical. The idea is that the responses are tapping into the same unitary characteristic. And if that’s the case, then any variability across items must be error, i.e., it must be noise. Here, of course, the signal is the quantity of the personality trait. This aggregation process is so ingrained in the personality concept that every single scale in the International Personality Item Pool, for 274 traits, reports a Cronbach’s alpha level. Cronbach’s alpha is the most common measure of internal consistency of item responses. The alpha level for the questions above is about 0.85—very consistent. By the way, this is one piece of evidence in favor of the traditional trait approach. If you write the items carefully, people will give you pretty consistent responses. This is evidence of a unitary trait across situations, right? Well, it might be, but there are other processes (see above; also think priming and cognitive biases) that can yield consistent responses to situationally different questions. Also note that the questions above, while sometimes referring to specific situations (“parties” are very popular in Extraversion questions), are often VERY generic, e.g., “I don’t like to draw attention to myself.” Respondents generally aren’t stupid. They can rapidly fall into a role*—”Oh, they’re asking me about my general characteristics. I’ll kind of sum up in my head what I think about how outgoing I am and answer all the questions that way”. Self-stereotyping raises its ugly head once again.
*A slightly scary side note: The idea that people can rapidly perceive and adopt a role shows up, of all places, in hypnotism. The counter-theory to the traditional “Hypnosis is an altered state of consciousness” is that “I’m hypnotized” is a role most people are familiar with (or at least used to be back when hypnosis appeared on television and in the popular press more often). And the role theory of hypnosis says that people know what to do and will voluntarily and willingly play the role of hypnotized person. And if they’ll do something dramatic and overt like that, then why not play the role of an extravert or introvert when filling out a personality instrument?
But the bottom line on the way traditional personality tests are scored and analyzed is that it fits perfectly the way the construct is traditionally imagined. There’s a unitary construct in your head, so if your answers to several questions about the same construct differ, then the problem must be in the questions. Or there’s just random noise in people’s outputs and that creates inconsistencies from question to question, item to item. And to be fair, there is some truth in that. Human neural networks—and that’s what human brains are, massive networks of networks, approximately one trillion processing nodes in total—have a noise level. The processing nodes in human (or any) brain are hybrid analog-digital processors and neurons work largely by varying their firing rate (yes, there is something called the “all or none” law that you may have learned in Introductory Psychology but it’s something of a simplification). As such, human “outputs” (AKA behaviors) do have a level of randomness.
But, and here’s the bottom line, the averaging process is designed to eliminate the “noise”. That’s the whole point of it. We’ve traditionally made the assumption that, roughly, the errors are, in fact, random and therefore (as random errors in any domain do) they sum to zero. So if you average across sufficient responses you get the “true” response, the true level of the construct. In introductory psychometric theory classes you even see an equation that looks like this: T = x + e where T is the “true score,” x is the observed response, and e is a randomly distributed error term. If the errors average to zero, then T = x. Voilà!
But all of that assumes that there is a unitary construct and the ONLY source of variability across personality test items (from the same scale) is noise. And as I mentioned above, that’s probably not true at all. Mischel actually goes a bit further and refers to the traditional approach as “Eliminating Context by Aggregation Across Situations.” In his view, the averaging process is DESIGNED to eliminate situation, to eliminate context. One quick side note—I’m using averaging here as the most common example of the aggregation process. Different personality tests often use different aggregation methods. For example, tests may weight questions (or subscales) rather than merely averaging. A truly sophisticated test using the traditional approach would use Item Response Theory—a very clever mathematical process developed by psychometricians and based on maximum likelihood—that can calculate the optimum weighting (and not incidentally, a very accurate function linking the amount of a characteristic a person has and the probability they will respond to a given question in a specific way). I don’t know if any existing personality tests use the Item Response Theory approach but they clearly should if they are based on the trait view of personality. But again, the problem is: What if that traditional view, that’s been ingrained in Western culture for 3000 years, is wrong? Then the psychometric approach we’ve been taking is wrong and a new approach is needed. An approach that recognizes that what we’ve been labeling as “personality” for millennia must take situation into account.
Wait a minute, so why a list of 19 traits? And why those traits?
First, you DO need a set of traits. Our whole argument is that, per Mischel, personality is actually a set, a list, a repertoire of specific trait-by-situation-interactions. And it’s the interactions that are important. Just like the paper color/font/reading speed example above, it makes absolutely no sense to talk about traits in isolation. But conversely, you can’t have paper that doesn’t have a color. You can’t have print without a font. The experiment HAS to start by choosing the best paper colors, the ones that are most likely to show you the effect you hypothesize. In other words, the first step in an effort to measure REAL personality has to be to choose the best possible list of traits. Yes, even though we purposely reviled (kinda) the trait approach (and its evil twin, the type approach) above.
Another side note… It would be perfectly possible to take that list of 19 traits and build a quite usable but traditional personality instrument. Moreover, in our expert and not-so-humble opinion, that instrument would be superior to the usual Big 5 measures like the NEO, and far superior to the MBTI and its ilk. And the reason for that is largely mathematical. If personality is a set of trait-by-situation-interactions, then it stands to reason that the broadest possible set of traits will capture the broadest amount of variance in behavior. Yes, you’ll be missing all of the variance due to situations, but at least the wider trait “net” will capture more of what is actually happening.
But here’s another important point (and a guiding principle in our selection of the traits): You can’t just pick traits randomly, you have to pick the traits with the target use in mind. In psychometrics, one of the most important principles (and one that I habitually have to hammer into the heads of my graduate students) is that there’s no such thing as a “valid” test. And recall that “validity” is the correct scientific term for the operational accuracy of a test. Tests are only valid (or not valid) “for the purpose for which they were designed”. You can make an important point in the wars over standardized testing based on this principle. Using standardized testing to measure teacher performance in K through 12 classrooms has generated a series of vitriolic and divisive debates. The one (and to my mind, most critical point) is that standardized tests were not designed to measure teacher performance and therefore, by themselves (note that important disclaimer), cannot be considered accurate and trustworthy measures of such performance. Notice that they might be valid for that purpose—but you’d have to conduct specific validity studies (for THAT purpose) to demonstrate it empirically. By the way, I’m not unilaterally opposed to standardized tests (far from it). They can (carefully and in very specific ways and situations) be used to help gauge teacher performance. Unfortunately, they are rarely used in those correct ways. Fortunately, both teacher groups and policymakers are starting to listen to industrial-organizational psychologists in that regard.
The point of that digression was: You have to start by picking traits—but you can’t pick the traits blindly, you have to have the end USE in mind for the instrument to be valid.
And of course that brings us to a conundrum. We don’t, at this moment, exactly know what the use of this proposed instrument will be. Most, if not all, personality test developers DON’T know exactly how their users will use the tests. We’re not the users. Are they businesses who want to use personality tests to recruit applicants or to coach current employees? Are they people who want to understand themselves better (“know thyself”) so they’ll be better parents, co-workers, bosses? Unfortunately, as personality test developers, we can’t know for sure. So we made our best estimate of what traits are more likely to describe job-relevant and academic performance-relevant characteristics.
Think about what you think measurement is. Go look up definitions of “measurement.” Most of them are generic and imprecise (and I’ll wager that some of the are just wrong). Instead of belaboring that point, let me just cut to the chase: Here’s the best definition of measurement I’ve ever heard (and I have heard many). From one of my favorite professors at University of Illinois, Fritz Drasgow: “Measurement is assigning numbers to a phenomenon such that the characteristics of the numbers reflect the characteristics of the phenomenon.”
Note the parts of that definition… First the part about “characteristics of the numbers…” What does that mean? You may remember from WAY back in school that there are all kinds of numbers with different sets of characteristics: integers, real numbers, irrational numbers, imaginary numbers, etc.—ask any mathematician for a complete list. And note the second part of the definition: “the characteristics of the phenomenon.” And especially note the part about how the characteristics of the numbers and the phenomena should MATCH (“reflect”). If you’ve ever taken a course in the measurement of psychological characteristics, then you heard about the basics of this on day one: remember that part about “nominal, ordinal, interval, and ratio” measurements. Those terms are all about different characteristics of the numbers that (pardon the expression) “characterize” a measurement. And remember, that’s only the first-day stuff. If you took a psychometrics course, you know it just gets more complex from there.
For example, say you’re watching a NASCAR race. Car number 20 goes by. But then car number 40 goes by. Forty is twice as much as twenty. Does that mean that car number 40 is twice as fast as car number 20 (and will shortly be passing car number 20)? Of course not—in this case the characteristics of the numbers DON’T match a primary characteristic (speed) of the phenomenon. In this case the characteristics of the numbers represent a different characteristic of the phenomenon—namely that these are two DIFFERENT cars (different owners, different drivers, different teams) and the only important characteristic of the numbers is that they are different. Likewise in psychological measurement, you have to be SUPER careful about what characteristics you are talking about and how the numbers correctly (or incorrectly) represent them. Going into all of the details of the psychometric properties of our (and other) personality tests would turn this white paper into a full-length textbook. And going into the inference processes that make up any psychological measurement (actually any measurement of anything) would require at least TWO textbooks. So we’re only going to cover some highlights, some “food for thought”…
Disclaimer: Note that I’m about to frame this issue in old-fashioned and frankly, backward, sexist, terms. But I’m doing it to (a) make a point and (b) frame it in exactly the same terms that I hear (over and over, even today). (Be sure and note that your class, men AND women, almost all went with the stereotype.)
—Problems in (all) measurement: reliability, validity, bogus precision, “objective” metrics, criterion deficiency and contamination, etc.
—Self-report measurement of personality—the good and the bad (actually, you touched on this earlier—you might want to just skip this).
Know Thyself
e.g., Meta-leadership
Employer: WE can’t tell an employer what characteristics they need, but we can help them measure them.
Teams: There is NO single magic formula for team composition. For that matter, sometimes teams are the worst way to approach a problem or goal—sometimes it’s better to skip the team altogether and assign the task or goal to an individual. But here’s one thing you CAN do when you’re trying to decide what characteristics you need on your team. Look at what the team NEEDS to do. And I mean that in two entirely separate ways: What knowledge, skills, and abilities does the team need to accomplish their tasks (note the plural) AND what knowledge, skills, and abilities does the team need to KEEP FUNCTIONING AS A TEAM? A team is a task/goal-achieving ENGINE. Like any engine, it has needs—gas, oil, water, spark. Are you providing the resources (including TIME) that the team needs. And like any engine, the different parts have to work together. Maybe some team orientation? AND like any engine, even the best ones, the team needs MAINTENANCE—down time to address team issues, maybe some training to improve communication, a team-friendly environment, you name it…
But any serious teams researcher will tell you that the mix of characteristics you need depends greatly on the task and task environment. It’s NEVER one-size-fits-all.
IMHO, it would behoove a good manager to peruse our list of characteristics as a starting point for what kinds of skills will be needed in the team. We think we have a good list—at least a good starting point for valuable team characteristics. But don’t forget, just because you have a hammer (and we think it’s a good one) doesn’t mean everything is a nail. Cast your net wide.
User’s Guide—How to Use ANY Personality Test Wisely
Using a personality test
There are generally two basic targets for personality test results:
- YOU. Know thyself (but also use that knowledge wisely – see below);
- Somebody else (employer, potential employer, trainer (business or personal), fellow team member, partner, spouse, you name it…
For any of these targets (including you) there’s a catch (the negative spin) and a magic ingredient (the positive spin) and they’re both the same: For personality test results to have a real and positive impact they HAVE to be used “wisely” = in other words, you HAVE to do your homework first, then you have to make good, rational decisions. Personality tests, unlike horoscopes, aren’t a fun little pastime/conversation-starter. They’re a power tool like a drill press or a chain saw. And like a power tool, you have to have a little education to use them wisely and safely. If you just “wing it” you could lose a body part. In personality testing the “winging it” part is only using your intuitive notions of what personality is, how personality tests work, etc. (which is why I went into so much detail above) AND how to use them to your own or your organization’s advantage. So consider the following as your basic instructions on HOW to use this powerful tool we call a personality test.
Here’s my take on how to properly use personality tests (and you’ll understand why I spent so much time and so many words above going into how they actually work).
If YOU are the “target”:
Don’t forget, EVERYTHING is an interaction (see above)—so in looking at your personality test results, DON’T look at them and go, “Oh, I’m like that.” That’s NOT what you should be thinking. Instead, think about how THAT, whichever trait it is and how much of it you have, interacts with the situations in your life – the different work situations, the different social situations, the different personal situations. (While we didn’t/couldn’t test for all of these situations—for the reasons I detailed above – you can still use the results from our test to give you insights into your behavior in those situations.) Sit down and think (and preferably REMEMBER) about how and to what degree you expressed that level of trait in each of the situations. It’s a hard task that will tax your memory (and try very hard not to self-stereotype while you’re doing it) but give it your best shot.
Ask yourself some questions about those traits and how you use them and when: Do my traits vary a lot or just a little across situations? If they only vary a little, is that a good thing – should I use that trait differently in different situations? In other words, maybe I need to be a little more adaptable (“plastic”). On the other hand, if a trait varies a lot across situations, is that a good thing? Should you be more consistent across situations? Does a trait that helps you at work (which you might not be using much at home) potentially help you at home? With your own cognitive and emotional states? With your interactions with your family? Or in other ways?
If you’re interested in a particular job or in getting better at your current job, which traits are helping there? Which ones may be hurting you at work? Is the quantity of each trait appropriate for each work situation (there’s that word again) you have to deal with? For example, maybe in some situations you need to up your level of Agreeableness, maybe in others you need to dial it down a notch. How about your Assertiveness – could it use a little fine-tuning?
Note that (unlike most personality tests) we’ve got a LONG (but unfortunately not exhaustive) list of traits. So you should have plenty to think about when you’re going over your results. And remember that while we can suggest some areas to look at, only YOU know what your specific situations and circumstances are (or if you’re an employer, what is specifically called for in your business). And the single best use of a personality profile is to sit down and carefully think through the implications for YOUR situations of each of the 19 traits. And remember, personality isn’t destiny.
And speaking of plasticity… I once got into a loud argument with a vice-president of human resources who was convinced that humans couldn’t and wouldn’t change and that the only way to tell who was going to be a good manager or executive was early testing. I tried to explain to her that humans were adaptable – even when it came to personality characteristics – and that if you had a good employee as a management candidate who had almost all of the requisite knowledge, skills, and abilities, then you’d be foolish (economically and organizationally) to throw that person away. Have them TRAIN on their one or two weaknesses and take advantage of the huge investment you’ve already made in them and their wealth of organizational knowledge and experience. (And a quick side note to my HR colleagues: yes, there are a few stars out there [ who have all the KSAs you need and have them in abundance]. But remember that they are few and far between, tend to be VERY expensive, and won’t stay long in your organization.)
Looking at other people’s personality test results:
And speaking of training, a little advice for the employers and business training and organizational development folks, etc….
As somebody who has studied selection (hiring and promotion) for most of his I/O Psychology career, I feel very confident in saying that training is always cheaper than selection. If you’re doing a cheesy selection process, then the costs for hiring someone can be fairly low up front but you almost always pay the price in the long run. And the price can be very steep (cf. the story above about employee illiteracy and the shirts). Obviously, sometimes you HAVE to hire but never forget the long-run costs (you WILL have to do some kind of training, implicit or explicit, starting with the selection process itself, through onboarding, through specific/continuing training, remediation if you made the wrong hiring choice, you name it…). So, the lesson here is, Don’t make the same mistake that a lot of people make when they’re looking at their individual personality test results (“Oh, I guess I’m like THAT”). Consider an employee’s personality profile (especially ours, which is situation-specific) as a training guide. If your job analysis (you DID do a job analysis, right?) shows that a specific trait is useful in a job, then the personality test results give you a picture of the distribution of that trait throughout a work unit/group. But (one more time) do NOT think of that trait as a fixed, unchangeable characteristic that a particular employee has or doesn’t have. “But Fred, you said earlier that personality, while it does change over decades, changes very slowly.” Yep, I did say that, and I meant it. But recall what else I said: Personality is NOT destiny. Personality is one of several ingredients in a particular situation that contributes to behaviors (and cognitions) in that situation. The WISE users of personality test results see a personality trait as ONE of the ingredients in a recipe for getting the best out of employees. And like any good chef, a good employer uses varying amounts of an ingredient depending on the dish and on what effect the chef wants out of that dish.
So, here’s a summary (but remember what I said about “executive summaries” at the beginning of this document).
If you’re looking at your own personality test results, use them to analyze. (I hate to use the term “strengths and weaknesses.” This implies that some traits are good and some are bad [and worse, that you only have one level of that trait—”type” theories, I’m looking at you]. One more time—don’t stereotype yourself.) Use them to analyze how you use your traits and how you use them relative to what you want to achieve.
If you’re looking at somebody else’s personality test results, look for potential and for training opportunities. You’ve got some insight into a menu of characteristics that can be used in LOTS of different ways and different situations. The smart play is to optimize the SITUATIONS to fit the menu, AND AT THE SAME TIME, help the person (yep, training and development again) improve their menu.
Don’t think in either/or terms. Humans are WAY more complicated than that. Humans have levels (remember the continuous distribution thing I talked about above), and remember that those levels vary by situation. TUNE the needs of the job to the appropriate level. And an IMPORTANT note from somebody who used to teach graduate-level job analysis: The optimum level of a particular trait may vary WITHIN a job. In plain English, the best level of a given characteristic (take assertiveness, for example) may change from situation to situation IN a single job. The good employees probably know to vary their level of assertiveness depending on whether they’re talking to a customer, to management, or to their fellow team members.
And a shameless plug for our test… Menus with more items generally have a broader appeal – especially if you have a bunch of people in your party who want different things. We made our menu as broad as we practically could – so it would be useful in lots of different circumstances and to lots of people.
Don’t forget the leadership part. Leadership is four toolboxes: people skills, technical skills, vision, and meta-leadership. We directly measure a LOT of traits in those toolboxes (assertiveness, agreeableness, business acumen, vision, social intelligence, etc.). We even included meta-leadership as a trait itself. But it goes beyond that—as a leader you can look at your own results, use those to inform your own assessment of your strengths and weaknesses, AND look at the profiles of your direct reports (for example) to help you form strategies to negate your weaknesses and bolster your strengths.
And don’t forget what I said earlier (multiple times!): Individual differences can only predict so much of the variance in work and life, even in the best of circumstances. This is why MANAGEMENT, MOTIVATION, ENGAGEMENT, RESOURCES, LEADERSHIP, etc. are all REAL things. If you know your I/O Psychology, I’d even throw in things like substitutes and neutralizers of leadership as an example of how non-personal (situational!) characteristics can modify, moderate, and even totally reverse individual differences. Ditto for motivation: Performance = Ability * Motivation * Resources/opportunity. Note that at best individual differences are only a part of that equation. And remember, as I said above, personality is only a subset of individual differences.
But because personality IS a part of those areas, they are all opportunities for the wise use of personality profiles. Think about employee engagement, for instance. If your company is having engagement issues (and who isn’t, nowadays?), then maybe you should look at your employees’ scores on Need for Autonomy AND how your managers are managing job autonomy (you do know that the right level of autonomy is generally considered to be critical in employee engagement, right?).
And while we’re talking about leadership, another important note: Just because you have a hammer, in this case a good personality test (ours, of course), doesn’t mean everything is a nail. For example, my colleagues (Andy Vinchur, Jeff Schippmann, and Phil Roth) and I, WAY back in 1998, did a massive meta-analysis of predictors of sales performance (“A Meta-Analytic Review of Predictors of Job Performance for Salespeople,” Journal of Applied Psychology, 83: 586–597). One of the more notable findings was that personality factors like extraversion did NOT predict sales performance very well. Spoiler alert: a test of knowledge of sales techniques did much better than a personality test.
Wisely using a personality test means thinking about it IN CONTEXT and with full understanding of its strengths and weaknesses. For example, you definitely wouldn’t want to rely on a personality test exclusively to hire salespeople. But you might have found (from your job analysis, of course) that for your product, knowing what’s going on in the industry at large and with your competitors gives your folks a distinct sales advantage. If some of your sales force are relatively low at Business Acumen, then you have an opportunity for improvement (by the way, don’t fire them, TRAIN them. It’s a lot cheaper and you’re not throwing away the knowledge and experience that they already have).
And what kind of business are you in? Are you trying to carve out market share or are you trying to innovate? Yeah, yeah, yeah, I know, you’re doing BOTH. Do both jobs need the same level of creativity? And here’s one of my points: You’ll be tempted to say, “Of course both jobs need creativity.” With innovation, it’s obvious, but you also want to think of new, creative ways to take market share from your competitors. That “sounds good on paper” but is it, really? And for that matter, is the focus of the creativity really the same? Coming up with a brand-new product is very different from figuring out how to beat the competition with your current product line. You want to do both but they’re definitely NOT the same process and the folks doing those jobs definitely DO need a different mix of skills and characteristics. Maybe a little more need for autonomy, need for cognition, or tolerance for ambiguity in the product development folks and maybe a little more assertiveness, perseverance, self-regulation, or team orientation in the market share folks.