Like a bad marriage, I have tried to make it work for years, but it is finally time to throw in the towel and get a divorce. I was never in love with personality psychology. My first passion was emotion research (states not traits) and social psychology. I soured on social psychology well before the replication crisis that showed many results were obtained with dishonest research practices because artificial lab studies don’t really tell us anything about real human behavior.

My interest in emotions in real life led me to daily diary and experience sampling studies, and I was fortunate to join Ed Diener’s Subjective Wellbeing (SWB) lab to conduct some of these studies. I became interested in SWB and life-satisfaction. For better or worse, SWB research was part of personality psychology because social psychologists focussed on artificial laboratory studies rather than real-world outcomes. Social psychologists only contribution was to make false claim about the validity of life-satisfaction judgments (Schwarz & Strack, 1999) that were amplified by Nobel Laureate Daniel Kahneman to propose a return to hedonism. He later admitted his mistake, but never really apologized to Ed Diener for his invalid criticism of life-satisfaction judgments as the best measure of human happinesses (there is not one happiness for social scientists to discover. Everybody has to define it for themselves, and scientists can only measure how happy people are given people’s own criteria for a good life. They cannot define happiness for them).

Personality psychologists were happy to include life-satisfaction measures in their studies to demonstrate that personality has real world consequences. However, the justification of personality psychology with prediction of real world outcomes also created an incentive to inflate effect sizes of personality and to dismiss other influences on life-satisfaction. This is the paradigm trap. Once you look at a question only from one perspective, you lose sight of other factors. As much as personality psychologists hated the slogan by social psychologists “the power of the situation” to make fun of small effect sizes in personality psychology, personality psychologists started to do the same thing to prove the “power of personality.”

A number of articles on wellbeing by personality psychologists illustrate the fallacy to think that people’s life-satisfaction is determined by their personality. You can find articles with titles like “Happiness is a personal(ity) thing” (Weiss et al., 2008) or “Most people’s life satisfaction matches their personality traits” (Mõttus et al., 2024). Most prominently, Lykken and Tellegen’s (1996) article “Happiness is a stochastic phenomenon” argued that stable personality traits explain 80% of the variance in wellbeing and Lykken claimed that trying to be happier is like trying to be taller. I hated this article and my motivated bias made me think there was an error in their computation of explained variance. I was wrong, but they made several other mistakes. First, they used a bunch of items from a personality questionnaire, averaged them, and then labeled this personality item sum score (PISS) “wellbeing”. The name implies that it is a measure of happiness or wellbeing, but it is just a name for the average of some personality items that were designed to measure personality, not happiness or life-satisfaction. Thus, an invalid measure of wellbeing was used to tell people that they cannot be happier.

The same mistake was repeated just recently when three items from a personality questionnaire were separated from the other personality items (“Am happy with my life,” “Feel that my life lacks direction,” “Have a dark outlook on the future”), averaged, and the 3-item PISS was labeled life-satisfaction, as if it is a standard life-satisfaction measure, which it is not. Other item were called a personality item (“I have a lot of fun” and the fact that this personality item is strongly correlated with the life-satisfaction PISS was used to argue that life-satisfaction is just personality (Mõttus et al., 2024). No evidence is provided that the “fun” item was not just another life-satisfaction item or that the “dark outlook” item is not a personality item. Welcome to shitty personality research in 2024. Rich Lucas, a former Ed Diener student, who knows better did not see any need to publish a comment on this terrible article (you see, why I no longer have psychology friends. You either pledge a paradigm oath, or you are a pariah. I chose pariah because I hate lies, deception, and shoddy pseudo-science.

The fundamental problem of personality psychology is that PISS are given labels that suggest they are measures of some construct. Like the 10-item PISS created by Rosenberg in the 1960s is called self-esteem and it is still used over 60 years later to measure self-esteem, but it is unclear how much scores on this PISS reflect people’s actual self-esteem. If we would not have to worry about measurement error, you could just increase your self-esteem by changing your responses to the items.

An article as old as the self-esteem scale explained how psychologists could validate their measures (Cronbach & Meehl, 1959), but construct validation research is boring and hard to publish. So nobody does it. with a few exceptions. Ed Diener actually did many studies to examine the validity of his 5-item life-satisfaction scale using experience sampling and informant ratings to show convergent validity. Most personality psychologists, however, treat self-report measures as if they are perfectly valid despite evidence to the contrary (self-informant correlations are around r = .4).

The careless use of labels for PISS often impedes real scientific progress. The biggest confusion was created again by Tellegen in his work with his students Watson and Clark. They created two 10-item PISSs, one called Positive Affect and the other called Negative Affect, but the Positive Affect measure does not measure what normal people would call pleasure, positive affect, or enjoyment. It includes items like alert, which you might feel when somebody breaks into your house at 2am. Hundreds of articles about wellbeing have used this misleading measure without every wondering whether the measure has any validity as a measure of happiness. One study even found that watching a distressing scene in the movie “Schindler’s list” increased Positive Affect, but most psychologists rely on labels to interpret results and are not bothered by findings that question construct validity. Construct validity is just not a thing, personality psychologists care about.

That is personality psychology in a nutshell. Write some items, show that the item average is reliable, give the PISS a name, and then correlate the shit out of it with other PISSs. Alternatively you can use exploratory factor analysis, demand two independent factors, call them Positive Affect and Negative Affect and then claim that – contrary to common sense – Positive Affect and Negative Affect are independent. This does not mean that happiness and sadness are independent or any other positive and negative feelings are independent. It simply shows something that was clear from the beginning. If you demand two orthogonal/independent factors, EFA will give you two independent factors. Trying to find a substantive theory for this discovery is silly because you didn’t discover anything. You create two PISSs that are independent by design. The surprising finding that Positive Affect and Negative Affect merely shows that you picked stupid labels for your PISSs and nothing more. Call me crazy, because the article that published this crap is the most highly cited article in the esteemed Journal of Personality and Social Psychology. So clearly, I must be crazy to criticize a measure that is so popular (well, science is not a popularity contest, cf. Galilelo vs. Catholic Church). Actually, their measure is slowly being replaced by a better measure developed by Diener and colleagues.

Another property of personality psychologists beloved exploratory factor analysis is that it never examines whether it fits the data. This is a problem because science advances by falsifying bad theories that do not fit data. So, using a method that always fits cannot advance a science. Fortunately, personality psychology is not a science and it doesn’t care about progress, except the creation of more PISSs that can be correlated with already existing PISSs.

In the 1990s, personality psychologists made the mistake to use an alternative method called Confirmatory Factor Analysis (CFA) that provides information about the fit to actual data. They were disappointed that the method suggested their beloved Big Five model was not consistent with the data. Fortunately, they found a solution to this problem. They decided that CFA should not be used (McCrae et al., 1996). The ban of methods that can reveal bad fit may explain why personality psychology has made no theoretical progress over the past two decades. Some people like to use Big Five PISSs and others use the HEXACO PISSs, and cool new kids use the Dark Triad or Tedrad PISSs. It doesn’t matter as long as the work creates some interesting PISS-PISS correlations that can be published.

Personality psychologists do not even have a clear definition of personality that guides their work. Personality can be a trait (an internal disposition that leads to different responses to the same situation), but it can also be a state that varies from situation to situation. In the end, any item that produces variation in responses across individuals can be a personality item. For example, one personality item is “I don’t like poetry,” but there are no items that ask whether you like rock music, or more specifically the Beatles. See, there is an infinite number of items that personality researchers can create and average. You can even average all items you can find and call it the general personality factor. No kidding, it is a thing among PISS researchers. If you agree strongly with items like “I have the best personality,” “I have many good qualities,” and “I am fantastic” you probably also score highly on the General Personality Factor. And if you score low on this factor, you probably score high on the general personality disorder scale. Again, I am not kidding. Pathological personality psychologists really believe that some people have all the disorders.

So it is time to say good bye to personality psychology. Maybe it was a mistake to become a psychologist. When I made my career choice, I was considering alternative fields like sociology or economics, but ironically I thought psychology is the most scientific way to study human behavior. In the end, I became a meta-psychologist or in German Über-psychologist. Meta-science is a new field that uses the scientific method to study scientists. As scientists are human, we can actually learn a lot about human behavior from studying psychologists. One of the biggest experiments (without a control group) is under way to see whether psychologists are able to recognize their mistakes and improve their research practices. Time will tell.

Saying good bye to personality psychology also does not mean that I will ignore personality traits as predictors of happiness. There is ample evidence that personality and genetic dispositions contribute to variation in life-satisfaction across people, especially in wealthy nations where social policies have reduced variation in people’s ability to fulfill basic human needs (Anusic & Schimmack, 2016). However, there is also ample evidence that social factors influence life-satisfaction ranging from unemployment and underemployment to being overworked and burnout. A happy marriage adds to the wellbeing of some, whereas unhappy marriages, divorce, bereavement, and involuntary singlehood reduce wellbeing. Although it is clear that personality and environmental factors often interact, there is hardly any research on these interactions, again, in part, because personality psychologists think only personality matters.

I am even still teaching a course on personality. This is only possible because I started a long time ago to write my own materials that are now published as a textbook. The book is short because it focusses on the few real scientific contributions that personality research has made. it also avoids the problem of PISS because it teaches students that PISS are imperfect measures of real personality traits that cannot be directly observed using the language of latent variable models. Fittingly, the book is called “The Science of Personality.” Unfortunately, most research that is published fails to follow the rules of science. Writing a book on personality is also easier these days because it has become common to publish open data and some impressive studies with large national representative samples include personality measures. The book uses these data with proper measurement models that avoid the pitfalls of PISS.