An interview
We talked with Daniel Dinner about his recently published paper titled, “Modelling the incremental value of personality facets: the domains-incremental facets-acquiescence bifactor showmodel”, which is available in the January-February issue of EJP. Daniel is a Professor at the University of Applied Labour Studies, Germany.
Read on to learn more about Daniel’s work on personality measurement and his other interests!
Q: Can you tell us a bit about yourself and what made you interested in personality psychology?
When I started in psychology, I was interested in psychological assessments. How do we get from an impression we have from somebody to a quantifiable number like IQ? A couple of years ago, when I started my PhD, I was interested in how we can measure cognitive ability. IQ tests seemed to work fine, but it always felt like there must be something more – for example, height is just one perimeter to describe the human body, and so cognitive ability may be just one part of the puzzle. I started to think about what else might be important to achieve educational success or life satisfaction. I concentrated on the Big Five personality traits first and discovered that this feels like a natural supplement for cognitive ability.
Q: What do you like to do outside of work?
Well, I like cooking, hiking and running. The latter two also feel like a natural supplement because I like to eat and then I need to train a bit. My grandfather was a winemaker and I tried to press my own wine a couple of years ago but, as it turns out, it’s like writing papers. You need a lot of practice. The first one did not turn out so well. The second one was quite good.
Q: Can you tell us about your study?
Yes, sure! A couple of years ago I worked within an Organisation for Economic Co-operation and Development (OECD) expert group on non-cognitive skills. In particular, the group was interested in whether socio-emotional skills and personality traits add value beyond cognitive skills. In other words, the group wanted to evaluate whether factors outside of intelligence can help us better explain differences in educational attainment and occupational success, and whether they could be used to better understand the association between education and cognitive skills.
From a methodological point of view, that question was quite easy to evaluate because cognitive ability and socio-emotional skills don’t overlap a lot. However, we then started to talk about whether specific personality facets add incremental value over global personality domains. And that was a bit trickier, because facets don’t only correlate on a measurement level and on a manifest level, but also on a theoretical, conceptual level. These facets constitute domains and these domains are an increment part of the facets.
We weren’t sure what the most appropriate methodological approach is and we didn’t come up with a solution for some time. Then I remembered latent trait theory, which I used during my PhD. Latent trait theory is a method that allows you to decompose variance that is specific for a trait and separate that variance from variance that is specific to the measurement occasion. I thought maybe we could do a similar thing with domains and facets. That is, I thought we could use a global domain that is consistent across the different facets (e.g., extraversion) and separate it from systematic variance that is specific to only a facet (e.g., assertiveness).
The first attempts at differentiating global domain-variance from systematic variance did not work out too well. As it turns out, the manifest personality items that we typically use are not pure indicators of personality traits. For instance, the item “I am full of energy” is supposed to measure extraversion, but it also taps into having a low level of negative emotional stability. Similarly, “I am someone who can be counted on” measures not only agreeableness but also conscientiousness. We had to find a way to deal with these overlapping sources of variance and we thought that maybe using a bifactor model in combination with explorative structural equation modelling would allow us to solve the problem. Basically, a bifactor model allows separating domain-level (e.g., extraversion) and incremental facet-level (e.g., assertiveness beyond extraversion) variance. Additionally, explorative structural equation models allow separating variables of different domains (e.g., extraversion and negative emotionality). As it turns out, the method worked, and we were able to separate the variance sources! With a traditional analysis, we would see that there is a correlation between all facets of conscientiousness and health. The bifactor model would show us that there is an association of health with global conscientiousness, and only with the facet “responsibility”., but not so much with the “organization” facet, etc.
Q: As you mentioned in the study, there is a growing interest in personality facets from researchers. Do you expect to see significant changes in the way we measure personality in the future?
Well, there are two approaches that I find particularly interesting. First, there’s a method with multiple choice items that allows us to deal with social desirability. Often in personality questionnaires, we make the assumption that people don’t lie. But that may not always be the case, especially in selection settings. From my point of view, personality researchers have made a very clever suggestion with making people choose between two equally desirable personality items and to code the answer in a dummy variable. We can then use these variables to control for social desirability (see for instance the amazing work on this by Anna Brown and Eunike Wetzel). This may be something we will see more often in the future.
Second, before I worked at the University of Applied Labor Studies, I worked at Gesis, the Leibniz Institute for the Social Sciences in Germany. They started to investigate how we can use big data (or found data) to observe interesting findings on personality measures. I believe the approaches we’ve been seeing in the last few years in machine learning and artificial intelligence will help us figure out if big (or found) data will help us to address some of the limitations of questionnaires.
Q: What advice would you give young researchers?
First, you should travel and see the world. This has nothing to do with your work itself, but traveling can help you broaden your horizons and see things from a different perspective. Especially as a researcher, you often have the opportunity to travel to foreign countries to visit a conference. Use this opportunity to see more of the world other than your desk or your university as often as you can! You should probably wait until the current pandemic is over, though.
Second, if you have the opportunity, work outside of university for a year or so. This is what I did after I finished my PhD. At first, I was afraid that if I decided to come back to science, I would have lost a year. But it was actually the other way around! I learned so much about how administration works, about how you can do things faster, and about how you can look at things from a different perspective. When I did come back to science, I was very grateful for the amount of freedom I have in university, something I would have never realized had I not been on the other side. In addition, in my experience you won’t close any doors if you go somewhere else for a year.
Third, when I talk to PhD students, a lot of them are motivated to be successful very quickly. They seem to be afraid they won’t be successful enough or that they won’t publish enough. I’ve been there and I know the questions that come to mind when you’re doing your PhD. However, I would like to encourage students to think about what inspires them instead of what scares them. Think about what you what to spend your life doing, instead of getting successful very fast. I think science will progress faster if there are less but better papers (including some of my own work). So, take your time, find out what inspires you and if you don’t know yet, you will find it out along the way. There is no need to rush, no need to think about expectations other people might have for you. Focus instead about what you want to do!
Q: Can you tell us which scholars have inspired you and influenced your work?
Prof. Klaus Fiedler, who taught me how to think critically. Prof. Dirk Hageman, who taught me how to work methodologically thoroughly. And Prof. Beatrice Rammstedt who introduced me to the world of large-scale assessments.
Q: What’s next for you?
Currently I am interested in measuring personality traits in the workplace. Most of my work has focused on the Big Five. When I started to work at the University of Applied Labor Studies, I started to get interested in examining whether measuring those traits in a specific context will be more useful for selection purposes and development purposes. Another interest of mine actually came from my PhD student. She came to me with a question – can people train themselves in doing cognitive ability tests? There is some work on whether people perform better at a cognitive ability test if they do the same test again and again.