3 reasons why we are so bad at predicting our chances of academic success

Vera Scholmerich  by Vera Schölmerich / Reading Time: 5 Minutes /

A question that I asked myself at the beginning and throughout my PhD is: “what are the chances that I will be able to pursue a career in academia after my PhD?”. Conversations with fellow PhD’ers at the coffee machine tell me that I am not alone in asking this question. In the absence of a friendly statistician popping by and informing us on our exact chances of advancing, we have to guesstimate what these might be. Humans use mental shortcuts (also called ‘heuristics’) to solve such problems. These shortcuts intuitively feel accurate, but actually provide us with very bad estimates. Borrowing from Daniel Kahneman’s international bestseller ‘Thinking, Fast and Slow’ (2011), I outline 3 major mental shortcuts that lead us astray.

Mental shortcut nr 1: Humans tend to focus on individual cases and neglect statistics, even if the latter are available (Kahneman, 2011, pp. 166). When I first started pondering on my chances of staying in academia, I did not go online to look for data on how many PhDs actually stay in academia. Rather, my first move was to look for individual examples I knew: which of my colleagues that had recently obtained their PhD continued in academia? Last week I heard a fellow PhD declare that our chances of staying in academia were virtually zero because “All of the six fellow PhD candidates that started together with me had to leave academia, so it must be impossible”.

Consequence of this mental shortcut: depending on the (often unrepresentative) individual cases we focus on, our predictions of success rates in academia are either too high or too low. This means that we might be overly optimistic, or pessimistic.

Illustration by David Parkins via nature.com

Mental shortcut nr 2: Human brains are wired to ascribe causal explanations to events (Kahneman, 2011, pp. 169). So when mental shortcut nr 1 leads us to focus on colleagues that managed to stay in academia, we look for reasons as to why this happened. We search for characteristics of these people – intelligence, nr of publications, how hard they worked, etc. “Leah did well because she put in so many long hours and because she is very clever”. The problem, however, is that we are bad at distinguishing between causality (the long hours Leah put in actually contributed to her promotion) and mere association (Leah, like many academics, is a workaholic, but this did not contribute to her promotion). In fact, the finding that most people who do well in academia tend to work long hours might say more about the type of people that work in this profession rather than a necessary attribute for success.

Consequence of this mental shortcut: we are bad at assessing what we need to do in order to advance in academia.

Mental shortcut nr 3: People underestimate the role of luck (Kahneman, 2011, pp. 177). Due to the mental shortcut nr 2, we prefer causal explanations. Causality leaves little to no room for the role of ‘luck’. Especially if you put in a lot of tears and sweat to achieve something, it is difficult to entertain the idea that part of your success was due to pure luck – the mood that a reviewer was in or whether that hotshot researcher from Oxford also decided to apply to the job that you want. When asked by Edge.org’s John Brockman what his favorite equation was, Kahneman replied:

Success = talent + luck

Great success = a little more talent + a lot of luck

Consequence of this mistake: we do not account for the unpredictability of success in academia and incorrectly entirely attribute success (or lack thereof) to the actions of individuals.

Fortunately, there is a way out:

Kahneman provided some tricks for making better predictions, which I have adapted to fit our particular question at hand:

Step 1: Start with the base rate of advancing in academia. This is what I should have done when I first starting thinking about this question (but never did until I started writing this blog…). In the Netherlands about 20% of PhD candidates stay in academia upon completion of their PhD (WOPI 2011). In other words, there is a 1 in 5 chance of advancing in academia.

Step 2: To determine your personal chances of staying in academia, adjust this rate up or down based on individual variables that influence the success rate.

This step is much more difficult in the Netherlands due to lack of accessible data and analyses. One crucial piece of information that we do have is the distribution of women/men in post-PhD positions. Not surprisingly, the percentages of women drop with each jump up the career ladder (women as assistant professors: 33%, associate professors: 21%, professors: 14%, see WOPI 2011). Hence, if you are a woman, your chances are much lower than 1 in 5.

(WOPI, 2011)

A crucial variable is missing, however: we don’t know how many of the 80% of PhDs leaving academia would have preferred to stay. For example, if almost all of the PhDs that left academia did not want to stay anyway, then the future looks quite bright for those eyeballing an academic career!

What are other important variables we need to take into account? My intuition tells me that the characteristics usually proclaimed as important for an academic career – namely being very clever and working long hours – are outdated. Academia is changing, and perhaps other skills are becoming more and more important, such as social skills, cooperation with others, productivity and being able to spot opportunities. However – I’ll leave it up to future research to figure this out – as these answers are also just the product of my mental shortcuts. Have any of the readers come across interesting research that tells us which characteristics we need to include in our guesstimation? I would be enchanted to hear from you.  


Vera Schölmerich MSc is a PhD Candidate at the Department of Obstetrics and Gynecology of Erasmus Medical Center and at the Department of Organization Sciences of the VU University Amsterdam. Wedged in between a medical and a social science faculty, Vera looks at how ‘social factors’ influence prenatal health and the organization of maternal health care.

Thumbnail image via clarabridge.com