This contribution summarizes the results of the article awarded the Prize for the best article in health economics published in 2023, awarded by the Health Economics Association in 2024 and published in Economic Magazine.
Motivation
All over the world, in both the public and private sectors, assessments of cognitive skills are a constant. These can range from university admission tests, job interviews, to job performance reviews. However, recent research has revealed that certain external factors, such as air pollution (Ebenstein et al., 2016), pollen exposure (Bensnes, 2016) or extreme temperatures (cho, 2017; park, 2020), can influence cognitive performance.
In this context, the question arises of our article: Could the time of day influence people’s cognitive performance? We know from the biological literature that the circadian rhythm – responsible for sleep and wake cycles – affects levels of alertness and mental focus at different times of the day (Cardinals, 2008). However, there is still a need to investigate how this translates into real contexts where cognitive skills are evaluated, such as academic or work environments.
Our empirical approach
To answer this question, we analyze an ideal case provided by the higher education system in the United Kingdom. University students have a particularity in their exam schedule: they can only take one test per day, at three possible times: 9:00 am, 1:30 pm, and 4:30 pm The fact that the time allocation is almost randomly, managed by the Student Services Office, offers us a unique opportunity to identify the causal effects of time of day on academic performance.
Key findings
Figure 1 illustrates our main results. We found that student performance follows an inverted U-shaped relationship throughout the day. The highest performance is observed on the 1:30 pm exams, with an improvement of 0.068 standard deviations (SD) compared to the 9:00 am morning exam, which we use as a reference. However, this effect decreases again for the 4:30 pm exams, showing that performance drops towards the end of the afternoon.
Figure 1: Impact of time of day on academic performance
These results remain consistent even when we apply different robustness checks. We consider (i) variations in the sample and specifications of the conditional independence assumption (CIA); (ii) the inclusion of student fixed effects, controlling for time-invariant personal differences; and (iii) day fixed effects, which allow us to control for fatigue accumulated during the application day.
To explore when these effects are strongest, we built on previous research on circadian rhythm, which indicates three key factors: (1) exposure to natural light (Kantermann et al., 2007); (2) type of cognitive task (Goldstein et al., 2007; Zerbini et al., 2017); and (3) age (Roenneberg et al., 2003). In line with this literature, we observe that the effects are stronger, specifically in terms of a larger magnitude of the estimated coefficients, during exams in winter, when morning sunlight is scarce. Furthermore, these effects mainly affect exams from STEM areas (science, technology, engineering and mathematics), which require a type of fluid intelligence, such as working memory, logical thinking and problem solving. Finally, we found that the youngest students (under 20 years old) are those who show the greatest variation in their performance depending on the time of day.
Recommendations
Our study has important implications for organizations that assess cognitive performance, especially in youth. The results suggest that exams or evaluations could be biased depending on the time of day they are administered. We recommend that, in cases where the outcome of these assessments is crucial, such as in college entrance exams or job interviews, organizations should reconsider the scheduling of assessments or adjust scores to reflect the impact of the circadian rhythm.
In summary, for young adults, cognitive tests involving problem solving should be scheduled in the early afternoon, especially during times of year when there is less exposure to sunlight. On the other hand, other forms of assessments that require less cognitive demand can take place at the beginning or end of the day, when the impact of time is less significant.
Touch to Unlock I do not even understand how I ended up here, but I assumed this publish used to be great