Air pollution affects students’ performance on high-stakes tests and long-term educational and career prospects, new research shows
High-stakes exams are the norm in many developed nations, and the results often help to determine whether secondary students seek a university degree, what calibre of university admits them, and what kinds of careers they pursue.
Yet my recent research with colleagues suggests that performance in these exams can be affected by forces that are random, transitory and have nothing to do with student quality. One such factor is the air pollution level at the test sites.
Our research, believed to be one of the first tests of how air pollution may affect cognition, reveals long-term consequences for educational attainment, labour market outcomes, and earnings
Students’ exposure to higher levels of pollution on the dates of their tests negatively affected their performances at those times. This appears to be the result of a temporary variation in performance, rather than a permanent reduction in intellectual ability. But because the tests play such a crucial role in their lives – helping to determine whether and where they go to university, what career choices they have, and their professional options and earnings - the effects of this random factor linger on for more than a decade, and possibly for the rest of their lives.
My recent work, with Avraham Ebenstein and Sefi Roth, examines the effects of fine particulate matter and carbon monoxide, two of the most dangerous forms of air pollution, on the results achieved by Israeli high school students on secondary school exit exams.
In our work, we look at the effects of pollution on results in the Bagrut, a matriculation certificate used in Israel that is a prerequisite for university admission and is an economically important educational milestone. Though the Bagrut is exclusive to Israel, many countries and some American states have similar high school matriculation exams. These include the French Baccalaureate, the German Certificate of Maturity (Reifezeugnis), and the Italian Diploma di Maturità. In many respects, the Israeli system is not so different from the system used in the UK (with reliance on a combination of AS- and A-level exams and GCSEs) and the US, with a reliance on SAT and ACT scores as critical in university admissions. For the Bagrut, Israeli students take a series of national exams in core and elective subjects following 10th and 11th grades, and then they take another large set of exams following 12th grade. The Bagrut includes seven mandatory subjects and one elective subject. The exams are given bi-annually during the two exam seasons. The Bagrut is a pre-requisite for university entrance. Failure to obtain certification will have a significant effect on higher educational attainment. Access to university majors is also determined by Bagrut performance. As a result, it is a gatekeeper in Israel for many lucrative professions.
Our work examines how outcomes on the exam were affected by levels of fine particulate matter and carbon monoxide exposure on the day of the exam at the test sites, as measured by the Air Quality Index. Fine particulate matter is a complex mixture of solid and liquid microscopic droplets found in the air that consists of various components including acids, metals, dust particles, organic chemicals and allergens. These fine particles can travel through small passageways, suggesting that high levels of pollution may affect test takers even indoors. In Israel, the main sources of particulate matter are sand storms, coal-burning power plants, and certain industrial processes.
Ambient air pollution has significant known consequences for human health and life expectancy. Short-term acute exposure decreases circulatory performance and leads to increased illness and hospitalization rates. Research has documented a link between carbon monoxide and higher incidents of respiratory and heart-related emergency room visits.
Though evidence documenting a link between cognition and ambient air pollution is extremely limited, our work documented a connection.
The effect of pollution is through exposure during several exams (eight to 10 on average) which are held at different dates in high schools . Our work estimated the effect of exposure during an exam on the test score on that exam. Then we estimated the effect on long-term outcomes (post-secondary schooling and earnings as adults).
We found that a 10-unit increase in the ambient concentration of fine particulate matter as measured by the Air Quality Index reduced Bagrut exams test scores by 0.46 points, and the same increase in ambient concentration of carbon monoxide reduced scores by 0.85 points. (The score in each exam ranges from 0 to 100 and the mean is around 70.) Alternatively, relative to a day with average air quality, one standard deviation increase in the fine particulate matter is associated with a .65 decrease in scores, and the same increase in carbon monoxide is associated with a 0.54 point decrease in scores.
Our results are largely driven by poor performance of test takers on very polluted days. The results suggest that modest pollution levels have only a marginal impact, but that very polluted days can have much larger effects.
The results also show that the impact of pollution is highest for certain sub-populations with high rates of asthma, or other pre-existing conditions that make individuals vulnerable to air pollution. For those with higher rates of asthma and respiratory illness, the effects of fine particulate matter are larger, suggesting the physiological impairment is a potential underlying mechanism. In contrast, the effects of carbon monoxide are largely consistent among Israeli subpopulations, suggesting that neurological impairment may be a mechanism.
We estimate that an additional 10 units of fine particulate matter is associated with 0.03 per cent decline in Bagrut certification, 0.03 per cent decline in probability of enrolment at university, a 0.15 decline in years of education at university and a 2 per cent decline in monthly income a decade later.
Our evidence shows that pollution exposure is only related to temporary variation in performance, rather than a permanent diminution of intellectual ability. We find that a student’s average pollution exposure at Bagrut exams dates during the period (May-July) in 11th grade has no correlation with his or her average Bagrut scores on exams taken in May-July of 12th grade, suggesting that the effects we estimate are capturing the consequences of short-term random effects, rather than reflecting actual quality gaps between students.
We believe our findings hold significant policy relevance and in several arenas.
Insofar as there are permanent wage consequences from completely random shocks to student performance, our results highlight the danger in assigning too much weight to a student’s performance on a single, high-stakes exam, rather than an overall academic record. If completely random variation in scores can still matter 10 years after a student completes high school, this suggests that placing too much weight on high-stakes exams like the Bagrut may not be consistent with meritocratic principles.
In addition, our results provide new and compelling evidence that cognition is affected by air pollution exposure. Our work suggests that the gain from improving air quality may be underestimated by the current, narrow focus on health impacts. Insofar as air pollution may lead to reduced cognitive performance, the consequences of pollution may be relevant for a variety of everyday activities that require mental acuity.
Our results underscore the need for pollutionregulating policies to take mental and physical human health into account. These results may also highlight a way in which individuals in highly polluted areas, such as those living in cheaper, industrial areas of cities, could have economic disadvantage exacerbated by pollution.
While our data is too recent to know whether the affected students will “catch up“ later with their peers who had lower pollution exposure, our results suggest that having excessive pollution during one’s Bagrut administration has a non-trivial effect on education and wages that are long lasting.
About the authors
Victory Lavy is an economics professor at the University of Warwick, Hebrew University of Jerusalem, the National Bureau of Economic Research, and research associate at CAGE.
Avraham Ebenstein is assistant professor at the Hebrew University of Jerusalem.
Sefi Roth is a PhD student at Royal Holloway, University of London.
This article is based on Air Pollution, Cognitive Performance, and the Long-term Impact of Transitory Shocks on Human Capital and Income, a working paper available here.