Thread by @NPestel83

🆕📰Now finally out as @iza_bonn DP:

My paper on the impact of indoor #airquality on cognitive performance of #chess players! 🏭♟️ (with @KunnSteffen @umsbe & #JuanPalacios @MIT)

Paper here: ftp.iza.org/dp12632.pdf

Summary 👇🏻(1/N)

We know that #airpollution is really bad for your #health.

Exposure to poor air quality may have harmful impacts on the 🧠.

Question: What are consequences for performance in complex cognitive tasks? (2/N)

We study the effect of air quality on cognitive performance using data from #chess tournaments. Why chess? ♟️

Because it's a complex, cognitively demanding task with strong incentives, involving strategic decision making under time pressure. (3/N)

In addition: In chess, the quality of cognitive performance can be objectively evaluated using a chess engine, which evaluates human players' actual moves by comparing them to moves deemed optimal according to its algorithm.♟️💻(4/N)

We obtained data from tournaments in Germany over 3 years (2017, 2018, 2019). Each tournament edition: 7 rounds over 8 weeks.

Overall: 121 players, 596 games and about 30,000 moves. We use digitized notation sheets of each individual game evolution. (5/N)

To measure the indoor environment we installed sensors inside the tournament venue.

We focus on the concentration of fine particulate matter (PM2.5), which may penetrate deep into the lungs and brains. We also look at carbon dioxide (CO2) concentration. (6/N)

Identification strategy exploits panel structure:

Observe performance of same individuals, playing multiple games against different opponents in same venue, weekday and time of the day, BUT under varying levels of indoor air quality, beyond the control of players. (7/N)

Our main outcome variable is a binary indicator for a move being annotated as an error by the chess engine. We also look at the magnitude of errors. (8/N)

Main result: indoor concentration of fine particles significantly deteriorates cognitive performance:

PM2.5 ⬆️by 10 μg/m3 (~3/4 of sd) ➡️ probability of meaningful error ⬆️by 2.1 percentage points (relative increase by 26.3%).

No evidence for CO2 effects in full sample. (9/N)

Nice feature of tournament setting: allows to exploit heterogeneity in time pressure: Players must complete 40 moves in 110 minutes ➡️time pressure before move #40.

Result: Impact of PM2.5 increases, now also effects of CO2 in subsample of moves just before time control. (10/N)

Many more details in the paper: role of outdoor pollution, controlling for traffic and a number of further sensitivity checks.

Please go and check(mate) it out... 😎 (11/N)

Overall: We find detrimental effects of poor air quality on cognition, particularly if decisions taken under time pressure.

Implications for high-skilled workers executing non-routine cognitive tasks under time pressure, gaining more importance in labor markets. (12/N) (end)