Issue 054 - Education - Human attention
How many books could teenagers read by replacing 30 minutes of daily screen time?
The OECD's 2025 PISA assessment found that reading performance among 15-year-olds had fallen to its lowest level since testing began in 2000. Students in OECD countries reported spending hours per day on recreational digital devices, and the OECD found associations between heavy digital use, distraction, and lower academic performance.
The problem
Estimate the total recreational screen time accumulated in one year by a single age cohort of U.S. 15-year-olds.
Then suppose each teenager replaced just 30 minutes per day of recreational screen time with reading.
Estimate how many hours of additional reading that would create per teenager per year, how many average-length books one teenager could theoretically read in that time, and how many books the entire U.S. cohort could collectively read in one year.
The purpose is not to assume that replacing screen time with reading would automatically reverse declining test scores. The Fermi problem is simply to establish the scale of the time involved.
Because Fermi problems target an order of magnitude, I normally use no more than two significant digits and write most calculations in scientific notation; the Fermi reference explains both conventions.
Before checking sources
Matt's first pass
First I figured out about how much time is spent recreationally consuming digital content by the average 15 year old. If it is 3.4 hours daily, over 365 days that is about 1,200 hours in a year.
annual recreational device time
~= 3.4 h/day x 365 day/year
~= 1.2 x 10^3 h/yearNext, I figured out how much time would be spent in a year if they spent half an hour daily. 0.5 hours x 365 days is about 180 hours in a year, a fraction of the digital doom-scroll time.
replacement reading time
~= 0.5 h/day x 365 day/year
~= 1.8 x 10^2 h/yearNext I assumed a reading rate of about 1.5 minutes per page, and an average book length of about 300 pages. That would mean about 40 pages an hour, about 7,200 pages in a year, and divided by an average book length of 300 pages, that would mean about 24 books read per year.
reading speed
~= 60 min/hour / 1.5 min/page
~= 40 pages/hour
pages per year
~= 40 pages/hour x 1.8 x 10^2 h/year
~= 7.2 x 10^3 pages/year
books per teenager
~= 7.2 x 10^3 pages/year / 300 pages/book
~= 24 books/yearNext, I assumed an equal distribution of the U.S. population across all ages up to about 80 years old, not correct, but just for estimating purposes. That would mean about 1.3% of the population is any given age in that range.
1.3% of 330 million is about 4.3 million, so let's assume there are about that many 15 year olds in the U.S. 4.3 million 15 year olds each reading 24 books per year would mean if this became a uniform practice in the U.S., 15 year olds would read about 100 million books in a year.
15-year-old cohort
~= 3.3 x 10^8 people / 80 age-years
~= 4.1 x 10^6 people
~= 4.3 x 10^6 people
cohort books
~= 4.3 x 10^6 teens x 24 books/teen-year
~= 1.0 x 10^8 books/yearCalibration Score
Matt's Calibration Score: 100 / 100
Higher is better: earn points for accurate pegs, sound models, correct math, and a result close to the sourced answer. The image shows percent full of it: 100 minus the Calibration Score.
Pegs: 30/30. The 30-minutes-per-day annualization, U.S. single-age cohort estimate, reading pace, and average book length were all well inside a reasonable reference range.
Model: 30/30. The model correctly turns a daily time transfer into annual reading hours, books per teen, and cohort-wide books.
Math: 10/10. The arithmetic is clean and keeps the powers of ten straight.
Result: 30/30. The roughly 100-million-book result is essentially on target for a corrected range around 100 million to 130 million books.
Grounding facts
AP reports that newly released PISA 2025 results show math, reading, and science scores at their lowest point in PISA history among highly developed economies, with U.S. teens' reading and math scores also at their lowest point since PISA began in the early 2000s.
The OECD's digital-age child well-being report, using PISA data, says 15-year-olds in OECD countries averaged substantial digital-device time, including leisure use around 1.1 hours per school day at school, 2.6 hours before and after school, and 3.9 hours per weekend day. That makes a 30-minute transfer a modest share of the broader recreational screen-time pool.
Census national population-by-characteristics tables provide single-year-of-age population estimates. For Fermi work, the simple peg of about 4 million people per U.S. single-year age cohort is a good starting point.
A 2019 meta-analysis by Marc Brysbaert estimated average adult silent reading in English around 238 words per minute, with broad ranges by text type. A normal trade book or novel is often on the order of 70,000 to 100,000 words, which makes 5 to 8 hours per book a useful Fermi range.
After checking sources
Check and recalibrate
Start with the replacement time. This part is fixed by the prompt:
replacement reading time
~= 0.5 h/day x 365 day/year
~= 1.8 x 10^2 h/yearMatt's book estimate can be checked either by pages or by words. A 300-page book at roughly 250 to 300 words per page is about 75,000 to 90,000 words. At 200 to 250 words per minute, that is roughly 5 to 8 hours per book.
hours per book
~= 8 x 10^4 words/book / 2.2 x 10^2 words/min
~= 3.6 x 10^2 min/book
~= 6 h/book
books per teenager
~= 1.8 x 10^2 h/year / 6 h/book
~= 3.0 x 10^1 books/yearThat word-based check gives around 30 books per teenager per year. If the teenager reads more slowly or chooses longer books, Matt's 24-book estimate is still very plausible:
longer/slower case
~= 1.8 x 10^2 h/year / 7.5 h/book
~= 2.4 x 10^1 books/yearThe cohort size check also comes out close to Matt's estimate:
U.S. 15-year-old cohort
~= 3.4 x 10^8 people / 80 age-years
~= 4.3 x 10^6 teensPutting those together:
cohort books
~= 4.3 x 10^6 teens x 24 to 30 books/teen-year
~= 1.0 x 10^8 to 1.3 x 10^8 books/yearThe best Fermi answer is therefore about 25 to 30 books per teenager per year, and around 100 million to 130 million books across one U.S. 15-year-old cohort.
The original 3.4 hours per day of recreational device time is roughly 1,200 hours per year. A 30-minute daily transfer uses only about 15% of that pool, but still creates enough time for a teenager to read dozens of average-length books in a year.
Post-check reflection
Matt's reflection
Pretty much nailed this one. What can I say, I know a lot about reading and got lucky with a population estimate. All of these approximations came from my personal reading habits and experience, so none of it was very surprising.
I wonder what impact it would have if we could successfully encourage teens to take up practices like regular sustained reading. I do not think it is reasonable to expect it as a uniform practice, and it is not especially clear to me that the benefits would be worth tremendous efforts. But it does look like a relatively small transfer of daily attention could lead to a substantial exposure to print media rather than digital.
Recommended memory peg
For attention-transfer problems, remember: 30 minutes per day is about 180 hours per year. For reading problems, a practical peg is about 6 hours per average book, so half an hour of daily reading can become roughly 25 to 30 books per year.
Reader results
Bars show how submitted estimates sort into the answer choices from the gut-check prompt.