What they did
Norway conscripts nearly all its young men, and every conscript has sat for the same cognitive test for decades. Bratsberg and Rogeberg used Norwegian administrative registers — which link conscripts to their families — to analyze cognitive scores across three decades of birth cohorts, from 1962 to 1991, covering over 730,000 men.
What they found
- Scores rose steadily through the mid-20th century — the well-documented "Flynn effect," roughly consistent with the broader international pattern of rising IQ scores across the century.
- Scores peaked with men born around 1975, then declined in the cohorts born after.
- Critically, the decline appears within families — a younger brother scoring lower than an older brother from the same parents. Because the brothers share both genes and household, this pattern rules out both genetic selection (a changing gene pool) and shifts in immigration as explanations.
"This establishes that the large changes in average cohort intelligence reflect environmental factors and not changing composition of parents, which in turn rules out several prominent hypotheses for retrograde Flynn effects." — Bratsberg & Rogeberg, PNAS 2018, abstract
What they did not find
The paper identifies the cause as "environmental" — a broad category that rules out genetics but does not pin down a specific mechanism. It does not test, measure, or claim that smartphones, social media, or screen time caused the decline. That connection is a separate argument, made by other researchers and popularized in different work (notably Jonathan Haidt's writing on adolescent phone use), and it's worth treating as a distinct claim with its own evidence rather than as part of what this study established.
Where this shows up
This study is frequently used as a cold open in videos and posts about technology and cognitive decline — accurately citing the data, then attaching a causal story the paper itself doesn't make. See Case 001 for a worked example.
Also worth knowing
Similar stagnation or decline in conscription and large-sample cognitive testing has been reported in several other developed countries since roughly the mid-1990s, including Denmark, Finland, France, and the UK, while patterns elsewhere (including Germany) show more stagnation than decline. The reversal is not unique to Norway — Norway is simply where the data quality made it easiest to prove conclusively.