Cognitive function and crossword puzzles: Which way does the causal direction go?

In a recent article in the Journal, Brooker et al analyzed a large online sample of older adults (aged 50‐93) and discovered that those who performed crossword puzzles more frequently had higher cognitive function than those who performed them less frequently across a large number of cognitive tests. From these findings, Brooker et al concluded that performing crossword puzzles helped prevent cognitive decline in older age. Their data were cross‐sectional, however, and it is impossible to draw causal inference from them. Proper causal inference requires controlled experiments or, as the authors themselves recognize, prospectively longitudinal data. The British Cohort Study (BCS) is a prospectively longitudinal study that included all babies (n = 17,196) born in Great Britain (England, Wales, and Scotland) during the week of 5 to 11 April 1970. All surviving members of the cohort who still resided in the United Kingdom (Great Britain plus Northern Ireland) were subsequently interviewed at Ages 5, 10, 16, 26, 39, 34, 38, and 42. Respondents' intelligence was measured at Age 5 with eight cognitive tests. The raw scores were factor‐analyzed and converted into the standard IQ metric with the mean of 100 and standard deviation of 15. At Age 16, BCS asked the respondents how frequently they “read” crossword puzzles in newspapers, and their responses were