A 1996 study published in the New England Journal of Medicine is still being cited as evidence that lack of sleep that results from beginning daylight saving time increases traffic accidents. The Los Angeles Times referenced this study as part of a March 11, 2013 article that warned of the ill effects that result from moving your clock forward. Among these: Traffic accidents, fatal traffic accidents, heart attacks, and “cyberloafing behavior.”
Despite the fact that after more than 15 years major new publications still rely upon it, the 1996 study lacks appropriate statistical rigor and overlooks some important causal considerations.
The study considers accident rates on the first Monday after the start of daylight saving time, the Monday from one week before, and the Monday from one week after for nine Canadian provinces (it excluded Saskatchewan because the province does not observe daylight saving time). The study concludes, “The loss of one hour’s sleep associated with the spring shift to daylight savings time increased the risk of accidents.”
The study has several problems. It considers only two instances of daylight saving time (1991 and 1992), does not control for any other factors (e.g. weather), and simply ignores accident rates in Saskatchewan, the one province that could have been used as a benchmark or control.
Further, the study attributes the increase in accidents to loss of sleep. This is certainly a possibility, but the study does nothing to address what portion of the population takes their lost hour from sleep time. Assuming the statistical correlation exists at all, the increase in accidents could be attributable to the panic associated with running late for work (a likely occurrence one day after clocks are set forward) rather than lost sleep. This would still be an ill effect of DST, but not an effect of lost sleep.