Table 3.
Summary results of quality assessment
| Author (year) | Data source | Sampling | Statistical reporting | Focus | ||||
| Official* | Short and long term† | Morning and evening‡ | Spring and autumn§ | Multiple road users¶ | Incidence/ mean/SD** | Other factors†† | Light rather than sleep‡‡ | |
| Askenasy26 | Y | N | N | Y | N | Y | N | N |
| Chu37 | Y | N | Y | N | N | Y | Y | Y |
| Coate18 | Y | N | Y | Y | Y | Y | Y | Y |
| Conte27 | Y | N | N | Y | N | Y | Y | N |
| Coren28 | Y | N | N | Y | N | Y | N | N |
| Crawley29 | Y | Y | N | Y | N | N | N | Y |
| Ferguson9 | Y | N | Y | Y | Y | N | N | Y |
| Green30 | Y | N | N | Y | N | Y | N | Y |
| Hicks25 | Y | N | N | Y | N | Y | Y | N |
| Hicks31 | Y | N | N | Y | N | Y | N | N |
| Huang16 | Y | Y | Y | Y | N | N | Y | Y |
| Lahti32 | Y | N | N | Y | N | N | Y | N |
| Lambe33 | Y | N | N | Y | N | Y | N | N |
| Meyerhoff17 | Y | Y | Y | Y | N | Y | N | Y |
| Sarma and Carey20 | Y | Y | Y | Y | Y | Y | Y | Y |
| Smith34 | Y | N | Y | Y | N | N | Y | Y |
| Sood and Ghosh22 | Y | Y | N | N | Y | Y | N | Y |
| Stevens (2005) | ? | N | Y | Y | Y | Y | N | Y |
| Sullivan21 | Y | N | Y | Y | Y | Y | Y | Y |
| Sullivan38 | Y | N | Y | Y | Y | Y | N | Y |
| Sullivan (2003 and 2004) | Y | N | N | N | Y | Y | N | Y |
| Sullivan6 | Y | N | Y | Y | N | Y | N | Y |
| Varughese36 | Y | N | N | Y | N | Y | N | N |
| Whitaker (1996) | Y | N | Y | Y | Y | Y | N | Y |
*Data derived from official collision data source such as police or national authority.
†Short-term and long-term analyses reported.
‡Separate analyses for morning and evening collisions is more sensitive to DST effects.
§Separate analyses for spring and autumn transitions can test hypothesised DST effects for each transition.
¶Separate analyses for different road users is important to the CET debate.
**Reporting of incidence and mean/SD would support a meta-analytic review if comparison periods were uniform across studies.
††Reporting data for factors that could explain, in whole or part, collision trends around DST transitions aids interpretation of light effects.
‡‡Papers that focused specifically on light transitions, rather than only on the impact of time changes on sleep duration and latency, were more relevant to our review (if they focused on both sleep and light, they were coded Y).
CET, Central European Time; DST. daylight saving time.