Objectives: To compare restraint-use practices and injuries among children in crashes with grandparent versus parent drivers. Methods: This was a cross-sectional study of motor vehicle crashes that occurred from January 15, 2003, to November 30, 2007, involving children aged 15 years or younger, with cases identified via insurance claims and data collected via follow-up telephone surveys. We calculated the relative risk of significant child-passenger injury for grandparent-driven versus parent-driven vehicles. Logistic regression modeling estimated odds ratios (ORs) and 95% confidence intervals (CIs), adjusting for several child occupant, driver, vehicle, and crash characteristics. Results: Children driven by grandparents comprised 9.5% of the sample but resulted in only 6.6% of the total injuries. Injuries were reported for 1302 children, for an overall injury rate of 1.02 (95% CI: 0.90-1.17) per 100 child occupants. These represented 161 weighted injuries (0.70% injury rate) with grandparent drivers and 2293 injuries (1.05% injury rate) with parent drivers. Although nearly all children were reported to have been restrained, children in crashes with grandparent drivers used optimal restraint slightly less often. Despite this, children in grandparent-driven crashes were at one-half the risk of injuries as those in parent-driven crashes (OR: 0.50 [95% CI: 0.33- 0.75]) after adjustment. Conclusions: Grandchildren seem to be safer in crashes when driven by grandparents than by their parents, but safety could be enhanced if grandparents followed current child-restraint guidelines. Additional elucidation of safe grandparent driving practices when carrying their grandchildren may inform future child-occupant driving education guidelines for all drivers.