Lauren Jones and Nicolas Ziebarth, "Successful Scientific Replication and Extension of Levitt (2008): Child Seats Are Still No Safer than Seat Belts", Journal of Applied Econometrics, Vol. 31, No. 5, 2016, pp. 920-928. All data files are tab-delimited ASCII. The analysis in this paper relies on the data from the Fatality Analysis Reporting System (FARS), a database of all fatal accidents occurring in the United States, maintained by the National Highway Traffic Safety Administration (NHTSA). We use data from years 1975 to 2011, all of which are freely available online at ftp://ftp.nhtsa.dot.gov/fars/. For each fatal crash, data are available for the accident, each vehicle involved, and each person involved. We rely mostly on the person-level data, conducting our analysis at the person-level, but we also use some crash- and vehicle-level characteristics that we merge into the person-level files. We use two main data files - one that contains one observation for each child aged 2 to 6 years, involved in a fatal crash (full sample); and another that only contains observations for children for whom a fatality occurred in the other vehicle (sample selection corrected). The first set contains 56,366 observations, and the second contains 15,510 observations. The files are called and FARS-data-full-sample.txt and FARS-data-selection-sample.txt. They are both ASCII files in DOS format, and they are zipped in the file ljnz-data.zip. Unix/Linux users should use "unzip -a". Here are the variables we use, which we provide in the both of the datasets: ------------------------------------------------------------------------ storage display value variable name type format label variable label ------------------------------------------------------------------------ age int %8.0g Age year float %9.0g accident year passgcar float %9.0g Lyesno passenger car suv float %9.0g Lyesno sports utility vehicle weekend float %9.0g Lyesno crash occurred on weekend frimp float %9.0g Lyesno direct frontal impact indfrimp float %9.0g Lyesno indirect frontal impact rearimp float %9.0g Lyesno direct rear impact indrearimp float %9.0g Lyesno indirect rear impact rsimp float %9.0g Lyesno right side impact lsimp float %9.0g Lyesno left side impact death float %9.0g Lyesno fatality modelyr float %9.0g model year childseat float %9.0g Lyesno person was in safety seat lapbelt float %9.0g Lyesno person was using lap belt lapshould float %9.0g Lyesno person was using lap and shoulder belt restraint float %17.0g Lrest restraint type used impchldst float %9.0g childseat improperly used impstblt float %9.0g seat belt improperly used ruralrd float %9.0g Lyesno accident occurred on a rural road row1 float %9.0g Lyesno person was seated in first row of car backright float %9.0g person seated in back right backleft float %9.0g person seated in back left backother float %9.0g person seated in back other male float %9.0g Lyesno person is male crashtm float %9.0g Lcrashtm vehicles1 float %9.0g Lyesno one vehicle crash vehicles2 float %9.0g Lyesno two vehicle crash missweight float %9.0g vehicle weight is missing thoulbs_I float %9.0g vehicle weight in thousands of pounds set to mean if missing numcrash float %9.0g number of ppl involved in crash drivebelt float %9.0g driver was waering seatbelt crashcar double %10.0g unique auto identifier splmU55 float %9.0g Lyesno speed limit less than or equal to 55mph lowviol float %9.0g previous speeding or accident violations highviol float %9.0g previous suspensions or dwi violations