*2.1. Crash Data*

The data used in this study were extracted from traffic accident records obtained from the Jordan Traffic Institute for the years 2014 to 2018. The available data provide information about more than 10,200 crashes on nine roadway segments (five rural and four suburban), which connect different areas with Amman (capital city of Jordan). The information includes crash date and time; driver's age, gender, and fault; pavement surface type and condition; lanes' direction, number, horizontal alignment, and grade; weather and lighting conditions; crash type and speed of vehicles involved; severity; and location coordinates. Some of the driver age records were incomplete (missing or mistakenly entered). The retention of cases with outliers (out-of-range values) can distort the results of the analyses; hence, such cases were either corrected or deleted [30]. It is worth mentioning here that distracted driving was listed among drivers' faults, but the records mentioned nothing about what caused the distraction. Therefore, it is assumed that distracted driving may be attributed to any of the known distraction types (visual, manual, cognitive, and auditory) or a combination of them. Although some of the drivers' faults

can be interrelated—for example, distraction can lead to degraded lane keeping, speed control problems, missed traffic signal, etc.—the results presented here depend solely on the provided information, and thus, there is no way to find such a link between the driver's faults. Therefore, the total sample size analyzed in this study was restricted to those cases that the police clearly indicated as being caused by distracted driving (n = 2472). Among the 2472 crashes that were caused by distracted driving, 910 (36.8%) occurred in suburban areas, and 1562 (63.8%) took place in rural areas. The details of the collected data and their categories are discussed in the next section.

#### *2.2. Data Analysis*

The IBM SPSS version 22 statistical software package was used to perform independent samples *t*-tests to determine whether there were statistically significant differences between drivers. The analysis conducted in this work consisted of two levels. In the first level, an analysis of drivers' faults was conducted by using descriptive statistics to discover the most common driver faults responsible for causing crashes on rural and suburban roadways. In the second level, on the other hand, an independent samples *t*-test analysis was used to examine the characteristics of distracted driving crashes and investigate the differences between crashes caused by distracted driving on rural and on suburban roadways based on human- and location-related factors.
