This study is a cross-sectional survey of 163 parent-teen dyads in Iowa. Data were drawn from the baseline survey for a randomized controlled intervention trial to increase parental involvement in teen driving. Teenagers who were within 3 months of obtaining their intermediate driver’s license and did not currently hold a minor school license (school permit) were eligible for participation. In Iowa, the intermediate license can be obtained after required driver’s education instruction and an instructional permit, which requires 6 months of supervised driving, are complete.¹⁷ This intermediate license does not allow unsupervised driving between 12:30 a.m. and 5:00 a.m., unless the teen has a waiver due to work or school-related extracurricular activities.¹⁷ For the majority of teens in this study, it represents the first legal opportunity to drive unsupervised, the only exception being those who previously held school permits.

Participants were enrolled from local high schools and employee lists from 2 large hospitals in Des Moines and Iowa City, Iowa. At the high schools, the principal sent a letter to parents that explained the study and provided information to enroll. A study booth was also set up at school events, such as conferences or sports events. The hospital administration sent a similar letter to hospital employees. Parents or teens who were interested in the study provided contact information and were then contacted by the study team to screen for eligibility and to arrange an in-person meeting to introduce the study and enroll the parent and teen by obtaining consent and completing the baseline surveys. Of those who expressed interest (N = 336), 163 parent-teen dyads were enrolled and included in analysis, 52 were ineligible, 48 had a parent or teen who declined to participate, and 73 were unable to be reached, resulting in a participation rate of 48.5%.

Once signed parent consent and signed teen assent were obtained, surveys were administered separately to the participating parent and teen. Questions from the survey that were examined in this study included: basic demographics (sex, age, grade (teen), race/ethnicity, marital status (parent)), if teen had ever driven unsupervised, how often teen currently drives unsupervised, current and future expectations of car privileges, how often teen currently drives to certain destinations, expectations on driving unsupervised to nine different destinations (eg, school, work, evening/weekend school-related activities) and expectations on being allowed to drive under 12 conditions (eg, after midnight, on high speed roads, etc.).

Car privileges and expectations were measured by asking “Do you have regular access to a car?”, “Do you have your own car?” and “If no, will you get one once you receive your license?” Unsupervised driving was measured by asking “Have you driven unsupervised (without an adult)? and “How often do you drive unsupervised currently?” Expectations of driving unsupervised were measured by asking “In the first 3 months after you get your intermediate license, how often do you think you will be allowed to drive unsupervised under the following conditions?”(eg, after midnight, without telling a parent who will ride with you) and “In the first 3 months after you get your intermediate license, how often will you be allowed to drive unsupervised for the following purposes?” (eg, to or from school, to go to friends’ houses). These same questions regarding conditions and destinations were asked to parents, but “how often do you think you will be allowed” was replaced with “how often will your teen be allowed”.

Descriptive statistics, including means and frequencies, were calculated for all variables of interest to identify distributional tendencies. Questions about driving expectations utilized Likert scales which ranged from one to 5: 1 = never, 2 = rarely, 3 = sometimes, 4 = frequently, and 5 = very frequently. Weighted Kappas were computed on each of these Likert scale questions to examine degree of agreement between parent and teen dyads on driving expectations. Weighted Kappa was used instead of Kappa because it takes into account the nature of ordinal data, while Kappa does not and is more appropriate for nominal data.¹⁸ A Weighted Kappa of 1 indicated perfect agreement, while a value of 0 indicated agreement that is no better than chance.¹⁹ This non-parametric test assigned less weight for parent-teen dyads who report scores that were further apart (eg, a 1 and 5 pair would receive lower weights than a 1 and 2 pair).

Linear regression models were used to examine how various teen and parent characteristics predicted parent and teen expectations of how often the teen would be allowed to drive unsupervised under various risky conditions and to different destinations. Predictors included parent sex, teen sex, number of times teen currently drives per week to various places (teen report), if the teen had already driven unsupervised (teen report), if the teen has regular access to a car (parent report), and if the teen has his/her own car (parent report). Four outcomes were examined with the linear regression models. First, the outcome of driving unsupervised in risky conditions was calculated as the sum of all risky conditions (eg, on high speed roads, after dark). Second, the outcome of driving destinations was calculated as the sum score for all driving destinations (eg, to or from work, to or from entertainment places such as the mall or a movie theater). These 2 sum scores (risky conditions and destinations) were created for both parent and teen reports; therefore we had 4 total outcomes (Table 4). Sample sizes for regression models ranged from 133 to 162 based on the number of participant responses.

Our sample included slightly more teenage girls than boys (females 53.4%, p = .39), and most teens were 15 years old (92%) and in the 10^th grade (80.4%) at the time of enrollment. Almost all of the teens (99%) had driven and 48.5% of those had already driven unsupervised, either illegally or because they had previously held school permits. The most frequent trip destinations (both supervised and unsupervised) were to or from school (Avg: 2.5 times/wk, Range: 0-10) and to school-related sports or athletics (Avg: 1.5 times/wk, Range: 0-10; Table 1).

More mothers participated in the study than fathers (82.8 % and 17.2%, respectively). Almost all of the parents had at least a high school diploma or beyond (98.8%) and most were married (82.8%). Both the teens and the parents were predominantly White (Table 1).

Over 80% of teens reported regular access to a car and over 40% had their own car. The teen responses were confirmed by 85% of parents reporting that their teen had regular access to a car and 40% of parents reporting that their teen had their own car. Among teens who did not have their own car, approximately half of parents and 41% of teens reported that the teen would have their own car once they obtained their intermediate license (Table 2).

For all types of driving conditions (eg after dark, in bad weather), teens expected to drive unsupervised more frequently than their parents. For example, on average, teens expected to drive more frequently on high speed roads (Mean = 4.0) than parents expected to allow teens (Mean = 3.3). For all of the items in this scale, the agreement between parents and their teens was low (eg, on high speed roads: 35% agreed, without telling parent who will ride with her/him: 42.5% agreed). The items with the lowest degree of agreement, according to weighted Kappa (κ_w), were those related to asking permission (κ_w = 0.07) and telling a parent when he/she will return (κ_w = 0.05). The highest agreement, although still poor, were items related to driving at night: between 9pm and midnight (κ_w = 0.23) and after midnight (κ_w = 0.24; Table 3).

Teens expected to drive unsupervised more frequently to different destinations (eg, to or from work, to go to friends’ houses) than their parents expected for their teens. Thus, agreement between parents and teens about frequency of driving unsupervised was low (Table 3). For example, on average, teens expected driving to run family errands more frequently than their parents (Mean = 3.8 and Mean = 3.2 for teens and parents, respectively). Running family errands also had the lowest degree of agreement between parents and teens (κ_w = 0.05). Driving unsupervised to or from school had the highest agreement, although agreement was still rather low (κ_w = 0.25).