Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jun 1.
Published in final edited form as: J Transp Health. 2021 Apr 16;21:101068. doi: 10.1016/j.jth.2021.101068

Associations Between Graduated Driver Licensing Restrictions and Delay in Driving Licensure Among U.S. High School Students

Federico E Vaca 1, Kaigang Li 1,2,3, James C Fell 4, Denise L Haynie 5, Bruce Simons-Morton 5, Eduardo Romano 6
PMCID: PMC8128141  NIHMSID: NIHMS1696855  PMID: 34012771

Abstract

Introduction:

Some of the most vulnerable groups of teens choose to delay driving licensure (DDL). We assessed longitudinal associations between state-level Graduated Driver Licensing (GDL) restrictions and DDL among U.S. high school students.

Methods:

Data from seven waves of the NEXT Generation Health Study (starting 10th-grade (2009–2010)), were analyzed in 2020 using Poisson regression. The outcome was DDL (delay vs. no-delay). Independent variables were driving restrictions (at learner and intermediate phases of licensure), sex, race/ethnicity, family affluence, parent education, family structure, and urbanicity.

Results:

Of 2525 eligible for licensure, 887 (38.9%), 1078 (30.4%), 560 (30.7%) reported DDL 1–2 years, >2 years, no DDL, respectively. Interactions between GDL restrictions during the learner permit period and covariates were found. In states requiring ≥30 hours of supervised practice driving, Latinos (Adjusted relative risk ratio [aRRR]=1.55, p<.001) and Blacks (aRRR=1.38, p<.01) were more likely to DDL than Whites. In states where permit holding periods were <6 months, participants with low (aRRR=1.61, p<.001) and moderate (aRRR=1.45, p<.001) vs. high affluence were more likely to DDL. Participants in single-parent households vs. both-biological parent households were also more likely to DDL (aRRR=1.37, p<.05). In states where permit holding periods were ≥6 months, participants with low (aRRR=1.33, p<.05) vs. high affluence were more likely to DDL. In states that allowed ≥3 passengers or no passenger restriction, participants living in non-urban vs. urban (aRRR=1.52, p<.05) areas were more likely to DDL, and in states that allowed only 1 or no passenger, participants living in non-urban vs. urban areas (aRRR=0.67, p<.001) were less likely to DDL.

Conclusions:

Our findings heighten concerns about increased crash risk among older teens who age out of state GDL policies thereby circumventing driver safety related restrictions. Significant disparities in DDL exist among more vulnerable teens in states with stricter GDL driving restrictions.

Keywords: delay in driving licensure, disparities, graduated driver licensing, longitudinal, teen driver, novice drivers

INTRODUCTION

Graduated Driver Licensing (GDL) programs were put in place to introduce new young drivers gradually and safely into the driving population, gaining needed driving experience before obtaining full independent driving privileges (Foss, Feaganes, & Rodgman, 2001; Governors Highway Safety Association, n.d.). It has been well documented that comprehensive GDL programs are associated with reduced fatal crash involvements by 20 to 40% among teen drivers in the U.S. (Chen, Baker, & Li, 2006; Shope, 2007; Williams, 2007; Zhu, Chu, & Li, 2009; Zhu, Cummings, Chu, Coben, & Li, 2013) as well as in other countries such as Australia, New Zealand, Northern Ireland, and Canada (Begg & Stephenson, 2003; Christie, Steinbach, Green, Mullan, & Prior, 2017; Scott-Parker, 2016). In the U.S., every state has a GDL program with three stages, a learner’s permit, intermediate license, and unrestricted license (Centers for Disease Control and Prevention (CDC), 2016; Masten, Foss, & Marshall, 2013). In 1996, the adoption of GDL policies in all states began with the incorporation of additional novice driver requirements and restrictions (e.g., mandated supervised practice driving hours, limiting night time driving and number of passengers) during the learning period (Williams, McCartt, & Sims, 2016).

Across the U.S., the minimum age to obtain a driver permit can range from 14 to 16 years old, a restricted license from 14 to 17 years old, and a full license typically from 16.5 to 18 years old (Williams et al., 2016; Williams, Tefft, & Grabowski, 2012b). Previous research has shown that GDL laws requiring older minimum licensing age were associated with reduced crash rates among young drivers (McCartt, Teoh, Fields, Braitman, & Hellinga, 2010). This has been attributed to an age effect in several studies (Mayhew, Simpson, & Pak, 2003; McCartt, Mayhew, Braitman, Ferguson, & Simpson, 2009). Alternatively, extended permit holding periods try to increase safety by providing youth with greater and longer opportunity for driving exposure (i.e., driving practice). As of 2010, 47 states require new drivers to hold a permit ranging from 6 to 12 months before obtaining a license (Williams, Tefft, & Grabowski, 2012a). A five-year survival analysis study conducted in North Carolina found a permit holding period of 9 to 12 months to reduce crash rates among teen drivers (Masten & Foss, 2010). Furthermore, a New Zealand study found young drivers who gained a license after 18 months had significantly less crash involvement compared to drivers who gained their license after a holding period of 12 to 18 months (Lewis-Evans, 2010). Even so, a study analyzing collision claim frequencies in young drivers found that mandatory holding periods for a learner’s permit did not reduce crash risk once they were licensed when age was not taken into consideration (Trempel, 2009). This suggests a reduction in crash frequency could be due to drivers simply being older or greater pre-independent licensure driving experience. Generally, most states require 40 to 50 hours of supervised practice driving for the young driver by a parent or guardian as a part of their GDL program (Foss, Masten, Goodwin, O’brien, & TransAnalytics, 2012). However, this is a source of controversy given that adult supervisors may choose to “sign off’ on the total number of required practice-driving hours as completed even if they were not. A naturalistic driving study in Virginia found 54% of young drivers did not truly complete all supervised hours before obtaining their driver’s license (Ehsani et al., 2017). It is possible that low adherence to required supervised practice driving hours contributed to research findings of no independent association between fatal crash rates among young drivers and increasing required practice hours (Masten, Thomas, Korbelak, Peck, & Blomberg, 2015; McCartt et al., 2010).

In the intermediate license stage, common GDL restrictions include passenger and nighttime driving restrictions. These restrictions are in place to allow young drivers to continue to practice driving and continue to gain experience mainly under lower crash risk conditions. Particularly for young drivers, there is a significant increase in more severe and fatal when passengers are present (Ouimet et al., 2015). Teenage passengers seem to pose the biggest risk to a vehicle crash when the driver is a teenager (Tefft, Williams, & Grabowski, 2012). The negative effect is often attributed to increased distraction from teenage passenger behaviors such as talking loudly, and sudden physical movements (e.g., “rough housing”). Recent research has found passenger restrictions to decrease teenage crash involvement (Masten et al., 2015; McCartt & Teoh, 2015). Nighttime driving restrictions have also been found to reduce crashes and injuries among teenage drivers (McCartt et al., 2010; Williams et al., 2012b). Approximately one third of 16-to-17 y/o drivers involved in a crash in the U.S. between the hours of 9:00 pm and 5:59 am (Shults, 2016). This high crash risk is often attributed to drowsy driving and alcohol use during the later night to early morning hours.

Recently, the rate of teens that delay driving licensure (DDL) has increased. Between 1996 and 2010 the percentage of U.S. high school seniors with a driver’s license decreased from 83% to 73% (Shults & Williams, 2013). This trend is concerning as GDL programs have been established to decrease overall crash risk in novice drivers but most of these programs only extend to drivers at the age of 18 years and younger (Curry, Metzger, Williams, Tefft, & Foss, 2018). As of 2017, only 7 states required some GDL restrictions for novice drivers ≥18 years old (Curry, Foss, & Williams, 2017). A study conducted in New Jersey found that older novice drivers (i.e., ≥ 21years) experience fewer crashes compared to younger novice drivers (age 17 to 20 years) under GDL policies (Curry, Metzger, Williams, & Tefft, 2017). However, another study found at-fault crashes to be positively associated with driver age at the time of licensure (Waller, Elliott, Shope, Raghunathan, & Little, 2001), so there are inconsistencies in the findings. While DDL (i.e., generally referring to drivers that age-out of state GDL policy) may lead to less driver crash risk due to the reduced driving exposure and restricting driving under high-risk conditions (Hartos, Eitel, & Simons-Morton, 2001), this reduction may be reversed later among youth that DDL and end up with insufficient driving exposure, experience, and instruction (Tefft, Williams, & Grabowski, 2013) because of missing important GDL safety practices. Black and Latino youth engage in DDL at higher rates than their White counterparts (Shults & Williams, 2013; Tefft, Williams, & Grabowski, 2014). Furthermore, youth with low income backgrounds are more likely to delay licensure compared to high income youth (Curry, Pfeiffer, Durbin, Elliott, & Kim, 2015; Tefft et al., 2014; Thigpen & Handy, 2018).

Currently, there is conflicting evidence on how GDL programs may be associated with DDL. A retrospective case study (Thigpen & Handy, 2018) investigated the factors influencing the increasing DDL among young adults and found that multiple factors including travel attributes, driver attitudes, parental influences, and GDL policies, have considerable influence on license possession and timing. Other studies have found GDL to have a slight influence on DDL among youth (Tefft et al., 2014). On the other hand, Australian researchers have found that rates of DDL had been increasing prior to the establishment of their GDL programs, with no significant increases after GDL was established (Raimond & Milthorpe, 2010). Given the mixed findings, the purpose of this study was to assess longitudinal associations of state-level GDL restrictions with DDL among U.S. high school students.

METHODS

DESIGN, SETTING, and PARTICIPANTS

Design

Data used in this project were from all seven waves (W1–7) of the longitudinal NEXT Generation Health Study (NEXT), a longitudinal study that followed a nationally representative cohort of U.S. 10th grade students (mean age = 16.3 y/o (se=0.03)) adolescents into emerging adulthood. Additional information about the NEXT study can be found elsewhere (Li, SimonsMorton, & Hingson, 2013; Vaca et al., 2020).

Setting

This study used primary sampling units (PSU, school districts) from the nine U.S. census divisions including 22 states: CA, CO, CT, FL, GA, ID, IL, KY, LA, MA, MI, MN, NV, NY, OH, OR, PA, SD, TN, TX, WI, and WV. Schools and classrooms were randomly sampled from the PSUs.

Participants

A total of 2785 students from 81 high schools participated in the NEXT study. W1 data collection began in the 2009–2010 school year and continued to survey participants yearly until W7 in 2016. From W1-W7, 91%, 88%, 86%, 78%, 79%, 84%, and 83% of the full sample (N=2785) completed the survey during the springtime each year. Parental consent and adolescent assent were obtained. When participants turned 18 years of age, their consent was obtained. The protocol for the study was approved by the Institutional Review Board of the Eunice Kennedy Shriver National Institute of Child Health and Human Development. Analyses were completed in 2020.

MEASURES

Outcome Variable

The dichotomous outcome variable was delayed driving licensure (DDL). At each annual assessment the NEXT survey repeated the following question, “Do you have a driver’s license?”, with the possible responses being, No license of any sort; Permit to take the classroom component driver education only; Permit allowing supervised practice driving with an instructor or licensed adult; License allowing independent, unsupervised driving (with or without temporary restrictions on late night driving, teen passengers, etc.). In this study, we only included participants who had a license allowing independent-unsupervised driving, and DDL is defined as the time in years elapsed between the year that a participant was legally eligible to be licensed and the age they were licensed. Specifically, DDL was calculated as year(s) elapsed between participant’s eligible year and licensed year, or eligible year and final assessment if not licensed and then dichotomized (No Delay vs. Delay).

Independent Variables

GDL Restrictions

We used state-level GDL law details per the Governors Highways Safety Association (Governors Highway Safety Association, n.d.) for the period when NEXT data were collected and GDL law strength was determined by a point-rating system where key safety-related licensing components were scored as described by McCartt et al. (McCartt et al., 2010). GDL components were coded such that zero points indicate no or low GDL restrictions and one or more points indicates higher restriction. Restrictions on the learner’s phase were coded as: minimum permit age (0 points=less than 16 years, 1 point=16 years or older), permit holding period (0 points=less than 6 months, 1 point=6 or more months), and permit required practice driving hours (0 points=less than 30 hours, 1 point=30 or more hours). Restrictions in the intermediate phase were coded as: limits on night driving points (0 points=no restriction, 1 point =after 10 p.m., 2 points=10 p.m. or earlier) and limits on young (age of 18 to 25 or younger, depending on state) passenger(s) allowed (0 points=3 or more passengers or no restriction, 1 point=2 passengers, 2 points=zero or 1 passenger).

Covariates

Covariates assessed in this study included sex (male vs. female), race/ethnicity (Latino, Non-Latino Blacks, others vs. Non-Latino Whites), family affluence, parental education, and onset age of initial alcohol use.

Family Affluence

Family socioeconomic status (SES) was estimated using the validated Family Affluence Scale (Currie et al., 2004) including number of vehicles (i.e., including car, van, and/or truck) owned by family (none = 0, one = 1, and ≥2 = 2), number of computers owned by family (none=0, one=1, two=2, >2=2), whether the student had his/her own bedroom (no=0, yes=1), and the number of family vacations in the last 12 months (not at all=0, once=1, twice=2, >twice=2). A composite score was calculated for each young person based on his or her responses to these four items and students were then categorized as low (score= 0–3), moderate (score=4, 5) and high affluence (score=6, 7) (Spriggs, Iannotti, Nansel, & Haynie, 2007).

Family Structure

Students’ family structure was categorized as: both biological parents; one biological parent, one stepparent; single parent, mother only; single parent, father only; and other.

Parental Education

We categorized the higher education level of both parents as less than high school diploma; high school diploma or general equivalency diploma (GED); some college, technical school, or associate degree; and bachelor’s or graduate degree.

Onset Age of Initial Alcohol Use

Onset age of initial alcohol use was measure by asking participants, “How old were you the first time you had a drink of an alcoholic beverage? (Please do not include any time when you only had a sip or two from a drink.)” at W1 to W7. The reported year at the earliest wave was used. Then, the participants were categorized into two group (≥14 vs. <14 y/o).

Urbanicity

Urbanicity was defined as participants’ school location at W1 using a seven-level scale ranging from metropolitan city to rural region. Those attending schools in an urban region were categorized as urban, and remaining categories were classified as non-urban (i.e., suburban/rural).

Statistical Analysis

Due to the fact that the DDL is a common (vs. rare) variable, Poisson regressions were conducted to assess the associations between the DDL and each GDL restriction (Cummings, 2009; Greenland, 2004). Valid confident intervals (CIs) were estimated with a robust estimator of variance (Wolter, 2007) and risk ratios (i.e., exponentiated coeffects) were estimated to indicate the risk (prevalence ratio) of DDL compared with risk (prevalence ratio) of no DDL (Cummings, 2009). Separate Poisson models were run for each demographic variable (race/ethnicity, sex, family affluence, family structure, parent education, and onset age of alcohol use) that included the interaction term of each demographic variable by each GDL restriction, controlling for other demographic variables. The association between DDL and demographic variables by GDL restrictions were reported (Table 3) if the interaction was statistically significant. Interaction contracts were used to examine the significant associations; for example, given states that required 30 or more hours of supervised practice driving, likelihoods of DDL in Latinos and in Non-Latino Blacks compared to the likelihood of DDL in Non-Latino Whites (as the reference) was calculated respectively and denoted by adjusted relative risk ratio (aRRR). Missing data were deleted listwise. We performed the Poisson regressions with Stata version 16 (StataCorp LLC, TX) and other analyses with SAS version 9.4 (SAS Institute, Cary, NC) and accounted for features of the complex survey design (i.e., stratification, clustering, and sampling weights) in all procedures.

Table 3.

Association between DDL (delay vs no delay) and demographic variables by GDL restrictions

DDL % DDL IRR 95% CI p
Learner’s phase: Required practice hours Race/ethnicity <.001#
 < 30 hours  Latino vs. Non-Latino White 84.1 vs. 80.5 0.94 0.64 1.37 .73
 < 30 hours  Non-Latino Black vs. Non-Latino White 37.5 vs. 80.5 0.37 0.07 1.94 .22
 ≥ 30 hours  Latino vs. Non-Latino White 91.6 vs. 55.0 1.55 1.30 1.85 <.001
 ≥ 30 hours  Non-Latino Black vs. Non-Latino White 84.5 vs. 55.0 1.38 1.10 1.74 <.01
Learner’s phase: Permit holding period Family affluence <.001#
 < 6 months  Low Affluence vs. High Affluence 84.0 vs. 45.5 1.61 1.35 1.93 <.001
 < 6months  Moderate Affluence vs. High Affluence 68.4 vs. 45.5 1.45 1.29 1.63 <.001
 ≥ 6 months  Low Affluence vs. High Affluence 94.0 vs. 64.0 1.33 1.00 1.77 <.05
 ≥ 6 months  Moderate Affluence vs. High Affluence 71.6 vs. 64.0 1.08 0.87 1.33 .50
Learner’s phase: Permit holding period Family structure <.05#
 < 6 months  Biological/stepparent vs. Both Biological 83.4 vs. 77.6 0.92 0.75 1.13 .41
 < 6 months  Single Parent Vs. Both Biological 85.5 vs. 77.6 1.37 1.00 1.90 .05
 ≥ 6 months  Biological/stepparent vs. Both Biological 66.2 vs. 59.8 1.11 0.95 1.28 .17
 ≥ 6 months  Single Parent Vs. Both Biological 79.4 vs. 59.8 1.17 0.74 1.84 .48
Intermediate phase restriction on young passenger(s) Urbanicity <.001#
 ≥ 3 passengers or no restriction  Non-urban vs. Urban 97.0 vs. 66.9 1.52 1.06 2.19 <.05
 2 Passengers  Non-urban vs. Urban 69.0 vs. 87.7 0.85 0.64 1.14 .26
 Zero or 1 passenger  Non-urban vs. Urban 60.3 vs. 97.6 0.67 0.58 0.78 <.001
Open in a new tab
#

indicate the overall p-value for the interaction.

IRR: incidence rate ratios. Interactions between demographic variables and restrictions were examined. For significant interactions, the association between DDL and demographic variables by level of GDL restriction are reported in this table.

RESULTS

The sample characteristics overall and by the outcome variable (DDL and no DDL) are shown in Table 1. Of 2525 eligible for licensure during the study period, 887 (38.9%) reported DDL by 1–2 years, 1078 (30.4%) by ≥2 years, and 560 (30.7%) reported no DDL (data not shown).

Table 1.

Sample characteristics

Overall DDL No DDL
N Weighted % N Weighted % N Weighted %
Race/ethnicity
 Latino 835 19.3 684 89.9 42 10.1
 Non-Latino Black 687 20.2 567 83.7 57 16.3
 Non-Latino White 1106 55.7 608 57.7 438 42.3
 Other 142 4.8 100 62.2 23 37.8
Sex
 Male 1255 45.5 862 68.6 257 31.4
 Female 1525 54.5 1103 69.9 303 30.1
Family affluence
 Low 920 24.9 732 87.2 60 12.8
 Moderate 1285 48.5 901 68.9 285 31.1
 High 578 26.6 332 55.2 215 44.8
Family structure
 Both biological parents 1320 52.2 871 62.1 359 37.9
 Biological & stepparent 419 19.2 282 67.7 85 32.3
 Single parent 516 19.5 383 79.6 74 20.4
 Other 269 9.2 211 82.8 24 17.2
Highest education of either parent
 High school or less 979 33.1 748 80.8 105 19.2
 Some college 924 39.8 633 66.6 229 33.4
 Bachelor or higher 626 27.1 387 58.6 202 41.4
Age at onset of alcohol use
 ≥14 years old 1745 76.1 1262 67.1 404 32.9
 <14 years old 508 23.9 387 70.9 97 29.1
Urbanicity
 Urban 905 13.9 729 83.0 61 17.0
 Non-urban 1629 86.1 1020 66.0 481 34.0
Open in a new tab

CI: confidence interval; the sum of Ns of DDL and no DDL is not equal to the overall N due to the missing data of DDL variable.

As shown in Table 2 of the eligible NEXT sample (N=2525), the percentage of participants residing in a state with learner phase restrictions was 20.1% (weighted, hereafter) for ≥16 yrs. minimum age, 17.9% for ≥6 month holding period, and 88.6% for ≥30 hours required practice. For intermediate phase restrictions, 20.2% of participants had a night driving limit of no later than 10 PM, and 22.7% and 63.8% had limits on young passengers of ≤2 and ≤ 1, respectively. The percentages of participants in states with different GDL restrictions by the outcome variable (DDL vs. no DDL) are also showed in Table 2.

Table 2.

Prevalence of DDL and GDL restrictions in NEXT participants

Overall DDL No DDL
N Weighted % N Weighted % N Weighted %
DDL
 No DDL 560 30.7 - - - -
 DDL 1965 69.3 - - - -
Learner’s phase minimum permit age
 < 16 years 2110 79.9 1664 69.9 446 30.1
 ≥ 16 years 415 20.1 301 66.9 114 33.1
Learner’s phase permit holding period
 < 6 months 1885 82.1 1393 67.2 492 32.8
 ≥ 6 months 640 17.9 572 78.7 68 21.3
Learner’s phase permit required practice hours
 < 30 hours 155 11.4 134 79.7 21 20.3
 ≥ 30 hours 2370 88.6 1831 67.9 539 32.1
Intermediate phase restriction on night driving
 After 10 p.m. 1940 79.8 1436 65.0 504 35.0
 10 p.m. or earlier 585 20.2 529 86.4 56 13.6
Intermediate phase restriction on young passenger(s)
 ≥ 3 passengers or no restriction 555 13.5 492 83.7 63 16.3
 2 Passengers 206 22.7 142 69.7 64 30.3
 Zero or 1 passenger 1764 63.8 1331 66.1 433 33.9
Open in a new tab

As shown in Table 3, significant interactions between GDL restrictions at the learner’s phase and covariates included race/ethnicity and required practice hours; family affluence and permit holding period; and family structure and permit holding period. Significant interaction between GDL restrictions at the intermediate phase was found between urbanicity and restrictions on young passenger(s). In states that required 30+ hours of supervised practice driving, Latinos (aRRR=1.55, 95%CI: 1.30, 1.85, p<.001) and Non-Latino Blacks (aRRR=1.38, 95%CI: 1.10, 1.74, p<.01) were more likely to DDL vs. Non-Latino Whites.

In states that required a permit holding period of <6 month, participants with low (aRRR=1.61, 95%CI: 1.35, 1.91, p<.001) and moderate (aRRR=1.45, 95%CI: 1.29, 1.63, p<.001) affluence were more likely to DDL vs. those in high affluence. In states that required a permit holding period of ≥6 month, participants with low affluence (aRRR=1.33, 95%CI: 1.00, 1.77, p<.05) were more likely to DDL vs. those with high affluence. In states that required a permit holding period of <6 months, participants in single parent households (aRRR=1.37, 95%CI: 1.00, 1.90, p<.05) were more likely to DDL vs. those living in households with both biological parents.

In states that allowed ≥ 3 passengers or had no passenger restrictions, participants living in non-urban areas (aRRR=1.52, 95%CI: 1.06, 2.19, p<.05) were more likely to DDL vs. those in urban areas, and in states that allowed only 1 passenger or no passengers, participants living in non-urban areas (aRRR=0.67, 95%CI: 0.58, 0.78, p<.001) were less likely to DDL vs. those in urban area.

Supplementary Table 1 shows bivariate association of DDL (DDL vs. no DDL) with demographic variables and GDL restrictions.

DISCUSSION

This study examined the associations of DDL among U.S. high school students with state-level GDL restrictions. Our findings are consistent with previous research showing Black and Latino youth are more likely than their White counterparts to delay licensure (Shults & Williams, 2013; Tefft et al., 2014; Williams, 2017) and additionally showed differences by SES.

Our analysis finds that the associations of race/ethnicity and DDL differed by states’ requirements for hours of supervised practice driving. Thus, race/ethnicity was differentially associated with DDL depending on the level of GDL restriction, suggesting that more restrictive GDL may be an impediment to licensure among more vulnerable teens.

Several family attributes were found to be significantly associated with GDL factors in our analysis. For instance, in states where the minimum holding period for a learner’s permit was <6 months, participants in single parent households were more likely to DDL compared to participants living in a household with both biological parents. The learner’s stage is designed to increase driving practice in as safe of an environment as possible. This includes supervised driving often completed by parents (Williams et al., 2012b). In a single parent compared to a two-parent household, potentially more limited time and vehicle availability/access could make it considerably more difficult to provide dedicated supervised practice driving. Further, it is conceivable that a single parent may find it more challenging to confidently assess a teens’ “developmental” readiness for driving at a younger age (i.e., ≤ 16 years) as compared to two parents making a collective-shared decision. Some parents may view obtaining a permit at these earlier ages to be unnecessary, inconvenient, fraught with parenting challenges and/or economic strains, and, ultimately, a developmental milestone that can wait (Simons-Morton & Hartos, 2003).

Another family attribute which was significantly associated with GDL factors in our analysis was family affluence. As suggested in previous studies, economic circumstances are inextricably tied to licensure such that family economic burdens may result in the delay of licensure (Schoettle & Sivak, 2014; Shults & Williams, 2013). We found that participants with low family affluence were more likely to DDL compared to those with high family affluence no matter the length of the holding period. This could possibly be due to the financial demands of vehicle ownership (Tefft et al., 2013). Moreover, previous research in the U.S. covering 1996–2010 showed that the U.S. economic recession (December 2007 – June 2009) led to reductions of high school senior licensure rates (Shults & Williams, 2013). In this study, low-income status showed similar effects on DDL (i.e., greater likelihood of DDL). Therefore, the economic cost of licensure may be a dominant factor for youth and their families with low affluence as they determine the timing of licensure.

The geographical attribute was differentially associated with DDL depending on the level of GDL restriction at the intermediate phase. In states where GDL was less restrictive (i.e., allowing 3 young passengers or no restriction on young passengers), non-urban teens were more likely to DDL compared to urban teens. In contrast, when comparing non-urban to urban teens living in states with more restrictive GDL (i.e., allowing only 1 passenger or no passengers), non-urban teens were less likely DDL. We speculate, that in general, non-urban families could have a higher financial burden to bear compared to urban families when it comes to teen licensure. As a result, the cost of licensure and associated vehicle costs (e.g., gas, registration, insurance) may heighten the burden of family finances among non-urban teens, resulting in DDL. In the states that have lower passenger restrictions, teens may choose to ride with peers which may reduce the immediate necessity for their own licensure. Coupling the financial burden of licensure along with an attenuated necessity of licensure, given the availability of carpooling with a driving peer, could facilitate an increase in DDL among non-urban adolescents. In contrast, despite the financial strain, some non-urban adolescents in states that allowed only 1 or no passenger may not be able to ride with peers due to strict passenger restrictions. Consequently, these teens may be forced to pursue timely licensure, avoiding DDL, so they can meet important transportation needs (e.g., school, sport/social activities, employment). More research is needed to explore this area and to bring greater clarity to these scenarios.

Graduated Driver Licensing (GDL) programs in different countries have been effective in reducing crashes and injuries among young drivers, where licensure is obtained at older ages (Begg & Stephenson, 2003; Christie et al., 2017; Senserrick & Williams, 2015; Williams, 2017). Given that GDL in the U.S. has mainly applied to those under 18 years of age, it is unclear if GDL could influence crash risk if more widely applied to older novice drivers. A recent study conducted in the state of Queensland, Australia, evaluated the effect of a strengthened GDL program that applied some GDL requirements to those under 25 years found substantial declines in all novice driver crashes and serious crashes as well as the resulting deaths and/or injuries (Senserrick, Boufous, Olivier, & Hatfield, 2018). These findings argue that requiring GDL of those at a later licensing age improve novice drivers’ performance, increase experience, and reduce fatal crashes. However, despite these implications, in the U.S. we know little of how an extended-age GDL program might affect differences and disparities among not only DDL but also definitively among unlicensed groups of older novice drivers.

Finally, while not evaluated in this study, an additional area that we believe is worthy of highlighting in the context of DDL is state level requirements for driver education. For some teens and their families, given their contextual setting (i.e., state residence, household income), driver education requirements may influence decisions related to DDL. Taking into consideration that enrolling and participating in a driver education course takes time and money, it’s probable that such a requirement could have a disproportionate impact on lower SES teens (e.g., consideration of interaction effects between affluence, race/ethnicity, family structure). A recent U.S. survey of young adults ages 18–24 years provided some suggestive evidence of the impact of driver education requirements on DDL. In this U.S. representative sample study, there were similar licensure rates for those youth that became licensed at ≤16 years regardless of state-specific driver education requirements. However, licensure rates at age 17 years were found to be higher in states that did not require completion of a driver education course before the new driver could be licensed. Further, in states that required a driver education course as prerequisite for licensure for youth age <18 years, but not for those new drivers ≥18 years, licensure at ≥ 18 years was more common. Interestingly, the study found that the largest proportion of “not- yet-licensed” new drivers were those that lived in states that required the completion of a driver education course for drivers ≥18 years. (Tefft & Foss, 2019). While we firmly believe that driver education is of major importance for novice drivers of any age, we also believe that additional study of the effects of state-level driver education course requirements on DDL would be beneficial in understanding identified disparities in delayed licensure.

We recognize that our study has limitations. First, the school-based recruitment limits the generalization to youth in school at 10th grade but not those who were not enrolled in school. Second, it would have been ideal to know the exact date on which participants obtained their driver’s license so that DDL could be more accurately estimated. However, these data were available in the NEXT survey data. As a result, we were limited to calculating and approximating the DDL variable. Third, only a limited number of covariates were collected and analyzed. This may exclude other important factors influencing DDL (e.g., youth time conflict, parents’ schedule, youth needs for transportation). Fourth, although the influence of economic factors is acknowledged, the information on the overall cost of licensing in different states is missing. Therefore, it is unclear whether and/or how a more accurate state-based economic burden measure (e.g., state-level licensing costs) may influence DDL.

CONCLUSION

Differences and disparities in DDL among racial/ethnic groups exist. This could be limiting the beneficial GDL safety effects of teen novice drivers with potentially more negative results among more vulnerable youth groups. Our findings cohere with previous research attributing some part of DDL to the economics of licensure. Furthermore, important family level factors were related to DDL. Additional investigation is warranted to disentangle the complexity of factors contributing to DDL. Advancing our understanding of disparities in DDL could inform policy and enhance efforts to further prevent injury among more vulnerable teen groups.

Supplementary Material

1

Highlights.

  • Graduated Driver Licensing (GDL) programs are associated with reduced fatal crash

  • Some of the most vulnerable groups of teens choose to delay driving licensure (DDL)

  • Teens aging out of state GDL may circumvent driver safety related restrictions

  • Significant disparities in DDL exist in states with stricter GDL driving restrictions

Acknowledgments

NIAAA - R21

Research reported in this publication was supported by the National Institute on Alcohol Abuse and Alcoholism of the National Institutes of Health under Award Number R21AA026346. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

NICHD - NEXT Generation Health Study

This project (contract HHSN275201200001I) was supported in part by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development; the National Heart, Lung, and Blood Institute; the National Institute on Alcohol Abuse and Alcoholism; the National Institute on Drug Abuse; and the Maternal and Child Health Bureau of the Health Resources and Services Administration.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Begg D, & Stephenson S (2003). Graduated driver licensing: the New Zealand experience. J Saf Res, 34(1), 99–105. [DOI] [PubMed] [Google Scholar]
  2. Centers for Disease Control and Prevention (CDC), N. C. f. I. P. a. C. N. (2016). Graduated Driver Licensing. Centers for Disease Control and Prevention, Retrieved from https://www.cdc.gov/parentsarethekey/licensing/index.html [Google Scholar]
  3. Chen L-H, Baker SP, & Li G (2006). Graduated driver licensing programs and fatal crashes of 16-year-old drivers: a national evaluation. Pediatrics, 118(1), 56–62. [DOI] [PubMed] [Google Scholar]
  4. Christie N, Steinbach R, Green J, Mullan MP, & Prior L (2017). Pathways linking car transport for young adults and the public health in Northern Ireland: a qualitative study to inform the evaluation of graduated driver licensing. BMC Publ Health, 17(1), 551. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cummings P (2009). The relative merits of risk ratios and odds ratios. Archives of pediatrics & adolescent medicine, 163(5), 438–445. [DOI] [PubMed] [Google Scholar]
  6. Currie C, Roberts C, Settertobulte W, Morgan A, Smith R, Samdal O, . . . Organization WH (2004). Young people’s health in context: Health Behaviour in School-aged Children (HBSC) study: international report from the 2001/2002 survey: Copenhagen: WHO Regional Office for Europe. [Google Scholar]
  7. Curry AE, Foss RD, & Williams AF (2017). Graduated driver licensing for older novice drivers: critical analysis of the issues. Am JPrevMed, 53(6), 923–927. [DOI] [PubMed] [Google Scholar]
  8. Curry AE, Metzger KB, Williams AF, & Tefft BC (2017). Comparison of older and younger novice driver crash rates: Informing the need for extended Graduated Driver Licensing restrictions. Accid Anal Prev, 108, 66–73. [DOI] [PubMed] [Google Scholar]
  9. Curry AE, Metzger KB, Williams AF, Tefft BC, & Foss RD (2018). PW 2529 Extending graduated driver licensing policy to older novice drivers: a critical analysis of current evidence. In: BMJ Publishing Group Ltd. [Google Scholar]
  10. Curry AE, Pfeiffer MR, Durbin DR, Elliott MR, & Kim KH (2015). Young driver licensing: examination of population-level rates using New Jersey’s state licensing database. Accid Anal Prev, 76, 49–56. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/25590921.doi: 10.1016/j.aap.2014.12.022 [DOI] [PubMed] [Google Scholar]
  11. Ehsani JP, Klauer S, Zhu C, Gershon P, Dingus T, & Simons-Morton B (2017). Naturalistic assessment of the learner license period. Accid Anal Prev, 106, 275–284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Foss RD, Feaganes JR, & Rodgman EA (2001). Initial effects of graduated driver licensing on 16-year-old driver crashes in North Carolina. Jama, 286(13), 1588–1592. [DOI] [PubMed] [Google Scholar]
  13. Foss RD, Masten SV, Goodwin AH, O’brien NP, & TransAnalytics L (2012). The role of supervised driving requirements in graduated driver licensing programs. Retrieved from
  14. Governors Highway Safety Association (n.d.). Teen and Novice Drivers Retrieved from https://www.ghsa.org/state-laws/issues/teen%20and%20novice%20drivers
  15. Greenland S (2004). Model-based estimation of relative risks and other epidemiologic measures in studies of common outcomes and in case-control studies. American journal of epidemiology, 160(4), 301–305. [DOI] [PubMed] [Google Scholar]
  16. Hartos JL, Eitel P, & Simons-Morton B (2001). Do parent-imposed delayed licensure and restricted driving reduce risky driving behaviors among newly licensed teens? Prevention Science, 2(2), 113–122. [DOI] [PubMed] [Google Scholar]
  17. Lewis-Evans B (2010). Crash involvement during the different phases of the New Zealand Graduated Driver Licensing System (GDLS). J Saf Res, 41(4), 359–365. [DOI] [PubMed] [Google Scholar]
  18. Li K, Simons-Morton BG, & Hingson R (2013). Impaired-driving prevalence among US high school students: Associations with substance use and risky driving behaviors. American journal of public health, 103(11), e71–e77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Masten SV, & Foss RD (2010). Long-term effect of the North Carolina graduated driver licensing system on licensed driver crash incidence: A 5-year survival analysis. Accid Anal Prev, 42(6), 1647–1652. [DOI] [PubMed] [Google Scholar]
  20. Masten SV, Foss RD, & Marshall SW (2013). Graduated driver licensing program component calibrations and their association with fatal crash involvement. Accident Analysis & Prevention, 57, 105–113. [DOI] [PubMed] [Google Scholar]
  21. Masten SV, Thomas FD, Korbelak KT, Peck RC, & Blomberg RD (2015). Meta-Analysis of Graduated Driver Licensing Laws. Retrieved from
  22. Mayhew DR, Simpson HM, & Pak A (2003). Changes in collision rates among novice drivers during the first months of driving. Accid Anal Prev, 35(5), 683–691. [DOI] [PubMed] [Google Scholar]
  23. McCartt AT, Mayhew DR, Braitman KA, Ferguson SA, & Simpson HM (2009). Effects of age and experience on young driver crashes: review of recent literature. Traffic injury prevention, 10(3), 209–219. [DOI] [PubMed] [Google Scholar]
  24. McCartt AT, & Teoh ER (2015). Tracking progress in teenage driver crash risk in the United States since the advent of graduated driver licensing programs. J Saf Res, 53, 1–9. [DOI] [PubMed] [Google Scholar]
  25. McCartt AT, Teoh ER, Fields M, Braitman KA, & Hellinga LA (2010). Graduated licensing laws and fatal crashes of teenage drivers: a national study. Traffic Inj Prev, 11(3), 240–248. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/20544567. doi: 10.1080/15389580903578854 [DOI] [PubMed] [Google Scholar]
  26. Ouimet MC, Pradhan AK, Brooks-Russell A, Ehsani JP, Berbiche D, & Simons-Morton BG (2015). Young drivers and their passengers: a systematic review of epidemiological studies on crash risk. J Adolesc Health, 57(1), S24–S35. e26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Raimond T, & Milthorpe F (2010). Why are young people driving less? Trends in licence-holding and travel behaviour. Paper presented at the Proceedings of Australasian Transport Research Forum. [Google Scholar]
  28. Schoettle B, & Sivak M (2014). The reasons for the recent decline in young driver licensing in the United States. Traffic injury prevention, 15(1), 6–9. [DOI] [PubMed] [Google Scholar]
  29. Scott-Parker B (2016). Review of the graduated driver licensing programs in Australasia. Journal of the Australasian College of Road Safety, 27(4), 15. [Google Scholar]
  30. Senserrick T, Boufous S, Olivier J, & Hatfield J (2018). Associations between graduated driver licensing and road trauma reductions in a later licensing age jurisdiction: Queensland, Australia. PLoS one, 13(9), e0204107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Senserrick T, & Williams A (2015). Summary of literature of the effective components of graduated driver licensing systems. 2020(August 19). [Google Scholar]
  32. Shope JT (2007). Graduated driver licensing: review of evaluation results since 2002. J Saf Res, 38(2), 165–175. [DOI] [PubMed] [Google Scholar]
  33. Shults RA (2016). Graduated driver licensing night driving restrictions and drivers aged 16 or 17 years involved in fatal night crashes—United States, 2009–2014. MMWR. Morbidity and mortality weekly report, 65. [DOI] [PubMed] [Google Scholar]
  34. Shults RA, & Williams AF (2013). Trends in driver licensing status and driving among high school seniors in the United States, 1996–2010. J Safety Res, 46, 167–170. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/23932698. doi: 10.1016/j.jsr.2013.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Simons-Morton BG, & Hartos JL (2003). How well do parents manage young driver crash risks? Journal of safety research, 34(1), 91–97. [DOI] [PubMed] [Google Scholar]
  36. Spriggs AL, Iannotti RJ, Nansel TR, & Haynie DL (2007). Adolescent bullying involvement and perceived family, peer and school relations: Commonalities and differences across race/ethnicity. JAdolescHealth, 41(3), 283–293. doi : 10.1016/j.jadohealth.2007.04.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Tefft B, & Foss R (2019). Prevalence and Timing of Driver Licensing Among Young Adults, United States, 2019 (Research Brief.). Washington, D.C: AAA Foundation for Traffic Safety. [Google Scholar]
  38. Tefft BC, Williams AF, & Grabowski JG (2012). Teen driver Risk in relation to Age and Number of Passengers. [DOI] [PubMed]
  39. Tefft BC, Williams AF, & Grabowski JG (2013). Timing of driver’s license acquisition and reasons for delay among young people in the United States, 2012. AAA Foundation for Traffic Safety, Retrieved from https://aaafoundation.org/wp-content/uploads/2018/01/TimingofDriversLicenseAcquisitionReport.pdf [Google Scholar]
  40. Tefft BC, Williams AF, & Grabowski JG (2014). Driver licensing and reasons for delaying licensure among young adults ages 18–20, United States, 2012. Injury epidemiology, 1(1), 4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Thigpen C, & Handy S (2018). Driver’s licensing delay: A retrospective case study of the impact of attitudes, parental and social influences, and intergenerational differences. Transport Res Pol Pract, 111, 24–40. doi: 10.1016/j.tra.2018.03.002 [DOI] [Google Scholar]
  42. Trempel RE (2009). Graduated driver licensing laws and insurance collision claim frequencies of teenage drivers. Arlington, USA: Highway Loss Data Institute. [Google Scholar]
  43. Vaca FE, Li K, Luk JW, Hingson RW, Haynie DL, & Simons-Morton BG (2020). Longitudinal associations of 12th-grade binge drinking with risky driving and high-risk drinking. Pediatrics, 145(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Waller PF, Elliott MR, Shope JT, Raghunathan TE, & Little RJ (2001). Changes in young adult offense and crash patterns over time. Accid Anal Prev, 33(1), 117–128. [DOI] [PubMed] [Google Scholar]
  45. Williams AF (2007). Contribution of the components of graduated licensing to crash reductions. Journal of safety research, 38(2), 177–184. [DOI] [PubMed] [Google Scholar]
  46. Williams AF (2017). Graduated driver licensing (GDL) in the United States in 2016: A literature review and commentary. J Saf Res, 63, 29–41. [DOI] [PubMed] [Google Scholar]
  47. Williams AF, McCartt AT, & Sims LB (2016). History and current status of state graduated driver licensing (GDL) laws in the United States. J Safety Res, 56, 9–15. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/26875159. doi: 10.1016/jjsr.2015.11.006 [DOI] [PubMed] [Google Scholar]
  48. Williams AF, Tefft BC, & Grabowski JG (2012a). Graduated Driver Licensing Research Review, 2010-Present (November 2012). Prev Med, 42, 316–326. [DOI] [PubMed] [Google Scholar]
  49. Williams AF, Tefft BC, & Grabowski JG (2012b). Graduated driver licensing research, 2010-present. J Saf Res, 43(3), 195–203. [DOI] [PubMed] [Google Scholar]
  50. Wolter K (2007). Introduction to variance estimation: Springer Science & Business Media. [Google Scholar]
  51. Zhu M, Chu H, & Li G (2009). Effects of graduated driver licensing on licensure and traffic injury rates in Upstate New York. Accident Analysis & Prevention, 41(3), 531–535. [DOI] [PubMed] [Google Scholar]
  52. Zhu M, Cummings P, Chu H, Coben JH, & Li G (2013). Graduated driver licensing and motor vehicle crashes involving teenage drivers: an exploratory age-stratified meta-analysis. Injury prevention, 19(1), 49–57. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1696855-supplement-1.docx (116.1KB, docx)

RESOURCES