The drink and drug driving behaviours of young Queensland drivers and attitudes toward apprehension

Abstract

Objective

This study aimed to explore the self-reported drink and drug driving behaviors of a sample of young drivers from Queensland (Australia) holding a learner license (N = 162, 27%), provisional 1 (P1) license (N = 251, 41.8%), or provisional 2 (P2) license (N = 188, 31.3%), as well as motorists’ corresponding perceptions about the likelihood of apprehension and their future intentions to offend.

Methods

The study involved 601 drivers aged between 16 and 24 years (302 males, 50.2% and 296 females, 49.3%) who all completed an online survey.

Results

Almost half the sample acknowledged engaging in drink and/or drug driving behaviors in the previous 12 months (N = 264, 44%), with a total of 773 drink driving and 8842 drug driving events reported. Offending was more common among P2 drivers (62% reported at least once), and least common among learner drivers (25% reported at least once). When predicting future drink driving and drug driving intentions, lower certainty of apprehension and greater past offending both emerged as significant predictors.

Conclusions

Young drivers may benefit from reminders throughout the GDL process regarding the laws and risks associated with impaired driving. Greater policing resources for drug and alcohol testing may be needed for young drivers, to establish a high certainty of apprehension, as early as possible in their driving experience.

Keywords:

• Young drivers
• drug driving
• drink driving
• apprehension perceptions

Introduction

It has been recognized that a young drivers’ risk of crash increases between the learner and provisional phase of licensure, when drivers are no longer supervised (Centre for Accident Research and Road Safety Queensland [CARRSQ, 2020). The independence gained may lead to increased time spent around peers (Scott-Parker et al. 2015), driving to new places where they may encounter new road or traffic conditions, or engagement in unsafe driving behaviors that they had not attempted while supervised (Watson-Brown et al. 2021). Importantly, young adulthood is also considered a time of identity formation (Schwartz et al. 2015) where young people may begin to experiment with drugs and alcohol (Duff 2003; Australian Institute of Health and Welfare (AIHW)), 2021). Despite education delivered to young people to prevent impaired driving, and the consequences associated with being detected (e.g., loss of license), studies have demonstrated a considerable proportion of young drivers report driving impaired (Stevenson et al. 2001; Scott-Parker et al. 2014; Freeman et al. 2021; Mills et al. 2021a). What remains unknown is how these behaviors develop; for instance, whether they increase in frequency and likelihood as young people progress from a learner driver, to the provisional licensure phases. This study aimed to explore this further.

All Australian states have a graduated drivers licensing (GDL) system which is designed to provide young people with gradual exposure and experience on the road. In the learner phase of the GDL system, young drivers gain experience on the road but must drive with a supervising, experienced driver (AustRoads 2020). In the provisional phase (also referred to as “probationary” phase, Victoria State Government 2022), when young drivers have passed a practical driving assessment and are no longer supervised, there are restrictions in place to limit high-risk situations (AustRoads 2020). In all Australian states, there are two provisional phases – Provisional 1 (P1) and Provisional 2 (P2). Some key elements to the P1 phase across all Australian states include restrictions on passengers between 11 pm and 5am, zero tolerance for the presence of any alcohol or illicit drug, and a lower demerit point threshold compared to open licensed drivers (AustRoads 2020). Progressing from the P1 to P2 phase results in the removal of some restrictions such as nighttime passenger restrictions.

To detect and deter impaired driving, police employ roadside drug and alcohol testing. In Queensland in 2021, there were 57,749 drug tests and 1.59 million breath tests conducted, with 20.4% of drug tests identifying a positive result, and 0.8% of breath tests detecting a drink driver (Bureau of Infrastructure, Transport, Regional Development and Communications [BITRE], 2021a; Bureau of Infrastructure, Transport, Regional Development and Communications [BITRE], 2021b). These operations, like other road safety countermeasures (e.g., speed cameras) are informed by the theory of classical deterrence. This is a theory of crime that proposes that the more certain, severe and swift a person perceives punishment to be, the less likely they are to offend (Homel 1988). Studies examining apprehension perceptions among young drivers have identified that perceptions of detection for drug driving are perceived to be low (Armstrong et al. 2005), and that some young drivers perceive the likelihood of detection to be lower for drug driving than drink driving (Barrie et al. 2011). Among a sample of 286 young drivers, 57 (26%) reported driving while feeling the effects of alcohol, with 42% of these participants reporting that they did not believe they would be caught by police, and 15% were not sure (Stevenson et al. 2001). However, it has also been identified that young people think they are more likely to be targeted by police for traffic crimes (Bates et al. 2021).

Despite zero-tolerance laws for impaired driving among young drivers, studies have found drink and drug driving behaviors to be common. In one study of 1076 Queensland young drivers aged 18-20 years, 6.3% of males and 1.3% of females reported drug driving, while 18.5% of males and 11.8% of females reported drink driving (Scott-Parker et al. 2014). In another sample of young drivers (average age 21 years) in Western Australia, 26% reported driving while feeling the effects of alcohol, and 18% reported driving while feeling the effects of a drug in the previous 12 months (Stevenson et al. 2001). Young drivers are also overrepresented in drink driving crashes (Queensland Government 2015) and also overrepresented in drug driving detections. Regarding the latter, 22% of those detected between 2015 to 2020 in Queensland were aged 16 to 24 years (Mills et al. 2021a), whereas young drivers make up only around 10-13% of driver population (Department of Transport and Main Roads 2022). One study compared drink driving across three levels of the GDL system, finding that being a learner driver decreased the likelihood of drink driving (Watling et al. 2018). However, the frequency of these drink driving events was not explored, and further, the exploration of drug driving behaviors across the three levels of the GDL system was not explored.

The overall aim of this paper was thus to explore perceptions of certainty of apprehension for drink and drug driving among a sample of young drivers, and related offending behaviors, while comparing these across the three levels of the GDL system (learner, P1, P2).

Specifically, the research had the following objectives:

1. Explore self-reported drink and drug driving among young drivers and evaluate whether percentages differ across the 3 GDL levels.

2. Investigate whether certainty of apprehension for drink and drug driving differs across the 3 GDL categories, and different levels of offending.

3. Explore the utility of perceptions of certainty and GDL status in predicting future offending intentions.

Method

Participants and procedure

The study involved 601 young Queensland drivers, who were aged between 16 and 24 years, and held either a learner license, a provisional 1 license (P1) or provisional 2 license (P2). The study was part of a larger project into drink and drug driving behaviors, perceptions of certainty of apprehension and exposure to roadside enforcement. The research was promoted on Facebook with Facebook advertisements, and participation involved an online survey (which was confidential and anonymous). This project received ethical approval from the University of the Sunshine Coast Human Research Ethics Committee (#A191276).

Measures

Demographic data

Demographic information was collected, including age, gender, and license type.

Drink and drug driving behavior

Drink and drug driving behaviors were assessed with the items, “During the past 12 months, approximately how many times did you…” (1) drive a motor vehicle within four hours after consuming an illegal drug and (2) drive when you thought you were over your legal alcohol limit. Both items were informed by previous studies (Davey et al. 2007; Watson-Brown et al. 2021; Mills et al. 2021b). Participants could enter a whole number. It is important to note that in the context of those holding a learner, P1 or P2 license, that any amount of alcohol would constitute being “over your legal limit”, as learner and provisional drivers have no alcohol allowance (unlike open licensed drivers who can have up to 0.05 g/L). Intentions to offend was also assessed, by asking participants the likelihood they would engage in drink and drug driving behaviors in the future, with the following response options: (1) definitely would not, (2) very unlikely, (3) unlikely, (4) likely (5) very likely. Previous drink driving and drug driving studies have used similar items and scales to assess future intentions to offend (Watling et al. 2010; Armstrong et al. 2018; Freeman et al. 2020; Mills et al. 2022c).

Certainty of apprehension

Certainty of apprehension for drink driving was assessed by asking all participants, “even if you don’t drink and drive, if you were to drive when you thought you were over your legal blood alcohol limit, how likely is it that you would be caught by the police?” Similarly, certainty of apprehension for drug driving was assessed with the item, “Even if you don’t drug drive, if you were to drive within four hours after taking an illegal drug, how likely is it that you would be caught by the police”. Participants could choose from the following response options: (1) very unlikely (2) unlikely, (3) likely, (4) very likely. These items were based on those from previous studies (Armstrong et al. 2018; Freeman et al. 2021), however, to encourage non-offenders to consider the likelihood of detection, the phrase ““even if you don’t drink/drug drive…” was added.

Data analysis

The data was analyzed using SPSS statistics program (version 27). Descriptive statistics and frequencies were utilized to assess drink and drug driving behaviors, and certainty of apprehension. The past offending variables breached normality assumptions and were utilized as two-condition categorical variables, differentiating offenders (1 or more instances reported) from non-offenders (0 instances reported). The continuous versions of these items were used only to report descriptive statistics. Analysis of Variance tests were used to compare learner, P1 and P2 drivers regarding future offending intentions and certainty of apprehension. Games-Howell post hoc tests were used when homogeneity of variance was breached, and LSD post hoc tests were used when this assumption was met. Finally, a linear regression was conducted to explore the predictive utility of gender (F/M), past offending behavior (N/Y), license type and certainty of apprehension on future offending intentions. While three different gender categories emerged (e.g., male, female, other), the cell size for the “other” was small (N = 3) and thus only the male/female categories were used in the regression analyses.

Results

Participant demographics

Participants were aged between 16 and 24 years (M = 17.98, SD = 1.35), comprising of 302 males (50.2%), 296 females (49.3%) and 3 people who reported a gender as “other” (0.5%). The majority of participants held a P1 license (N = 251, 41.8%), followed by P2 (N = 188, 31.3%) and then learners (N = 162, 27%).

Research objective 1: Explore self-reported drink and drug driving among young drivers and evaluate whether percentages differ across the 3 GDL levels

Drink and drug driving of overall sample

As displayed in Table 1, a total of 165 participants (27.5%) reported a drink driving event in the previous 12 months. A total of 773 drink driving events were reported by these participants. When considering drink driving offenders, the average number of offenses reported was 4.68 (SD = 10.28). There were 191 participants (31.8%) who reported driving within four hours of consuming an illegal drug in the previous 12 months, with a total of 8842 events being reported (M = 46.29). Regarding future intentions, the sample reported it was unlikely (on average) they would drink drive in the future (M = 1.62, SD = .92) or drug drive in the future (M = 2.03, SD = 1.38).

Table 1. Self-reported past drink and drug driving, across GDL-system levels.

	Learners		P1		P2		Total
Offending	N	%	N	%	N	%	N	%
Drink driving
No	145	89.51	185	73.71	106	56.38	436	72.55
Yes	17	10.49	66	26.30	82	43.62	165	27.45
Drug driving
No	124	76.54	180	71.71	106	56.38	410	68.22
Yes	38	23.46	71	28.29	82	43.62	191	31.78
Neither	121	75%	144	57%	72	38%	337	56%
Drink driving only	3	2%	36	14%	34	18%	73	12%
Drug driving only	24	15%	41	16%	34	18%	99	17%
Drink and drug driving	14	9%	30	12%	48	26%	92	15%
Total	162	27%	251	42%	188	62%	601	100%

Drink and drug driving behaviors across 3 GDL groups

Table 1 displays descriptive statistics of past drink and drug driving (no/yes) for learner, P1 and P2 drivers. A crosstabs analysis revealed that there was a statistically significant association between past offending and license type for drink driving, x²(2) = 48.23, p < .001 and drug driving, x²(2) = 18.74, p < .001.

Further analysis was undertaken to explore drink and drug driving behaviors by grouping those who reported (1) no offending, (2) only drink driving, (3) only drug driving, and (4) both drink and drug driving events, and comparing them across the 3 GDL-system categories (see Table 1). A significant association was identified, x²(6) = 61.23, p < .001 between offending categories and GDL levels. The proportion of participants who reported offending increased across the GDL levels.

Research objective 2: Investigate whether certainty of apprehension for drink and drug driving differs across the 3 GDL categories, and different levels of offending

Certainty of apprehension of overall sample and GDL levels

Table 2 displays certainty of apprehension perceptions for drink and drug driving, for the general sample, and across the three GDL levels. Participants considered it “unlikely” on average, that they would be detected for drug driving or drink driving. A paired samples t-test revealed that there was a significant difference between perceptions of certainty of apprehension for drink and drug driving, t(600) = −5.32, p < .001, with detection for drink driving perceived to be slightly more likely than drug driving. A Games-Howell post hoc test revealed that for drink driving certainty, Learner drivers reported a greater score than P2 drivers (p = .004), and P1 drivers reported lowers scores compared to P2 drivers (p = .014). No difference was identified between the Learner and P1 groups. For the drug driving certainty item, Learner drivers reported lower scores compared to P2 drivers (p = .013).

Table 2. ANOVA comparing learner, P1 and P2 drivers on future intentions to offend and certainty of apprehension for drink and drug driving.

	Learners		P1		P2		ANOVA
	M	SD	M	SD	M	SD	F/Welch	p
Drink driving certainty	2.30	.76	2.24	.67	2.06	.64	6.33	.002
Drug driving certainty	2.16	.88	2.02	.77	1.91	.75	4.04	.018
Future drink driving	1.48	.77	1.54	.85	1.86	1.07	7.87	<.001
Future drug driving	1.8	1.19	2.03	1.39	2.25	1.49	5	.007

Certainty of apprehension, based on offending levels

When looking at drink driving certainty of apprehension perceptions (Table 3), a significant difference was found across the four different offending groups (“neither”, “drink driving only”, “drug driving only”, “both drink and drug driving”), F(3,597) = 5.03, p = .002. A LSD post-hoc test revealed the “neither” group reported a significantly greater drink driving certainty score, when compared to the “both drink and drug driving” group (p = .015) and the “drink driving only” group (p < .001). The “drink driving only” group also reported a significantly lower score for certainty of apprehension compared to the “drug driving only” group (p = .018).

Table 3. ANOVA comparing certainty of apprehension for drink and drug driving across offending levels.

	Drink driving certainty of apprehension		Drug driving certainty of apprehension
	M	SD	M	SD
Neither	2.27	.73	2.13	.81
Drink driving only	1.97	.6	1.97	.76
Drug driving only	2.22	.6	1.86	.8
Both drink and drug driving	2.08	.65	1.85	.73

For perceptions of drug driving certainty of apprehension, a significant difference was also found across the four offending groups, Welch’s F(3, 194.385) = .002. A Games-Howell post-hoc test revealed the “neither” group reported a significantly greater score of drug driving certainty of apprehension when compared to the “drug driving only” group (p = .018) and the “both drink and drug driving” group (p = .008).

Research objective 3: Explore the utility of perceptions of certainty and GDL status in predicting future offending intentions

Two linear regression analyses were conducted to understand the factors associated with greater intentions to (1) drink drive in the future and (2) drug drive in the future. Variables of gender, license type, past drink driving offending and certainty of apprehension for drink driving were entered as predictor variables (see Table 4). The same analysis was repeated for drug driving and the corresponding drug driving items. The regression model predicting future drink driving intentions was found to be significant, F (4, 593) = 43.16, p < .001 (R Square = .225), as was the drug driving model, F (4, 593) = 129.12, p < .001 (R Square = .466). Past offending and lower certainty of apprehension were both predictive in each model. A sensitivity analysis revealed that when past drink and drug driving (no/yes) were used as the outcome variable in a logistic regression, with gender, license type and certainty of apprehension used as predictor variables, all variables were significant predictors of past offending, with P2 license type, lower certainty of apprehension, and male gender predicting the yes offending condition.

Table 4. Linear regression predicting future drink and drug driving behaviors.

	Future Drink Driving						Future Drug Driving
	β	B	Lower	Upper	SE	sr^2	β	B	Lower	Upper	SE	sr^2
Gender (F/M)	0.03	0.05	−0.08	0.18	0.07	0.0008	0.052	0.144	−0.023	0.311	0.085	0.002601
License Type	0.02	0.03	−0.06	0.12	0.05	0.0004	0.001	0.002	−0.107	0.111	0.055	0.000001
Past drink/drug driving (N/Y)	0.44***	0.91	0.75	1.06	0.08	0.1739	0.661***	1.953	1.771	2.134	0.092	0.401956
Certainty drink/drug driving	−0.09*	−0.12	−0.21	−0.02	0.05	0.0074	−0.046	−0.08	−0.183	0.024	0.053	0.002025

Note: *p < .05, **p < .01, ***p < .001.

Discussion

This study had the overall aim of investigating perceptions of certainty of apprehension for drink and drug driving among young drivers, and related offending behaviors, across the three levels of the GDL system (Learner, P1, P2). The results reveal that 44% of the sample had driven after consuming drugs and/or alcohol at least once in the past 12 months, with drug driving events found to be more common and frequent than drink driving. A previous study with young adults identified that drink driving was more common (26%) than drug driving (18%) (Stevenson et al. 2001). Thus, while the rates of drink driving remain consistent with this study, reports of drug driving are greater. It is difficult to ascertain whether this finding reflects differences in the item used to assess impaired driving, or actual differences in the sample. For instance, Stevenson and colleagues assess drug driving as “feeling the effects of the drug” whereas the current study assessed drug driving as driving within four hours.

The findings support the general notion that as restrictions ease for young drivers, offending behaviors increase. Learner drivers were the least likely to report offending (both drink and drug driving), which was similarly found in a 2018 drink driving study (Watling et al. 2018). Such a finding may be expected given that learner drivers have the supervision of their parents, caregiver or driving instructor, and thus may be less likely to break the rules. However, it is noteworthy that 25% of learner drivers in the present study still reported one or both offending behaviors in the previous 12 months. In previous studies, some learner drivers have been found to report driving unsupervised (Scott-Parker et al. 2013), or even driving before having a learner license (Scott-Parker et al. 2012). In fact, pre-licensed driving has been found to be related to other risky driving behaviors as learner and provisional drivers (Scott-Parker et al. 2012).

The finding that Provisional 2 licensed drivers (mean age 19 years in current sample) reported the greatest level of offending may reflect increased drink and drug taking behaviors as a result of being of legal age to drink (i.e., 18 years in Australia, Australian Government 2020), and thus attend venues where drugs and alcohol are consumed. Provisional licensure also comes with greater independence and makes it easier for young people to see their friends (Scott-Parker et al. 2015). It has been found that peers displaying attitudes and approval toward drink driving, may influence greater engagement in offending behaviors (Davey et al. 2005; González-Iglesias et al. 2015). Thus, when considering these social changes that arise from provisional licensure, it may not be surprising that reports of drink and drug driving increase.

While learner drivers reported slightly higher certainty of apprehension scores for drink and drug driving, scores across all three levels of the GDL system still indicated a perception of “unlikely”. Low certainty of apprehension is associated with a lack of exposure to police testing (Homel 1988; Mills et al. 2022c) and punishment avoidance. According to Stafford and Warr (1993), each time a motorist offends without detection, they are provided with more evidence that the likelihood of detection is low. Participants reported a total of 773 past-12-month instances of drink driving and 8842 instances of drug driving, which highlight the frequency of offending, most presumably undetected by police.

While some mixed evidence has emerged regarding the utility of certainty for predicting offending behaviors (Freeman et al. 2010; Szogi et al. 2017; Freeman et al. 2020; Hasan et al. 2022; Truelove et al. 2023), the current study found that lower certainty of apprehension for drink driving and drug driving was found to predict a greater intention to offend in the future. Past offending was also a significant predictor of future offending which has emerged in previous research, and illuminates the habitual nature of offending (Freeman et al. 2020). Neither license type nor gender predicted future offending intentions, however these variables did predict past-12 month offending (no/yes) for both drink and drug driving. Regarding license type, the finding that P2 licensed drivers were more likely to report past offending may demonstrate the aforementioned intersection of social, developmental and GDL system changes (Cook et al. 2017; Watson-Brown et al. 2021).

Limitations and future directions

There are some limitations to this research that should be acknowledged. Firstly, it is difficult to ascertain how accurately the sample used in the present study reflects the wider population of young drivers. An approximately even sample of males and females were involved in the study, however, due to recruiting participants on Facebook, young drivers who do not use social media are not represented. While drink and drug driving behaviors were assessed, the exact drink and drug taking behaviors of participants, including the degree of substance dependency, was not assessed. However, studies have identified greater drug and alcohol use to be significantly predictive of greater drink and drug driving offending behaviors (Cook et al. 2017; Mills et al. 2021b; Hasan et al. 2022). Including drug and alcohol disorder scales in future research may be important to identify the factors most related to future offending intentions. Future research may also benefit from qualitative studies to examine how perceptions of certainty for drink and drug driving, and related offending behaviors change across the GDL levels. Examining protective factors against driving under the influence will also be important to understand how to best deter young drivers from impaired driving.

Implications and conclusion

The findings of this study highlight that at a time when their risk of crash is greatest, a large proportion of young drivers are driving after the consumption of drugs and/or alcohol. The earlier that young drivers can experience police drug and alcohol testing, the more potential there may be to enhance certainty of apprehension perceptions and decrease offending behaviors. Importantly, given that offending behaviors appear to increase with reduced restrictions, it is vital to consider how to effectively remind drivers of these risks as they progress into decreasing driver regulation stages. For example, text or email-based reminders throughout the duration of the GDL may be an option to remind young drivers of the risks of impaired driving. At the very least, this study indicates that some cohorts may be dismissive of the risks of impaired driving, at a time when they are least experienced to identify and respond appropriately to such risks. This problem appears further compounded by the correspondingly low levels of perceived certainty of apprehension for both drink and drug driving. Any effort to increase perceptions of certainty of apprehension and reduce offending behaviors in the earliest years of driving experience are important steps for improving young driver safety.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by the Motor Accident Insurance Commission.