Two Edge Departures in 18 Seconds: Vermont Clarifies Reasonable Suspicion for DUI Stops on Unmarked Roads

Introduction

In State v. Karen Norton, 2025 VT 56, the Vermont Supreme Court addressed when an officer has reasonable and articulable suspicion to conduct a DUI traffic stop based on a brief observation of erratic driving on a road without edge markings. The central question was whether an officer’s observation—over approximately eighteen seconds—of a vehicle’s passenger-side tires twice leaving the paved portion of the roadway, coupled with the officer’s training and experience, sufficed to justify an investigatory stop for suspected impaired driving under the Fourth Amendment and Article 11 of the Vermont Constitution.

The case pits the State’s public-safety rationale for prompt DUI intervention against a defendant’s claim that the officer lacked adequate grounds—particularly where the dashcam video was not crystal clear and the road had no “fog line.” The trial court denied a suppression motion; the defendant entered a conditional guilty plea to DUI and appealed. The Vermont Supreme Court affirmed, over a dissent.

Summary of the Opinion

The Court (Eaton, J.) held that the officer had reasonable suspicion to stop the vehicle for suspected DUI based on:

Observation that the vehicle’s passenger-side tires left the paved portion of the roadway twice in quick succession on a straight road at night;
Additional movements (returning to lane, approaching the center line, and then again leaving the paved surface); and
The officer’s articulated training and experience (DUI, ARIDE, DRE) that such departures are an “indicator of possible impairment.”

The majority emphasized the totality of the circumstances and the recognized need to address the imminent danger of impaired driving. Because reasonable suspicion existed to investigate DUI, the Court did not reach the alternative argument that a motor vehicle violation occurred under 23 V.S.A. § 1038(1) (lane discipline). The Court also reaffirmed that appellate review of a trial court’s findings, including findings based on video evidence, is for clear error.

Justice Cohen dissented, concluding that the video did not show erratic driving, that the finding that the vehicle left the paved surface twice was clearly erroneous, and that the officer’s brief observation—at most, slight intra-lane movement—was insufficient to support reasonable suspicion of impairment.

Analysis

Precedents Cited and Their Influence

State v. Pratt, 2007 VT 68: The Court reaffirmed that erratic driving can create reasonable suspicion of DUI. In Pratt, repeated intra-lane weaving over five miles justified a stop. The majority here analogized Norton’s conduct (twice leaving the paved surface) as even more indicative of erratic operation than the intra-lane weaving in Pratt, while cautioning—as Pratt did—that slight, isolated weaving alone may be insufficient.
State v. Boyea, 171 Vt. 401 (2000): Cited for the dictionary definition of “erratic” (“having no certain course; wandering; moving”) and for the recognition of the imminent risks posed by drunk drivers. The majority applied this definition to conclude that twice leaving the paved surface on a straight road qualifies as erratic.
State v. Lamb, 168 Vt. 194 (1998): Emphasized public-safety risk as a critical factor in assessing the reasonableness of stops. The Court invoked Lamb’s “instrument of death” language to underscore the need for swift intervention.
State v. Mara, 2009 VT 96A and State v. Santimore, 2009 VT 104 (mem.): Reinforced the “paramount importance” of promptly removing impaired drivers and balancing privacy intrusions with public safety.
State v. Harris, 2009 VT 73: Reiterated the State’s burden to justify a stop.
State v. Calabrese, 2021 VT 76A and State v. Rutter, 2011 VT 13: Clarified the mixed standard of review—clear error for factual findings and de novo for whether those facts meet the legal standard for a stop.
Swett v. Gates, 2023 VT 26: Notably applied to video evidence; the Supreme Court does not make its own findings from the video but reviews the trial court’s video-based findings for clear error. This approach materially supported the majority’s deference to the trial court’s credibility determinations where the dashcam was ambiguous.
State v. Davis, 2007 VT 71 (mem.): Distinguished. In Davis, the officer did not explain how observed movement related to impairment. Here, the officer expressly tied the departures from the paved surface to her training and experience as an indicator of impairment.
Out-of-circuit cases cited by the defense: U.S. v. Gregory, 79 F.3d 973 (10th Cir. 1996), and U.S. v. Freeman, 209 F.3d 464 (6th Cir. 2000). The majority found these distinguishable—each involved a single, isolated encroachment into the shoulder/emergency lane and, in Freeman, a higher probable-cause standard. Norton involved two departures in quick succession and the lesser reasonable-suspicion standard.

Legal Reasoning

The majority’s reasoning proceeds on three main pillars:

Totality of Circumstances and Erratic Driving: The officer observed a sequence—departure from the paved surface, return and approach to the center line, followed by a second departure—on a straight roadway at night, all within roughly eighteen seconds. This pattern, not a single deviation, was deemed “erratic,” satisfying the threshold for reasonable suspicion of impairment. Importantly, the Court rejected any requirement that an officer must see “intra-lane weaving,” “veering,” or “swerving” per se; leaving the paved surface twice in short order is sufficient.
Officer Training and Articulation: The officer’s DUI/ARIDE/DRE training and extensive traffic-stop experience, coupled with a clear articulation that leaving the paved roadway is an indicator of impairment, provided the necessary link between observed behavior and suspected criminal activity. This connection distinguished the case from Davis, where no such explanation was offered.
Standard of Review and Video Evidence: Applying Swett v. Gates, the Court deferred to the trial court’s factual finding—based on the officer’s testimony—that the vehicle twice left the paved surface, even though the dashcam did not “perfectly clearly depict” it. The trial court found the testimony credible and consistent with what could be seen on the video. Absent clear error, the appellate court would not reweigh the video.

The majority also emphasized public safety: reasonable suspicion is a modest standard, and the dangers of impaired driving justify prompt intervention without requiring prolonged observation. The Court expressly did not reach whether 23 V.S.A. § 1038(1) (lane control) applied on this roadway, resolving the case solely on DUI suspicion.

Impact and Forward-Looking Consequences

Norton clarifies several points that will shape future DUI-stop litigation in Vermont:

Short, Multi-Factor Observation Can Suffice: Two quick departures from the paved surface on a straight road, observed over a brief period, can establish reasonable suspicion when coupled with an officer’s training-based articulation of impairment indicators. Prolonged tailing is not required.
No “Intra-Lane Weaving” Requirement: The Court expressly rejects a categorical need for weaving or swerving within the lane. Erratic driving is broader and includes leaving the paved roadway twice in quick succession without an apparent benign cause.
Deference on Ambiguous Dashcam: When video is inconclusive, trial courts may credit officer testimony if consistent with observable footage; appellate courts will review such factfinding, including video-derived findings, only for clear error. This is a significant procedural signal for suppression litigation.
Unmarked Roads and § 1038(1) Left Open: The Court avoided deciding whether a “single-lane” statute applies to roads marked only by a center line (no fog line). Expect further litigation on what counts as “clearly marked lanes” under § 1038(1), especially in rural Vermont.
Public Safety Emphasis Under Article 11: The Court aligns Article 11 with the Fourth Amendment here, reiterating the strong public-safety interest in swiftly removing impaired drivers. This alignment may guide trial courts to credit prompt stops where multiple articulable facts suggest impairment.

Practical Implications

For law enforcement: Carefully document each observed deviation and explicitly link those observations to training and impairment indicators in reports and testimony. Even brief observation, if specific and multi-factor, can suffice.
For defense counsel: Focus suppression challenges on undermining factual findings—especially where dashcam is ambiguous—by highlighting road geometry, lighting, absence of fog lines, vehicle dynamics, and alternative explanations. The dissent’s approach provides a roadmap for arguing clear error where video and testimony diverge.
For trial courts: Make explicit credibility findings and explain how video evidence supports or does not contradict testimony. Appellate deference depends on well-reasoned findings tied to the record.

Complex Concepts Simplified

Reasonable and Articulable Suspicion: A low threshold requiring specific, articulable facts that, taken together, suggest criminal activity may be afoot. It is more than a hunch, but much less than proof by a preponderance of the evidence or probable cause.
Erratic Driving: Not limited to weaving within a lane. It broadly means driving without a “certain course”—for example, drifting off the paved surface twice in quick succession with no obvious explanation.
Totality of the Circumstances: Courts consider all facts together—time of night, road configuration, pattern of movement, officer experience—rather than any single fact in isolation or the duration of observation alone.
Standards of Review:
- Factual findings (including those informed by video): reviewed for clear error—appellate courts do not reweigh evidence if the findings are supported by the record.
- Legal conclusion (do the facts amount to reasonable suspicion?): reviewed de novo.
Dashcam Evidence Deference (Swett v. Gates): Vermont appellate courts will not substitute their own view of video for the trial court’s supported findings. This differs from jurisdictions where appellate courts freely interpret video evidence.
ARIDE and DRE Training: Specialized law-enforcement training (Advanced Roadside Impaired Driving Enforcement; Drug Recognition Expert) that equips officers to recognize signs of impairment and helps satisfy the requirement to articulate why observed driving indicates possible DUI.
Conditional Guilty Plea: A plea that preserves the right to appeal a specified pretrial ruling (here, the denial of the suppression motion).
23 V.S.A. § 1038(1): Vermont’s lane-discipline statute requiring travel within a single lane where the roadway has “two or more clearly marked lanes.” Whether a road with only a center line but no edge lines qualifies remains unresolved in this case.

The Dissent’s Perspective

Justice Cohen would have reversed. Key points:

The dashcam shows approximately sixteen seconds of straight driving; it does not clearly depict leaving the pavement or weaving.
The trial court’s finding that the vehicle left the paved surface twice is clearly erroneous given the video.
Even assuming slight intra-lane movement, the brief observation period and minimal deviation are insufficient under Pratt’s caution that slight weaving alone does not justify a stop.
Because neither the officer nor the trial court relied on § 1038(1), the only possible basis is suspected impairment—and that basis was inadequate on this record.

The dissent signals that not all brief, ambiguous observations will pass muster; it emphasizes rigorous scrutiny of factual findings where video evidence exists.

Conclusion

State v. Norton establishes a clear, practical rule for Vermont: two quick departures from the paved surface on a straight, unmarked road, observed over a short period and credibly explained by an officer as indicative of impairment based on training and experience, satisfy reasonable suspicion for a DUI stop under both Article 11 and the Fourth Amendment. The Court underscores public safety and the totality of the circumstances while confirming that appellate courts defer to trial courts on video-informed factfinding absent clear error.

At the same time, the dissent highlights a limiting principle: where video does not corroborate alleged erratic driving and the deviations are minimal and brief, reasonable suspicion may be lacking. Future litigation will likely explore the precise contours of § 1038(1) on unmarked roads, the weight of ambiguous dashcam footage, and the extent to which short observational windows suffice—issues that Norton frames but does not finally resolve.