Can 10 Minutes of VR Teach a Basic Wet-Lab Skill?

Author: Julia Kirshner, PhD

Summary

Hands-on technical skills are learned through practice, but practice requires both equipment and an expert instructor, and in many classrooms and workforce development programs both are in short supply. Interactive and immersive training approaches, including virtual reality (VR), are being developed to enable learners to rehearse a task before they touch real equipment; the question is whether those approaches actually teach what they promise, that is, whether the trained behavior shows up when the person faces the real task.

To address this, we ran a pilot study at two 2026 conferences (SkillsUSA, n = 48; AWE USA, n = 15 participants) in which each visitor to our booth was observed interacting with a real micropipette before and after a short VR session that introduced micropipetting technique while teaching the learner how to determine a patient’s blood type. Before the VR session, 10 of 43 first-time micropipette users (23%) held the pipette and used the plunger correctly; after VR, 39 of 43 users (91%) did both correctly. Twenty-five of those novices then ran a real ABO/Rh blood-typing protocol, and 22 (88%) completed the procedure and correctly identified an unknown blood type. The effect held across both demographically different participant populations.

This pilot was designed to evaluate whether a skill learned in a short VR training session transfers to real-life proficiency, not to compare VR to other learning modalities. Taken together, the data suggests that VR training has real potential to teach hands-on laboratory skills, especially for learners with limited access to traditional training or those needing to refresh, upskill, or reskill their technique.

Methods

Sixty-three visitors to the Fantastic Voyage Labs booth were enrolled at SkillsUSA 2026 (June 2–4, n = 48) and AWE USA 2026 (June 16–18, n = 15). The two participant populations differed substantially: SkillsUSA participants were largely secondary and post-secondary CTE students, spread across many trade disciplines and not concentrated in lab-adjacent tracks; AWE participants were XR and technology professionals.

Each participant moved through five steps:

1. Intake. Prior pipette experience, lab experience, and basic demographics were captured by self-report.
2. Pre-VR observation. The participant was asked to pick up a real micropipette from the stand and hold it as if about to use it. An observer scored two items, hand position on the pipette and plunger operation, on a three-point scale (YES, both grip and plunger were correct / NO, neither were correct / PARTIAL, either grip or plunger were correct, but not both).
3. VR intervention. The participant completed the Destination Lab blood-typing scenario, which guided them step-by-step through micropipetting technique while demonstrating how to determine a patient's blood type (Figure 1A). Session start and end timestamps, an observer-coded completion rating, and a self-reported comfort rating were captured.
4. Post-VR observation. The pre-VR observation was repeated immediately after the headset came off.
5. Real-bench task. The participant optionally performed the full ABO/Rh blood-typing protocol with real pipettes and reagents (Innovating Science IS3101 kit), scored on five procedural steps: (1) dispense simulated blood, (2) add anti-A, (3) anti-B, and (4) anti-Rh reagents; (5) mix; and one interpretation step reading the blood type by comparing the agglutination pattern developed during the procedure to a reference card (Figure 1B).

Observations were entered in real time through a purpose-built web app. Confidence intervals around the reported percentages were computed as Wilson score intervals, which give the 95% confidence interval for each percentage and remain accurate when the sample is small or the percentage is near 0% or 100%.

Results

Sample composition. Forty-three of the 63 participants (68%) reported never having used a pipette, 17 (27%) had used one a few times, one used pipettes regularly, and two had no intake recorded (Figure 2A). The "never used" group was the main focus of the study, and it was the majority of the participants at both conferences: 34 of 48 at SkillsUSA (71%) and 9 of 15 at AWE (60%). Comfort in the headset was not a limiting factor (Figure 2B): 60 of the 61 participants with a recorded rating were comfortable, and the single poor rating came from a phlebotomist who became briefly dizzy and stopped.

Baseline technique. Among the 43 novices, only 10 (23%) picked up the instrument with both a correct grip and correct plunger operation. Twelve (28%) had correct grip and twelve (28%) operated the plunger correctly, but these were not the same people: 10 had both, 2 had grip only, 2 had plunger only, and 29 had neither. The errors were consistent and recognizable (Figure 3). Participants held the pipette like a syringe, a hammer, or a pen, and some flipped it upside down; on the plunger, the most common failure was pressing it with the wrong finger or confusing the plunger with the eject button.

Pre-VR "correct" rating captured appearance, not capability. Of the 12 novices whose baseline grip looked correct, the observer notes that not one could attach a tip or knew how to carry out the pipetting procedure. Several participants arrived at the correct-looking grip by visible trial and error, picking up the instrument wrong, pausing, and correcting themselves, and at least one was observed watching a nearby participant who had done it right. In practical terms, no novice could actually use the pipette before VR, so every post-VR figure below understates the size of the change.

Technique after a single VR session. Immediately after the headset came off, correct grip among novices rose from 12/43 (28%) to 40/43 (93%) and correct plunger operation from 12/43 (28%) to 40/43 (93%). The share of novices who had both grip and plunger correct in the same session rose from 10/43 (23%, 95% CI 13–38%) to 39/43 (91%, 95% CI 78–96%); the two confidence intervals did not overlap (Figure 4). The remaining errors were few and specific: after VR, one novice had grip only correct, one had plunger only, and two were incorrect on both.

The same pattern held across the entire participant population rather than just the novices (Figure 5). Before VR, 30 of 63 were scored incorrect on grip and 31 of 63 incorrect on the plunger, with a further 4 partial on each. Afterward, 56 of the 60 participants were correct on grip and 57 of 60 on plunger; the remainder were 2 partial and 2 incorrect on grip, and 1 partial and 2 incorrect on plunger. Put another way: 39 of 63 participants were not fully correct on both measures before VR, and of those, 35 (90%) improved on at least one and 33 (85%) reached fully correct technique on both.

Replication across two audiences. The improvement was not specific to one population (Figure 6). At SkillsUSA (n = 34 novices), correct grip went from 11/34 (32%) to 31/34 (91%) and correct plunger from 10/34 (29%) to 32/34 (94%); at AWE (n = 9 novices), grip went from 1/9 (11%) to 9/9 (100%) and plunger from 2/9 (22%) to 8/9 (89%). Among participants who were not fully correct before VR, 27 of 30 (90%) improved at SkillsUSA and 8 of 9 (89%) at AWE. The two samples differed in age, occupation, and technology fluency, and they produced the same result. The AWE numbers are small enough that the panel shows the direction of the effect rather than its exact size, but the direction was unambiguous.

Transfer to the real bench. Thirty-four participants (54%) went on to run the actual ABO/Rh blood-typing protocol with real pipettes, tips, and reagents; 25 of them were novices. Thirty-one of the 34 (91%) completed the procedure and correctly identified the unknown blood type, and among the novice subset, 22 of 25 (88%) were correct (Figure 7A). All 34 participants executed all five procedural steps correctly, which meant no one in the study erred on execution of the technique itself. In every case the misread was of the agglutination pattern, not an error in handling the pipette or reagents. Identification accuracy was 24/25 (96%) at SkillsUSA and 7/9 (78%) at AWE.

Time on task. The VR module took a mean of 9.6 minutes (median 9.8, range 3.1–14.1) across the 34 participants with complete timers. The bench protocol afterward took a mean of 3.3 minutes (median 2.8, range 1.0–8.2) across the 19 participants with usable protocol timers (Figure 7B); one additional sub-1-minute recording (0.9 min) was excluded as a likely timer error, since the protocol cannot be completed in under a minute. Novices were not slower on the bench than the rest of the sample: their mean protocol time was 3.4 minutes, essentially the same as the 3.3-minute overall mean. Prior pipette experience gave no measurable speed advantage.

Prior experience is a weak signal. Of the 17 participants who reported having used a pipette "a few times," 5 still showed incorrect grip or plunger operation at baseline, and among those who reached the bench, the experienced group did not outperform the novices on either accuracy or speed. Self-reported exposure to a pipette was not evidence that anyone had ever been taught to hold or use one.

Discussion

We set out to ask whether a short, unsupervised VR session could prepare someone who had never held a micropipette to use one on the real bench. Within the limitations of this pilot design, the answer is yes. Wherever instructor time, equipment access, or reagent budgets limit how many learners can be brought to a baseline level of competence, a ten-minute VR session can carry a first-time user through the physical and procedural fundamentals before they ever touch a real instrument, freeing scarce in-person instruction for the parts of the task that still require human judgment. VR also enables unlimited, repeated practice without consuming instructor time, reagents, or wear on real equipment, so learners can rehearse the same steps as many times as they need to reach proficiency. The next question is no longer whether this works, but how long the effect lasts and what else VR can teach at this scale.

The finding we did not anticipate was the split between execution and interpretation (Figure 7A). All 34 bench participants performed all five procedural steps correctly; every failure was a misread of the agglutination result. VR taught the physical and procedural parts of the task efficiently and left the interpretive part, reading the result, largely untouched.

Two limitations shape what these numbers can claim. Both cohorts self-selected by approaching the booth, so the consistent result across two dissimilar samples is stronger than a single-population result but is not a substitute for random sampling. And the post-VR observation was taken as the headset came off, so whether the technique survived an hour, a day, or a week was not tested. The pre-VR "correct" rate measured appearance only; a novice holding the pipette the right way could not necessarily use it properly, which meant the true capability change was larger than the headline numbers suggested.

Three directions would strengthen the result:

• Assess skill retention. Delayed measurements at 24 hours, 1 week, 1 month, and 6 months would test whether the technique persists; how long the effect lasts is what educators and workforce programs most need to know before adopting a VR training tool.
• Evaluate whether interactive or immersive training is superior to other learning modalities. A randomized control arm (no training, video-only, text-only, or in-person instruction) would allow comparative claims.
• Establish durability of the skill transfer effect over a larger and more diverse population of learners. Recruiting from existing classroom groups and workforce programs, rather than conference booths, would broaden generalization. Deciding scoring rules in advance and having a second scorer independently rate a sample of participants would let us check how reliably observers agree.