Can Older Adults Underestimate Alcohol-Related Test Slowing? A Placebo Trial

What does this experiment establish?

A participant’s feeling of being more or less impaired did not reliably mirror that participant group’s relative performance on the Trail Making Test. The investigators deliberately separated reported intoxication, reported impairment and measured completion time. Those are not interchangeable outcomes. The distinction is particularly useful when testimony relies on a person’s statement that alcohol did not affect them, or conversely that they felt markedly intoxicated.

The key comparison was not a courtroom determination of fitness to drive. It was an experimental contrast between two age ranges, alcohol versus placebo assignment, and two assessment times. Although the older group showed slower performance in certain comparisons after alcohol, the same study did not show a uniform worsening for every age group, test or time point. It is therefore inappropriate to present this trial as proof that every older adult becomes impaired at a particular breath reading, or that younger adults benefit from alcohol.

Who participated, and how were the groups constructed?

Investigators enrolled 68 medically screened, moderate social drinkers. The 42 older participants were aged 50–74 years, while the 26 younger participants were aged 25–35 years. Within each age stratum, participants received alcohol or a sensory-matched placebo under double-blind conditions. This was a between-subject alcohol/placebo comparison, not a study in which every participant received both beverages on different days.

These are small and unequal groups, a major reason to resist individual prediction. Participants had been screened for relevant illnesses and problematic drinking, so the results cannot be assumed to describe frail older adults, people with major neurological disease, or every person who drinks in ordinary life. Medication use was not absent: the researchers noted stable medication use among slightly more than 40% of the alcohol-exposed participants in both age categories.

What was administered and how was alcohol assessed?

Alcohol doses were designed to approximate a moderate social-drinking exposure and achieve an investigator-specified target on the breath-alcohol curve. Placebo beverages mimicked aspects of the alcohol beverage to reduce expectancy differences. Participants were tested as breath alcohol was rising and again as it was falling, at approximately 25 and 75 minutes after beverage administration. The intervention and timing matter: a performance difference at one phase should not automatically be carried forward to another phase.

An Intoxilyzer 400 breath instrument supplied the alcohol readings. The article reports breath-alcohol concentration (BrAC) in its stated mg/100 mL convention. Whatever that instrument’s reporting convention, these were not directly measured venous whole-blood ethanol concentrations obtained by forensic headspace gas chromatography. The assay’s matrix and measurement method cannot be silently changed in a later legal narrative.

At the first testing time, the older alcohol group had a reported mean instrument reading of 39 mg/100 mL (SD 11), compared with 32 mg/100 mL (SD 9) in younger alcohol recipients; this age difference was statistically significant (p = .02). Reported peak group means were 56 (SD 12) and 50 (SD 11), respectively, with no significant age difference at peak. Mean BrAC averaged across measured times did not differ significantly by age (p = .74). Investigators included the early BrAC difference as a covariate in certain performance analyses; the reported age contrasts persisted, but adjustment does not establish equal concentrations or responses for individuals.

What did the Trail Making Test actually measure?

Participants completed two paper-and-pencil forms of the Trail Making Test. Form A emphasizes visual scanning, sequencing, attention and motor speed; Form B adds greater sequencing and cognitive-switching demands. The researchers timed completion in seconds. Longer time means slower task completion on that particular test. It is not itself a direct measure of blood alcohol, reaction time behind a steering wheel, lane departure, safe braking distance or legal impairment.

In broad terms, older participants required more time than younger participants on both forms. Task completion was generally faster at the later session, a pattern that may reflect changing exposure, practice or other time-related influences. The design does not permit attributing every between-session difference solely to alcohol elimination.

Which alcohol-versus-placebo differences were statistically supported?

On the ascending phase, older alcohol recipients took a mean of 81.06 seconds on Trails B versus 67.26 seconds among older placebo recipients; the study reports p = .03 for that within-age comparison. In younger adults, Trails A was completed faster by the alcohol group (21.54 versus 32.13 seconds, p = .008). Faster completion of this simple task in a small group does not demonstrate generalized cognitive enhancement, superior driving, or the absence of adverse effects on another task.

The investigators’ initial three-way age-by-beverage-by-phase interactions were not statistically significant (Trails A p = .12; Trails B p = .44). Subsequent phase-specific analyses identified differences the authors considered informative in an exploratory pilot study. Their ascending-phase Age × Beverage interaction was significant on Trails A (p < .005), while the corresponding Trails B interaction approached but did not reach conventional significance (p = .051). Describing that last interaction as established would overstate the results.

Each cell is a group mean, not an individual result, forensic BAC estimate or diagnostic cutoff. The primary paper also reports variability and formal comparisons; visually comparing table means cannot substitute for its statistical analysis. In the older groups, the later Trails B times of 64.00 and 59.52 seconds did not yield the same statistically supported alcohol-versus-placebo difference reported earlier.

How did perceived intoxication differ from perceived impairment?

Participants rated both how intoxicated they felt and how much they believed the beverage affected their performance. These were subjective rating scales, not measured ethanol concentrations or independent motor-performance tests. A witness’s self-evaluation can be relevant evidence of perception, but it is not a laboratory assay or a validated replacement for performance assessment.

During the earlier assessment, the older alcohol group’s mean perceived-impairment rating was 3.41, and the younger alcohol group’s was 4.09. The source reports a significant main effect of age on perceived impairment across assignment groups (p < .001). That main effect should not be re-described as a separately established older-alcohol-versus-younger-alcohol pairwise result. At the later assessment, older alcohol recipients rated perceived impairment at 2.82 versus 1.43 for older placebo recipients (p < .001 for that group contrast), even though their objective trail-making comparison at that phase was not statistically different.

In other words, a lower self-rating does not invariably accompany faster observed performance, and a higher self-rating need not coincide with a statistically detectable group difference on these particular tasks. The important finding is discordance between two kinds of evidence, not proof that one self-report is necessarily dishonest. An individual’s expectations, momentary feelings, awareness of errors and prior experience can all affect reporting; this experiment does not determine which explanation applies in a particular case.

Does rising-versus-falling breath alcohol change the inference?

Yes. The phase of the concentration-time curve is part of the experiment’s exposure context. At approximately 25 minutes, alcohol was on the ascending part of the measured curve; by approximately 75 minutes, participants were assessed on its descending part. The age-related separation in objective performance was more apparent in the earlier observations. Subjective impairment showed a different pattern later, especially among older participants.

But the trial does not validate a formula that maps an individual’s BrAC, clock time or ascending/descending status to an exact Trail Making time. The authors compared means across study arms, not individual counterfactual performances. Nor did they administer continuous breath and blood measurements across a real-world drinking or driving event. Without a known time course and appropriate case evidence, a single result cannot identify precisely which physiological phase someone occupied.

See the site’s independently published pharmacokinetics and interpretation guide for general discussion of absorption, distribution and time-related uncertainty.

Why must the breath reading stay separate from blood BAC?

The primary study used breath instrumentation to track its experimentally targeted alcohol exposure. Breath measurements and whole-blood laboratory concentrations are different analytical observations. A numerical breath reading reported in a blood-equivalent convention cannot be converted into a directly measured venous whole-blood specimen merely by labeling it “BAC.” Any proposed correspondence depends on the instrument method, reporting conventions and known uncertainty; no direct whole-blood gas-chromatography assay was reported for these participants.

The site’s breath-alcohol measurement guide addresses breath sampling and method limitations. The distinct blood ethanol chromatography guide describes laboratory analysis of an actual blood specimen. Neither source makes the Gilbertson task results an individual BAC correction factor.

What may a forensic toxicologist fairly say in litigation?

A defensible description is that a controlled human experiment found certain age- and phase-dependent differences between observed trail-making completion times and perceived impairment following an alcohol or placebo beverage. The earlier older-alcohol versus older-placebo Trails B difference was significant, and the later subjective ratings did not mirror the absence of a supported later group-level performance difference. Such evidence cautions against treating “I did not feel affected” as proof of preserved performance.

That conclusion must remain bounded. A specific person’s testimony, driving behavior, medical condition, exposure history and specimen result require their own evaluation. This article cannot assign a scientifically validated individual probability of impairment, quantify decrement in vehicle control, reconstruct a BAC, or determine who caused a collision. A group means table is not a diagnostic standard, and a task slowing in seconds is not a legal impairment threshold.

Likewise, an expert should not use the younger ascending Trails A observation as a blanket claim that alcohol improves performance. It was a limited task-specific comparison in 11 alcohol-exposed young adults, and the authors did not demonstrate widespread improvement across more complex functions. Both adverse and seemingly favorable individual outcomes are uncertain outside the study protocol.

Which limitations materially restrict individual conclusions?

Small group sizes: the youngest alcohol cell contained only 11 people, while 17 older participants received alcohol. Smaller, unequal arms can produce unstable effect estimates and limit subgroup inference.

Limited behavioral measures: two versions of one test were used, not a battery proving real-world driving competence, balance, divided attention, crash avoidance or occupational safety. Trail making requires visual and motor output but is not a simulator or on-road study.

Exploratory analyses: the initial three-way interactions were nonsignificant, and some highlighted results arose in subsequent phase-specific analyses explicitly justified by the authors’ pilot design. A borderline p value is not a positive finding.

Selection and generalizability: these were screened moderate social drinkers, not a representative population of all older adults, people with alcohol use disorder, medically complex patients or people taking unfamiliar drugs. The investigators documented stable medication use in substantial minorities.

Timing and learning: repeated administration permits practice and phase effects to coexist; differences between 25 and 75 minutes cannot be uncritically assigned to ethanol metabolism.

Analytical boundary: the alcohol values were breath-instrument readings, not individually verified forensic whole-blood BAC or a combined blood/breath pharmacokinetic validation. No personal retrospective correction, impairment threshold or causal attribution was calibrated.

Frequently asked forensic questions

Can a person feel unimpaired while objectively performing worse?

Yes, self-assessment and measured task performance can diverge. In this trial, the mismatch was especially informative for older adults during the earlier test phase. This finding does not establish that every person who denies impairment was impaired.

Did the researchers prove that older adults were impaired at a universal BAC?

No. The experiment involved group comparisons at targeted breath readings and tested a limited behavioral task. It did not validate a universal individual threshold, nor did it directly measure each participant’s venous whole-blood ethanol concentration.

Were the older and younger groups exposed to identical measured alcohol levels?

Not at every time point. The older group had a somewhat higher average first breath reading (39 versus 32 in the study’s mg/100 mL convention); peak group means did not differ significantly. Covariate adjustment did not turn the readings into identical personal exposures.

Did alcohol improve younger adults’ driving ability?

No such conclusion was tested. The younger alcohol group completed the relatively simple Trails A faster than its placebo control at the earlier session. Driving and complex divided-attention performance were not evaluated by that comparison.

Does a higher self-reported impairment score prove worse observed performance?

No. At the later testing phase, older alcohol recipients reported higher perceived impairment than older placebo recipients without a statistically supported difference in their trail-making performance. Self-report and tested function must be evaluated separately.

Can an expert back-calculate blood BAC from the trial’s trail-making times?

No. Trail-making time is a behavioral outcome, not a blood-alcohol assay. The trial provides no validated function for converting seconds, subjective ratings or age into a forensic whole-blood ethanol concentration.

Could these findings establish that alcohol caused an individual’s crash?

No. Accident causation requires case-specific evidence about driving, environment, timing, exposures, witnesses and alternative explanations. This experiment establishes limited group-level laboratory findings, not a crash reconstruction.

Primary research and evidence boundaries

Original paper: Gilbertson R, Ceballos NA, Prather R, Nixon SJ. Effects of acute alcohol consumption in older and younger adults: perceived impairment versus psychomotor performance. Journal of Studies on Alcohol and Drugs. 2009;70(2):242–252. https://doi.org/10.15288/jsad.2009.70.242. Read the complete original article and study tables.

The study distinguishes assignment to alcohol or placebo, breath-instrument alcohol readings, timed task performance, and subjective impressions. Neither the primary paper nor this analysis substitutes for the full case record, validated laboratory documentation, and context-specific impairment assessment.