Does Oral THC Reduce Willingness to Work for Rewards? A Controlled Study of 58 Women

Answer: In a randomized, double-blind, repeated-session laboratory experiment involving 58 women in the primary analysis, 15 mg of oral delta-9-tetrahydrocannabinol (THC) significantly reduced choices of a more demanding task offering a larger possible cash reward compared with placebo. The lower 7.5-mg dose did not produce a statistically significant individual comparison with placebo. THC also slowed button pressing and reduced trial completion. The findings concern an experimentally defined choice between effort and monetary reward; they do not establish a blood-THC threshold, a chronic lack of motivation, or impairment in a particular driver, employee, or witness.

The original primary study is Wardle, Pabon, Webber, and de Wit, Delta-9-tetrahydrocannabinol reduces willingness to exert effort in women, Psychopharmacology (2022), DOI 10.1007/s00213-021-06032-1. Its strongest contribution is an unusually concrete distinction between choosing to expend effort, physically executing that effort, and reporting that one feels drug effects.

What was the actual question the investigators tested?

The investigators asked whether a measured administered oral dose of THC changed participants’ willingness to select a harder, potentially better-paying task. The question was not whether participants could work in ordinary life, whether cannabis caused a psychiatric disorder, or whether THC impaired operation of a motor vehicle. The experiment operationalized willingness as the probability of choosing one of two specified button-pressing assignments after seeing the reward amount and chance of receiving it.

A person can choose an easier task for many reasons: less perceived reward, aversion to motor effort, a slower response, fatigue, medication effects, or deliberate preference. The experimental design varied the offered payoff and probability so those influences could be studied statistically. It did not isolate a singular neurological construct called “motivation,” and it could not directly observe anyone’s motivation outside the study.

Who participated, and why are 60, 58, and 40 different counts?

The investigators enrolled 60 healthy women between 18 and 35 years of age with some previous cannabis experience but without frequent recent use. Two women who consistently chose only the easy task were excluded under the protocol’s task-analysis rule, leaving 58 participants for the effort-choice modeling. Exclusion is analytically important: the final findings do not describe the response of those two always-easy-choice participants.

Before the study’s COVID-related protocol change, 40 participants completed placebo, 7.5-mg THC, and 15-mg THC sessions. A further 20 completed only placebo and 15-mg sessions. Two women lacked 15-mg effort-task observations because of adverse reactions; their other available observations were retained in the mixed model. Consequently, the 7.5-mg descriptive condition has a smaller denominator than the placebo and 15-mg comparisons. The source reports 7,452 choice observations from 58 analyzed individuals in the mixed-effects model; each trial is not an independent person.

The authors’ narrative gives the three-session count as 40, whereas a demographic table includes an incongruous subgroup heading. The protocol and the actual dose-specific Table 2 footnote support using 40 for the 7.5-mg task condition. It would be unsound to treat every row as an independent 58-person randomized parallel arm.

What did the women actually receive?

Participants swallowed opaque capsules containing pharmaceutical dronabinol formulated to supply 7.5 mg THC or 15 mg THC, or matching dextrose placebo. Treatment order was randomized and concealed during the repeated-session protocol, with visits separated by at least three days. Sessions lasted approximately four hours. The 7.5-mg condition was present only in the earlier three-session cohort, so the full set of dose contrasts was not equally observed across all 58 analyzed participants.

Those milligram values are nominal oral administered amounts. They are not measured concentrations in blood, plasma, serum, breath, urine, or oral fluid, and cannot be converted into a patient-specific forensic blood value without additional observations and assumptions. Oral absorption, first-pass metabolism, timing, individual prior exposure, and distribution all matter. The experiment did not collect serial THC plasma concentrations to estimate a pharmacokinetic curve or a per-participant concentration-response relationship.

How did the effort-for-reward task work?

The Effort Expenditure for Rewards Task, or EEfRT, offered an easy and a hard button-pressing option on each trial. The easy task required 30 presses using the dominant index finger within seven seconds for a potential one-dollar reward. The hard task required 100 presses using the nondominant little finger within 21 seconds, with possible rewards exceeding one dollar. The exact hard-task payoff varied from trial to trial.

Participants were told that a successfully completed trial might pay with a stated probability of 12%, 50%, or 88%. They therefore weighed reward size, chance, and effort. Choosing the hard option was the prespecified behavioral choice outcome; finishing the selected task and how quickly buttons were pressed were related but distinct endpoints. A person could choose the hard task yet fail to complete it, or choose the easy task despite being physically able to complete the hard one.

The high-reward task did not resemble a driving maneuver, courtroom examination, occupational safety inspection, or prolonged employment requirement. It was a structured, short-duration manual choice task with variable monetary incentives. The defensible result is therefore about this paradigm rather than a global determination that a THC-exposed individual is unmotivated.

What happened to hard-task choices under placebo and THC?

In the study’s descriptive Table 2, participants chose the hard task in 41% of placebo trials, 38% of 7.5-mg trials, and 35% of 15-mg trials, with respective reported standard deviations of 16, 21, and 24 percentage points. These percentages are descriptive summaries under unequal dose-condition participation; they are not proportions of women diagnosed as impaired. Their apparent small numerical separation should be interpreted alongside the repeated-observation statistical model, not substituted for it.

The generalized linear mixed-effects analysis found a significant linear THC-dose term for choosing the difficult task, with an odds ratio (OR) of 0.46, 95% confidence interval 0.26–0.81, p = 0.008. The reported 15-mg-versus-placebo pairwise odds ratio was 0.45. In contrast, the 7.5-mg-versus-placebo comparison was not statistically significant (OR 0.68, p = 0.21), and 7.5 mg and 15 mg did not significantly differ from one another in the reported pairwise contrast (OR 0.66, p = 0.18).

An odds ratio of 0.45 is not a 55-percentage-point drop in the proportion of women who work, nor a claim that 55% of the participants were impaired. It compares odds of a particular repeated laboratory choice under modeled conditions. Baseline choices, individual variability, reward characteristics, and repeated observations influence how odds translate into probabilities.

Did higher rewards and greater winning probabilities change the THC pattern?

Yes, but the details resist a simplistic conclusion that THC suppresses all goal-directed behavior equally. Across the task, difficult choices rose when possible payment or probability of winning rose. Descriptively, the hard option was selected in only 9% of trials at the 12% win-probability condition, compared with 43% at 50% and 61% at 88%. Higher reward magnitude was likewise associated with more hard choices.

The mixed model included interactions among dose, reward amount, and nonlinear probability terms. One reported linear-drug by quadratic-probability by reward-amount interaction was statistically significant (OR 2.15, 95% CI 1.01–4.58, p = 0.046). The authors interpreted THC’s effect as more evident where the potential expected value of selecting the harder task was moderate to high. This is a model interaction under the tested reward schedules, not an independently validated algorithm for predicting an individual’s decisions after taking a known amount of THC.

At low win probability and low offered reward, a particular comparison was nonsignificant (OR 1.03, p = 0.96); low probability with high reward was also nonsignificant (OR 0.56, p = 0.19). Significant contrasts occurred in some moderate- and high-value combinations. The number of interaction tests and the wide uncertainty of some estimates warrant caution about overinterpreting every plotted difference as a reproducible biological rule.

Did THC also slow the motor task?

Yes. The authors reported a dose-related slowing in button-pressing speed (linear drug effect p = 0.04). That matters because the hard task was physically more demanding and performed with the nondominant little finger. A person who anticipates pressing more slowly may rationally avoid a harder time-limited option even if the underlying subjective valuation of money is unchanged.

The investigators included keypress speed as a covariate in an additional model. They reported that the drug-related hard-choice result persisted. This supports the interpretation that simple measured button speed did not statistically explain the entire choice association. It does not experimentally separate motivation from attention, fatigue, movement initiation, expectations of failure, intoxication-related discomfort, or all other features of decision-making.

The number of trials completed also fell under THC: the source reported a modeled OR of 0.26 for successful completion with its linear drug contrast, p < 0.01. Completion is particularly vulnerable to motor speed and task persistence. It should not be merged with the hard-choice odds ratio as though both estimate the same endpoint or prove a specific executive deficit.

What physiological and subjective effects were measured?

Heart rate and blood pressure were monitored repeatedly during the visits, alongside the Addiction Research Center Inventory marijuana-effects scale. Participants endorsed more marijuana-like sensations after both active doses. These questionnaire scores indicate self-reported subjective drug effects. They are not chemical THC measurements, medical diagnoses, or objective tests of legal incapacity.

The study also found THC-associated increases in heart rate at both doses. In contrast, it did not find significant THC effects on systolic or diastolic blood pressure across the tested session interval. The significant heart-rate and subjective changes support pharmacological activity of the administered capsules, but they cannot independently quantify each woman’s circulating THC concentration or tell whether her performance fell below a legal or occupational threshold.

Several endpoints therefore changed in different ways: reward/effort choice, motor speed, task completion, self-reported drug sensations, and heart rate. That multidimensional pattern is more informative than treating all drug-related observations as interchangeable evidence of “impairment.”

Was THC directly measured in blood, plasma, urine, or breath?

No contemporaneous quantitative cannabinoid concentration series was reported as part of the effort-task analysis. The researchers obtained arrival urine and breath screening and collected blood for other study screening/hormonal purposes. That does not make the trial a study of analytically confirmed venous whole-blood THC concentrations. An administered 15-mg dronabinol capsule is an exposure intervention; it is not a blood test result.

The distinction among analytical biological matrices is critical. Whole blood contains both cells and liquid components. Plasma and serum are separated liquids with different collection and processing; their cannabinoid concentrations are not automatically interchangeable with whole-blood values. Urine screening typically answers a different question about excreted analytes or cross-reactive metabolites. Breath alcohol screening does not quantify THC in blood, and physiological heart-rate change is not an analytical substitute.

Where an actual specimen must be interpreted in litigation, method identity, chain of custody, collection timing, assay specificity, calibration, uncertainty, and confirmatory drug testing matter. None of those safeguards can be replaced with this study’s milligram administration label, laboratory choice percentage, or self-report scale.

Does the study establish a THC concentration threshold for impaired motivation?

No. A valid concentration threshold would require measurement of a named cannabinoid analyte in a specified specimen at appropriate times, paired with independently defined functional outcomes in a suitable population. This study experimentally varied oral dose but did not report measured postdose THC or THC-metabolite concentrations sufficient to fit a quantitative blood-level/effort-choice curve. Without those measurements, there is no basis to identify a nanogram-per-milliliter cutoff.

Nor could a numerical task-choice threshold alone establish legal impairment. The incentives, finger used, required presses, individual motor ability, baseline preferences, and chance of winning all influence observed choices. Laboratory group-average effects can coexist with substantial variation among participants and overlap between placebo and drug conditions. Translating this experimental observation into a conclusion about a particular person requires additional individualized evidence rather than assuming the group odds ratio applies to that person.

What does the study tell us about the “amotivational syndrome” claim?

It provides evidence of a short-term change in effort selection after controlled oral THC in the tested women. It does not demonstrate that people who use cannabis become chronically apathetic, have reduced work ethic, or experience permanent motivational decline. Participants were studied during discrete laboratory sessions after assigned capsules, not tracked through months or years of occupational behavior.

Indeed, the paper’s design deliberately focused on acute administration. The investigators discuss previous work comparing occasional and more frequent users, but this study itself was not a longitudinal test of persistent motivational change, receptor tolerance, or recovery after sustained abstinence. An acute response to 15 mg of a purified oral drug cannot simply be extended to smoked products of variable composition or to long-term use patterns.

Calling a difficult-task choice “motivation” can be useful scientific shorthand if the operational meaning remains visible. In expert analysis, the precise description—selection of a 100-press versus 30-press task under stated reward probabilities—is preferable to broad psychological labels that conceal what was actually observed.

Why does the all-female design matter?

The investigators recruited women specifically; the analyzed participants were 18–35 years old, medically screened, and not heavy recent cannabis users. This selection limits generalization to men, older adults, people with significant illness, long-term daily THC exposure, or workers using multiple medications. It would also be improper to read the all-female cohort as evidence that the same THC dose affects men less or more: the study contained no simultaneous male comparator and did not formally estimate such a sex difference.

Additional design constraints include varying prior cannabis exposure, modest overall sample size despite thousands of repeated trials, exclusion of consistently easy-choice participants, and reduced 7.5-mg participation following a protocol change. Blinding, randomized dose order, and placebo improve causal interpretation for the intervention in the tested sample, but they do not erase selection or generalizability limitations. Statistical modeling helps with repeated measures and missing sessions; it does not create absent biological measurements.

How should a forensic expert use this evidence?

First, identify the precise proposed inference. If the question is whether controlled oral THC can alter a laboratory decision to exert effort for reward, this study supplies affirmative evidence at 15 mg with caveats. If the question is whether a person’s postincident blood THC concentration proves unwillingness to work, the study cannot answer it. If the question is whether a subject could operate a vehicle safely, no instrumented driving, collision events, hazard detection, or field impairment criteria were studied.

Second, preserve the important nonsignificant findings. The individual 7.5-mg-versus-placebo comparison did not meet the stated significance criterion; neither did the contrast between active doses. Systolic and diastolic pressure did not significantly change. Some reward/probability combinations yielded null results. A careful expert should disclose those nulls, the observational uncertainty, and the original paper’s reporting inconsistencies rather than select only the strongest contrasts.

Third, separate four distinct propositions: a known drug dose was administered; participants endorsed characteristic drug effects; their choices and manual-task performance changed in measured ways; and a particular person’s real-world capacity or causal role in an incident might be disputed. The first three are experimental observations. The fourth is not established by this study alone.

Frequently asked questions about THC, effort, and forensic interpretation

1. Did every woman choose fewer hard tasks after THC?

No. The article reports aggregated descriptive percentages and mixed-effects associations across repeated choices. It does not establish that every participant moved in the same direction or that an individual effect could be predicted reliably.

2. Was 7.5 mg oral THC proven to have no motivational effect?

No. Its pairwise hard-choice comparison with placebo was nonsignificant (OR 0.68, p = 0.21). Nonsignificance is not proof of exact equivalence, especially with fewer participants receiving this condition.

3. Is a 0.45 odds ratio the same as 45% motivation?

No. It is a modeled relative odds measure for choosing the high-effort task under the experiment’s conditions. Motivation is not measured on a 0-to-100 clinical scale in this experiment.

4. Could button-pressing impairment explain the choice result?

Motor slowing was observed. Adjustment for the measured keypress-speed variable did not eliminate the choice association, but the design cannot exclude every behavioral or expectancy mechanism.

5. Did blood testing establish a legal THC level associated with these choices?

No. No paired quantitative forensic whole-blood THC measurement or validated blood concentration–task cutoff was reported. Administered milligrams and subjective symptoms cannot substitute for those missing assays.

6. Does this demonstrate chronic cannabis-related amotivation or workplace disability?

No. The experiment assessed acute laboratory choices after oral capsules, not permanent symptoms, job performance, or a formal functional-disability diagnosis.

7. Can these results establish that THC caused a particular accident or driving violation?

No. There was no driving trial or accident-causation analysis. Such an inference would require case-specific toxicology, timing, competing explanations, observed behavior, and the relevant legal standard.

Bottom line: a measured reward-effort choice, not a forensic impairment test

Wardle and colleagues demonstrated a statistically significant decrease in hard-task selection after the higher tested oral THC dose in women, while also finding motor slowing, fewer completed trials, increased subjective drug effects, and higher heart rate. Several contrasts—including the individual 7.5-mg choice comparison and blood-pressure changes—were nonsignificant. That combination is scientifically more informative than either declaring THC universally demotivating or dismissing the task effect because button speed changed.

The original primary report is Wardle et al., DOI 10.1007/s00213-021-06032-1. Its appropriate evidentiary role is to inform an expert about one controlled dose-related behavioral phenomenon, while preserving the absence of individual blood concentration data, external validity, and direct evidence concerning legally defined impairment or an incident’s cause.