Reactions to Warnings on Hookah Ads: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Reactions to Warnings on Hookah Ads (NCT03723616), a randomized na study. The primary outcome is Emotional Reactions to Ads.

2. Clinical Question

3. Trial Design

This was a randomized, single-masked, parallel study recorded as completed. The registry lists Duke University as the sponsor and an enrollment of 2588. The recorded study start date is 2018-10-05, and the primary completion date is 2021-03-26. The phase is listed as NA.

The registry description distinguishes an earlier stimulus-development process from the larger randomized online trial. Phase 1 and Phase 2 are described as small pilots used to inform and select study stimuli; Phase 3 is described as the larger randomized controlled online trial. The record does not report the detailed procedures of those pilots or explain masking beyond the single-masked designation.

The posted primary outcome has six ad conditions: social-allure ads without warnings, with text-only warnings, and with text-plus-graphic warnings; and product ads in those same warning formats. The posted comparison results use Social Allure Ads Only as the control for each listed contrast. This means the product-ad conditions are compared with social-allure ads without warnings, not with a product-ad-only reference condition. Those comparisons therefore reflect differences across both ad type and warning format where the product ads are involved.

The registry lists perceived risk, global attitude toward waterpipe tobacco smoking, urge to smoke, and avoidance or acceptance of warnings measured by eye tracking as secondary outcomes. Numeric secondary results are posted for perceived risk, global attitude, and urge to smoke. Results for the eye-tracking outcome are not reported in the registry record. The protocol’s original sample-size justification and the detailed statistical analysis plan are also not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Participants were randomized to one of three health warning label (HWL) arms in a 1:2:2 ratio: 1) ads only, 2) ads with text HWLs only, and 3) ads with text + graphic HWLs within the Social Allure Ads and Product Ads groups.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants who completed the study.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Participants who completed the study. Baseline Characteristics information was collected prior to further stratification into one of three health warning label (HWL) arms in a 1:2:2 ratio: 1) ads only, 2) ads with text HWLs only, and 3) ads with text + graphic HWLs.

6. Primary Endpoint

The registry lists the primary outcome as Emotional Reactions to Ads, measured over up to 45 minutes.

The primary outcome is emotional reactions to advertisements, measured for participants who completed the study. Participants rated how each ad made them feel on bipolar scales covering anxiety, worry, fear, and whether the feeling was good or bad. The ratings were combined into a composite by averaging the items. Scores range from 1 to 7, with higher mean values representing more negative emotional reactions to viewing the ad.

A composite can summarize several related responses in a single outcome, rather than treating each feeling as a separate primary result. Here, the items cover different aspects of emotional response to an advertisement. The measure therefore describes participants’ reported reactions to the ad stimuli, not a clinical outcome or a measure of later behavior. The registry record does not report the detailed scoring rules beyond the stated averaging, nor does it describe a threshold for interpreting a particular score as meaningful.

The outcome timeframe is listed as up to 45 minutes. The posted arm summaries give a mean score and its standard error for each condition. The comparisons express the mean difference relative to Social Allure Ads Only. Interpretation should keep the scale direction in view: a higher mean indicates a more negative emotional response, so a positive difference relative to the control corresponds to more negative reported reactions in the compared condition.

7. Results Posted in the Registry

• Social Allure Ads With Text-only Health Warning Labels (HWLs): mean 3.46 versus 2.81 on control; difference 0.65 (95% CI 0.45 to 0.85), standardized effect d = 0.45; z-test p = <0.001.
• Social Allure Ads With Text + Graphic Health Warning Labels (HWLs): mean 3.74 versus 2.81 on control; difference 0.93 (95% CI 0.73 to 1.13), standardized effect d = 0.65; z-test p = <0.001.
• Product Ads Only: mean 2.83 versus 2.81 on control; difference 0.02 (95% CI -0.23 to 0.27), standardized effect d = 0.01; z-test p = 0.876.
• Product Ads With Text-only Health Warning Labels (HWLs): mean 3.60 versus 2.81 on control; difference 0.79 (95% CI 0.58 to 1.00), standardized effect d = 0.57; z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Perceived Risk as Measured by Questionnaire

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Social Allure Ads With Text-only Health Warning Labels (HWLs): mean 4.93 versus 4.71 on control; difference 0.22 (95% CI 0.05 to 0.39), standardized effect d = 0.18; z-test p = 0.011.
  • Social Allure Ads With Text + Graphic Health Warning Labels (HWLs): mean 5.03 versus 4.71 on control; difference 0.32 (95% CI 0.15 to 0.49), standardized effect d = 0.26; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Social Allure Ads With Text-only Health Warning Labels (HWLs) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Social Allure Ads With Text + Graphic Health Warning Labels (HWLs) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Global Attitude on Waterpipe Tobacco Smoking (WTS)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Social Allure Ads With Text-only Health Warning Labels (HWLs): mean 3.83 versus 4.14 on control; difference -0.31 (95% CI -0.55 to -0.07), standardized effect d = -0.18; z-test p = 0.011.
  • Social Allure Ads With Text + Graphic Health Warning Labels (HWLs): mean 3.61 versus 4.14 on control; difference -0.53 (95% CI -0.77 to -0.29), standardized effect d = -0.31; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Social Allure Ads With Text-only Health Warning Labels (HWLs) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Social Allure Ads With Text + Graphic Health Warning Labels (HWLs) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Urge to Smoke

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Social Allure Ads With Text-only Health Warning Labels (HWLs): mean 3.28 versus 3.34 on control; difference -0.06 (95% CI -0.34 to 0.22), standardized effect d = -0.03; z-test p = 0.677.
  • Social Allure Ads With Text + Graphic Health Warning Labels (HWLs): mean 3.06 versus 3.34 on control; difference -0.28 (95% CI -0.56 to 0.00), standardized effect d = -0.14; z-test p = 0.052.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Social Allure Ads With Text-only Health Warning Labels (HWLs) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Social Allure Ads With Text + Graphic Health Warning Labels (HWLs) and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes provide related but not identical perspectives on responses to the ads. The posted numeric results include questionnaire-rated perceived risk, global attitude toward waterpipe tobacco smoking, and urge to smoke. The registry also lists avoidance or acceptance of warnings assessed by eye tracking, but results for that outcome are not reported in the registry record.

At a broad level, the reported secondary summaries do not all move in the same way or with the same clarity. The perceived-risk comparisons for social-allure warnings are positive relative to the no-warning control, while global-attitude comparisons are negative. The urge-to-smoke contrasts are less clearly separated from zero in the unadjusted calculations. These patterns can be considered alongside the primary emotional-reaction results: the outcomes represent different constructs, so a pattern in emotional response does not necessarily imply a parallel pattern in perceived risk, attitude, or urge.

The secondary outcomes can help describe the range of immediate reactions measured during the study, but they do not replace the primary endpoint. They also should not be combined into a single overall conclusion without a prespecified method for doing so. The record does not describe such a combined analysis.

Unadjusted secondary results carry less weight because multiple outcomes and contrasts create more opportunities for chance findings. The registry record does not report a multiplicity correction or a confirmatory hierarchy for these secondary analyses. Their estimates are useful for generating questions and describing the posted data, but the p-values should not be interpreted as if each were the sole planned test. The eye-tracking result’s absence also means that one of the listed secondary domains cannot be compared with the reported questionnaire summaries.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Perceived Risk as Measured by Questionnaire
• Global Attitude on Waterpipe Tobacco Smoking (WTS)
• Urge to Smoke
• Avoidance / Acceptance of Health Warnings as Measured by Eye Tracking

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: it is a mean score on a scale, not a count of events. The posted summaries report a mean and standard error for each condition. A mean gives the average score among the participants included in that condition’s outcome summary. The standard error describes the precision of the estimated mean; it is not the spread of individual participants’ scores. A small standard error generally indicates a more precisely estimated mean, while individual scores may still vary substantially.

What does a mean difference represent?

Each listed comparison expresses the difference between a warning or product-ad condition and Social Allure Ads Only. A positive mean difference indicates a higher score in the compared condition. For this outcome, higher values represent more negative emotional reactions. Thus, a positive difference corresponds to a higher average negative-reaction score relative to the control. The difference is measured in the original scale units, which makes it interpretable alongside the range of 1 to 7.

The mean differences are recalculated, unadjusted contrasts derived from the posted group summaries. They do not incorporate covariate adjustment or a correction for making multiple comparisons. The registry record lists zero reported analyses, so it does not give a sponsor-reported model or analysis result to compare with these recalculations. The record also does not report whether the original analysis used additional adjustments, how repeated ratings of ads were handled, or how any dependence among ratings was addressed. Those details are not reported in the registry record.

How should the confidence intervals be read?

A confidence interval gives a range of values compatible with the data under the method used to calculate it. It communicates uncertainty in the estimated mean difference, not the range of individual responses. A narrower interval indicates greater precision for the estimated contrast; a wider interval indicates less precision. If an interval includes zero, the data are compatible with no mean difference as well as with differences on either side of zero at the interval’s stated confidence level. If it lies wholly on one side of zero, the estimated difference is more clearly separated from zero under that calculation. The interval does not show the probability that the true effect lies within its endpoints, and it does not resolve limitations in the design or analysis.

The posted p-values are unadjusted. A p-value summarizes how compatible the observed data are with a null model of no difference, assuming the calculation’s model and conditions. It is not the probability that the null hypothesis is true, nor does it measure the size or practical importance of a difference. A small p-value alone cannot show that an effect is important, free from bias, or applicable beyond the study conditions.

Why does multiplicity matter?

Several conditions are contrasted with the same control, and the registry lists several outcomes. When many hypotheses are examined, the chance of at least one apparently notable result can increase if no multiplicity strategy is used. The record does not report a multiplicity adjustment or a hierarchy for secondary outcomes. Consequently, unadjusted secondary p-values should be treated as exploratory rather than as a set of confirmatory tests. The primary endpoint has a central role in the registry record, but its detailed prespecified testing strategy is not reported.

General statistical context: a randomized design can support comparisons between assigned conditions when the assignment and analysis are appropriately implemented. Randomization does not by itself guarantee that every potential source of bias is removed, particularly if outcomes are missing, analyses depart from the plan, or the assignment procedure is not fully described. The registry record identifies the study as randomized, but the allocation procedure, missing-data handling, analysis population, and detailed model specifications are not reported in the registry record.

What can the standard errors tell us?

The arm-level standard errors accompany the means and contribute to uncertainty estimates for contrasts. They quantify sampling precision under the calculation, rather than describing the full distribution of participant responses. General statistical context: if observations are clustered or repeated within individuals, a method may need to account for that dependence; otherwise standard errors can misstate precision. The registry record does not provide enough detail to determine how the primary analysis handled the fact that participants rated ads or how many ratings contributed per participant.

The trial’s reported enrollment is 2588, while the primary outcome is defined among participants who completed the study and the arm summaries have their own participant counts. Enrollment should therefore not be treated as the denominator for every outcome comparison. The record does not report the number of enrolled participants with missing primary outcome data, reasons for missingness, or the method used to handle such data. These omissions limit assessment of whether the posted summaries represent all randomized participants or a selected completion group.

11. Confidence Intervals and Precision

The primary outcome summaries include standard errors for each condition, and the pairwise estimates include 95% confidence intervals. The interval width conveys how precisely each mean difference is estimated under the unadjusted calculation. For social-allure ads with text-only warnings, the estimated difference from control was 0.65 and the interval ran from 0.454 to 0.846. For social-allure ads with text-plus-graphic warnings, the estimated difference was 0.93 and the interval ran from 0.734 to 1.126. Both intervals are above zero, indicating that the calculated differences are separated from zero under this method.

For Product Ads Only versus the social-allure no-warning control, the estimated difference was 0.02 and the interval ran from -0.231 to 0.271. Because that interval includes zero, the result is compatible with no difference as well as with differences in either direction. The intervals for product ads with text-only warnings and product ads with text-plus-graphic warnings ran from 0.582 to 0.998 and from 1.012 to 1.428, respectively. Both are above zero in these unadjusted calculations.

Precision is not the same as importance. A confidence interval can be relatively narrow while still leaving uncertainty about how meaningful a score difference is, and a result with a small p-value does not itself establish practical importance. The registry record does not report a threshold for a meaningful change on this composite scale. Nor does it provide adjusted intervals that account for the set of comparisons. The intervals should therefore be read as uncertainty estimates for the individual unadjusted contrasts, not as a complete account of uncertainty across all outcomes and comparisons.

The means and standard errors describe the reported summaries; they do not show the distribution of individual responses. The registry record does not report additional distributional information or the detailed model used for the sponsor’s analysis. Interpretations of precision are consequently limited to the posted summaries and the recalculated comparisons.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm for 80% power and participants per arm for 90% power. They are not total sample sizes or total event counts. The calculations use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning scenario targets the stated fraction of the observed effect: 100%, 75%, or 50%. These are planning illustrations based on the observed effect, not observed power. The original trial’s design assumptions and sample-size justification are not reported in the registry record.

In general, the 80% power column describes the number of participants per arm estimated to give an 80% chance of detecting the assumed effect under the planning assumptions, if that effect is present and the calculation’s assumptions hold. The 90% power column describes the corresponding planning target for a 90% chance. Greater planned power usually requires a larger sample under otherwise similar assumptions. The figures are conditional on the assumed effect size and do not guarantee a particular result.

The effect-size scenarios should be read as alternatives for planning, not as separate observed findings. A smaller effect is harder to distinguish from random variation and therefore needs more participants per arm to reach the same power target. This is why the planning figures rise when the assumed effect is reduced from the full observed effect to a smaller fraction. When an observed contrast is very small, the illustrative participant requirements can become very large; such projections are especially sensitive to the observed estimate and should not be treated as a recommendation for a future study without an independent design rationale.

These calculations do not account for details not included in the stated assumptions, such as missing outcome data, any planned adjustment for multiple comparisons, or complications in the original trial design. The registry record does not report whether these illustrations match the trial’s original planning process; they are separate from the study’s enrollment and observed results.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.650; pooled standard deviation σ = 1.444.
2. n = 2 × (1.96 + 0.842)² × 1.444² / 0.650² = 77.5, rounded up to 78 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.65                      # observed difference in means
sigma <- 1.44397876254004                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(3723616)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 78 (80% power), 104 (90% power)
 75% of observed effect: n per arm = 138 (80% power), 185 (90% power)
 50% of observed effect: n per arm = 310 (80% power), 415 (90% power)
power.t.test: 78.4 
Simulated power at n = 78 per arm: 0.797 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists multiple secondary outcomes and the primary results include several contrasts against a shared control. The posted p-values are unadjusted. No correction for multiple testing, primary-versus-secondary testing hierarchy, or prespecified order of comparisons is reported in the registry record. When many outcomes or contrasts are examined, interpreting each p-value in isolation can overstate how much evidence the collection of results provides. This is particularly relevant for secondary findings, which are often used to explore patterns rather than to make a single confirmatory claim.

The record reports no analyses, and a detailed statistical analysis plan is not reported in the registry record. It is therefore not possible to determine from the registry information whether the displayed comparisons match the prespecified analysis, whether covariates were included in the sponsor’s model, or how the shared control and multiple ad ratings were handled. No discrepancy is flagged, but the absence of reported sponsor analyses means these recalculated unadjusted comparisons should not be mistaken for a full account of the sponsor’s analysis.

The primary population is described as participants who completed the study. The record does not report how many randomized participants failed to complete the study, whether primary outcome data were missing, why any data were missing, or how missing data were handled. It also does not report whether the analysis followed an intention-to-treat principle or another analysis population definition. Without these details, the extent to which the posted outcome summaries preserve the benefits of random assignment cannot be fully assessed.

These omissions do not demonstrate that the analyses were inappropriate; they mark aspects that cannot be evaluated from the registry record. The distinction matters because the posted means and recalculated comparisons describe the available summaries, while conclusions about a prespecified testing strategy, missing-data assumptions, and the full analysis population require information not reported in the registry record.

15. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The registry’s serious adverse event summaries report zero serious events in each listed condition and zero deaths in each condition. The other adverse event counts are also recorded as zero across the listed groups. These are the reported counts for the study record; they should not be interpreted as proof that no adverse event could occur or that the study assessed every possible harm.

The displayed at-risk counts differ by condition, as do the participant counts associated with the outcome summaries. The serious-event figures are counts, not estimates of comparative risk with confidence intervals. The registry record does not provide event descriptions because no serious-event terms are listed, and it does not describe the method or completeness of adverse-event ascertainment in the information available for these results.

The trial concerned responses to advertisements and health warnings, and the posted primary endpoint is a short-term questionnaire measure. The safety record should be read within that study context. It does not establish safety or risk outside the study’s procedures, population, or observation period.

16. What the Result Does, and Does Not, Mean

The posted primary results describe immediate reported emotional reactions to waterpipe tobacco advertisements. On the composite scale, higher scores represent more negative reactions. The warning conditions had higher mean scores than Social Allure Ads Only in the posted comparisons, while Product Ads Only was close to that control and its unadjusted confidence interval included zero. The results therefore speak to reactions to the particular ad conditions measured in this study, not to longer-term behavior or health outcomes.

The findings do not by themselves establish that a warning changes smoking behavior, prevents tobacco use, or changes health. Those outcomes are not the primary endpoint described in the registry record. The study measured reported reactions up to 45 minutes, and the record does not report follow-up outcomes that would show whether those reactions persisted or translated into later actions.

The comparison structure also matters. Product-ad conditions were compared with social-allure ads without warnings. Those contrasts mix differences in ad type with warning format, so they cannot isolate the effect of warning format among product ads using the posted contrasts alone. Similarly, the reported comparisons do not directly test text-only warnings against text-plus-graphic warnings. Such direct contrasts would require their own analysis.

Statistical evidence and practical importance are distinct. The confidence intervals and p-values describe the uncertainty and compatibility of the unadjusted mean differences under the calculations used. The registry does not report a threshold for a meaningful difference on the emotional-reaction scale. It is therefore not possible to determine from these results alone how consequential a given score difference is for participants or for decisions about advertising policy.

Finally, the registry record does not report enough detail to assess the full analysis plan, missing-data handling, or multiplicity strategy. The results are informative as posted summaries, but their interpretation should remain tied to the measured short-term outcomes and the limitations of the record.

17. Why This Trial Matters Statistically

This trial illustrates why the definition and direction of an outcome matter. The primary measure combines ratings of anxiety, worry, fear, and good-or-bad feelings, and higher scores are explicitly defined as more negative reactions. Without that direction, a higher mean could be misread. The scale’s meaning provides essential context for interpreting the contrasts.

It also shows the importance of identifying the actual comparator. The registry’s primary contrasts use Social Allure Ads Only as the common control. Product-ad results therefore do not answer the same question as a product-ad comparison against product ads without warnings. A contrast is defined not only by the outcome and intervention, but also by the reference condition.

The results offer an opportunity to distinguish statistical precision from strength of evidence across a family of analyses. Confidence intervals clarify the uncertainty around each unadjusted difference, while the absence of reported multiplicity handling and sponsor analyses limits how confidently the collection of comparisons can be treated as confirmatory. Reporting estimates and intervals makes the size and direction of results visible in a way that p-values alone cannot.

For clinical researchers and biostatisticians, the record also emphasizes the value of transparent reporting on analysis populations, missing outcomes, planned comparisons, and the handling of multiple outcomes. These details are needed to connect registry summaries with the original study plan and to assess whether an observed pattern is robust. The trial’s endpoint concerns immediate reactions to ad stimuli; keeping that scope clear prevents extrapolating the results to behavioral or health outcomes not measured in the primary endpoint.

18. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record gives limited information about participant eligibility, recruitment, and the detailed online procedures. Although it describes the target context as young adults and defines the primary outcome population as participants who completed the study, it does not report further population characteristics in the information available for these results. Generalizability beyond the participants and ad stimuli used in the study therefore cannot be assessed fully from this record.

The registry describes pilot phases used to inform and select stimuli, but it does not provide their detailed methods or explain how the final stimuli were selected. The posted comparison structure also uses Social Allure Ads Only as a common control for product-ad conditions. Consequently, those product-ad contrasts do not isolate warning-format differences within product ads. The posted results likewise do not give direct pairwise tests between warning formats.

Important analysis details are not reported in the registry record, including the prespecified analysis plan, analysis population, missing-data methods, and handling of multiple comparisons. The record lists no reported analyses, while the displayed unadjusted comparisons are recalculated from summary statistics. The absence of those details limits assessment of whether the results align with the planned analysis and whether the uncertainty estimates account for all relevant design features.

The primary measure captures short-term self-reported emotional response, with a timeframe of up to 45 minutes. The record does not report a threshold for a meaningful scale difference or evidence that these ratings predict later behavior. Numeric results for the listed eye-tracking secondary outcome are not reported in the registry record. These limits should shape interpretation of the range and practical meaning of the posted findings.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Reactions to Warnings on Hookah Ads — NCT03723616.
• Registry record last updated 2025-12-18; status last verified 2025-12.
• Results first posted 2025-12-18.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

NCT03723616, titled “Reactions to Warnings on Hookah Ads,” is recorded as a completed, randomized, single-masked, parallel study sponsored by Duke University. The registry lists an enrollment of 2588. Its primary outcome is emotional reactions to ads, reported as a composite mean on a scale from 1 to 7, with higher scores representing more negative reactions. The outcome timeframe is up to 45 minutes, and the primary population is participants who completed the study.

The posted means are higher for warning conditions than for Social Allure Ads Only in the primary comparisons. Product Ads Only is close to that control, with an unadjusted interval that includes zero. The comparisons are unadjusted recalculations from posted summary statistics. The registry record does not report sponsor analyses, the original sample-size justification, detailed missing-data methods, or a multiplicity strategy. The results describe short-term ad reactions and should not be interpreted as evidence of later behavior or health outcomes.