Revolutionizing Normative Re-education: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Revolutionizing Normative Re-education (NCT04356261), a randomized na study. The primary outcome is Change From Baseline Daily Drinking at 4 Months.

2. Clinical Question

The trial’s stated aims also included identifying an optimal feedback dose, examining person-level moderators, and evaluating voluntary engagement and sustainability outside the randomized trial. The posted primary results address the drinking outcome at 4 months; they do not, by themselves, settle those broader aims. In particular, the registry record does not report a single overall finding that establishes an optimal dose or describes the voluntary-use evaluation.

3. Trial Design

The registry lists a completed, randomized, single-masked, parallel-group behavioral trial sponsored by Loyola Marymount University. The phase is listed as not applicable. Enrollment was 1143. The record gives a start date of 2021-08-15 and a primary completion date of 2024-10-25. It describes three groups: Control PNF, Light Dose of Alcohol PNF, and Heavy Dose of Alcohol PNF.

Control participants received personalized normative feedback about non-alcohol and health-related topics during all weekly rounds. Participants in the light-dose group received alcohol-use feedback in 33% of weekly rounds, while the heavy-dose group received alcohol-use feedback in 67% of weekly rounds. The intervention was described as a weekly game centered on students’ perceptions of peers’ attitudes and behaviors in campus-relevant domains. The randomized comparison therefore concerns the content and frequency of alcohol-related feedback within the game, against an active feedback control, rather than against no intervention.

The registry’s project description refers to a planned large-scale multi-site trial, but the record does not establish the number of sites or report site-specific results. The game was intended to run across 12 weeks, but detailed information about adherence, allocation procedures, masking procedures, or protocol deviations is not reported in the registry record. The registry also does not report the original sample-size justification, the planned statistical analysis plan, or whether interim analyses were conducted.

Descriptive primary-outcome statistics include participants who completed the 4-month follow-up: 293 in Control PNF, 289 in Light Dose of Alcohol PNF, and 311 in Heavy Dose of Alcohol PNF. The record says regression analyses included participants with complete data on all covariates and that model sample sizes varied slightly across outcomes. Details about missing-data handling beyond those statements are 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 study condition follow-up the baseline survey (Started) and prior to any follow-up survey (Waves 2-4). Waves represent when follow-up surveys were collected (number of months post baseline). Follow-up surveys were collected 4-, 9-, and 12-months post baseline.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Descriptive statistics reflect all 893 participants who completed the 4-month follow-up assessment (293 Control, 289 Light Dose, 311 Heavy Dose). Regression analyses included participants with complete data on all covariates. Sample sizes in regression models varied slightly across outcomes due to missing covariate data (approximately 889 participants total across most analyses).

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: 1143 Participants were assigned to condition.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline Daily Drinking at 4 Months, measured over baseline, 4 months.

The primary endpoint is titled change from baseline in daily drinking at 4 months and is measured in drinks per week. The registry describes the Daily Drinking Questionnaire as assessing the number of drinks consumed in a typical week over the past three months. This gives the endpoint a defined behavioral focus and a period of reference, rather than asking only for a broad impression of drinking.

A continuous measure can describe differences in average drinking across randomized groups and can retain more gradation than a simple yes-or-no outcome. The registry reports group means and standard deviations for the primary outcome. The descriptive results cover 293 Control PNF participants, 289 Light Dose participants, and 311 Heavy Dose participants who completed the 4-month follow-up. Regression analyses are described separately as using participants with complete data on all covariates.

The outcome name indicates change from baseline, and the stated timeframe is baseline and 4 months. The posted data do not explain in detail how the change score was constructed or how baseline values entered the sponsor’s regression models. The sponsor’s reported analyses use negative binomial regression and report incidence rate ratios, which are not the same summary as an unadjusted difference between the posted group means. That distinction matters when reading the descriptive values alongside the sponsor’s inferential results.

The endpoint is useful for assessing the trial’s stated drinking-related question, but it is only one outcome at one follow-up. It does not alone describe peer-norm perceptions, longer-term outcomes, engagement, or the experience of students outside the randomized trial.

7. Results Posted in the Registry

• Light Dose of Alcohol PNF: mean 1.79 versus 2.11 on control; difference -0.32 (95% CI -1.20 to 0.56), standardized effect d = -0.06; z-test p = 0.477.
• Heavy Dose of Alcohol PNF: mean 2.18 versus 2.11 on control; difference 0.07 (95% CI -0.71 to 0.85), standardized effect d = 0.01; z-test p = 0.861.

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.

Change From Baseline Descriptive Peer Drinking Norms- Drinking Norms Rating Form at 4 Months

• 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 deviations.
• Result:
  • Light Dose of Alcohol PNF: mean 5.75 versus 7.00 on control; difference -1.25 (95% CI -2.48 to -0.02), standardized effect d = -0.16; z-test p = 0.046.
  • Heavy Dose of Alcohol PNF: mean 5.76 versus 7.00 on control; difference -1.24 (95% CI -2.41 to -0.07), standardized effect d = -0.17; z-test p = 0.039.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Light Dose of Alcohol PNF 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 Heavy Dose of Alcohol PNF than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline Descriptive Peer Drinking Norms- Drinking Norms Rating Form at 9 Months

• 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 deviations.
• Result:
  • Light Dose of Alcohol PNF: mean 5.45 versus 5.47 on control; difference -0.02 (95% CI -1.15 to 1.11), standardized effect d = -0.00; z-test p = 0.972.
  • Heavy Dose of Alcohol PNF: mean 5.36 versus 5.47 on control; difference -0.11 (95% CI -1.24 to 1.02), standardized effect d = -0.02; z-test p = 0.849.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Light Dose of Alcohol PNF 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 Heavy Dose of Alcohol PNF and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline Descriptive Peer Drinking Norms- Drinking Norms Rating Form at 12 Months

• 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 deviations.
• Result:
  • Light Dose of Alcohol PNF: mean 6.08 versus 6.91 on control; difference -0.83 (95% CI -2.13 to 0.47), standardized effect d = -0.10; z-test p = 0.209.
  • Heavy Dose of Alcohol PNF: mean 5.65 versus 6.91 on control; difference -1.26 (95% CI -2.53 to 0.01), standardized effect d = -0.16; z-test p = 0.052.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Light Dose of Alcohol PNF 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 Heavy Dose of Alcohol PNF and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary results cover selected peer-drinking-norm measures at 4, 9, and 12 months. Across the unadjusted summaries, the pattern is not uniform over follow-up: the estimates at the earlier assessment point generally favor lower descriptive peer-drinking norms in the alcohol-feedback groups, while estimates at later assessments are closer to zero or remain uncertain. The reported intervals and sponsor analyses do not establish a consistent separation across outcomes and time points.

These norm measures are related to the trial’s proposed mechanism, but they are not interchangeable with the primary drinking endpoint. A change in perceived peer drinking is not itself a measure of an individual’s own drinking. The secondary findings can therefore add context to the primary result, but they cannot substitute for it or establish that changes in norms led to changes in drinking. The primary result at 4 months remains uncertain, and the secondary pattern does not resolve that uncertainty.

The registry lists 18 outcome measures and 35 analyses, while 9 secondary outcomes are posted. The range of measures, follow-up times, and reported analyses increases the chance that an apparently noteworthy result could occur by chance if many comparisons are considered. Unadjusted secondary results consequently carry less weight than a clearly prespecified primary analysis, particularly when the registry record does not describe a multiplicity strategy or outcome hierarchy. The record does not report whether the secondary analyses were adjusted for multiple comparisons.

Interpretation is further limited by differences in the participants contributing to each follow-up and by the statement that regression sample sizes varied slightly across outcomes. The record does not give a full account of missing observations or analysis populations for each secondary outcome. The broad reading is therefore one of mixed, time-specific estimates rather than a single consistent secondary conclusion.

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

• Change From Baseline Descriptive Peer Drinking Norms- Drinking Norms Rating Form at 4 Months
• Change From Baseline Descriptive Peer Drinking Norms- Drinking Norms Rating Form at 9 Months
• Change From Baseline Descriptive Peer Drinking Norms- Drinking Norms Rating Form at 12 Months
• Change From Baseline Descriptive Peer Drinking Norms- Modified Quantity, Frequency, Peak Index at 4 Months
• Change From Baseline Descriptive Peer Drinking Norms- Modified Quantity, Frequency, Peak Index at 9 Months
• Change From Baseline Descriptive Peer Drinking Norms- Modified Quantity, Frequency, Peak Index at 12 Months
• Change From Baseline Injunctive Peer Drinking Norms (Adapted From Baer, 1994) at 4 Months
• Change From Baseline Injunctive Peer Drinking Norms (Adapted From Baer, 1994) at 9 Months

9. Statistical Interpretation

10. Statistical Methods Explained

The registry’s primary descriptive summaries are means and standard deviations. A mean gives the average value in a group; the standard deviation describes how spread out individual values are around that average. The posted primary means are 2.11 drinks per week for Control PNF, 1.79 for Light Dose, and 2.18 for Heavy Dose. The corresponding standard deviations are 4.97, 5.84, and 4.81. The wide spread relative to the means indicates that participants’ reported values varied considerably. A group mean alone does not describe that variation or the distribution’s shape.

What does a mean difference represent?

The recalculated mean difference is the control-group mean subtracted from the relevant feedback-group mean, using the posted summaries. A negative value means the feedback group’s posted mean is lower; a positive value means it is higher. These are unadjusted comparisons, not regression estimates. For the primary endpoint, the recalculated differences are -0.32 for Light Dose versus Control PNF and 0.07 for Heavy Dose versus Control PNF. Their confidence intervals span zero, so the summaries do not pin down a difference in one direction with precision.

The effect-size values, -0.059 for Light Dose and 0.014 for Heavy Dose, express the unadjusted differences relative to a pooled standard deviation. Such standardized measures can help describe the magnitude of an observed contrast on a scale that accounts for the outcome’s spread. They are descriptive quantities, not proof of a causal or clinically important effect. The registry does not report a threshold for interpreting these effect sizes as important.

Why are there two kinds of analysis?

The sponsor’s reported primary analyses use negative binomial regression and report incidence rate ratios. Regression estimates can account for modeled covariates, whereas the simple recalculations use only the posted group-level summaries. An incidence rate ratio compares modeled rates; it is not a difference in mean drinks per week. The registry record does not report the full model specification, the covariates included, or the mapping of the two posted estimates to the individual dose-versus-control contrasts. These details limit direct reconciliation of the estimates.

Negative binomial regression is generally used for count-like outcomes when variability exceeds what a simpler count model would accommodate. This extra variability is called over-dispersion. In general, allowing for over-dispersion can produce more suitable uncertainty estimates when observations vary more than a restrictive model expects. The registry identifies the method but does not report diagnostic checks or its rationale for choosing it. The endpoint is described in drinks per week, while the sponsor reports incidence rate ratios; the precise model outcome construction is not reported in the registry record.

How should confidence intervals and p-values be read?

A confidence interval expresses the uncertainty around an estimated effect under the analysis assumptions. A wider interval indicates less precision. For a difference, zero is the value representing no difference; for an incidence rate ratio, 1 represents no ratio difference. The posted primary recalculated intervals include zero, and the sponsor’s reported incidence rate ratio intervals include 1. This means the results are not precise enough to rule out the corresponding null value at the reported confidence level.

A p-value describes how compatible the observed data are with a specified null model, under the analysis assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or importance of an effect. The sponsor reports p-values of .155 and .845. The recalculated unadjusted comparisons report p-values of 0.477 and 0.861. These should not be treated as interchangeable because the methods and estimands differ.

Randomization supports causal comparisons in principle by creating groups whose assignment is determined by chance. However, interpretation also depends on outcome measurement, follow-up, missing data, adherence, and the analysis. The registry record does not report enough detail to evaluate all of these factors. It reports a single-masked design, but does not say who was masked or how masking was maintained.

Finally, the registry lists 18 outcome measures and 35 analyses, while 9 secondary outcomes are posted in the record. When many outcomes or contrasts are examined, some results can appear noteworthy by chance. Unless the analysis plan and multiplicity procedures are known, an isolated secondary finding deserves more caution than a result from a prespecified primary analysis. The registry record does not report whether multiplicity adjustments or a hierarchy of outcomes were used.

11. Confidence Intervals and Precision

The primary unadjusted mean-difference intervals are broad relative to their point estimates. For Light Dose versus Control PNF, the estimate is -0.32 drinks per week, with a 95% confidence interval from -1.202 to 0.562. For Heavy Dose versus Control PNF, the estimate is 0.07, with a 95% confidence interval from -0.711 to 0.851. Each interval crosses zero. The data therefore do not locate either unadjusted difference precisely on one side of no difference.

The sponsor’s negative binomial results are also uncertain: the reported incidence rate ratio intervals are 0.481 to 1.123 and .640 to 1.441. Each spans 1, the no-ratio-difference value. The reported estimates and intervals are not evidence that the groups are identical; they show that the fitted analyses leave a range of plausible values that includes the null and values on either side. The record does not identify which of the two model estimates corresponds to which alcohol-feedback dose in its posted analysis entries.

Precision depends on the amount of information, the variability of the outcome, and the analysis. The primary descriptive groups contain 293, 289, and 311 participants, but the standard deviations are large relative to the means. The registry also notes that the regression models use complete covariate data and that sample sizes vary slightly across outcomes. These features matter when interpreting uncertainty and when comparing model results with the descriptive summaries.

Confidence intervals are not guarantees that the true value lies within a particular interval, and they do not account for every possible source of bias or model misspecification. The registry does not report enough information about missing-data methods, model diagnostics, or analysis planning to assess those issues fully. Overall, the posted primary estimates are compatible with no difference, but their intervals also leave room for differences in either direction.

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 describe hypothetical sample sizes for future studies based on the observed unadjusted effects; they are not the trial’s original design calculations. For these continuous outcomes, the figures are participants per arm, not total events. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set at the stated fraction of the observed effect. The trial’s own original sample-size assumptions are not reported in the registry record.

The 80% power column gives the participants per arm estimated to provide an 80% chance of detecting the specified effect if that effect is real and the planning assumptions hold. The 90% power column describes the corresponding sample size for a 90% chance. These are design-planning quantities, not a measure of the probability that the trial’s conclusion is correct. The figures are not observed power.

Rows for 100%, 75%, and 50% of the observed effect show planning for progressively smaller effects relative to the observed estimate. Smaller effects are harder to distinguish from random variation, so they require more participants under the same assumptions. When an observed effect is very close to zero, the resulting planning figures can become extremely large and unstable; they should not be read as a reliable forecast of what a future trial must enroll. The figures also inherit uncertainty from the observed effect and do not account for design features that are not listed in the planning assumptions.

Because the calculations compare two arms at a time, they do not automatically represent a sample-size calculation for a full three-arm trial or account for multiple comparisons. The table and curve are best read as illustrations of how sample size changes with the assumed target effect, not as evidence that the completed trial had a particular level of power.

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 = -0.320; pooled standard deviation σ = 5.419.
2. n = 2 × (1.96 + 0.842)² × 5.419² / -0.320² = 4,502.5, rounded up to 4,503 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 <- -0.32                      # observed difference in means
sigma <- 5.41948521540561                     # 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(4356261)
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 = 4,503 (80% power), 6,028 (90% power)
 75% of observed effect: n per arm = 8,005 (80% power), 10,716 (90% power)
 50% of observed effect: n per arm = 18,011 (80% power), 24,111 (90% power)
power.t.test: 4503.5 
Simulated power at n = 4,503 per arm: 0.796 (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. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analyses as negative binomial regression, with incidence rate ratios of 0.735 (95% confidence interval 0.481 to 1.123; p = .155) and .960 (95% confidence interval .640 to 1.441; p = .845). The analysis entries list all three groups, but do not make clear which estimate is the comparison for each alcohol-feedback dose against Control PNF. Both intervals include 1, the null value for an incidence rate ratio.

The recalculated comparisons use the posted group means and standard deviations to estimate unadjusted mean differences. They give -0.32 for Light Dose versus Control PNF and 0.07 for Heavy Dose versus Control PNF, with intervals that include zero. These estimates are not a reproduction of the sponsor’s regression model: they use a different measure of effect and do not adjust for covariates. The registry record does not report enough model detail to derive the sponsor’s estimates from the descriptive statistics.

The distinction is methodological rather than a recorded numerical discrepancy. The registry does not flag a discrepancy between the sponsor’s results and the recalculations. The sponsor’s own regression analysis is the primary model-based result; the recalculations offer a transparent, unadjusted view of the posted summaries. Neither set of results should be described as proof of no effect, and neither provides precise evidence of a difference in the posted primary outcome.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists 18 outcome measures and 35 analyses, and 9 secondary outcomes are posted. Multiple outcomes, time points, group comparisons, and models create a multiplicity issue: even when there is no underlying effect, testing many hypotheses can increase the chance of at least one apparently unusual result. General statistical practice can address this through prespecified outcome hierarchies, adjusted significance thresholds, or other procedures, but the registry record does not report whether any of these were used.

The primary endpoint is identified as change from baseline in daily drinking at 4 months. However, the registry record does not provide the full statistical analysis plan, the original protocol’s analysis details, or a clear mapping of the sponsor’s two reported regression estimates to individual group contrasts. It also does not report whether the sponsor used a prespecified testing sequence or adjusted for multiple comparisons. As a result, the inferential status of secondary findings and the family of comparisons covered by any reported p-values cannot be assessed from the record.

For the primary outcome, descriptive statistics are reported for participants completing the 4-month follow-up. Regression analyses included participants with complete data on all covariates, and the record notes that model sample sizes varied slightly across outcomes. It does not describe the amount or pattern of missing outcome data beyond the reported follow-up counts, nor does it state whether imputation, weighting, or sensitivity analyses were used. The reasons for incomplete follow-up and the relationship between missingness and participants’ drinking are not reported.

The record also does not fully define the analysis population for every outcome or explain how randomized participants were handled if they did not complete the intervention or follow-up. Randomized assignment is an important foundation for comparison, but missing data and analytic exclusions can affect how well the final comparison preserves that foundation. These unreported details do not show that the analyses are invalid; they limit how thoroughly readers can evaluate bias, multiplicity, and generalizability.

16. 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 lists serious adverse events and other adverse events by randomized group. For serious adverse events, 388 participants were at risk in Control PNF, 374 in Light Dose of Alcohol PNF, and 381 in Heavy Dose of Alcohol PNF; the number affected was 0 in each group. The record also lists 0 deaths in each group. For other adverse events, the at-risk counts are the same and the number affected is 0 in each group.

These figures mean that no events in the listed categories were recorded among the participants counted as at risk in the registry’s safety entries. They do not establish that no events occurred outside the registry’s definitions, reporting period, or recorded population. The record does not provide event-term details, a fuller description of safety monitoring, or explanations of how the at-risk populations were defined.

Safety counts should be interpreted separately from the drinking and norms outcomes. A zero count in a registry entry is a report about recorded events, not evidence that an intervention has no possible risks. The entries also do not provide a basis for comparing unreported types of events or estimating event rates beyond the counts and denominators listed.

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

The posted results describe a randomized comparison of a game-based personalized normative feedback approach with an active control. On the primary drinking measure at 4 months, the unadjusted mean differences are small in absolute terms relative to the uncertainty in their intervals, and both intervals include zero. The sponsor’s reported negative binomial analyses also have confidence intervals that include the no-difference value for an incidence rate ratio. Taken together, the posted primary data do not establish a clear difference in daily drinking between the alcohol-feedback groups and the control.

This is not the same as showing that the groups are equivalent or that the intervention has no effect. The intervals remain compatible with effects in either direction, and the record does not report an equivalence design or an equivalence margin. A non-significant result cannot establish that two approaches produce the same outcome. Nor do the results justify a medical recommendation or a general conclusion about alcohol interventions.

The findings apply most directly to the measured outcomes and participants represented in the reported analyses. The primary descriptive results are from 4-month follow-up completers, and the regression analyses used complete covariate data. The record says sample sizes varied slightly across outcomes. The broader project description also includes peer norms, longer follow-up, moderators, and voluntary engagement; the primary endpoint alone cannot answer all those questions.

The secondary results offer context but do not form a consistently precise pattern across time. They involve multiple measures and analyses, and multiplicity procedures are not reported. They should therefore be treated as exploratory context unless the full analysis plan establishes otherwise. In addition, important details about missing data, model specification, and the mapping of sponsor estimates to dose contrasts are not reported in the registry record.

The soundest conclusion is limited: the posted results show uncertain estimates for the primary comparison and do not precisely determine the direction or size of a difference. They neither prove benefit nor prove lack of benefit.

18. Why This Trial Matters Statistically

This trial is statistically interesting because it compares two doses of alcohol-related personalized feedback embedded in a game with an active feedback control. That structure separates the alcohol content and its frequency from the broader experience of receiving weekly personalized feedback. It also makes the choice of contrast important: a comparison against an active control answers a different question from a comparison against no intervention.

The primary endpoint is continuous and reported with means and standard deviations, while the sponsor’s model reports incidence rate ratios from negative binomial regression. These are different summaries of the outcome. Reading them together illustrates why the estimand, model, and scale of effect need to be explicit before interpreting whether two analyses agree. The registry record does not report enough model detail to reconstruct the sponsor’s regression estimates from the descriptive summaries.

The trial also illustrates the distinction between evidence of no clear difference and evidence of equivalence. The primary intervals include the null, but they also leave room for non-null effects. Precision is therefore central: the point estimate alone is not a sufficient account of the result. The planning illustrations further show how strongly required sample size can depend on the effect targeted, although those calculations are not a substitute for the original design rationale.

Finally, the project has a broad set of aims and many listed outcomes. The primary drinking result, secondary peer-norm findings, and engagement aims address related but distinct questions. A careful statistical reading keeps those outcomes separate, recognizes the added uncertainty from multiple analyses, and avoids turning a mixed set of estimates into a single overarching claim.

19. 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.

Several trial-specific details needed for a full assessment are not reported in the registry record. Although the project description refers to a multi-site trial, the posted record does not establish the number or locations of sites or provide site-level estimates. The record identifies the design as randomized, single-masked, and parallel, but does not say who was masked, describe allocation concealment, or explain how masking was maintained.

The primary endpoint is titled change from baseline, but the record does not explain the exact change-score construction or how baseline measurements were included in the sponsor’s regression model. The sponsor reports negative binomial regression and incidence rate ratios, but the full model specification, covariates, model diagnostics, and the mapping of reported estimates to specific dose comparisons are not reported. This makes it difficult to connect the adjusted model results directly to the descriptive mean differences.

The registry reports follow-up completers for the primary descriptive outcome and says regression analyses included participants with complete covariate data. It does not describe missingness patterns, reasons for non-completion, imputation, sensitivity analyses, or a complete intention-to-treat approach. The record also does not report intervention adherence or how exposure to the intended weekly rounds was measured.

The study’s project aims include moderators, feedback dosage, engagement, and sustainability, but the posted results summarized in the registry do not provide a complete account of those aims. The original sample-size justification, interim analyses, multiplicity approach, and complete statistical analysis plan are also not reported in the registry record. These gaps constrain interpretation; they do not by themselves establish that the conduct or analysis was inappropriate.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Revolutionizing Normative Re-education — NCT04356261.
• Registry record last updated 2026-05-01; status last verified 2026-03.
• Results first posted 2026-05-01.
• 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.

23. Record Summary

NCT04356261, “Revolutionizing Normative Re-education,” is listed as a completed randomized, single-masked, parallel behavioral trial sponsored by Loyola Marymount University. The registry reports enrollment of 1143 and compares Control PNF with light- and heavy-dose alcohol-related personalized normative feedback in a game format.

The primary outcome is change from baseline in daily drinking at 4 months, measured in drinks per week. Descriptive means are 2.11 for Control PNF, 1.79 for Light Dose, and 2.18 for Heavy Dose among follow-up completers. The sponsor’s negative binomial regression results have intervals that include the incidence-rate-ratio null value. The unadjusted recalculations also have intervals that include zero.

The posted evidence does not precisely establish the direction or size of a primary difference. Secondary results are mixed across follow-up, and the registry record does not report key details about multiplicity, missing-data methods, or the full regression specification.