Attention Bias Modification Training for Young People: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Attention Bias Modification Training for Young People (NCT02270671), a randomized na study. The primary outcome is Threat Bias Measurement at 4-week Post-intervention and 12-week Follow-up.

2. Clinical Question

The clinical question is whether changing where attention is directed during a computer task is associated with a different threat-bias measurement in socially anxious adolescents than the placebo task. The registry describes a two-phase study: participants first completed a wellbeing and emotions screening questionnaire, and those meeting the stated social-anxiety criterion could then be invited to the four-week training program. The primary outcome was assessed after the intervention and again at follow-up. Fear of negative evaluation, anxiety-related disorders, and depression symptoms were listed as secondary outcomes at corresponding follow-up periods. The record does not establish that every screened participant entered the randomized comparison, and the number meeting the invitation criterion is not reported in the registry record.

3. Trial Design

This was a randomized, single-masked, parallel-group trial. The registry lists 130 participants as the enrollment and identifies the status as completed. The trial record gives a start date of 2014-10 and a primary completion date of 2015-05. The masking description does not identify which people were masked, and further details about random-sequence generation and allocation concealment are not reported in the registry record.

The intervention was attention bias modification training (ABMT); the comparator was labeled placebo. Both protocols used face-pair presentations in a computer task. The ABMT protocol is described as directing the target probe to the neutral-face location in angry-neutral trials. In the placebo protocol, angry-face location, probe location, and actor were described as counterbalanced. The available protocol descriptions are incomplete, so details beyond these features are not reported in the registry record.

The study summary describes screening followed by a four-week training phase for participants who met a social-anxiety cut-off. This indicates that screening and participation in the intervention phase were distinct steps; the registry does not give the numbers screened, eligible, invited, or randomized separately. The record reports an overall enrollment of 130 and adverse-event denominators of 64 for placebo and 66 for intervention, but those figures alone do not describe participant flow or confirm the analysis population for each outcome.

The primary outcome was a continuous measure of threat bias in milliseconds, described as a mean with standard deviation. Measurement was scheduled after the intervention at week 4 and at 12-week follow-up. Secondary outcomes included fear of negative evaluation, anxiety-related disorders, and depression symptoms at post-intervention and follow-up. The registry record lists five outcome measures and five analyses, but does not report enough detail to identify the exact statistical model, analysis population, handling of repeated measurements, or prespecified contrast.

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.

Overall Study

Reasons for not completing this period:

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: Comparison of baseline data for participants who completed the intervention/placebo training.

6. Primary Endpoint

The registry lists the primary outcome as Threat Bias Measurement at 4-week Post-intervention and 12-week Follow-up, measured over Post-intervention (week 4), and 12-week follow-up.

The primary endpoint is threat bias measured with a dot-probe task. Participants saw angry-neutral and neutral-neutral face pairs and responded to a probe. The registry defines the threat-bias score as the mean reaction time on neutral NT trials minus the mean reaction time on threat NT trials. Reaction times were measured in milliseconds, and the endpoint was summarized as a mean and standard deviation.

By the registry's definition, a score >0 indicates bias toward threat. A score <0 indicates that a participant was slower to respond to threatening stimuli than to neutral stimuli. These signs describe the direction of the task-based score; they do not by themselves establish whether an observed group difference is important to participants or clinically meaningful. Interpretation would require the estimated difference between groups, its uncertainty, and a clear account of the analysis and measurement time point. Those results are not reported in the registry record as a single primary-outcome comparison.

The task provides an operational measure of how response times differ across threat-related and neutral stimulus conditions. It is used to quantify attentional bias in a standardized task, rather than to directly measure all aspects of social anxiety. The registry lists assessments after the four-week intervention and at 12-week follow-up, so the outcome concerns both an immediate post-intervention time point and a later follow-up. A detailed account of how scores at those time points were combined or compared is not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does randomization contribute?

In general, randomization assigns participants to groups through chance rather than through a clinician's or participant's choice. Its statistical purpose is to reduce systematic differences between groups at the start of a trial, making the assigned groups more comparable on average. Here, the registry identifies the study as randomized and parallel, but does not describe sequence generation, allocation concealment, or the process used to enroll participants into each arm. Those specific safeguards are not reported in the registry record.

A parallel design means participants are assigned to one group and evaluated against the other group, rather than crossing through both interventions. That is the design label recorded for this trial. The record also states that the study was single-masked, but does not identify who was masked. Without that detail, the potential influence of awareness of assignment on task administration, participant behavior, or outcome assessment cannot be evaluated from the registry description.

How should a continuous endpoint be understood?

General statistical explanation: a continuous outcome can take values along a measurement scale, rather than being limited to categories such as yes or no. This trial's primary measure is reaction time in milliseconds, summarized as a mean and standard deviation. A mean describes the average score in a group. A standard deviation describes the spread of individual scores around that average; it is not the uncertainty in the estimated group mean and is not a measure of the difference between randomized groups.

For a treatment comparison, the central quantity is usually a contrast between groups at a specified time point, or a contrast defined by a statistical model. A confidence interval expresses the range of values reasonably compatible with the observed data and the model's assumptions. It helps show precision: a narrower interval indicates greater precision than a wider interval. The registry record does not report a single primary-outcome contrast or confidence interval, so neither the magnitude nor precision of a primary effect can be assessed from the entry.

What is the score's direction?

The registry defines the score using the mean of neutral NT trials minus the mean of threat NT trials. Its stated interpretation is that scores >0 indicate bias toward threat, while scores <0 indicate slower responses to threatening stimuli than to neutral stimuli. This is a task-specific sign convention. It should be kept separate from a between-group effect: even if a group average has a particular sign, a comparison requires knowing the other group's result and the method used to estimate the difference.

The endpoint uses task reaction times as an indicator of attention, but a task score is not identical to a broad clinical outcome. The registry also lists measures related to fear of negative evaluation, anxiety-related disorders, and depression symptoms. These outcomes address different aspects of participants' experiences and should not be treated as interchangeable with the reaction-time endpoint. The record does not report the scoring details for those secondary measures.

Why can repeated time points complicate interpretation?

General statistical explanation: when an outcome is measured more than once, results may be reported separately at each time point, as changes over time, or as a model-based comparison that uses repeated observations. These approaches answer different questions and can make different assumptions about correlations between measurements from the same participant. The registry specifies post-intervention and 12-week follow-up assessments but does not state how those measurements were combined or compared. It also does not report whether the main contrast was defined at one time point or across time.

More than one outcome and more than one time point can create multiplicity: there may be several chances to identify a nominally unusual result when many comparisons are examined. General statistical practice may address this through prespecified primary comparisons, adjustment procedures, or a clear distinction between confirmatory and exploratory analyses. Which, if any, approach was used in this trial is not reported in the registry record. The record lists five outcome measures and five analyses, but those counts do not specify the comparison structure or adjustment method.

What do sample-size illustrations mean?

General statistical explanation: sample-size planning estimates how many participants may be needed to detect a target difference under stated assumptions. A planning illustration based on an observed effect is not observed power, and it does not establish that the trial was adequately powered for its actual question. The record describes assumptions for planning illustrations, including two-sided significance level 0.05, equal allocation, and a normal approximation. It also states that the target effect is a stated fraction of the observed effect. The specific planning results and the trial's own sample-size justification are not reported in the registry record.

The record does not report the expected effect used in the original design, a formal sample-size justification, or whether any planned assumptions were updated. Consequently, enrollment of 130 cannot by itself establish the precision or adequacy of the primary comparison. The results needed to judge precision—an estimated contrast and uncertainty interval—are not reported as a single primary-outcome comparison.

Which analysis details are essential?

To interpret a randomized trial's estimates, readers need to know which participants were included, how missing measurements were handled, whether the analysis followed randomized assignment, and what model or test was used. For a continuous endpoint measured at multiple time points, relevant details may include how baseline scores were treated, whether repeated observations were modeled jointly, and how the group contrast was defined. These are general considerations, not claims about the analysis used in this trial. The registry record does not report these details.

Finally, numerical findings should be read with their uncertainty and alongside the prespecified endpoint and follow-up schedule. A p-value alone, if reported, would not describe the magnitude or clinical importance of a difference. The registry entry does not provide a single primary-outcome estimate, interval, or p-value for interpretation. It therefore supports description of the planned measure and design, but not a statistical conclusion about the observed primary effect.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists five outcome measures and five analyses. The primary outcome includes measurements after the intervention and at 12-week follow-up, while the listed secondary outcomes include fear of negative evaluation, anxiety-related disorders, and depression symptoms at post-intervention and follow-up. Examining several outcomes or time points can create multiplicity because multiple comparisons offer multiple opportunities for chance findings. General statistical approaches include prespecifying a primary contrast, adjusting for multiple comparisons, or distinguishing confirmatory from exploratory analyses. The method used in this trial is not reported in the registry record.

The record does not report a statistical analysis plan, the specific tests or models used, whether results were adjusted for baseline values, or how repeated measurements were handled. It also does not state which analyses were considered primary if more than one time point or contrast was examined. The listed counts of outcomes and analyses do not answer these questions.

Missing-data details are also not reported in the registry record. The number with an observed primary endpoint at each time point, reasons for missing measurements, and methods for handling missing values are not described. The analysis population is not identified, including whether all randomized participants were analyzed in their assigned groups. Therefore, the overall enrollment and safety denominators should not be assumed to equal the analysis sample for every outcome. These omissions limit assessment of attrition, potential bias from missing outcomes, and the robustness of any group comparison.

11. Safety Results

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

Other (non-serious) adverse events

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

The adverse-event record lists 64 participants at risk in the placebo group and 66 in the intervention group. It records 0 serious adverse events and 0 other adverse events for each group. These are the event counts recorded for the listed groups; the registry entry does not provide event descriptions because no affected participants are listed.

The death count is not reported in the registry record. A zero count for the serious and other adverse-event categories should be read as no events recorded in those categories, not as proof that no events occurred outside the reporting system or observation period. The record does not describe the duration or completeness of safety follow-up, the method for collecting events, or whether all randomized participants contributed to safety monitoring.

These figures are descriptive. They do not establish that the intervention is safe or unsafe, and they do not compare risks with enough detail to evaluate rare events. The denominators indicate the groups contributing to the adverse-event record, but participant-level follow-up and event-collection procedures are not reported in the registry record.

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

The trial addresses a focused statistical question: whether adolescents selected for a social-anxiety training phase had different threat-bias measurements after ABMT than after a placebo computer task. It also scheduled assessment at 12-week follow-up and listed several secondary symptom outcomes. The registry record is useful for identifying the intervention, comparator, endpoint, and broad design.

It does not, however, report a single primary-outcome comparison. The direction and size of any between-group difference, its uncertainty, and its statistical significance are therefore not available for interpretation from the registry entry. The schedule of assessments is not a result, and the descriptions of task protocols do not establish an effect. No conclusion about whether the groups differed on threat bias or secondary outcomes can be drawn from the outcome information reported in the registry record.

Even if an estimated difference were available, its meaning would depend on the score's direction, the time point, the analysis method, and the precision of the estimate. The endpoint is a task-based reaction-time measure, so it would not by itself answer every question about social anxiety or participants' broader functioning. The registry's secondary outcome names indicate additional domains of interest, but their numerical results and interpretation are not reported in the registry record.

The record also does not identify the analysis population, missing-data approach, multiplicity strategy, or the identity of the person or people masked. These omissions do not negate the randomized design label, but they constrain assessment of how the comparison was conducted and how robust any findings might be. The enrollment and safety counts do not replace outcome-specific sample sizes or effect estimates.

Accordingly, the record supports a description of the trial's question and measurement strategy, not a conclusion about the size, precision, or clinical importance of an effect. That distinction is central: a trial can be completed and have a defined primary endpoint while its registry entry still lacks enough reported outcome detail for a statistical reading of the comparison.

13. Why This Trial Matters Statistically

This trial is statistically notable for linking an intervention's proposed mechanism to a task-based measure. ABMT is described as changing the location to which attention is directed during angry-neutral face trials, while the primary endpoint uses reaction-time differences across threat and neutral conditions. That creates a direct measurement question about attentional bias, distinct from the broader secondary outcomes concerning fear of negative evaluation, anxiety-related disorders, and depression symptoms.

The design also raises common issues in clinical-trial interpretation. Participants passed through a screening stage before possible invitation to the training phase, so the population described by the trial is narrower than all young people. The registry does not report how many people screened, qualified, entered randomization, or contributed data at each follow-up. This matters for understanding who the results would describe and whether missing outcome data may affect interpretation.

The primary endpoint was scheduled at more than one time point, and several secondary outcomes were listed. That makes the definition of the primary comparison and the handling of multiple analyses important. The registry record lists five outcome measures and five analyses but does not describe a multiplicity strategy or a statistical analysis plan. Without a clearly reported estimate and its uncertainty, the measurement design cannot be connected to a precise account of the observed group difference.

The record therefore illustrates why a complete statistical report needs more than a design label and outcome definition. Readers also need arm-level outcome results, a defined contrast, uncertainty measures, analysis populations, and information about missing data and multiplicity. Those elements are not reported in the registry record as a single primary-outcome comparison. The study remains a useful example of how mechanism-oriented outcomes can be specified while the interpretability of results depends on transparent reporting.

14. 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 entry contains limited detail about the randomized process. It identifies random allocation and single masking but does not report the sequence-generation method, allocation concealment, or who was masked. The descriptions of both task protocols are incomplete, so the precise implementation cannot be reconstructed from the registry record.

The target population also requires careful definition. The study summary describes screening and invitation based on scoring above a social-anxiety cut-off, but the number screened, eligible, invited, and randomized is not reported in the registry record. The record gives overall enrollment but does not provide participant flow by study phase or outcome time point. It is therefore not possible to evaluate recruitment, retention, or the representativeness of participants who completed assessments.

The primary endpoint has two listed assessment times, but no single primary-outcome comparison is reported. Numerical results for the primary and secondary outcomes, uncertainty estimates, and analysis methods are not reported in the registry record. Missing-data handling, analysis populations, baseline adjustment, and multiplicity procedures are also unspecified. These gaps prevent a detailed assessment of the reported estimates or their robustness.

Finally, the safety record lists zero serious and zero other adverse events in the stated groups, but deaths are not reported in the registry record and the safety-monitoring period and collection procedures are not described. The adverse-event counts should therefore be interpreted only as the registry's recorded figures, not as a comprehensive account of all possible events.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Attention Bias Modification Training for Young People — NCT02270671.
• Registry record last updated 2018-11-20; status last verified 2018-11.
• Results first posted 2018-11-20.
• 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.

18. Record Summary

NCT02270671, titled Attention Bias Modification Training for Young People, is a completed randomized, single-masked, parallel-group trial sponsored by University College Dublin. The registry lists enrollment of 130, a start date of 2014-10, and primary completion in 2015-05. The study compared a computer-based attention bias modification training protocol with a placebo protocol in young people described as socially anxious adolescents.

The primary outcome was threat bias measured in milliseconds after the intervention and at 12-week follow-up. Secondary outcomes included fear of negative evaluation, anxiety-related disorders, and depression symptoms. The registry lists five outcome measures and five analyses, but does not report a single primary-outcome comparison or sufficient analysis detail to assess its magnitude and precision. The adverse-event entry records 0 serious and 0 other events in each group; death counts are not reported in the registry record.