This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Delays to Influence Snack Choice (NCT02359916), a registered phase not stated study. The primary outcome is Proportion of Healthy Snacks Purchased.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02359916 |
| Phase | Phase not stated |
| Condition | Food Choice |
| Allocation | Not Stated |
| Design model | Not Stated |
| Masking | Not Stated |
| Enrollment | 32662 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Rush University Medical Center |
| Results posted | Yes |
2. Clinical Question
Population
The study concerns snack vends at vending locations. The primary-outcome description says each vend was treated as an analyzed participant and that no data were collected from individual human subjects. The record lists enrollment as 32662, but that figure should not be read as the number of human participants in the primary analysis.
Intervention
The tested approaches involved delivery delays for less healthy snacks, discounts for healthy snacks, and higher prices for less healthy snacks. The record describes conditions combining delays with discounts or taxes, as well as conditions using each approach alone.
Comparator
Equal pricing without delays appears in the arm descriptions, including Baseline conditions. However, the registry warns that this was not a randomized comparison, says allocation was not stated, and cautions that the arms are not a like-for-like control comparison.
Primary question
The primary endpoint is the proportion of vending sales coming from healthy snacks. The broad question is whether the proportion of healthy snacks purchased differs across the experimental conditions. The registry does not report numerical results or an analysis establishing the size, precision, or direction of any such difference.
The motivating idea in the registry is that a brief delay before receiving a less healthy snack might change the immediacy of its reward and influence choice. Discounts and higher prices represent separate ways of changing the relative cost of healthy and less healthy snacks. These are study questions, not conclusions: the record describes the approaches being examined but does not establish that any approach changed purchasing. In particular, the stated comparison warning limits how confidently differences between conditions can be attributed to the intervention itself.
3. Trial Design
The registry identifies the study as “Delays to Influence Snack Choice,” with the official title “Time Over Money? A Novel System to Influence Snack Machine Choices.” It lists Rush University Medical Center as the sponsor and Food Choice as the condition. The record gives a start of 2015-06, primary completion of 2016-08, and status of COMPLETED. These registry entries describe the study record; they do not supply a numerical outcome analysis.
Seven arm descriptions are listed. They include equal pricing without delays, discounts on healthy snacks, delays for less healthy snacks, combinations of delays with discounts, higher prices for less healthy snacks, and combinations of delays with higher prices. Two arms are described with equal pricing and no delays, while the other conditions vary the pricing or delay approach. The registry also says Baseline 1 and Baseline 2 were combined into a single Baseline group for statistical analysis. It does not explain further how the listed arms map to that combined group.
The record explicitly warns that this was not a randomized comparison, that allocation was not stated, and that the arms are not a like-for-like control comparison. Consequently, the arm labels alone do not establish that conditions were assigned in a way that supports causal interpretation. The registry describes the primary outcome over a timeframe of 28 weeks per vending location, while its endpoint description says the proportion was calculated for each experimental condition over 4 weeks. How those timeframe statements relate to each other is 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.
Only the first 6 of 7 groups are shown.
Overall Study
| Stage | Baseline 1 | Discount Only | Delay Only | Delay + Discount | Tax Only | Delay + Tax |
|---|---|---|---|---|---|---|
| STARTED | NA | NA | NA | NA | NA | NA |
| COMPLETED | NA | NA | NA | NA | NA | NA |
| NOT COMPLETED | NA | NA | NA | NA | NA | NA |
Analysis population for the primary outcome, as defined in the registry: In this table, each vend is treated as an analyzed participant. No data were collected from individual human subjects. Analyses included all available data on healthy and regular snack vends. Data from Baseline 1 and Baseline 2 were combined into a single arm/group ("Baseline") in statistical analyses.
5. Primary Endpoint
The registry lists the primary outcome as Proportion of Healthy Snacks Purchased, measured over 28 weeks per vending location.
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | not applicable |
| Unit | proportion of vends from healthy snacks |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is the proportion of healthy snacks purchased, described as the proportion of vends from healthy snacks. In plain terms, it expresses healthy snack vends relative to the relevant vending activity being analyzed. It is a measure of the composition of purchases, rather than a measure of the number of human participants who chose a snack. The registry states that each vend was treated as an analyzed participant and that analyses included all available data on healthy and regular snack vends.
A proportion is useful when the research question concerns the share of choices belonging to a category. It can distinguish a change in the mix of healthy and less healthy purchases from a change in the total volume of vending activity. That distinction is general: a share alone does not reveal total sales or revenue. The registry lists Total Daily Vending Revenue in US$/Day as a secondary outcome, but does not provide a numerical result for it in the record details.
The endpoint’s purpose is to assess whether purchasing patterns differ among the conditions involving prices and delays. The record does not give the numerical proportion for any arm, the comparison estimates, uncertainty intervals, or a statistical test. Therefore, the endpoint’s definition is available, but its observed magnitude and precision are not reported in the registry record.
6. Why No Re-analysis Is Shown
7. Statistical Interpretation
The registry establishes the study’s topic, arms, outcome definition, and analysis population description, but it does not report numerical findings for the primary endpoint. No arm-specific proportions, contrasts, confidence intervals, or p-values are stated in the record. The direction and magnitude of any observed difference in healthy snack purchasing are therefore not reported in the registry record. The completion status does not, by itself, provide evidence about the outcome.
Interpretation is also limited by the registry’s warning about the comparison structure. It says this was not a randomized comparison, allocation was not stated, and the arms are not a like-for-like control comparison. A numerical difference between arms, if one were available, would describe an association between condition and vending pattern. Without information establishing comparable groups or an appropriate adjustment strategy, that difference could not automatically be attributed to the delay, discount, or higher price.
The endpoint is a proportion of vends, not an outcome collected from individual human subjects. The registry says no individual human-subject data were collected for this analysis. Thus, the unit described for the primary measure is vending activity, and the enrollment figure should not be treated as a count of people contributing individual-level outcome data. The record does not report how repeated vends at a location were handled statistically, or whether variation among vending locations was accounted for.
The timeframe is not completely clear from the registry description: the primary outcome is associated with 28 weeks per vending location, and the endpoint description says calculations were made over 4 weeks for each experimental condition. The relationship between those statements is not reported in the registry record. This matters because the period used to define an outcome affects what the proportion represents and how much vending activity informs it.
In short, the record supports a description of the research question and its measurement, not a conclusion about whether the interventions changed snack choices. The statistical interpretation would require the posted arm-level results, an account of the comparison and analysis, and a clear description of the outcome period. Those details are not reported in the registry record.
8. Statistical Methods Explained
The registry identifies the primary outcome as a proportion and says each vend was treated as an analyzed participant. It does not report the statistical model, the comparison procedure, uncertainty estimates, or the method used to account for variation across vending locations. The concepts below are general explanations of what statistical methods would need to address; they are not claims about methods used in this study.
What does a proportion measure?
Generally, a proportion represents the part of a defined total belonging to a category. Here, the category is healthy snacks, and the endpoint is described as the proportion of vends from healthy snacks. The precise calculation depends on the relevant numerator and denominator and on how vends are classified. The registry describes healthy and regular snack vends as included in the analysis, but does not give a fuller classification rule or an arm-level numerator and denominator.
A proportion is not the same as a count. Two settings can have the same share of healthy snack vends while having different overall vending activity. Conversely, total sales or revenue can change without an equivalent change in the healthy-snack share. This is one reason the record’s separate secondary outcome, Total Daily Vending Revenue in US$/Day, asks a different question from the primary endpoint.
Why does the unit of analysis matter?
Generally, statistical inference depends on identifying what was measured and what observations can reasonably be treated as independent. The registry’s description treats each vend as an analyzed participant while also stating that no data were collected from individual human subjects. Vends may be recorded within the same vending location, so they may share conditions or patterns. If observations within a location are related, treating every vend as independent could understate uncertainty. Appropriate analysis might need to account for grouping, but the registry does not report whether or how that was done.
The number of vends is not the same as the number of independent locations, and the enrollment field does not resolve that distinction. The record does not report the number of vending locations, the distribution of vends across locations, or the degree of similarity among vends from the same location. Without those details, the precision of an arm-level proportion cannot be assessed from the registry description alone.
What does over-dispersion mean in general?
For count or proportion data, a simple model may assume that observations vary according to a particular probability structure. Over-dispersion means the observed variability is greater than that structure expects. In a vending setting, general sources of extra variation could include clustering or differences among locations, but the registry does not identify these as observed features or state that over-dispersion was assessed. If extra variability is present and ignored, standard errors can be too small and confidence intervals too narrow. A method that accommodates extra variability can better represent uncertainty, provided its assumptions suit the data.
What would an interval or test tell the reader?
Generally, a confidence interval describes the range of effect values reasonably compatible with the data and the assumptions of the analysis. Its width conveys precision: wider intervals indicate less precise estimation, while narrower intervals indicate greater precision. A hypothesis test addresses compatibility with a specified null hypothesis under an analysis model; a p-value is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of an effect. The registry does not report intervals or tests for the primary outcome, so neither precision nor statistical evidence for a contrast can be evaluated from the record.
What can the planning illustrations establish?
The sample-size assumptions describe a planning illustration computed from an observed effect, not the trial’s own design. They state a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect defined as a stated fraction of the observed effect. The record specifies that these illustrations are planning exercises, not observed power. For time-to-event tables, the stated convention is total events; for other outcomes, it is participants per arm. These assumptions do not demonstrate that the original study used equal allocation, had a particular design target, or had a given probability of detecting an effect. The registry does not report its own sample-size justification, or an analysis plan that would establish such claims.
9. Multiplicity, Missing Data and Analysis Populations
The record lists several intervention approaches and multiple arms, alongside a primary outcome and a secondary outcome of Total Daily Vending Revenue in US$/Day. When an analysis examines many arms, outcomes, or pairwise comparisons, the chance of at least one apparently notable result can increase. General methods for addressing this include prespecifying a limited set of primary comparisons or adjusting inference for multiple testing. The registry does not report whether multiplicity was addressed, which comparisons were planned, or whether a hierarchy of outcomes was specified.
Missing-data handling is also not described in enough detail to evaluate. The primary outcome description says analyses included all available data on healthy and regular snack vends. That wording does not specify whether any location-period observations or vending records were missing, why any records might be unavailable, or how missing information would affect the calculated proportions. The registry does not report an imputation method, sensitivity analysis, or a missing-data assumption.
The analysis population is described at the level of vending activity: each vend was treated as an analyzed participant, and the record states that no data were collected from individual human subjects. Baseline 1 and Baseline 2 were combined into a single Baseline group in statistical analysis. The record does not further define inclusion rules for vending records, exclusions, or the handling of incomplete locations or periods.
These omissions do not establish that a particular approach was inappropriate; they mean the registry record does not allow readers to assess it. A complete statistical account would clarify the planned comparisons, multiplicity strategy, missing-data procedures, and analysis population. Such details are not reported in the registry record.
10. What the Result Does, and Does Not, Mean
The study asks whether vending conditions involving delays and price changes are associated with the proportion of healthy snack vends. Its primary measure is a share of vending activity, not a direct measure of a person’s health, body weight, or cardiometabolic outcomes. The registry’s background presents these approaches as potentially relevant to prevention efforts, but the listed endpoint concerns snack purchasing rather than those downstream outcomes.
The record does not provide a numerical primary result, so it cannot establish the direction or size of a difference among conditions. It also does not provide uncertainty estimates with which to assess precision. A completed study status indicates that the registry marks the study completed; it is not itself a finding that an intervention changed purchasing.
The record’s warning that the comparison was not randomized, that allocation was not stated, and that arms were not like-for-like is central to interpretation. Even if a difference in the outcome were reported elsewhere, the design information in this record would not justify treating that difference alone as proof that the intervention caused it. Other explanations related to how conditions and locations were organized cannot be evaluated from the details reported.
General statistical interpretation also distinguishes the measured unit from the target of a broader claim. The primary description concerns vends and vending locations, while noting that no individual human-subject data were collected. It therefore does not directly show how an individual person responded, nor does it establish long-term changes in diet or health. The registry does not report those outcomes.
The defensible conclusion from this record is limited: it documents a study of snack vending conditions and defines a primary purchasing-share outcome, while leaving numerical results and key analytical details unreported. It neither demonstrates benefit nor demonstrates absence of an effect.
11. Why This Trial Matters Statistically
This study is statistically interesting because its outcome is defined through vending activity rather than individual participants. The registry explicitly says each vend was treated as an analyzed participant and that no data were collected from individual human subjects. That distinction is important: the apparent volume of recorded activity does not automatically tell a reader how much independent information there is. If multiple observations arise from the same location, their dependence can affect uncertainty.
The design also raises a general distinction between measuring a difference and identifying its cause. The registry lists a range of conditions involving delays, discounts, and higher prices, but cautions that the comparison was not randomized, allocation was not stated, and arms were not like-for-like. A statistical contrast can quantify observed differences, but causal interpretation requires a design and analysis that address competing explanations. The record does not supply enough detail to judge that question.
There is also a useful measurement distinction between the primary outcome and the secondary revenue outcome. The healthy-snack proportion describes the composition of vending, whereas revenue describes a different aspect of activity. Neither endpoint alone necessarily captures the other. Making that distinction clear helps prevent readers from treating a change in share as equivalent to a change in total purchasing or revenue.
Finally, the record illustrates why estimates need uncertainty and transparent analysis descriptions. Without numerical results, intervals, or a reported method for handling location-level dependence and multiple comparisons, readers cannot assess how precise or robust a reported contrast would be. The study’s topic is relevant to environmental approaches to food choice, but the statistical record supports careful description of the question rather than a claim about the effect.
12. 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 principal limitation of the registry information is that it does not report numerical primary-outcome findings. Arm-level proportions, comparisons, uncertainty intervals, and test results are not reported in the registry record. As a result, the direction, magnitude, and precision of any observed association cannot be assessed from these details.
The comparison structure is a further limitation. The registry says this was not a randomized comparison, allocation was not stated, and the arms are not a like-for-like control comparison. It does not explain how vending locations or conditions were assigned, whether the same locations experienced more than one condition, or how the analysis addressed possible differences among locations. These details are not reported in the registry record.
The outcome timeframe is described in two ways: 28 weeks per vending location and calculations over 4 weeks for each experimental condition. Their relationship is not reported in the registry record. The record also does not report the number of locations, vending activity by location, snack-classification rules, or how repeated observations and any incomplete data were handled.
The primary analysis is described as using all available healthy and regular snack-vend data, but this does not fully define the analysis population or missing-data approach. The statistical model, treatment of clustering or over-dispersion, and strategy for multiple arms or outcomes are also not reported. The record lists an adverse-event summary with zero counts in the displayed categories, but it also states that no individual human-subject data were collected for the primary outcome; the scope and interpretation of adverse-event collection are not explained. These limitations describe what cannot be established from the registry record, not proof that the study lacked additional procedures.
13. Related Tutorials
14. Related Statistical Calculators
15. Sources
- ClinicalTrials.gov: Delays to Influence Snack Choice — NCT02359916.
- Registry record last updated 2018-02-23; status last verified 2018-01.
- Results first posted 2018-01-26.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
16. Record Summary
The registry record describes a completed study of snack vending conditions involving delays, discounts on healthy snacks, and higher prices for less healthy snacks. It names the primary outcome as the proportion of healthy snacks purchased and describes each vend as an analyzed participant, with no individual human-subject data collected. It also lists Total Daily Vending Revenue in US$/Day as a secondary outcome.
The record warns that the comparison was not randomized, allocation was not stated, and arms were not like-for-like. It describes both a 28 weeks per vending location timeframe and calculations over 4 weeks per experimental condition; the relationship between these statements is not reported. Numerical outcome findings and key details of the statistical analysis are not reported in the registry record. The available information therefore supports a description of the research question and its measurement, but not a conclusion about the size, direction, or cause of any difference in snack purchasing.