Interpreting a Small Dietary Survey
A community health team ran a two-week food frequency questionnaire with 48 adults attending a local health centre to decide whether to fund a "reduce sugary drinks" program. Results showed 29 of 48 participants (60%) reported drinking one or more sugar-sweetened beverages daily. The team also ran three focus groups (n=14 total) in which several participants said they "don't really count" cordial or sports drinks as sugary drinks. The survey was distributed at the health centre on weekday mornings, and the team notes that 38 of the 48 participants were women aged over 50. The program budget is limited and the team must decide whether to proceed, and if so, how to target it.
1. Evaluate the strength of the evidence for the claim that 60% of the local adult population drinks sugary drinks daily.
Read the worked answer
The 60% figure is not a valid estimate of the local adult population and should not be presented as one.
Sampling problems:
- Convenience sample: people recruited at a health centre on weekday mornings are not representative of local adults. They are more likely to be retired, not in full-time work, and already engaged with health services.
- Selection bias by sex and age: 38 of 48 were women over 50. Sugary drink consumption is typically higher in younger people and in men, so this sample likely underestimates consumption.
- Small size: 48 people gives a wide confidence interval. A rough calculation puts the 95% interval for 60% at roughly 46–74%, which is too wide for confident planning.
Measurement problems:
- Self-report and social desirability: people report what they think is expected, especially in a health setting.
- Definition ambiguity: the focus groups showed that some participants do not count cordial or sports drinks, so the questionnaire may have underestimated total sugary drink intake.
- No validation: there is no comparison against a food diary or sales data.
Conclusion: the figure is useful as a signal that sugary drinks are common, but it cannot support a population prevalence claim. The team should describe it as "60% of survey respondents", not "60% of local adults".
2. The team has a limited budget. Recommend how to proceed and what additional data would most improve the decision.
Read the worked answer
Recommendation: proceed with a targeted pilot, not a population-wide campaign, and collect better data alongside it.
Reasons:
- The evidence is too weak to justify a large spend on a broad campaign, but it is strong enough to justify action. Both the survey and the focus groups point the same way: sugary drinks, including cordial and sports drinks, are commonly consumed.
- A pilot generates local data while delivering something useful, which is better value than spending the whole budget on more survey work.
What the pilot should target:
- Cordial and sports drinks specifically, because the focus groups identified these as a blind spot. Messaging that only says "soft drink" will miss them.
- Settings where the target group actually is - schools, sports clubs, workplaces - rather than the health centre, which reached an unrepresentative group.
Additional data, in priority order:
- A short, well-worded survey with a representative sample (e.g. recruited through schools, workplaces and clubs, with quotas for age and sex), using a clear definition list of what counts as a sugary drink.
- Sales or procurement data from local outlets or the health centre café as an objective cross-check.
- A small number of interviews to understand why people choose these drinks, which shapes the message.
Decision rule: set a threshold in advance - for example, proceed to a full program only if the pilot shows measurable change in the targeted setting.
3. A colleague wants to publish a poster stating "60% of our community drinks sugary drinks every day." Explain the ethical and professional issues with this and propose accurate wording.
Read the worked answer
Ethical and professional issues:
- Misrepresentation: the statement converts a finding about 48 survey respondents into a claim about the whole community. This is a form of overgeneralisation and is misleading, even if unintentional.
- Potential harm: exaggerated prevalence can stigmatise the community, alarm residents, and be picked up by media or funders as fact. If later contradicted, it damages the team's credibility and future work.
- Poor basis for decisions: if the figure is treated as established, the program may be designed for the wrong group (for example, aimed at older women when younger men may be the higher consumers).
- Professional standards: health promoters have a duty to represent evidence accurately, state limitations, and avoid claims the data cannot support.
Accurate wording options:
- "In a survey of 48 adults attending our health centre, 60% reported drinking sugary drinks daily. This group was mostly women over 50 and may not represent the wider community."
- "60% of survey respondents reported daily sugary drink consumption. Cordial and sports drinks were often not counted as sugary drinks."
- "Early local findings suggest sugary drinks are common; further data are being collected."
Accurate wording does not weaken the case for action. It makes the case more credible and protects the team if the figure is questioned.