The UCL Smoking Toolkit Study (STS) is a long-running monthly survey of adult smokers and recent ex-smokers in England, designed to track smoking behaviour, quit attempts, and use of cessation services. This recipe-style guide walks you through the essential steps to conduct a similar study, from defining your sample to analysing the data, with realistic timelines and practical tips for success.
Ingredients
- 1 clear research question (e.g., “How does e-cigarette use affect quit success?”)
- 1 target population: adults aged 18+ in England who smoke or have quit within the last 12 months
- 1 sampling frame: list of households or random digit dialling numbers (approx. 1,800 contacts per wave)
- 1 structured questionnaire (30–35 items, including demographics, smoking status, quit attempts, and product use)
- 1 data collection platform (e.g., Qualtrics, SurveyMonkey, or paper forms)
- 1 data analysis software (e.g., R, SPSS, or Stata)
- 1 ethical approval from a recognised review board
- 1 informed consent form (approved by the ethics board)
- 1 team of trained interviewers (3–5 people)
- 1 pilot test with 10–20 respondents
- 1 data cleaning protocol (e.g., check for missing values, outliers)
- 1 statistical weighting plan (to adjust for sampling bias)
Instructions
- Define your research question and target population. Be specific: the UCL STS focuses on adults in England, but you can adapt to your region. Write a one-sentence aim that will guide every other step.
- Obtain ethical approval. Submit your study protocol, questionnaire, and consent forms to your institutional review board. This can take 2–4 weeks, so start early.
- Develop the questionnaire. Include core items: age, gender, smoking status (daily, occasional, non-smoker), number of cigarettes per day, quit attempts in the past 12 months, use of nicotine replacement therapy, e-cigarettes, or other aids, and intention to quit. Keep it under 10 minutes to reduce drop-out.
- Pilot the questionnaire. Test with 10–20 people from your target population. Check for clarity, missing response options, and technical glitches. Revise as needed.
- Build your sampling frame. Obtain a list of households or phone numbers. If using random digit dialling, generate numbers randomly. Ensure your frame covers both landlines and mobile numbers to avoid bias.
- Train interviewers. Explain the purpose, how to read questions verbatim, how to record answers, and how to handle refusals or sensitive responses. Practice with role-play.
- Collect data. Conduct interviews by phone, face-to-face, or online. Aim for a response rate of at least 50% to maintain representativeness. For a monthly wave, plan to collect data over 2–3 weeks.
- Enter and clean data. Double-check entries for errors. Remove duplicates, flag missing values, and code open-ended responses. Create a codebook to document variables.
- Weight the data. Compare your sample to known population benchmarks (e.g., age, sex, region) and create weights to adjust for non-response. This is crucial for accurate estimates.
- Analyse the data. Use descriptive statistics (frequencies, means) and inferential tests (chi-square, logistic regression) to answer your research question. For example, compare quit success rates between e-cigarette users and non-users.
- Interpret and report. Write up your findings, including limitations (e.g., self-report bias, cross-sectional design). Share results with stakeholders or publish in a peer-reviewed journal.
Tips
- Substitutions: If you cannot access a random digit dialling frame, use an online panel (e.g., Prolific) but note that results may not be nationally representative.
- Keep the questionnaire short: Longer surveys increase drop-out and incomplete responses. If you need more items, consider splitting into two waves.
- Common mistake – ignoring non-response bias: If certain groups (e.g., younger adults) are less likely to respond, your estimates will be skewed. Always weight and report response rates.
- Common mistake – leading questions: Avoid phrasing like “Do you agree that vaping is safer than smoking?” Use neutral wording: “How would you compare the health risks of vaping to smoking?”
Serving Suggestions
Present your findings as a clear report with tables and graphs, highlighting key trends over time. Consider creating an infographic for policymakers or a press release for the public. If you are doing monthly waves, publish a trend line showing changes in quit attempts or product use.
Variations
- Regional focus: Limit your sample to a specific city or county to study local patterns.
- Longitudinal design: Follow the same individuals over time to track changes in behaviour, rather than a cross-sectional snapshot.
- Add biomarkers: If budget allows, collect saliva or breath samples to verify smoking status (e.g., cotinine levels).
- Different population: Adapt the questionnaire for youth, pregnant women, or people with mental health conditions.
Storage and Reheating
Store your raw data in a secure, encrypted folder with access restricted to the research team. Keep a copy on a separate drive or cloud service. Cleaned and anonymised data can be deposited in a public repository (e.g., the UK Data Service) after publication. For analysis, keep your code and codebook in the same folder. Revisit your data periodically to check for errors or to answer new questions – but do not reuse the same data for a different study without ethical approval.
Frequently Asked Questions
How long does the whole process take? From ethics approval to final analysis, plan for at least 3–6 months. The monthly data collection itself takes 2–3 weeks, but preparation and analysis are the bulk of the work.
Can I use existing data instead of collecting my own? Yes, the UCL STS data are publicly available for free on their website. You can download the dataset and start analysing immediately, which saves months of work.
What if I don’t have funding for a large sample? A smaller sample (e.g., 500) is still useful for exploratory analysis, but confidence intervals will be wider. You can also combine waves or use oversampling for key groups.
How do I ensure my sample is representative? Use probability sampling (e.g., random digit dialling) and weight your data to match census demographics. Report your response rate and discuss any limitations.
What ethical issues should I consider? Obtain informed consent, assure confidentiality, and allow participants to skip questions or withdraw at any time. If you collect sensitive data, ensure secure storage and anonymise before sharing.
Can I adapt the questionnaire for my country? Absolutely. Translate and culturally adapt the items, but validate the changes with a pilot test. Be aware that smoking patterns and product availability differ across countries.
Related
- Steaming Asparagus Spears: Easy How-To
- Bagel Kitchen Pacific Grove Style Bagels Recipe
- Vacuum Sealing Clothing: A Step-by-Step Guide
More recipes to cook
Safety notes: safe minimum internal temperatures from USDA FSIS; raw and undercooked fish guidance from the U.S. FDA. Food safety notes
Spotted an error? Tell us and we will correct it (corrections policy).