Skip to content
No Vape

Why one vaping image experiment had two random choices

Distinguish assignment to a branded image set from the later choice of one image for a survey question.

No Vape · Updated

Stack of books on a table
Illustrative image · Laura Ohlman / Unsplash

Map the images to the two survey questions

Tattan-Birch and colleagues embedded an experiment in a survey in England, Canada and the United States. One condition saw four images of standardized vaping devices, while the other saw four images with branding. The interest question concerned the devices displayed within the participant’s condition and allowed selection of more than one. For perceived harm, the survey displayed one of the four devices from that condition at random. These two responses therefore concern different presentations. A person’s answer about interest in any item from a set is not four separate ratings of harm.

Tobacco Control via UCL Discovery: Impact of standardising the colour and branding of vape devices on product appeal among young people: a randomised experiment in England, Canada and the United States

Reconstruct a fictional participant’s view

Suppose a fictional experiment has sets A and B. An assigned participant sees four cards from set A and selects two as interesting. Later the survey shows the third card from A for a separate rating. The participant never sees set B. The answer about two cards belongs with the first question, while the rating belongs to the single displayed card. A tally that gives that person four ratings would invent three responses. A summary that calls the later card selection a change from A to B would also misdescribe the assignment. The two uses of chance can coexist without creating four additional study arms.

Read a headline with the original task in view

Before describing an outcome, identify the assigned set, the images visible at that point and the response the study actually collected. If a claim says participants “tried” a device, check whether the source instead asked about interest in trying a pictured device. If it says the study measured harm, check whether the source assessed perceived harm. Those distinctions come from the actual question and require no conclusion about the products. If a short summary omits which image the rating concerned, return to the procedure and outcome sections. When a figure combines replies by assigned condition, identify which single-image question fed that figure. A combined result across images should stay a condition summary unless the authors actually provide separate estimates for each image. Keep the experimental viewing task separate from behavior or exposure outside the survey.

Tobacco Control via UCL Discovery: Impact of standardising the colour and branding of vape devices on product appeal among young people: a randomised experiment in England, Canada and the United States

Sources and scope

Sources inform the guidance. The worksheets and examples are No Vape editorial suggestions, not tested treatments.

  1. Tobacco Control via UCL Discovery — Impact of standardising the colour and branding of vape devices on product appeal among young people: a randomised experiment in England, Canada and the United States
  2. No Vape — No Vape — app features and store links

Written with AI assistance and checked against the cited sources. No clinical review is claimed. This guide covers practical planning, not individual medical treatment. How we research and write