VAPER Study FAQ

Who participated in the study?

American adults (aged ≥21 years) who used e-cigarettes at least 5 days per week were recruited via advertisements posted in 404 Craigslist catchment areas covering all 50 states. In wave 5, adults who had previously joined our study and were using e-cigarettes less or not at all were included.

How many people were included in the study and when did you collect data?

In total, 5 waves of data were collected between 2020 and 2023 and each wave included between 1209 and 1312 participants. Those lost to follow-up each wave were replaced. Specific details for each wave of data collection can be found below:

  • Wave 1: May 2020-Sept 2020; n=1209
  • Wave 2: Dec 2020-April 2021;  n=1218
    • Follow-up rate: 52%
  • Wave 3: Sept 2021-Nov 2021; n=1254
    • Follow-up rate: 57%
  • Wave 4: July 2022-Sept 2022; n=1247
    • Follow-up rate: 72%
  • Wave 5: Feb 2023-April 2023; n=1312
    • Follow-up rate: 74%

What exactly did you collect data on?

Some of the primary measures included in the VAPER Study were:

  • E-cigarette use and dependence
  • The specific device characteristics of each device used by participants, such as device brand, model, type, wattage, voltage, resistance, and modifiability (e.g., adjustable airflow, power, coils)
  • The specific liquid characteristics of each liquid used by participants, such as the liquid brand, flavor, nicotine concentration and formulation, and propylene glycol and vegetable glycerin content
  • Use of other tobacco products, including cigarettes and cannabis

For a more complete list of measures, you can access the full questionnaires.

How did you account for the high variation and customizability of e-cigarette products?

The questionnaire measures and skip logic were designed to accommodate marketplace heterogeneity and user customization (e.g., different skip logic pathways for different device types and customizations). Photos of devices, their visual displays (when applicable), and liquids allowed for the collection of key data not easily self-reported by participants, such as nicotine concentration.

How did you avoid bots and unreliable data?

Several strategies were applied to maximize the odds that participants who received incentives were not bots and were likely to possess an e-cigarette. For example, we required participants to submit a photo of their e-cigarette device and requested photos of their liquids, which were reviewed to ensure the participant was in possession of an e-cigarette and the photos weren’t pulled from the internet. Additionally, incentives were delivered by mail to new participants (to confirm their address and delay gratification) and electronically to returning participants. We also conducted identity checks of all participants, employed a CAPTCHA test, and included attention-checking questions, among other strategies.

What did you learn about recruiting and maintaining participants?

Participant recruitment and retention can be a substantial challenge in longitudinal cohort studies, particularly those conducted online. Participants of the VAPER Study were initially recruited via social media channels (e.g., Facebook, Instagram); however, recruitment was slow and expensive. Subsequently, recruitment was transitioned to Craigslist, which proved much quicker and more affordable for this sample. Additionally, we implemented a range of strategies to improve participant retention, such as higher valued incentives, annual gift card raffles, various messages and frequencies for survey invitations and reminders, and a webpage for sharing study updates and results with participants. A randomized experiment was also conducted which demonstrated that participants receiving a $30 gift code upon completion of the survey were more likely to complete the survey than those receiving a $15 gift code both before and after completing the survey.

What did you learn about the web-based survey methodology?

Relative to existing e-cigarette cohort studies, this study methodology has some advantages, including efficient recruitment of a lower-prevalence population and collection of detailed data relevant to tobacco regulatory science (e.g., device wattage). The web-based nature of the study requires several bot- and fraudulent survey–taker–related risk-mitigation strategies, which can be time-intensive. When these risks are mitigated, web-based cohort studies can be successful.

Where can I learn more about this study and its peer-reviewed publications?

If you are interested to learning more about the study, we suggest taking a look at our publications and presentations. These will provide more information about the study’s methodology, research questions, and conclusions. If you still have questions, you may contact Jeffrey Hardesty.