Table 3.
Key Findings, Implications, and Suggested Approaches.
| Key Findings | Measurement Implications | Suggested Approaches |
|---|---|---|
| Specification Error | ||
| * The cannabis market is changing quickly with new methods, devices, forms, and cannabinoid concentrations | - Out-of-date definitions - Misreported use |
- Use detailed and updated definitions of
marijuana/cannabis - Update product definitions as products change - Employ cannabis typology to capture multiple forms of cannabis consumption e.g. definition: The next question is about use of cannabis, which is also called marijuana. Cannabis is often used for fun or medical purposes, and people can smoke, eat, or vape it. When answering please include cannabis in all forms, including flower/bud, hashish, hash oil, tinctures, any food or drink containing cannabis (aka edibles), and products with CBD and no psychoactive effects. |
| Low/no THC products are used in different situations, for different purposes, and with different physical effects than high THC products | - Imprecise measurement of drug use - Imprecise measurement of THC exposure |
- Differentiate between low/no CBD product use and
other cannabis use - Explicitly ask about or exclude CBD-only products - Alternatively, differentiate between medicinal vs. recreational use e.g. When you use a vaporizer with hash oil or cannabis concentrates, do you typically use hash oil or concentrates that are: (1) CBD only (no THC) (2) Low THC / High CBD (3) Equal amounts of THC and CBD (4) High THC / Low CBD (5) I don’t know |
| * Inconsistent reporting of CBD-only products | - Measurement error of unknown magnitude | - Specify whether CBD-only products are included or
excluded in cannabis definitions - Ask about CBD-only products separately from other cannabis e.g. When answering please include cannabis in all forms, including flower/bud, hashish, hash oil, tinctures, any food or drink containing cannabis (aka edibles), and products with CBD and no psychoactive effects. |
| * Blunts are not always considered co-use | - Underreported
co-use (particularly when a global co-use question is used to screen and trigger skip logic) - Underreported blunt use |
- Ask about blunt use to all marijuana users, including those who report no cannabis-tobacco co-use |
| Methods of co-use vary | - Underreported co-use that varies across method (e.g. underreports by users who are chain tobacco smokers/vapers who intersperse cannabis (unintentional overlapping)) | - Specify methods of co-use e.g. co-administration: In the past 30 days, have you used cannabis and tobacco or nicotine together at the same time in the same product (e.g. in a spliff, blunt, or moke)? |
| Motivations for co-use vary | - Underreporting co-use if measures ask only about specific motives | - Ask about co-use methods and co-use motives
separately - Specify which co-use motives are of interest (e.g. the item below asks about sequencing only) The next question is about using nicotine or tobacco after smoking or vaping cannabis. This is sometimes called “chasing.” In the past 30 days, how often did you smoke a cigarette, cigar, or cigarillo after smoking or vaping cannabis? (1) Every time I smoked/vaped cannabis (2) Some of the time (3) None of the time |
| Social use is common, introduces new ways of cannabis consumption, and affects dosage | - Social use/experimentation increases methods of
exposure to drug and potentially to risks from the delivery method (e.g. trying
dabbing) - Increases poly-cannabis use - Past 30-day measures include both regular (preferred) and social products - Product-specific measures of use will be noisy |
- Differentiate between regular or preferred
products and social/occasional products - Interpret measures with this finding in mind |
| Measurement Error | ||
| Response editing was low | - In this population, survey responses likely reflect use | - Work is needed to investigate in other contexts, such as where cannabis is not legal or more stigmatized |
| Dosage varies between and within devices and between and within cannabis forms | - Global measures of use will be noisy - Even product-specific measures of use will be noisy |
- Differentiate between products - Specify CBD/THC content - Specify which cannabis form is used in vaporizers - Use THC biomarkers to enhance measures of exposure |
| Dosage can be uncertain; Product knowledge can be uncertain (e.g. about cannabinoids and strains) |
- Respondents may have difficulty answering detailed questions - Difficulty may be higher for individuals who share with friends, purchase on the black market, or have less domain knowledge |
- Tailor questions to the level of expertise of the
target population (e.g. questions of the general population might be less
detailed than questions of daily cannabis users; questions in states without
legal cannabis might be less detailed than questions in states with legal
cannabis) - Include “I don’t know” option for CBD/THC concentration and dosage |
*
indicates findings we encourage every survey researcher to address in their measures
For each source of potential specification or measurement error, we describe the primary implications for survey measurement and potential solutions. Survey researchers must balance maximizing accurate measurement and minimizing survey fatigue by prioritizing the needs of their research question.