Abstract:
|
When individuals respond to a survey about illicit drug consumption, they tend to underreport the amount of drugs they consume. To adjust for underreporting of marijuana use, researchers multiply the proportion of individuals who reported using by a constant factor, such as the ONDCP's 1.3. Researchers have suggested that different factors should be used for different demographics, but no values have been suggested. Although the current adjustments are simple, they don't account for changes in reporting over time. We determine whether reporting is changing over time by estimating changes in a quantity that is a proxy for changes in reporting. Our data was from the 1979-2013 NSDUH survey, a US-representative cross-sectional survey. By using domain estimation and weighted least squares regression we found that there is no significant changes in reporting in this period. So, although it might seem impossible to estimate any part of measurement error, we were still able to show that individuals' responses about their marijuana use have not changed despite drastic changes in stigma about the drug use.
|