Content
Citation(s) & Footnote(s)
Book(s)No access
Identity and complete supplied coverage
Michael Siegel, William DeJong, Timothy S. Naimi, Erin K. Fortunato, Alison B. Albers, Timothy Heeren, David L. Rosenbloom, Craig Ross, Joshua Ostroff, Sergei Rodkin, Charles King, Dina L. G. Borzekowski, Rajiv N. Rimal, Alisa A. Padon, Raimee H. Eck and David H. Jernigan. “Brand-Specific Consumption of Alcohol Among Underage Youth in the United States.” Alcoholism: Clinical and Experimental Research 37(7), 1195–1203 (2013). DOI 10.1111/acer.12084. Publication metadata checked against PubMed; supplied early-layout PDF is numbered 1–9 with placeholder volume/issue, not final journal pagination.
All nine supplied pages, six tables with captions/footnotes, acknowledgments and 41 bibliography entries read. Tables 1–6 visually inspected on sheets 5–7. No figures or appendices supplied; no supplement is cited in the article. Canonical PMC access returned a browser challenge, so no claim of canonical full-version collation is made.
Coverage and findings
Sheet 1 abstract, authors/affiliations and rationale; 2 surveillance arguments/prior work/design; 3 recruitment, eligibility, response rates, exclusion and instrument; 4 standard drinks, estimation, winsorization, weights, validation and misclassification; 5 demographics/nonresponse and results (Tables 1–2); 6 type/prevalence results (Tables 3–4); 7 concentration and nonresponse interpretation (Tables 5–6); 8 discussion, funding and references; 9 remaining references.
The design surveyed 1,032 US respondents aged 13–20 who reported drinking in the prior 30 days; one implausible responder was excluded, leaving 1,031 for analysis. The prerecruited probability-based panel offered Internet access to households without it. Older respondents were recruited directly; younger teens required parental involvement. Conditional response estimates were 43.3% and 44.4%; these are not complete population recruitment yields. Respondents reported use, frequency and typical daily amount across 898 listed brands in 16 categories, with write-ins.
Prevalence is the proportion of sampled recent drinkers reporting a brand. Consumption share is estimated brand drinks divided by all estimated drinks, with weights and a 20-drinks-per-brand-per-day cap. These are distinct estimands: Bud Light appears at 27.9% prevalence but 6.4% of reported drinks; the top 25 account for 48.9% of reported drinks (Table 5). Bourbon, whiskey and Scotch are separate survey bins; their prevalence percentages cannot be added to obtain unique whiskey drinkers.
Critical assessment and table audit
Strengths include a much broader brand list than favorite-brand questions, explicit write-ins, sensitivity testing with/without winsorization, transparent weights and nonresponse comparisons. Type-specific pilot concordance supports feasibility but does not by itself validate every brand-level volume estimate. Self-report, recall, shared bottles/mixed drinks, variable ABV within brand families, average drink sizes and differential response remain limitations.
Table 2 shows higher nonrespondent representation among Black/lower-income/offline groups. The narrative says nonrespondents were older, but the displayed age distribution has more 20-year-olds among respondents (42.4% versus 37.6%); the direction sentence appears inconsistent with the table. No comparable nonresponse assessment was possible for younger teens. Demographic weighting cannot eliminate unmeasured preference/consumption differences.
The statement that selection bias cannot threaten the basic concentration finding is stronger than the evidence warrants: unobserved brand patterns could change concentration. The result is descriptive, not proof that advertising caused consumption; exposure was not tested here. The adult comparison uses 2010 commercial spirits-market data rather than a simultaneously recruited adult survey with identical measurement.
Table 6's title says six categories but displays seven. Its wine heading says 5.7% (all wines), while Barefoot's 1.3% overall / 25.7% within-category figures are more consistent with a denominator near the 5.2% table-wine share than 5.7%; rounding alone cannot reconcile that row. Preserve the published numbers and flag denominator ambiguity. Small cumulative rounding differences elsewhere are not automatically errors.
Study category labels are not controlling legal product definitions: Jack Daniel's is grouped as bourbon; cocktails are a separate flavored-beverage group; other flavored products appear under bourbon. Do not copy these bins into the Academy's bottle catalog. The 90 misclassifications occur among write-in reports by 198 youth, not 90% or a population-wide error rate. Sample demographics are unweighted (Tables 1–2); brand/type results are weighted (Tables 3–6).
Academy contribution and limits
Use for responsible alcohol education and adult learners' research literacy: distinguish people from drinks, self-report from sales, category labels from legal identities, and descriptive association from causation. A proposed exercise asks why the same brand can have 27.9% prevalence and 6.4% consumption share, and why a survey of recent drinkers cannot estimate drinking prevalence among all youth.
This historical underage sample is not an adult-market opportunity estimate, evidence of present-day brand popularity, or a basis for youth-targeted promotion. The study's public-health aim and NIAAA funding (R01 AA020309-01; additional Pleiades support acknowledged for King) should remain visible. Exact fieldwork dates are not supplied here; do not invent them from publication year.
Cross-source connection: 
Full supplied source examined; no raw respondent data or 898-item questionnaire is supplied, so no independent reanalysis is claimed. Originals preserved. Proton hash identity, exact saved-versus-canonical collation and Notion browser rendering remain separate open checks.
