Department of Community Medicine and Health Care , The University of Connecticut School of Medicine, Farmington, Connecticut.
Alcohol Clin Exp Res. 2013 Oct;37(10):1787-93. doi: 10.1111/acer.12146. Epub 2013 May 17.
This study evaluated the use of a modified Delphi technique in combination with a previously developed alcohol advertising rating procedure to detect content violations in the U.S. Beer Institute Code. A related aim was to estimate the minimum number of raters needed to obtain reliable evaluations of code violations in television commercials.
Six alcohol ads selected for their likelihood of having code violations were rated by community and expert participants (N = 286). Quantitative rating scales were used to measure the content of alcohol advertisements based on alcohol industry self-regulatory guidelines. The community group participants represented vulnerability characteristics that industry codes were designed to protect (e.g., age <21); experts represented various health-related professions, including public health, human development, alcohol research, and mental health. Alcohol ads were rated on 2 occasions separated by 1 month. After completing Time 1 ratings, participants were randomized to receive feedback from 1 group or the other.
Findings indicate that (i) ratings at Time 2 had generally reduced variance, suggesting greater consensus after feedback, (ii) feedback from the expert group was more influential than that of the community group in developing group consensus, (iii) the expert group found significantly fewer violations than the community group, (iv) experts representing different professional backgrounds did not differ among themselves in the number of violations identified, and (v) a rating panel composed of at least 15 raters is sufficient to obtain reliable estimates of code violations.
The Delphi technique facilitates consensus development around code violations in alcohol ad content and may enhance the ability of regulatory agencies to monitor the content of alcoholic beverage advertising when combined with psychometric-based rating procedures.
本研究采用改良 Delphi 技术与先前开发的酒精广告评级程序相结合,评估其在美国啤酒协会准则中检测内容违规的应用。其相关目标是估计获得电视广告中代码违规可靠评估所需的最少评估员人数。
选择 6 个可能存在违规行为的酒精广告,由社区和专家参与者(N=286)进行评级。使用定量评分量表根据酒精行业自我监管准则衡量酒精广告的内容。社区组参与者代表了行业准则旨在保护的易受伤害特征(例如,年龄<21);专家代表了各种与健康相关的专业,包括公共卫生、人类发展、酒精研究和心理健康。酒精广告在相隔 1 个月的 2 次评估中进行评估。完成第 1 次评估后,参与者被随机分配到接收 1 个组或另 1 个组的反馈。
研究结果表明:(i)第 2 次时间点的评分方差普遍降低,表明反馈后达成了更大共识;(ii)专家组的反馈比社区组的反馈更能在形成组共识方面产生影响;(iii)专家组发现的违规行为明显少于社区组;(iv)代表不同专业背景的专家在识别违规行为的数量方面没有差异;(v)由至少 15 名评估员组成的评级小组足以获得可靠的代码违规估计。
Delphi 技术有助于围绕酒精广告内容中的代码违规达成共识,并且当与基于心理计量学的评级程序结合使用时,可能会增强监管机构监测酒精饮料广告内容的能力。