Advertising exposure measurement in health behavior campaigns: In search of best practices

Theoretical Background and research questions/hypothesis: Advertising recall has been a primary method to evaluate the effectiveness of social marketing campaigns designed to influence knowledge, awareness and health behaviors. However, there is little agreement among practitioners on what constitutes an acceptable level of recall or standard methods to capture recall. To investigate the best practices for measuring advertising recall, and to assess how to evaluate claims of campaign effectiveness, NORC at the University of Chicago conducted a literature search and review. The literature search and review focused on answering the question: What are the best practices to measure advertising using aided and unaided recall in the context of campaigns or promotional efforts for health behavior change?

Methods: The search and review process included identification of articles and publications in peer-reviewed literature and in grey literature. The search was conducted using an iterative process and implemented a combined approach using search engines and reference lists from identified publications. Articles and reports were organized into campaign evaluation and methods papers. Researchers further organized the literature on campaign evaluation by health behaviors which included both national and international campaigns. Sixty-two campaign evaluation articles were identified during the search process which were reduced to 28 in the final review designed to evaluate medium to large scale campaigns using self-report population level survey methods.

Results: Findings revealed the need for further scholarship and dialogue to establish standard methods and best practices in the evaluation of health campaigns and social marketing approaches to promote health behavior change. First, there is inconsistent terminology throughout the practice and this creates difficulty for practitioners attempting to disentangle unaided from aided recall as benchmarks of performance. Second, existing studies exhibit a wide range of results as to what constitutes effective performance with unaided recall typically yielding lower levels of self-reported exposure. Third, measuring unaided recall is expensive, time-consuming and error prone. Fourth, recognition is another measure of exposure that has been shown to co-vary with recall and can provide additional evidence of exposure. However, recognition does not appear to be widely used at this time. Some scholars have indicated that recognition indicates more cognitive certainty of exposure to a communication as it provides a confirmation based on the actual advertising or message available to the respondent. Finally, feasible approaches to administering valid measures that minimize bias must be identified to support consistency of practice and comparability of results across health topic areas.

Conclusions: As the field of health communication and social marketing grows, practitioners and researchers, particularly those with limited funding levels, need a guide for practical implementation and benchmarking of exposure measurement in order to assess the performance of social advertising campaigns.

Implications for research and/or practice: It is important that guidance is made available through accessible channels for feasible application among a variety of practitioners. Guidance should include the assessment of trade-offs that influence the quality and validity of the findings as well as approaches to coding judgments for use with open-ended methods.