Development and Testing of a Model and Instruments for the Collection of Comparative Communication and Marketing Data

Theoretical Background and research questions/hypothesis: The success of public health interventions relies not only on delivering accurate, accessible and actionable information to the public, but also on having accurate and actionable information about the public. Public health agencies, however, often have limited, qualitative information on people’s beliefs, perceptions, values and behaviors. The purpose of this project was to develop and test a construct-based tool for collecting population-level data on communication and marketing variables common to multiple health topics and that affect health behavior and outcomes. The project is based on the People and Places framework (Maibach, Abroms, and Marosits 2007), a streamlined ecological model of public health to inform communication and marketing.  Key constructs from the framework were used to create a population survey with core and topic-specific questions that could be fielded by telephone or the web. The main research questions were 1) what are the mode effects for telephone and Internet based surveys on communication and marketing 2) how do participant responses compare and differ across health topics, 3) how feasible is it to develop sets of questions across health topics using a common set of communication and marketing constructs?   

Methods:  The research team developed and tested a multi-part telephone and web survey with communication and marketing variables for core attitudes and behaviors and three topics: seasonal influenza, H1N1 pandemic influenza, and climate change health effects. The survey instrument was tested with participants and revised in 3 phases: cognitive (39 participants), pilot (275 participants) and field testing (4235 participants). The cognitive interviews were conducted in English, Spanish, Mandarin and Vietnamese, and the pilot and field tests in English and Spanish. The Asian languages were not used in the pilot and field tests due to programming challenges.  

Results:  The three-stage testing produced an instrument that was consistent with the model and validated in the field. Mode effects consistent with the literature and for certain types of questions were confirmed. Significant differences in 15 of the 18 common questions across the 3 health topics were identified. Seasonal and pandemic flu responses depended on age and education. The feasibility of conducting multi-topic surveys based on shared communication and marketing constructs was demonstrated.    

Conclusions: 

The findings suggest that a construct-based approach can be effectively used as a survey template to field common questions across a range of health topics. 

Implications for research and/or practice: 

The primary benefits of such a template include having a systematic, theory-based and efficient process to inform public health communication and social marketing efforts. This approach can also be cost effective, as it distributes the costs associated with a nationally representative data collection effort across multiple program areas. Further cost savings would accrue as new health topic modules would only require a minimal amount of pretesting.