ParticipACTION was a social marketing and communications organization that promoted physical activity in Canada for approximately 30 years before its official conclusion in 2001. We found that approximately 8% of Canadians were still aware of ParticipACTION unprompted and 82% were aware when prompted. In comparison to rates of unprompted awareness of 30% in 1989 and 17% in 1994  and 15% in 2004 , this reflects a definite decay in "top of mind" awareness of ParticipACTION between 1994 and 2007. This is also reflected by the fact that middle-aged and older Canadians were more aware of ParticipACTION than younger Canadians. Even when prompted, only 47% of people between the ages of 20-24 years were aware of the organization compared to 90% awareness among those 45-64 years of age.
The second purpose of this study was to determine if knowledge gaps in awareness of ParticipACTION existed and if these gaps were differentially related to household education and perhaps income. Both education and income were significant correlates of awareness among Canadians. The odds of people being aware of ParticipACTION were greater if they were more educated and reported higher income. These findings are consistent with the deficit model of knowledge gap phenomena  which suggests that the higher the education of an individual, the higher his/her motivation to attend to and comply with health messages.
The final purpose of this study was to determine if awareness of ParticipACTION was associated with physical activity related beliefs, intentions, and physical activity. We found good support for the use of HOEM  in that outcome expectations, self-efficacy, and intention varied significantly by awareness regardless of the awareness condition (unprompted, prompted). However, the sizes of the effects in the ANCOVAs were very small suggesting that other factors are influencing the beliefs and intention for physical activity apart from awareness. We also found that awareness of ParticipACTION was associated with LTPA status. However, this association was more apparent for the prompted condition. Consistent with the HOEM and other social cognitive models [28, 29], self-efficacy and intention were strong correlates of LTPA status. In fact, much of the association between awareness and LTPA status appeared to be attenuated with the addition of these constructs into the model for both awareness conditions. Thus the relationship between awareness and physical activity may be mediated by beliefs about physical activity. However, given the cross-sectional nature of the current data, this observation can only be verified in a longitudinal design (e.g., Bauman et al. ). We are not certain why awareness of ParticipACTION was independently associated with LTPA status in the prompted condition but not in the unprompted condition. By prompting people, it is possible that ParticipACTION was made a more salient behavioural target for people. Alternatively, it is possible that some individuals who were aware of ParticipACTION prior to its demise in 2001 were also aware of other organizations that continued to be in operation over the period (e.g. fitness studios like Curves, Heart and Stroke Foundation). In that case the other organizations may be more salient to those respondents at the time of the survey, which could influence their propensity to recall ParticipACTION relative to others without prompting, thereby attenuating the relationship between unprompted recall of ParticipACTION and current LTPA.
ParticipACTION was recently re-launched thus presenting a unique opportunity to better understand its potential to influence beliefs, intentions and behaviour through strategic evaluation . Our findings have both theoretical and practical importance for the new campaign. First, the HOEM is a useful theoretical framework for evaluating the mechanisms of influence of media campaigns on population-level physical activity and related beliefs. It appears that awareness of such campaigns is a central initial indicator of campaign effectiveness. But it was also apparent that, while awareness and knowledge of ParticipACTION and its messages on the part of the Canadian population is necessary, it is not sufficient if the campaign is to be effective in changing beliefs and behaviour [5, 6]. Second, education and income are critical determinants of knowledge about physical activity messages. Thus, levels of education and income are important factors to consider when designing and investigating population-based health promotion campaigns for physical activity. In particular, theoretical-based interventions that include information campaigns to influence knowledge and beliefs about physical activity should be tailored to be more salient for the intended audience . Tichenor and colleagues  identified communication skills, prior knowledge, relevant social contact, selective exposure and information storage, and the nature of the mass media system as explanations for education-related knowledge gaps. Thus, people with higher levels of education are more likely to expose themselves to mass media (e.g., read the newspaper), comprehend the information provided, and then have the social networks within which the messages are reinforced and valued. Therefore, attempts to lessen the knowledge gap about ParticipACTION could target specific segments of the population  and take into consideration these underlying processes.
In the past, the more information-rich media was consumed by better educated people while television was the source of information for the less educated segment of the population . Bonfadelli  suggests that knowledge gaps that occur in relation to 'old media' such as television and print news appear to stem largely from differences in interest to the topic. However, when considering the knowledge gap and internet use, there are new factors to consider. For example, compared to traditional media forms, the supply of information is conceivably unlimited on the internet and is more heterogeneous than the information journalists have access to. Additionally, internet use requires more technical savvy and is more dependent on access than old media. Therefore, the internet could act to promote individualized information seeking and increase heterogeneity within the audience . Data from two Swiss studies  found that internet use was associated with education level, income, sex and age. The author also found that diffusion and adoption of internet took place earlier and faster in high SES groups than in lower SES groups. Gaps in the usage of internet were related to variations in internet access and variations in content specific use. Gaps in the way the internet was used were primarily based on education. For example, people with higher education used the internet more for information seeking while people with lower levels of education used the internet significantly more for entertainment purposes. Therefore, physical activity promotion campaigns that use the internet will not be immune to the knowledge gap. If future ParticipACTION campaigns are planned for the internet, then strategies need to be put into place to overcome the potential knowledge gap phenomenon.
Strengths of the study include the validated survey and sampling methodology from an existing surveillance system (i.e., PAM) and the application of theoretical frameworks to guide the selection of questions and data analysis. However, this study is not without limitations that should be acknowledged. The use of single-item measures for self-efficacy, outcome expectations, intention, and the awareness variables, along with the lower response rate, may limit the reliability of our findings. As mentioned previously, the cross-sectional design prevents any discussion of cause and effect and limits the formal testing of the hierarchy of effects in the HOEM. While we were able to control for some demographic factors that may have accounted for the observed associations, other variables such as perceived health status and immigrant status may have been confounders for associations between intention and, or, LTPA and awareness. For instance, recent immigrants to Canada may be unaware of ParticipACTION regardless of their intentions to be physically active and levels of education or income. Exposure to media and media consumption are important variables to consider with regard to the effectiveness of a media campaign and the knowledge gap. Unfortunately, we had no such measures in our study. Finally, Kang  suggests that both awareness knowledge (i.e., the existence of the campaign) and depth of knowledge (i.e., actual message and contents) should be assessed when evaluating health media campaigns. With our items, we measured awareness knowledge but not depth knowledge. Future studies on the topic are advised to measure both types of knowledge while exploring the context in which campaign messages are received, interpreted, and used by audiences .
In summary, Canadians remain aware of ParticipACTION even though its campaigns have not been running for approximately 10 years. Those Canadians who are aware of ParticipACTION are more likely to hold positive beliefs about physical activity and to be physically active. However, a definite decay in awareness has occurred and younger adults are less likely to be aware of the campaign even when prompted. Knowledge gaps in awareness are associated with levels of education and household income. Thus, future physical activity promotion campaigns (i.e., the new ParticipACTION) need specific strategies to target different segments of the population, especially people who are living in deprived conditions with lower levels of education. Otherwise, such campaigns will actually increase the knowledge gap between these groups resulting in less opportunity to take advantage of programs and information .