Obesity is a major public health threat in the US.¹ It is linked to many chronic diseases including diabetes, coronary heart disease, stroke, hypertension, and certain types of cancer, and is expected to account for $147 billion in annual costs.² According to the Centers for Disease Control and Prevention (CDC), 34.9% of all Mississippians are classified as obese; the single highest obesity rate in the nation.^3,4 Obesity is caused by various factors including unhealthy eating⁵ and inactivity behaviors⁶; these behaviors can be influenced by characteristics of the environment (e.g., aesthetics of a neighborhood, grocery store access).^{7,8, 9}

Laws, regulations, and guidelines as public health policy can profoundly impact health behavior and status.¹⁰ Numerous policy interventions have been identified as effective in promoting physical activity, healthy eating, and obesity prevention.^11–22 In addition to nationally implemented policies, mounting evidence shows that implementation at the local level is critical to achieving changes in risk factors and health outcomes.^10,23–25 A 2009 IOM report describes the importance of local government action in obesity prevention and lays out policy recommendations likely to directly affect children.²⁴

When making decisions about supporting an issue, policymakers often prioritize constituents’ needs or opinions above other factors.²⁶ Further, the IOM report, Preventing Childhood Obesity: Health in the Balance²⁷, emphasized that policy changes occur more quickly “when there is strong social consensus behind them.”^28,29 It is therefore important to provide timely data showing public support of evidence-based approaches. Obtaining information on public health topics, however, requires a large and diverse sample of respondents to provide such information. Most public health data collection assesses factors such as health status, socioeconomic, and demographic data, but misses the information that may be of particular importance to influence policymaker priorities. There is a lack of public opinion data, particularly at the state and local levels, to help local-level decision makers tailor policies that would most efficiently and effectively impact obesity rates. The lack of data available to advocates also hampers efforts to communicate constituent concerns to policymakers. Peer reviewed studies reporting on such perceptions are needed to help shape such communication efforts. Putting these studies through peer review and making them available in the scientific literature (not just the popular press literature) enhances the opportunity for researchers and policymakers to make use of these findings. This work aims to fill that gap by reporting the opinions of a random sample of residents from 10 counties in Mississippi.

The overall goal of this study was to determine the public support for policies related to nutrition and physical activity through administration of a phone survey in 10 counties in Mississippi. The study protocol has been reviewed and approved by the Washington University Institutional Review Board.

The final sample contained 2,800 individuals, roughly half (1,401) of whom were from the five counties with the lowest obesity rates in Mississippi. The sample was 71% female (2,000). Fifty-one percent of the sample (1,432) was white, 45% (1,263) were Black or African American, and nearly 4% (105) were some other race. Additional demographic characteristics of the sample can be found in Table 1, with additional demographic information on supplemental Table 1.

Among the total sample, support for healthy eating and activity policies was highest for: Requiring at least 30 minutes of physical activity or physical education classes every day for children in kindergarten through 12th grade (93%) and lowest for: Taxing soda and soft drinks and using the money for public education campaigns to fight obesity in children (65%). Support for local funding policies was highest for Recreation centers (86%) and lowest for Swimming pools (74%) (Table 2).

The levels of support for each policy, stratified by race, obesity level of the county, and gender, are shown in Table 2. Support for nearly every policy was greater among African Americans, females, and those living in counties with higher levels of obesity. In general, the trends across BMI, education, income, and age were not linear (data not shown), in that support did not increase or decrease consistently across categories.

For the sum of all 10 healthy eating and activity policies, summary scores ranged from 10 to 40, with a mean of 17.1 (SD=6.0) and for local funding the sum of the four items ranged from 4 to 8, with a mean of 4.7 (SD=1.3). Females were significantly more likely than males to support both types of policies (OR= 1.69, 95% CI: 1.28–2.24 and OR=1.30, 95% CI: 1.06–1.60) (Table 3). When this model was adjusted for all demographic factors listed in the methods, the association was no longer significant for the healthy eating and activity policies (adjusted OR=1.42, 95% CI: 0.98–2.06) but remained borderline significant for the local funding policies (adjusted OR=1.39, 95% CI: 1.04–1.85). Compared to Caucasians, Blacks/African Americans were significantly more likely to support both healthy eating and activity policies (adjusted OR=4.18, 95% CI: 2.48–7.05) and local funding policies (adjusted OR=2.84, 95% CI: 1.95–4.15). Age was associated with support for local funding for all age categories compared to those 65 years and over, but this was significant after adjustment only for the age groups 25–34yrs; 35–44yrs; and 45–54yrs.

Those earning less than $15,000 were more likely to support both types of policies than those earning at least $75,000. However, the only association that was significant before and after adjustment was between those earning $15,000-$24,999 and those earning at least $75,000 for supporting the healthy eating and activity policies (adjusted OR=2.32, 95% CI: 1.19–4.52).

Before, but not after, adjustment, respondents living in the 5 counties with the highest levels of obesity were significantly more likely to support both types of policies than those in the counties with the lowest obesity prevalence. Respondents with at least one child in the home were significantly more likely to support policies related to local funding (OR=1.94, 95% CI: 1.53–2.44). Similarly, obese individuals were more likely to support policies related to local funding than those with a normal BMI between 18.5–24.9 kg/m². Neither of these associations was significant after adjustment. Only those in the lowest education category (less than high school) were significantly more likely than those with at least a college education to support healthy eating and activity policies; this association was no longer significant after adjustment.

Findings from this study indicate support among Mississippi residents for policies aimed at promoting healthy eating and activity in their communities, particularly among those living in counties with the highest levels of obesity. Other large-scale polls^32–35 have shown similar levels of support. In late 2010, the California Endowment funded a poll of 1,005 registered voters³⁴ and found that levels of support for banning advertising were similar in Mississippi and California (66% and 64%, respectively).³⁴ Public use of school facilities when school is not in session had 88% support in California³⁴ and 80% support in Mississippi. On the other hand, Mississippi residents had higher support for banning the sale of soda in school (70% strongly or somewhat supported) than those in New York State (59% support)³² (April 2010 Quinnipiac University RDD survey of 1,381 New York State registered voters). Similarly, 65% of Mississippi respondents expressed support for taxing soda and soft drinks to fund public education campaigns for childhood obesity while only 56% of respondents in the California survey supported this policy.^34,35 A slightly different version of this question (where the use of tax funds was not specified) was asked in a December 2009 Columbia Broadcasting System (CBS) poll,³³ which included a RDD sample of 1,048 adults nationwide; only 38% favored the policy. Finally, in the Quinnipac University poll,³² which suggested the money raised from such taxes would be used to fund health care only, 48% supported the policy.

Support among Mississippians has remained relatively constant for some policies since a 2006 poll by the Mississippi Center for Health Policy.³⁵ For example, support for requiring schools to send annual BMI scorecards to parents was 68% in 2011 and 66% in 2006; this was much higher than the national level of support (51%), at the time (2006). A 2006 survey found that support for eliminating vending machines in elementary schools was slightly greater in Mississippi³⁵ (78.5%) than the rest of the nation at that time (71%); in this 2011 survey, 74% of Mississippians supported the policy.

It is often difficult for researchers to conduct polls assessing policy perceptions due to financial and methodological constraints. Funding and research study timelines are often out of synch with the need for information about timely issues.³⁶ The surveys mentioned above were not published as part of the scientific literature. In the case of these surveys, dissemination through popular, rather than scientific press, allows practitioners and policymakers the opportunity to use the knowledge gleaned from these investigations in a more timely fashion. While inclusion in the popular press is important, publication of these findings in the scientific literature aids in their use by researchers and policymakers. This type of publication allows for peer-review, which adds credibility, and allows them to be searchable in databases such as Pubmed.

Many of the trends across demographic characteristics seen in the Mississippi population were also seen in similar surveys conducted elsewhere. In Mississippi, support for taxing soda and soft drinks had greater support among African Americans (76%) than Caucasians or any other race (55% and 61%, respectively). Similarly, in California,³⁴ the highest support was among African Americans and Latino/Hispanics (70% and 72%, respectively) compared to 50% and 53% for Caucasian non-Hispanics and Asian-American/others.

Findings from this Mississippi study have a number of research and policy implications. This study identified several characteristics as being related to policy support. Future research efforts should investigate the use of such components to target social marketing campaigns, such as those aimed at building support for policy change. A recent IOM report, Accelerating Progress in Obesity Prevention: Solving the Weight of the Nation,³⁷ recommends social marketing as a method of changing public opinion. An essential piece of this method, is that these campaigns be highly targeted (in terms of audience and behavior).³⁸

In addition to the research implications, this study has the potential to inform policy. Policymakers rely heavily on constituent needs and opinions in making crucial policy decisions.²⁶ Practitioners can use this information to tailor communication about public health interventions, including policy approaches. Local health leaders can determine priorities for education and awareness. In addition, this information, showing clear support for policies encouraging healthy eating and physical activity, can be powerful in the hands of advocates working with policymakers to achieve change.²⁶

An important use for these findings is to provide those looking to achieve policy change with information regarding population subgroups most likely to support (and those most likely to oppose) policy changes related to healthy eating and physical activity. This study found that African Americans and females were more likely to support healthy eating and activity policies and local funding policies. Similarly, those in younger age groups were more likely to support local funding than those in the oldest age group. These are individuals who may need to be specifically targeted, perhaps with different messages, in order to change their perception.

Study limitations include a cross-sectional design, which limits causal inferences. Secondly, though the study was designed to be representative of the counties sampled (the five with the highest and five with the lowest obesity levels in Mississippi), it is not representative of state or national samples, which may limit generalizability. The response rate for the sample was low, however the individuals were sampled for variability (by selecting those in the counties with the highest and lowest levels of obesity), and the aim was not to try to estimate area-wide prevalence. Third, related to sampling, we had a low response rate and though an attempt was used to survey by cell-phone as well as landline, there may have been bias in those choosing to participate in the survey.

These results paint a picture of the perceptions of residents concerning public health policies related to healthy eating and physical activity. Though patterns were found by demographics such as race and gender, overall support was quite high for all policies. In practice, individuals and groups that support such policies should be recruited as advocates. Those opposing the policies may be targeted with messages aimed at reducing their resistance. These findings can help move public opinion, a key driver of policy change.