Most people believe that climate change is happening and that human activities are the predominant cause (http://go.nature.com/eIj9uN). Support for policies to cut greenhouse-gas emissions and reduce the carbon footprints of communities varies, however.

National surveys that continue to focus on basic questions around climate science obscure the complexities of public opinion on climate change. To understand this variation, social science needs to move beyond national data and find new ways to quantitatively unpack the public's occasionally incoherent views on climate change.

In this issue, Peter Howe and colleagues show how national polls can hide local trends (page 596). They show that while 63% of US citizens believe global warming is happening, agreement at the county level ranges from 43% to 80%. Such variation translates into different levels of support for tackling the problem. For instance, in states such as California, up to 10% more citizens are in favour of regulating CO2 than the US average. In other states, such as Wyoming and North Dakota, up to 14% fewer citizens support such policies than the national average. Understanding such diversity is essential if politicians are to design mitigation and adaptation strategies at state and local levels, they argue.

But it's unlikely that simply knowing how opinion varies will be enough to design robust climate policies. Policymakers must also better understand how subnational socio-economics affects particular communities.

To this end, Ilona Otto and colleagues call for standardization of open-access, geo-referenced subnational datasets (page 503). Such data can generate better understanding of the interconnectedness of global environmental changes and social impacts, allowing policies to be tailored to a particular locale.

It's not just exogenous differences, such as where people live, that drive variation in public attitudes and responses to climate change. Endogenous differences in the way groups access and interpret scientific information also have an effect.

It is well known that the media, scientific literacy, and personal ideology can colour public attitudes to climate change. In this issue, P. Sol Hart and colleagues examine how these factors interact (page 541). They find that conservatives who pay attention to scientific news demonstrate greater knowledge of climate change, and perceive greater harm. But more knowledgeable conservatives were less likely to support climate mitigation policies. The study shows how exposure to scientific and political news can both polarize and bring together politicized views on climate change.

Understanding the nuances of public opinion in this way remains important as it affects the probable success of policies. Alexa Spence and colleagues analyse one example of this (page 550): UK citizens' attitudes to demand-side management. They find a large proportion of people indicated an unwillingness to share energy data, which could block the implementation of crucial 'smart' energy technologies. Significantly, respondents concerned about climate change were more likely to accept such perceived intrusions.

These studies show that detailed analyses of issues beyond headline attitudes, and at subnational levels are critical for casting light on what the public really thinks about climate change.