A home with solar panels on its roof is shown in a residential neighborhood in San Marcos, Calif. (REUTERS/Mike Blake)

Who chooses to install solar panels on their roof? You might assume that the people who do so are probably fairly rich (an average installation can cost around $35,000, prior to tax credits or other incentives), and most assuredly, politically liberal. They can afford it, and it fits their values to boot.

According to a new study, though, politics and income may not be such important factors after all. Examining the spread of solar photovoltaic residential installations in Connecticut, two researchers at Yale and the University of Connecticut found instead that the single most important factor driving whether a given house installed solar was peer influence -- whether other houses nearby had recently done so. In other words, much like with buying a Prius, it looks like installing solar has a lot to do with how you want people around you to think of you. "People have called it green envy before, where you want to be green so that you can show off your greenness effectively," says Yale's Kenneth Gillingham, a professor at the School of Forestry and one of the study authors.

The research, just out in the Journal of Economic Geography, was conducted by Gillingham along with the University of Connecticut's Marcello Graziano. Using data from Connecticut's Clean Energy Finance and Investment Authority, which provides incentives for solar installations, they geocoded the location of 3,843 residential solar photovoltaic units in the state between the years 2005 and the end of September 2013, tying each one to a "block group" as designated by the U.S. Census. (In general, a block group is comprised of between 600 and 3,000 people.) They also recorded the date when the initial application for an installation was made.

And hence the dramatic finding: The installation of one additional solar photovoltaic rooftop project within the past six months in a given area increased the average number of installations within a half mile radius by .44, or almost one half. As the spatial area widened, meanwhile, the influence of peer solar installations steadily decreased, a finding quite consistent with a theory of peer influence. Within a mile radius, the installation of one solar panel in the prior 6 months increased installations by .39, and within a 1 to 4 mile radius, by .12. This dovetails with past work by Gillingham and a colleague in California, which found that within a given zip code, one new solar installation increased the odds of another by .78 percent.

But while prior installations seemed to have a big influence on future ones, Graziano and Gillingham failed to find nearly as much of an influence for other socioeconomic and demographic factors. That included income, political party registration, and the unemployment rate. "I was expecting it to all be the wealthy communities," says Gillingham. "Pretty much, solar is going to be in the wealthiest places in Connecticut. But the pattern wasn’t that way at all." It's not that a lot of poor people were installing solar panels, of course--but a lot of middle income people were.

Here's a figure from the study showing the distribution of solar panels in relation to income across Connecticut. As you can see, many moderate income areas have plenty of solar:

Image credit: Kenneth Gillingham. Reproduced with permission. Based on "Spatial Patterns of Solar Photovoltaic System Adoption: The Influence of Neighbors and the Built Environment," Journal of Economic Geography, 2014.

Indeed, the study found that tiny Durham, Connecticut, with a population of less than 8,000 people, had the most solar systems -- 143 up through September of 2013. Durham is the very dense, orange colored cluster in south central Connecticut in the map above. It is certainly no hotbed of political liberalism: In the 2012 election, it split its votes almost evenly between Barack Obama and Mitt Romney, ever so slightly favoring the Republican candidate. And while people there are relatively well off, they are nowhere near as rich as in some other places in Connecticut, such as Westport, where the median household income is $ 152,586.  The population in Westport is more than three times that of Durham, and yet while it is also a relative solar leader, it can’t match the smaller, less wealthy town. "It’s not a poor community but I wouldn’t call it a very wealthy community," says Gillingham of Durham.

Granted, Durham also had some help. Despite the fact that when it comes to solar radiation, Connecticut gets less than many other states, it has nonetheless been a model for the growth of solar energy. Not only does the state's Clean Energy Finance and Investment Authority provide a variety of financial incentives to get homeowners to build rooftop solar panels (it has, thus far, supported 7,617 solar projects for residences). It has also promoted a program of "Solarizing" towns, which involves increasing solar adoption by giving better pricing if more people in an area sign up.  The program is "designed to leverage peer influence," says Gillingham.

What does this mean for the spread of renewable energy technologies, and solar in particular? In addition to initiatives like the Solarize program, Gillingham says the research suggests that it can be very important for houses who have just installed solar panels to let the installer put up a yard sign, especially if the installation is on the back of the house. "A common technique is to put a big sign in the front yard saying, 'This house went solar,'" says Gillingham. That then rubs off on neighbors, proving that while there may be many good economic and policy reasons to support clean energy, in the end, humans are also social animals, and motivated by peer and group effects.

"You want to conserve, and be environmental, but you want to do it in a conspicuous way," says Gillingham.