Community impacts research tries to understand how people experience the development in their everyday lives, if/how these experiences form into patterns across communities, and the extent to which these patterns can be causally linked to development of natural gas resources. In other words, we need to conduct the research such that we can say that development of unconventional natural gas (and all that entails) caused the changes that are seen in communities, such as increased crime rates, strains on local government budgets, decreased unemployment rates, increased homelessness, etc.
Here are the critical challenges we have faced as we’ve researched community impacts of Marcellus Shale development:
Correlation vs causation
Saying that a particular issue (such as a rise in crime rates) is directly caused by Marcellus development is difficult to document. Crime rates can increase for many reasons: 1) Unrelated economic changes can create greater opportunities for crime or can create greater desperation among residents; 2) heightened public awareness can result in more reporting of crimes; 3) changes in laws or enforcement procedures can change how or how likely crimes are reported.
How could Marcellus Shale activity lead to changes in crime rates? Establishing a causal relationship means comparing places with differing levels of Marcellus development and at multiple points in time and statistically controlling for some of these other factors (such as differences in population across places and time). It also means tracing through interviews and administrative data the ‘path’ by which Marcellus leads to increased criminal activity. Is it simply the out-of-state workers behaving badly? Is it local people working in the industry who have more money and more opportunities to commit crimes? Is it that more people means more opportunities for crime to occur, regardless of who committed the crime? Simply finding that crime is higher in places with Marcellus development (a correlation) is not enough to establish a causal relationship.
We have come across many instances where data simply do not exist or are not available for the years that we need. For example, the US Census is conducted every 10 years, giving us data in 2000 and 2010. The data from 2000 is a bit too early, in that there are about 6 years between that data point and the beginning of development activity. There are data available through the American Community Survey, but these data are 3- or 5-year averages, depending on the size of the place.
In other cases, data are not released in a timely fashion, such that in some topics the most recent data available is from 2009, which does not provide an adequate time series after the onset of development. For other topics, data simply do not exist. For example, there is no systematic data tracking health impacts, even for workers in the industry, because most health care providers have not collected such information historically. So while we use these data, we have to be very careful in interpreting the results.
Differences across the region
The Marcellus region is quite varied in its ecological, social, political, and economic histories, so it is quite difficult to apply research findings from one area to another. Some research and stories in the media study Bradford County, and simply apply the findings from Bradford County to other counties experiencing Marcellus Shale development. The characteristics of Bradford County (such as low population density, far from population centers, existing government structure and culture, economic history, landholding patterns) make this county more unique than representative of the rest of the state. It is not appropriate to assume that the impacts felt in Bradford County will be the same across the rest of the region.