Robert Lichter S, Stephen J. Farnsworth and Justin Rolfe-Redding
This study examines a unique dataset of polling results from two comparable surveys, which asked the same questions of US. atmospheric and ocean scientists in 1991 and 2007. Using this measure of change in scientific opinion over time, the data shed light on the evolution of a consensus over a critical and dynamic 16-year period in the history of climate science. In 1991, considerable uncertainty and disagreement existed among climate scientists. With the passage of time and additional research, by 2007 a clear consensus had emerged, along with increased confidence in the opinions expressed. This is the path that scientific progress is expected to take, and it argues against charges that the climate science community rushed to judgment or engaged in groupthink on global warming. The period 1991 to 2007 was a time of considerable expansion in the scientific community?s conclusions about global warming, as reflected in the reports of bodies such as the U.N.?s Intergovernmental Panel on Climate. In 1990, the IPCC issued its First Assessment Report [1,2], expressing confidence that global warming was occurring, but noting uncertainty on the anthropogenic origin of observed warming. By 2007 (just prior to data collection in the survey reported here), the initial sections of the Fourth IPCC assessment 2 were released, expressing ?very high confidence? that human activity was causing global warming (p. 3), and describing evidence of warming as ?unequivocal? (p. 5). In addition to the synthesis reports, results of surveys of scientists can offer another measure of the state of scientific opinion. In 1990, an international sample of ?global environmental change? scientists showed strong support for action to address climate change, but lower scientific certainty-65% thought there was a more than 50% chance of 2 degree Celsius warming over the subsequent 100 years [3]. A further survey indicated a divide among scientists over the existence of anthropogenic warming [4]. A 1996 international survey of climate scientists and meteorologists indicated fairly high certainty in long-range warming sufficient to justify policy action [5]. The scientists expressed limited trust, however, in climate models and uncertainty on the specific impacts of climate change. Between 2006 and 2008, three surveys of scientists with different sampling frames all provided a more recent picture of scientific opinion [6?8]. They demonstrated broad confidence in the methods, and consensus on the findings of climate science, in areas that are prerequisites to basic agreement on human causation. For example, 90% of one sample concurred that temperatures have risen over pre-industrial levels [9]. While confidence in scientific understanding of certain elements of the climate system was high (e.g., the role of albedo), lingering uncertainty remained on some of the predictions and other fine points of the workings of the climate system (e.g., the role of cloud cover [6]). These surveys, however, fail to provide a rigorous longitudinal perspective. Each employed its own survey frame, question wording, and sampling and administration methodology. While broad trends may be discernable from examining such disparate datasets, and worthy efforts have been conducted in this regard [10-13], an applesto- apples comparison would provide a far more precise measure of changes in scientific opinion. The current study addresses the need for longitudinal data on scientific opinion about climate change by comparing the results of two surveys of US atmospheric and ocean scientists-one administered in 1991 and one in 2007-designed with similar methodological profiles and virtually identical question wordings, allowing for superior comparison of the survey results from these two points in time. This study also examines a richer array of topics than prior longitudinal analyses.
