Global Warming’s Six Americas Screener Manual
This manual was developed to assist interested parties in using the Global Warming’s Six Americas audience segmentation typology. The segmentation typology is fully described in Maibach, Leiserowitz, Roser-Renouf, & Mertz (2011).
In brief, the segmentation analysis was performed by subjecting 36 variables – drawn from four categories of variables: global warming beliefs, issue involvement, behaviors, and preferred societal responses – to Latent Class Analysis (Magidson & Vermunt, 2002a, 2002b). The resulting six audience segments – which form a continuum – were named the Alarmed, Concerned, Cautious, Disengaged, Doubtful and Dismissive. A description of these audience segments can be found in the Maibach et al. paper, and in a variety of reports located in the resources sections of George Mason University Center for Climate Change Communication website (http://climatechange.gmu.edu) and the Yale Project on Climate Change Communication website (http://environment.yale.edu/climate-communication).
Also described in the Maibach et al. article are two survey tools – a 36-item instrument and a 15-item instrument – that we developed for our own use, and for use by other researchers to identify the Six Americas in new, independent data sets. These tools were created using linear discriminant functions (Hair, Anderson, Tatham & Black, 1992; Tabachnik & Fidell, 1989) to identify Six America segment status.
The discriminant analysis using the 36-item instrument correctly classifies 90.6% of the sample (as compared to the original Latent Class Analysis results); accuracy varies by segment, ranging from 79% to 99%. The 15-item instrument correctly classifies 84% of the sample, ranging by segment from 60% to 97%.
Both of these instruments – along with codebooks, and SAS and SPSS scripts that run the discriminant functions – are provided in this manual. Additional SAS files that include the discriminant functions are needed in order to run the SAS scripts, and should be downloaded along with this manual.
Hair, J.F, Anderson, R.E. Tatham, R.L., & Black, W.C. (1992). Chapter 3, "Multiple Discriminant Analysis" in Mulivariate Data Analysis with Readings. New York: Macmillan Publishing Company. pp 87-152.
Maibach, E. W., Leiserowitz, A., Roser-Renouf, C., & Mertz, C. K. (2011). Identifying like-minded audiences for climate change public engagement campaigns: An audience segmentation analysis and tool development. PLoS ONE. 6(3): e17571. doi:10.1371/journal.pone.0017571
Magidson, J., Vermunt, J. K. (2002a) Latent class models. In D. Kaplan (ed.)The Sage Handbook of Quantitative Methodology for the Social Sciences.Thousand Oaks, CA: Sage. Pp. 175-198.
Magidson, J., Vermunt, J. K. (2002b) Latent class models for clustering: A comparison with K-means. Can. J. Marketing Research, 20, 37-44.
Tabachnick, B.G. & Fidell, L.S. (1989). Chapter 11. "Discriminant Function Analyses", in Using Multivariate Statistics. New York: Harper Collins Publishers. pp. 505-596.