This project describes the development of the Climate Adjective Rating Scale (CARS), a psychological measure used to examine how people use language to characterize their experience of the weather and climate. The author searched multiple dictionaries along with published weather and climate summaries to identify all of the possible English words (most typically adjectives) that were used to describe weather and climate (or that could function in this capacity). Terms that were identified were then cross-referenced with several thesauri to ensure that an inclusive and comprehensive listing of terms was developed. The goals of this project were to develop this comprehensive listing of terms and then to understand the underlying latent dimensions that may reveal the various ways that people symbolize their experience of weather and climate conditions with words.

The CARS consisted of an initial listing of 153 adjectives that were evaluated using a five-point rating scale (1 = Never describes my experience of the weather/climate of this place to 5 = Always describes my experience of the weather/climate of this place). The study participants were instructed to think of what the climate of their current local has been like over time and in general. If they just moved to the locale, they were asked to think about the climate patterns at their former residence. The participants were 292 undergraduate students from the University of Georgia who responded to the CARS via the Internet.

Responses to the CARS were factor analyzed to explore the latent dimensions that may explain the correlations among the 153 items. A maximum likelihood factor extraction method was used along with a factor rotation that allowed for the factor intercorrelation. Fourteen factors were retained for interpretation that used 188 items (some items loaded on more than one scale). Given the number of factors and the likelihood of additional superordinate latent dimensions, a second-order factor analysis was performed. The two second-order factors corresponded to Bad and Fair weather, respectively. One first-order factor emerged that did not load onto either second-order factor. This factor comprised adjectives that pertained to the stability and predictability/unpredictability of the weather and climate. The factor names and structural relationships for the first- and second-order factors are depicted in Figure 1.

This research is significant because it represents the first attempt to systematically examine how various weather and climate terms that are used by meteorologists, climatologists, and the lay public is cognitively organized in the minds of weather information consumers. Further data needs to be gathered so as to make a confirmatory test of the initial factor structure. Data also needs to be gathered from climatologically different locations (and across different seasons) would also add variability to the adjective ratings that may well reveal additional or different latent dimensions. Overall, the underlying linguistic dimensions identified in the study may well correspond to the way people extract, symbolize, and use information provided in forecasts, watches, and warnings. In other respects, the data that the CARS can provide about a locale may be useful to climatologists and geographers who are interested in characterizing the human-experiential dimensions of the climate of that location.