This paper follows on from Jarrett, Takacs and Ferry (2011) which reported the first stage in development of a high school level concept inventory (CI) for the science of climate change: the climate change concept inventory (CCCI). In order to develop a reliable and valid instrument, it is necessary to follow appropriate procedures. This paper details the process of CI item development; reports statistical results of initial field trials and outlines how these will be used to further refine the CCCI. Item difficulty, discrimination, and point biserial coefficient were calculated for each item. Cronbach's alpha and test-retest data were used to assess reliability. Results suggest that about half of the items were too difficult for high school students. However, item discrimination and test reliability values were close to acceptable values, which suggests that most students were not simply guessing answers. Although it was initially designed for use in high schools, a group of undergraduates trialled the CI. Statistical analyses of scores suggest that for this group, the items performed better, and well within acceptable values. Given these favourable results and the fact that introductory-level climate change is increasingly taught at universities, further trials with undergraduates are taking place. It is intended that the final CI will be made available as a formative assessment instrument. The current version is available from the authors on request.