This paper reports on the evaluation of local usage statistics of a specific set of chemistry journals at the University of Denver in Colorado, USA. The objective of the study is to demonstrate that commercial publishers in chemistry charge considerably more for their journals than those from the non-commercial sector. There are three variables that are considered in this research: institutional cost of the journal, cost per Thomson Scientific Impact Factor (IF), and cost per article downloaded. In the defined set of 65 journals, it is found that commercial publishers charge more for their journals by a ratio of 2.59. Commercial publishers cost more per impact factor by a ratio of 3.20 and significantly more per local article download by a ratio of 6.16. This research demonstrates that local usage statistics should be used as part of the journal evaluation process. [ABSTRACT FROM AUTHOR]