I am working on determining cell viability following a resazurin-based toxicity assay. Some of the wells on the 96 well plate contained media, but no cells which I am using as a control. I have averaged out the fluorescence values for these wells and have subtracted this value from all of the other values to account for background signal. However, in doing so, I have some wells with negative values. I am able to calculate % cell viability per treatment, but the statistics are thrown of as a result of these negative values. How do I account for this?