MITRE compiled metal loss data and relevant design/operating information for sixty-six alloys in twenty-nine FBC systems. After all of the data were assembled and reviewed, twelve alloys were selected for statistical analysis. Regression methodology was used to investigate the relationship between metal loss and design, operating, and performance parameters of FBC systems for a given alloy. Both inferential and exploratory methods were applied to analyze the data. Based on results of these analyses, the following general conclusions can be drawn: Exposure time, specimen temperature, bed temperature, and coal and sorbent compositions appear to be related to metal loss for most of the alloys, as indicated by the univariate analyses. With the exception of specimen temperature, a high degree of correlation is not reflected in individual variable plots (scatterplots). The regression analyses resulted in twelve models - one for each alloy. The diagnostic analyses of all twelve models indicate that the models are appropriate for the data. While the models differ in terms of the types of variables in them, there are a number of similarities. Based on the study results and data quality, MITRE concludes that a statistically-designed test program is needed before the effects of identified important variables (such as coal and sorbent compositions) can be confirmed and further quantified. 98 refs.