This work presents a preliminary study on the production of solid dispersion using compressed carbon dioxide as antisolvent following similar principle of supercritical carbon dioxide antisolvent precipitation. In our study, liquid carbon dioxide was used as antisolvent to induce co-precipitation of a model ingredient, piroxicam, and excipient, polyethylene glycol, in order to minimize aggregation of final products. A factorial design of experiment was conducted to evaluate the effect of operational variables on the solubility of the resulting products. The influence of several factors, including temperature, pressure, concentration ratio of drug and polymer in methylene chloride, injection rate of the solution, etc., on the solubility of the solid dispersion was studied. Preliminary analysis of the results provides the information of relative importance among the factors which are assumed independent of each other. Such a design of experiment could offer a more efficient way of screening for the optimal operating conditions, which is beneficial for both bench-top studies and possibly subsequent scaling up processes.