Solar technology has advanced substantially in recent years, and policymakers interested in reducing climate change hope these improvements will continue. Nevertheless, only one measure, patent counts, has been widely used to identify the rate of improvement, and scant research has assessed the effectiveness of efforts to increase it. In this paper, the number of articles published on solar energy during each month in 1986-2009 is calculated to create a more detailed supplement to patent counts, and these article counts are used to assess certain relevant subsidies. A combination of counting articles manually and naive Bayesian classification is used to overcome the difficulties of identifying which articles are relevant. The resulting monthly article counts are modeled as a function of two major subsidy types. The largest U.S. subsidy for renewable energy, the Production Tax Credit, appears to be as much as half as effective at encouraging solar technology research as federal solar research subsidies, despite the fact that the tax credit is given to electricity producers rather than to researchers and goes mostly to non-solar energy sources. This finding may be considered a preliminary example of the use of monthly solar article counts.