Nonprot organizations have become increasingly engaged in disseminating information. Unlike federal statistics that are collected with rigidly proscribed procedures, nonprots have an advocacy role that may in uence the information they release and the interpretations they derive from that data. While federal statistics are released to inform opinion, nonprot statistics are released to in uence opinion. Using as cases the Feeding America county food insecurity statistics, the United Way ALICE estimates of working poor, and the MIT Living Wage Calculator, we show how the desire to in uence policy can lead nonprots to release data characterized by awed methodology and misleading or implausible values. Because errors or biases in federal data have been widely researched and publicized through peer review, users can adjust to the known noise in the data. Flaws in data reported by the nonprots are not made transparent, and so users will make bad decisions if they assume the information is correct.