Abstract
Four- step methodology is presented for quantitatively addressing the need to undertake accident- reduction research experimentation. Statistical analysis and sampling requirements are developed first. This is followed by a determination of the minimum accident- rate reduction that would economically justify nationwide deployment of the new traffic control device treatment. The cost- effectiveness of alternative experimental designs is then evaluated. The final step is a trade- off analysis of the value of information to be derived versus the cost of obtaining the information. A case study application of the methodology is also presented.