The nonindustrial private forest sector controls 58 percent of the nation’s commercial forest area. Timber output from the NIPF has been a major forest policy issue throughout this century. A long-run regional timber output projection model was developed for the NIPF sector. The model was based upon two key assumptions: (1) forest management intensity and timber output are inversely related to the forest landowner’s alternative rate of return, and (2) for the NIPF sector, the alternative rate of return is inversely related to the size of forest holding. The timber output model was based on a cross-classification of NIPF land by its physical characteristics (e.g., forest type) and economic characteristics (e.g., landowner’s alternative rate of return). The stratification objective was to create strata with a uniform timber output response to changes in the above characteristics. If the total forest resource is thought of in terms of a matrix, each characteristic serves as a stratification dimension, creating a matrix composed of “homogeneous timber output response cells.” The values of each cell can be used as observations in a regression analysis projecting NIPF timber output. For the NIPF, the size of forest holding would be a primary stratification dimension. Due to the relationships described above, stratification by size of forest holding constitutes stratification by the landowner’s alternative rate of return. Also, since the forest resource is physically uniform, the expected rate of return (marginal value growth percent) of the forest will be uniform within a cell. NIPF landowners base their timber output decision on the relationship between the expected and alternative rates of return, resulting in a convenient forest policy analysis tool. The model was applied to a study region, Virginia’s Coastal Plain, with satisfactory results. / Ph. D.