National forest assessments (NFA) are best conducted with sufficiently accurate and scientifically defensible estimates of forest attributes. This section discusses the statistical design of the sampling plan for a forest inventory. This includes the process used to define the population to be sampled and the selection of a sample intended to satisfy the NFA precision requirements. An experienced statistician should be a member of any team designing a national forest inventory. However, if such an expert is not available, this section provides guidance and recommendations for relatively simple sample designs that reduce risk and improve chances for success.