Rapid Rural Appraisal techniques are frequently used for wealth-ranking of households in rural communities. The main justifications of such approaches are that they are fast, allow prestratification of households in preparation for surveys and that they have been shown to have at least some empirical validity as means of stratifying households by socio-economic status. Sometimes, however, the participatory process may fail to produce consistent wealth categories, or people refuse to accept labels such as ‘poor’ and ‘rich’ and/or are reluctant to assign such categories to households within their village. A household survey in Tanzania was conducted in 12 villages across eastern Tanzania in 2008-2010. In all villages a participatory wealth ranking exercise was conducted and used for stratification of households. Sample households (n=40 per village) were allocated proportionally to wealth categories. The structured household survey emphasized assets, income, and perceived effects of Participatory Forest Management. The value of individual non-productive assets was estimated through a follow-up survey and combined with results from the household survey to estimate total value of assets for each household. In this study the original, partly inconsistent wealth ranking is examined and compared with estimated incomes and asset values. Next, alternative wealth classification rules are introduced and resulting classification outcomes are examined and discussed.