The Income and Expenditure Survey (IES) conducted by Statistics South Africa (Stats SA) between September 2005 and August 2006 was the third of its kind, after similar surveys in October 1995 and October 2000. The main purpose of the IES is to collect and provide information on income and expenditure patterns of a representative sample of households, so as to update the basket of goods and services required for the compilation of the Consumer Price Index (CPI). Nonetheless, these surveys have also become an important source of information for poverty and inequality analysis, mainly because of the absence of other detailed datasets containing income and expenditure data. There are, however, important reasons why these datasets cannot be unquestioningly compared. This paper attempts to show why. The IESs conducted in 1995 and 2000 used the recall method. In the recall method, a single questionnaire was administered to a household at a selected dwelling unit in the sample, and the responding household was required to recall income and expenditure either during the month prior to the survey or for the twelve months prior to the survey. However, in the IES conducted in 2005-2006, the diary method was used extensively for the first time in order to record the household’s daily acquisitions on a daily basis. In addition to the adoption of the diary method, the 2005-2006 IES is also different from the previous IESs in many aspects, such as sampling design, questionnaire structure, number of visits to the households, additions of some new expenditure items in the questionnaire, categorization of income and expenditure items, etc., and the focus of this paper is to look at how different the three IESs are, so as to assist researchers and policy makers when they try to analyze the IES data.