Rank data occurs quite regularly in the context of market research studies, opinion polls, sports, etc. In this paper, we discuss the problem of estimation of the true rank when the rank given by the respondents are subject to error. Various error structures are discussed like, transposition errors, cyclic errors etc. and methods to analyze data with these error structures are derived. Analysis of data with a completely general error structure is also discussed. It is seen that in all the cases considered, the posterior marginal distribution of the true rank can be conveniently obtained using Sampling Importance Resampling (SIR) technique. A real life rank data set is analysed to illustrate the methodology. Note: For softcopy of this paper, please contact the authors - email: arnab@iima.ac.in