Symmetric and asymmetric rounding: A review and some new results

摘要: Using rounded data to estimate moments and regression coefficients typically biases the estimates. We explore bias-inducing effects of rounding, thereby reviewing widely dispersed often half forgotten results in literature. Under appropriate conditions, these can be approximately rectified by versions Sheppard’s correction formula. discuss conditions under which approximations are valid also investigate efficiency loss caused rounding. The rounding error, corresponds measurement error a model, has marginal distribution, approximated uniform but is not independent true value. In order take account preferences (heaping), we generalize concept simple that asymmetric consider its effect on mean variance distribution.