An Application of Total Survey Error Estimation in a Small Scale Survey

Survey is a method of gathering information from a sample of units or all of the population. In sample survey, sampling distributions of estimators are affected by errors. These errors may push estimate in a specific direction or inflate its variation. ‎Difference between the estimated and the parameter that is intended to estimate‎, ‎is called survey error‎. Total ‎survey error ‎(TSE)‎ refers to totality of errors that can arise in the design‎, ‎collection‎, ‎processing‎, ‎and analysis of survey data‎. ‎It includes sampling error and nonsampling error. Components of nonsampling error are specification‎, ‎coverage‎, ‎measurement‎, ‎nonresponse‎, ‎and processing errors.‎An estimate of TSE is useful to compare the accuracy of data from alternative modes of data collection or estimation methods‎, ‎and to optimize the allocation of resources for the survey design‎. ‎The most common metric for quantifying TSE is the mean squared error (MSE)‎. If the gold standard measurements are available on all sample units‎, ‎an approximate estimate of MSE can be used to quantify TSE‎. ‎The preferred approach is to decompose the TSE into components associated with the various sources of errors in surveys.‎ Then total bias error and variance components can be estimated. In this article, data from a small scale survey, which has been conducted in a faculty of university in Tehran, is used to estimate TSE.