Dealing with nonresponse is a very important topic, since nonresponse is present almost in all surveys, and can cause biased estimation. Nonresponse is dened as the failure to provide the required information by a unit selected in a sample. We distinguish between unit nonresponse and item nonresponse. Unit nonresponse implies that we have no information at all from the sampled unit. Item nonresponse means that the sampled unit does not ll some of the survey items. Each unit selected in the sample has associated a sampling weight and a response probability to answer the questionnaire. The response probability is unknown and should be estimated. The main method to deal with the unit nonresponse is to use reweighting. This method adjusts the initial sampling weights by the inverse of the estimated response probabilities, providing new weights. We focus on unit nonresponse adjustment in survey data and estimate the response probabilities using an item response model called the Rasch model. This model uses a latent parameter. We believe that this latent parameter can explain a part of the unknown behavior
of a unit to respond in the survey. No information about the nonrespondents and no auxiliary information are required in the proposed method. Theoretical aspects and simulations are used to support our theory.