In shake table experiments of structures, a lot of accelerometers are generally used to identify structural response while only a few displacement transducers are installed because large stiff frames are generally needed to measure displacement. Theoretically, displacement can be calculated from integrations of acceleration. However, low frequency components of measurement errors, noise and drift components resulted from permanent drift are amplified and unrealistic large displacement can be observed. To reduce low frequency errors, low-cut-filters are empirically used. This report discusses about errors of measured displacement and evaluates accuracy of the displacement resulted from integration of measured acceleration.