Abstract:
Non-parametric bootstrap can be implicated at small size of random sample from unknown distribution when statistical methods are used. The paper deals with the problem, how many non-parametric bootstrap replications have to be done at estimation of bias abd which extent of errors might be caused.