http://hdl.handle.net/11089/12185 2026-04-04T17:48:39Z 2026-04-04T17:48:39Z Bednarski, Tadeusz http://hdl.handle.net/11089/14879 2018-02-01T11:19:59Z 2014-01-01T00:00:00Z

ON NONRESPONSE CAUSALITY TESTING IN ROTATING PANEL DESIGNS UNDER THE COX MODEL Bednarski, Tadeusz High survey nonresponse in unemployment duration studies may have a strong effect on inference if  exit from unemployment affects the chance of nonresponse (nonresponse causality). In rotational studies  large part of the nonresponse results from panel attritions. A method to test the  presence of the causality mechanism for rotating panel designs is proposed and its asymptotic consistency is proved under the Cox regression model. An application to real labor data and a simulation study are shown.

2014-01-01T00:00:00Z Kończak, Grzegorz http://hdl.handle.net/11089/14880 2018-02-01T11:19:59Z 2014-01-01T00:00:00Z

ON THE RESAMPLING METHOD IN SAMPLE MEDIAN ESTIMATION Kończak, Grzegorz Bootstrap is one of the resampling statistical methods. This method was proposed by B. Efron. The main idea of bootstrap is to treat the original sample of values as a stand-in for the population and to resample with replacement from it repeatedly. Bootstrap allows estimation of the sampling distribution of almost any statistics using only very simple methods. This paper presents a modification of a resampling procedure based on bootstrap sampling. The proposal leads to sampling from population with density function f(x), where f(x) is estimated based on the kernel estimation. The properties of the method were analyzed in the median estimation in Monte Carlo study.The proposal could be useful for the parameters estimation in the case of a small sample. This method could be used in quality control procedures such as control charts or in the acceptance sampling.

2014-01-01T00:00:00Z Pekasiewicz, Dorota http://hdl.handle.net/11089/14882 2018-02-01T11:19:59Z 2014-01-01T00:00:00Z

PROPERTIES OF SELECTED SIGNIFICANCE TESTS FOR EXTREME VALUE INDEX Pekasiewicz, Dorota The paper presents two tests verifying the hypothesis about the shape parameter of the generalized distribution of maximum statistic. It is called the extreme value index. The inverse of the positive index is called  the tail index and determines the degree of fatness of the tail. The asymptotic properties of the Pickands and the Hill estimator of the shape parameter are used to construct the test statistics. Simulation studies of the properties of these significance tests allow us to formulate some conclusions regarding their applications.

2014-01-01T00:00:00Z Chmielińska, Magdalena http://hdl.handle.net/11089/14881 2018-02-01T11:19:58Z 2014-01-01T00:00:00Z

Using the Stable Distribution to Monitor a Process, which Distribution is Unknown Chmielińska, Magdalena The control chart is a tool of statistical quality control, which is widely used in production. The fulfillment of its basic assumptions, guarantees flawless assessment of correctness of the monitored process. The purpose of this paper is to pay attention to the need to verify the assumptions of the used method and the effects of its unauthorized use, in case of not meeting its assumptions. In this paper a method that uses a family of stable distributions to estimate the unknown probability density of monitored diagnostic variable, is proposed. The estimated density function is the basis for determining the control limits.

2014-01-01T00:00:00Z