Browsing by Subject "Survival analysis (Biometry)"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
Item A survival analysis approach to employee turnover: its application and advantages(Texas Tech University, 1998-08) Walker, Kimberly DeannCurrent investigations of employee tumover have focused on the ability of researchers to predict tumover through measuring variables such as employee and job satisfaction, organizational commitment, and job involvement. The analysis technique typically used in these studies, logistic regression, does not incorporate time as a variable of interest and also may not be appropriate for use in studies where the dependent variable is binary. Survival analysis, on the other hand, allows time to be included as a variable of interest and is specifically designed for use with a binary dependent variable such as employee tumover. Using employee information from three separate corporations, this study demonstrates the benefits of applying survival analysis techniques to tumover data and examines the differences in the type of information gained from survival analysis versus logistic regression. Results showed that, unlike logistic regression, survival analysis is capable of indicating differences between active employees, early leavers(those with a longevity of zero to four years), middle leavers (those with a longevity of four to eight years), and later leavers (those with a longevity of more than eight years) based on predictor variables collected from employee records from each company. Corporation One logistic regression results indicated that salary, payrate, gender, age, absences, and average number of previous jobs were significant predictors of whether hourly employees remained with the company or left. In contrast, the discriminant function analysis based on the survival groups for that sample showed that active employees or late leavers were distinguished from early and middle leavers on the variables of payrate, salary, and absences for canonical one. Further, canonical two and the corresponding class means showed that middle and late leavers were separated from early leavers and active employees based on absences, number of partial days worked, and employee title. Results for corporations two and three showed a similar pattem of results. All results showed that logistic regression only indicates the relationship of predictor variables to the occurrence of tumover, while survival analysis produces information on both the timing and occurrence of tumover, as well as the differences between active employees, early leavers, and late leavers on selected predictor variables.Item Covariance estimation based on asymptotic normal estimating function(Texas Tech University, 1998-12) Hwang, Ming-WoeiSurvival analysis is the analysis of data that corresponds to the time from a welldefined origin of time until the occurrence of some particular event or end point. In medical research, the origin of fime could be the time of birth, or the recruitment of an individual into an experiment study, such as clinical trial to compare two or more treatments. If the end point is the death of a patient, the resulting data are literally survival times. The two-sample accelerated life model has been a general model for comparing censored survival times from two groups. This model assumes that the survival time of an individual in one group is distributed as a multiple of the survival time of an individual in the other group. Therefore, the probability of an individual on the new treatment surviving over time is the probability of an individual under the standard treatment surviving over time 01, where ^ is an unknown positive scale.Item On minimum distance estimation in two-sample scale problem with right censoring(Texas Tech University, 1995-08) Wu, KeThe estimation of the treatment effect in the two-sample problem with right censoring is of interest in survival analysis. In this paper we consider the scale change model in which a specific treatment extends the life of the patient, in the sense that the lifetime is multiplied by a certain scale factor. For the Koziol-Green (K-G) sub-model of the right censoring model, this research establishes large-sample properties of the minimum Z/2-distance estimator of the scale parameter, similar to that studied in Koul and Yang (1989) for the general right censoring model. The maximum likelihood estimator (MLE) for the survival function under the K-G model is used in the minimum L2-distance estimation of the scale parameter rather than the product limit estimator (PLE) of Kaplan and Meier (1958) as used in Koul and Yang (1989). Several properties of the estimator such as consistency and asymptotic normality are established. A representation of the estimator that facilitates its computation is derived. Furthermore, it is shown that the minimum L2-distance estimator using the MLE of the survival function is asymptotically more efficient than that using the PLE of the survival function under the specified K-G sub-model.Item On using the kernel method for functional estimation with current status data(Texas Tech University, 1999-12) Taylor, Scott A.To implement Yang's (1999) kernel estimator method, an important issue is the choice of bandwidth. It is well-known that the performance of kernel estimators depend upon the bandwidth. Furthermore, Yang's (1999) method uses the ratio of two kernel estimators which may be more sensitive to the choice of bandwidth. Our goal is to investigate the choice of bandwidth and evaluate the performance of the corresponding kernel estimators. The NPMLE will also be included in the studies for comparison. By comparing the mean square error of the kernel estimator against the NPMLE of Groeneboom and Wellner (1992), the kernel estimator was found to behave quite well under the given testing conditions. This thesis is organized as follows: The NPMLE method for functional estimation is introduced in Chapter II. Yang's kernel estimator method is introduced in Chapter III. In Chapter IV, we explain the procedures for evaluating the best choice of bandwidth. We analyze and display the results from the computer simulations in Chapter V. Finally, conclusions are drawn in Chapter VI.Item Parametric inference using combined weighted log-rank estimators for regression with survival data(Texas Tech University, 2002-05) Waldrop, Jaclyn MIn this paper, we will look at combining asymptotically normal tests. It is desirable to combine these candidate tests when several are availal)le that have nice properties such as efficiency and robustness. After looking at individual weight functions that perform well for Yang's regression estimator, we determine the combined function for these individual functions. Simulations are constructed that test for empirical coverage probabilities and empirical mean lengths. Analyzing these results allows us to determine if the combined test is beneficial in the study of survival analysis.