• Proceedings of the Korean Statistical Society Conference
• /
• 2000.11a
• /
• pp.241-246
• /
• 2000

A simple random sample is taken from a population and a particular survey item is subject to nonresponse that corresponds to random subsampling of the sampled values within adjustment cells. Our object is to estimate Bayesian probability interval of the population mean.

• Proceedings of the Korean Association for Survey Research Conference
• /
• 2001.06a
• /
• pp.57-70
• /
• 2001

In this paper we consider a various estimation methods including the post-stratification estimation, regression estimation and calibration estimation or a generalized raking estimation under a unit nonresponse. All of them have a common type of calibration estimation based on the post-stratification for a categorical auxiliary variables.

• Communications for Statistical Applications and Methods
• /
• v.16 no.1
• /
• pp.103-113
• /
• 2009

Weighting is a common form of unit nonresponse adjustment in sample surveys where entire questionnaires are missing due to noncontact or refusal to participate. A common approach computes the response weight as the inverse of the response rate within adjustment cells based on covariate information. In this paper, we consider the efficiency and robustness of nonresponse weight adjustment bated on the response propensity and predictive mean. In the simulation study based on 2000 Fishry Census in Korea, the root mean squared errors for assessing the various ways of forming nonresponse adjustment cell s are investigated. The simulation result suggest that the most important feature of variables for inclusion in weighting adjustment is that they are predictive of survey outcomes. Though useful, prediction of the propensity to response is a secondary. Also the result suggest that adjustment cells based on joint classification by the response propensity and predictor of the outcomes is productive.

• Clinical and Experimental Pediatrics
• /
• v.56 no.2
• /
• pp.75-79
• /
• 2013

Purpose: It has been reported that 10% to 20% of children with Kawasaki disease (KD) will not respond to intravenous immunoglobulin (IVIG) treatment. In this study, we aimed to identify useful predictors of therapeutic failure in children with KD. Methods: We examined 309 children diagnosed with KD at the Kyungpook National University Hospital and the Inje University Busan Paik Hospital between January 2005 and June 2011. We retrospectively reviewed their medical records and analyzed multiple parameters in responders and nonresponders to IVIG. Results: Among the 309 children, 30 (9.7%) did not respond to IVIG. They had significantly higher proportion of neutrophils, and higher levels of aspartate aminotransferase, alanine aminotransferase (ALT), total bilirubin, and N-terminal fragment of B-type natriuretic peptide than did responders. IVIG-nonresponders had a significantly longer duration of hospitalization, and more frequently experienced coronary artery lesion, and sterile pyuria. No differences in the duration of fever at initial treatment or, clinical features were noted. Conclusion: Two independent predictors (ALT${\geq}$84 IU/L, total bilirubin${\geq}$0.9 mg/dL) for nonresponse were confirmed through multivariate logistic regression analysis. Thus elevated ALT and total bilirubin levels might be useful in predicting nonresponse to IVIG therapy in children with KD.

• Proceedings of the Korean Association for Survey Research Conference
• /
• 2000.06a
• /
• pp.1-14
• /
• 2000

We consider statistical methods for nonresponse problem in social and economic sample surveys. To create a complete data set, which does not include item nonresponse data, imputation methods are generally used. In this paper, we introduce some imputation methods and compare them with one another. Also, we consider some problems, which occur when an imputed data set is treated as a response data set. Due to the imputed values, the true variance of the estimator after imputation is increased by the imputation variance. However, since usual naive variance estimator constructed from the imputed data set does not estimate the imputation variance, the true variance of the estimator after imputation tends to be underestimated. Theoretical reason is investigated and serious results are explained through a simulation study. Finally, some adjusted variance estimation methods to compensate for underestimation are presented and discussed.

• Survey Research
• /
• v.10 no.1
• /
• pp.169-186
• /
• 2009

The propensity score method is used to minimize the bias level in social survey, which comes from nonresponse. The theoretical concept and the background of the propensity score method is discussed first. The propensity score method was first applied in the epidemiology observational study. I have summarized the process of the three propensity score methods that were used to reduce estimation bias in this study. Matching by propensity score is applied to the relatively large control group. Subclassification has the advantage of using whole control group data and regression adjustment is applied to multiple covariates as well as propensity score of each unit is computable and usable. Lastly, the application procedures of propensity score method to reduce the nonresponse bias is suggested and its applicability to real situation is reviewed with the existing data.

• Survey Research
• /
• v.1 no.2
• /
• pp.1-14
• /
• 2000

We consider statistical methods for nonresponse problem in social and economic sample surveys. To create a complete data set, which does not include item nonresponse data, imputation methods are generally used. In this paper, we introduce some imputation methods and compare them with one another. Also, we consider some problems, which occur when an imputed data set is treated as a response data set. Due to the imputed values, the true variance of the estimator after imputation is increased by the imputation variance. However, since usual naive variance estimator constructed from the imputed data set does not estimate the imputation variance, the true variance of the estimator after imputation tends to be underestimated. Theoretical reason is investigated and serious results are explained through a simulation study. Finally, some adjusted variance estimation methods to compensate for underestimation are presented and discussed.

• Pediatric Infection and Vaccine
• /
• v.21 no.1
• /
• pp.29-36
• /
• 2014

Purpose: The objective of this study was to find the predictors and generate a prediction scoring model of nonresponse to intravenous immunoglobulin in patients with Kawasaki disease. Methods: We examined 573 children diagnosed with KD at the Severance Children's Hospital between January 2009 and december 2012. We retrospectively reviewed their medical records. These patients were divided into 2 groups; the experimental group (N=433) and the validation group (N=140). Each group were divided into 2 groups the intravenous immunoglobulin nonresponders and the responders. Multivariate logistic regression analysis identified predictive factors of intravenous immunoglobulin nonresponders which make predictive scoring model. We practice internal validation and external validation. Results: Multivariate logistic regression analysis identified male, cervical lymphadenopathy, changes of the extremities, platelet, total bilirubin, alkaline phophatase, lactate dehydrogenase, C-reactive protein as significant predictors for nonresponse to intravenous immunoglobulin. We generated prediction score assigning 1 point for (1) male, (2) cervical lymphadenopathy, (3) changes of the extremities, (4) platelet (${\leq}368,000/mm^3$), (5) total bilirubin (${\geq}0.4mg/dL$), (6) alkaline phophatase (${\geq}227IU/L$), (7) lactate dehydrogenase (${\geq}268IU/L$), (8) C-reactive protein (>77.1 mg/dL). Using a cut-off point of 4 and more with this prediction score, we could identify the intravenous immunoglobulin nonresponder group. Sensitivity and specificity were 52.5% and 82.4% in experimental group and 37.8% and 81.8% in validation group, respectively. Conclusion: Our predictive scoring models had high specificity and low sensitivity in Korean patients. Therefore it is useful in predicting nonresponse to intravenous immunoglobulin with Kawasaki disease.

• Journal of the Korean Statistical Society
• /
• v.8 no.1
• /
• pp.3-14
• /
• 1979

When we estimate the population proportion of the individuals in the population for the attribute or the characteristic, we consider the sample survey. We can consider many methods of the sample survey, as mail questionnaire, visits, personal calls, etc. When we have the list of units in the population, we usually make use of the mail questionnaire. It is economical and free from the investigator's effect on the respondent, but it has some objections. The principal objection is that it involves a large nonresponse rate that might cause a singificant bias in the result. The bias arises from the different in the characteristics under investigation between those who respond and those who do not respond.

• The Korean Journal of Applied Statistics
• /
• v.16 no.1
• /
• pp.33-44
• /
• 2003

Typically, it should be use auxiliary variable for calibrating the survey nonreponse in census or sampling survey. Where, if the dimension of auxiliary information is large, then it nay be spend a lot of computing time, and difficult to handle data set. Also because the variance estimator depends on the dimension of auxiliary variables, the variance estimator becomes underestimator. To deal with this problem, we propose the variable selection methods for calibration estimation procedure in unit nonreponse situation and we compare the efficiency by simulation study.