• The Korean Journal of Applied Statistics
• /
• v.25 no.6
• /
• pp.933-945
• /
• 2012

In this paper we used the KLIPS(Korean Labor and Income Panel tudy) data that surveyed from 2006(wave 9) to 2009(wave 12). Other previous studies are concerned with the panel attrition in the early wave, but this study classifies the response pattern and investigates some factors that influence panel attrition when the panel tends to stabilize. It was revealed that panel attrition was influenced by relocation and housing type through the logit model. Besides it was appeared that panel attrition was affected by the monthly living expenses and the overall household income through the decision tree.

• Journal of the Korean Data and Information Science Society
• /
• v.22 no.1
• /
• pp.1-8
• /
• 2011

In panel studies in which the same respondents are interviewed repeatedly over the long term, panel attrition may cause the problems in the reliability of the result and the representativeness of the sample in panel study. In this article, we explore the risk factors of sample attrition in the first 11 waves of the Korean Labor and Income Panel Study (KLIPS) data covering the years 1998-2008, for which the survival analysis techniques such as life-table method and Cox proportional hazard model based on the time to the attrition of each respondent as the survival time of the respondent are applied.

• The Korean Journal of Applied Statistics
• /
• v.22 no.5
• /
• pp.981-993
• /
• 2009

It would cause a serious problem in the panel data when panel attrition is concentrated on certain socioeconomic groups. Using the GOMS, this study investigates whether there exists non-random attrition bias in the data and seeks for feasible solutions to minimize the bias. The results of logit analyses show that panel attrition in the GOMS results mainly from surveying system but not from the surveyed. Therefore, the result suggests to develop well-organized management skill and systems as well as to construct weighting methods.

• Asia-Pacific Journal of Business
• /
• v.11 no.4
• /
• pp.291-301
• /
• 2020

Purpose - The purpose of this study was to investigate several household characteristics related to panel attrition, examining how they may have conditioned the panel data in the Korea Welfare Panel Study (KOWEPS). Design/methodology/approach - We studied the cause of the differences in household income between the original panel and the new panel in KOWEPS. Findings - To summarize our findings, whereas it is highly likely that a low-income household or a household without health insurance will remain in the panel, it is highly likely that a high-income household or a household of more than three members will be taken off the panel. Research implications or Originality - The proportion of low-income household tends to decrease over the years, which appears to result from an overall increase in household income. Such changes are reflected in the pattern in which older panels have higher estimates of household income than newer panels.

• Survey Research
• /
• v.10 no.1
• /
• pp.73-89
• /
• 2009

Even though panel surveys are very useful in estimating the change between time points, they suffer from sample distortion as survey rounds proceed due to panel attrition and conditioning. This study is to report the statistical aspects of KBS-MBMR's five-rounds panel survey for the 2007 Presidential Election of Korea. Main findings are as follows. First, the attrition propensities were higher among women, among the young and the old aged, and among "None"/"Don't Know"/"Won't Say"(DK) respondents for the question asking preferred candidate. Second, there existed the conditioning effect that pushes the respondents to choose one candidate in hurry. Third. repeated measurements of the survey did not influence panel respondents on the choice of preferred candidate. if DK respondents are set. aside, and on the final voting intention of panel respondents.

• Journal of the Korean Statistical Society
• /
• v.31 no.4
• /
• pp.469-482
• /
• 2002

This paper suggests a new method of compensating for wave nonresponse in panel survey, which combines weighting adjustment and imputation. By deleting less frequent nonresponse patterns, we can get simplicity. A new mean estimator under the new combining method is provided and a limited simulation study employing a real data is conducted.

• Journal of Korean Society of Transportation
• /
• v.18 no.6
• /
• pp.63-75
• /
• 2000

Stated Preference (SP) data has been regarded as more useful than Revealed Preference (RP) data, because researchers can investigate the respondents\` Preference and attitude for a traffic condition or a new traffic system by using the SP data. However, the SP data has two bias: the first one is the bias inherent in SP data and the latter one is the attrition bias in SP panel data. If the biases do not corrected, the choice model using SP data may predict a erroneous future demand. In this Paper, six route choice models are constructed to deal with the SP biases, and. these six models are classified into cross-sectional models (model I∼IH) and dynamic models (model IV∼VI) From the six models. some remarkable results are obtained. The cross-sectional model that incorporate RP choice results of responders with SP cross-sectional model can correct the biases inherent in SP data, and also the dynamic models can consider the temporal variations of the effectiveness of state dependence in SP responses by assuming a simple exponential function of the state dependence. WESML method that use the estimated attrition probability is also adopted to correct the attrition bias in SP Panel data. The results can be contributed to the dynamic modeling of SP Panel data and also useful to predict more exact demand.

• The Korean Journal of Applied Statistics
• /
• v.32 no.3
• /
• pp.375-389
• /
• 2019

Sample attrition according to a long-term tracking reduces the representativeness of the sample data in a panel study. Most panel surveys in South Korea and other countries have prepared response adjustment weights in order to solve problems regarding representativeness due to sample attrition. In this paper, we divided the panel data into continuous response group and non-continuous response group according to response patterns and considered a weighting adjustment method to reduce the bias of the non-continuous response group. A simulation indicated that the proposed composite estimation type weighting method, which reflected the characteristics of non-continuous response groups, could be more efficient than other weighting methods in terms of reducing non-response bias. As a case study, the proposed methods are applied to the Korean Longitudinal Study of Ageing (KLoSA) data of the Korea Employment Information Service.

• Korean Journal of Metals and Materials
• /
• v.48 no.9
• /
• pp.831-841
• /
• 2010

Indium tin oxide (ITO) nanopowders were synthesized by coprecipitation and the sol-gel method to prepare a stable dispersion of ITO nano-colloid for antistatic coating of a display panel. The colloidal dispersions were prepared by attrition process with a vibratory milling apparatus using a suitable dispersant in organic solvent. The ITO coating solution was spin-coated on a glass panel followed by the deposition of partially hydrolyzed alkyl silicate as an over-coat layer. The double-layered coating films were characterized by measuring the sheet resistance and reflectance spectrum for antistatic and antireflective properties.

• Korea journal of population studies
• /
• v.29 no.3
• /
• pp.159-188
• /
• 2006

This study attempts to create a longitudinal dataset by linking tdata on the identical individuals across the monthly sample household management lists of the Economically Active Population Survey(EAPS) and the Household Income and Expenditure Survey(HIES). Using the data constructed through such process, the study also tryies to analyze the duration of longitudinal responses and the characteristics of nonrespondents. Between 1998 and 2002, longitudinal response rates had declined to 46% of total EAPS and 34% of total HIES. The fact that nonresponse was not a random phenomenon leads to concerns about the representativeness of the remaining sample. Using Cox's proportional hazard model the study revealed that the duration of longitudinal responses is affected by the ownership of house and the age of the respondent.