• Survey Research
• /
• v.9 no.3
• /
• pp.71-91
• /
• 2008

Workplace Panel Survey(WPS) is the representative panel survey of workplace in Korea. WPS was newly sampled in 2005 and is to be used for the subsequent biennial survey. The main survey is divided into a questionnaire for human resources(HR) manager, a questionnaire for labor relations manager and a questionnaire for representatives of unions. The population of WPS 2005 included workplaces across the country with 30 or more employees. The WPS 2005 was composed of 1,905 workplaces including 290 workplaces in the public sector. The sample was selected by the stratified random sampling. Weighting process for the survey data was introduced to compensate for differential sampling and non-response rates. Personal interviews were conducted using the Computer Assisted Personal Interviewing(CAPI) system during visits by interviewers, along with survey via mail and e-mail concerning employment and financial issues. The CPAI system introduced for the WPS 2005 can by used for automatical detection for errors and inconsistencies which may occur during the survey process. The CAPI system played an important part in enhancing the reliability of the survey data.

• The Korean Journal of Applied Statistics
• /
• v.22 no.6
• /
• pp.1315-1329
• /
• 2009

The panel survey is need to have a more concern about a response due to a secession and non-response of a sample. And generally a population is not fixed and continuously changed. Thus, the rotation sample design can be used by the method replacing the panel research. This paper is the study of comparison to equal weight method, Duncan weight, Design weight method, weight share method in rotation sample design. More specifically, this paper compared variance estimators about the existing each method for the efficiency comparison, and to compare the precision using the relative efficiency gain by the Coefficient Variance(CV) after getting the design weight from the actual data.

• Journal of the Korean Data and Information Science Society
• /
• v.21 no.6
• /
• pp.1021-1029
• /
• 2010

In 2009 the former three surveys on small and medium enterprises were combined into one survey to reduce the response burden of enterprises. In this report, we study a sample design for 2009 survey on actual state of small and medium enterprises (SMEs). However, the differences between the manufacturing industries and knowledge-based service industries are so large that we need to consider separate populations in the survey. The total sample size is decided as 10,000 in new survey design for integration. We allocate the sample sizes for the first stratum based on CV and then allocate the sample sizes for the second stratum using root proportional formula. Also we calculate survey weights and propose the formula for the estimators as well as standard errors using weights for each characteristic.

• The Korean Journal of Applied Statistics
• /
• v.11 no.2
• /
• pp.461-475
• /
• 1998

In panel survey in which the sample is selected by stratified random sampling, if the sampling units shift from a stratum to others in time, then the movement should be incorporated in the estimation procedures. Dealing with the problem caused by the movement of units across stratum in the updated stratification sample, the bias of the conventional estimator neglecting the movement is investigated, arid the bias-adjusted estimators are proposed. The variance estimator of the suggested estimators is also derived. It is illustrated via a simulation study that the proposed estimators beat the conventional estimator in the sense of bias and mean squared error In particular, when the Neyman allocation is applied in stratified sampling, the proposed estimator is shown much more effective to this end.

• Proceedings of the Korean Association for Survey Research Conference
• /
• 2003.06a
• /
• pp.19-28
• /
• 2003

Weighting adjustment is a method of improving the efficiency of the estimator by incorporating auxiliary variables at the estimation stage. One commonly used method of weighting adjustment is the poststratification, which is a special case of regression estimation but is relatively feasible in terms of actual implementation. If too many auxiliary variables are used in the poststratification, the bias of the resulting point estimator is no longer negligible and the final weights may have extreme weights. In this study, we propose a method of weight ing adjustment that compromises the efficiency and the bias of the point estimator. A limited simulation study is also presented.

• Proceedings of the Korean Association for Survey Research Conference
• /
• 2001.11a
• /
• pp.77-85
• /
• 2001

In this paper we study the calibration estimator and its variance estimator for the population total using a bootstrap method according to the levels of an auxiliary information having strong correlation with an interested variable in nonresponse situation. At this point, we find tire calibration estimator in case of auxiliary information for population and sample, and then we drive the bootstrap variance estimator of it. By simulation study we compare the efficiencies with the Taylor and Jackknife variance estimators.

• The Korean Journal of Applied Statistics
• /
• v.31 no.5
• /
• pp.621-636
• /
• 2018

Research on the application of informative sampling technique has been conducted in order to reduce the influence of non-response. Chung and Shin (Korean Journal of Applied Statistics, 30, 993-1004, 2017) showed that the estimation accuracy improved when using exponential response rate information for the parameter estimation if the distribution of errors included in the super population model follows normal distribution. However this method divides the stratum into equally spaced substrata to obtain the sample weight of the informative sampling technique and shows that the accuracy of the estimation improves as the number of substrata increases. In this study, with the given number of total sample size, the optimal substratum boundary points are calculated using equal space, quantile, and LH algorithm; consequently, the results using those methods are compared through simulation. We also studied the criteria to determine the number of substrata and substratum boundaries that can be used in practice with various types of auxiliary variable distributions.

• Journal of Digital Convergence
• /
• v.15 no.10
• /
• pp.559-570
• /
• 2017

This study has the objective of finding out the impact of veblen effect and bandwagon effect on the choice behavior of hair salon. A structured survey questionnaire was used targeting 1,000 women over the age of 20. For the sample, proportionate stratified sampling method was utilized for the generalization of the findings. For analysis, frequency analysis, factor analysis, reliability analysis and regression analysis were performed by utilizing SPSS ver.19.0. Two hypotheses were set for the achievement of the research purpose. Hypothesis 1. Veblen effect will affect the hair salon choices. Hypothesis 2. Bandwagon effect will affect the hair salon choices. It was verified that the veblen effect and bandwagon effect greatly affects the patterns in the choice of hair salons. It appeared that the research hypotheses of 'Veblen effect will affect the hair salon choices.' and 'Bandwagon effect will affect the hair salon choices' could be partly adopted. The results of the study show that most of the users choose the hair salon with their friends, acquaintances or people they interact with.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.5
• /
• pp.3276-3283
• /
• 2015

The purpose of this study was to ascertain whether some socio-economic characteristics affect the service users' perception on the change of social condition in healthcare services. The research target for this study was fixed on the sample members in national sample design and the data of this study was used 37,648 effective samples collecting using Probability Proportional to Size. The results of this study using the causal relationships model are as follows: The variables including type of female, age, years of schooling show positive signs on the service users' perception on the social condition change of healthcare. And the variables including household income, temporary worker, casual worker, urban residents show negative signs on the service users' perception on the social condition change of healthcare. In conclusion, the socio-economic characteristics affect the service users' perception on the social condition change of healthcare. And the service user's perception on the social condition change of healthcare was affected positively on the beneficiary side.

• Survey Research
• /
• v.7 no.2
• /
• pp.53-69
• /
• 2006

Weights can be made and imposed in both sample design stage and analysis stage in a sample survey. While in design stage weights are related with sample data acquisition quantities such as sample selection probability and response rate, in analysis stage weights are connected with external quantities, for instance population quantities and some auxiliary information. The final weight is the product of all weights in both stage. In the present paper, we focus on the weight in analysis stage and investigate the effect of such weights imposed on the weighted mean when estimating the population mean. We consider a finite population with a pair of fixed survey value and weight in each unit, and suppose equal selection probability designs. Under the condition we derive the formulas of the bias as well as mean square error of the weighted mean and show that the weighted mean is biased and the direction and amount of the bias can be explained by the correlation between survey variate and weight: if the correlation coefficient is positive, then the weighted mein over-estimates the population mean, on the other hand, if negative, then under-estimates. Also the magnitude of bias is getting larger when the correlation coefficient is getting greater. In addition to theoretical derivation about the weighted mean, we conduct a simulation study to show quantities of the bias and mean square errors numerically. In the simulation, nine weights having correlation coefficient with survey variate from -0.2 to 0.6 are generated and four sample sizes from 100 to 400 are considered and then biases and mean square errors are calculated in each case. As a result, in the case or 400 sample size and 0.55 correlation coefficient, the amount or squared bias of the weighted mean occupies up to 82% among mean square error, which says the weighted mean might be biased very seriously in some cases.