• The Korean Journal of Applied Statistics
• /
• v.22 no.2
• /
• pp.341-353
• /
• 2009

We consider sample-size determination problem motivated by comparative clinical trials where patient outcomes are characterized by a bivariate outcome of efficacy and safety. Thall and Cheng (1999) presented a sample size methodology for the case of bivariate binary outcomes. We propose a bivariate Wilcoxon-Mann-Whitney(WMW) statistics for sample-size determination for binary outcomes, and this nonparametric method can be equally used to determine sample sizes of ordinal outcomes. The two methods of sample size determination rely on the same testing strategy for the target parameters but differs in the test statistics, an asymptotic bivariate normal statistic of the transformed proportions in Thall and Cheng (1999) and nonparametric bivariate WMW statistic in the other method. Sample sizes are calculated for the two experimental oncology trials, described in Thall and Cheng (1999), and for the first trial example the sample sizes of a bivariate WMW statistic are smaller than those of Thall and Cheng (1999), while for the second trial example the reverse is true.

• The Korean Journal of Applied Statistics
• /
• v.22 no.6
• /
• pp.1249-1263
• /
• 2009

The power function in sample size determination has to be characterized by an appropriate statistical test for the hypothesis of interest. Nonparametric tests are suitable in the analysis of ordinal data or frequency data with ordered categories which appear frequently in the biomedical research literature. In this paper, we study sample size calculation methods for the Wilcoxon-Mann-Whitney test for one- and two-dimensional ordinal outcomes. While the sample size formula for the univariate outcome which is based on the variances of the test statistic under both null and alternative hypothesis perform well, this formula requires additional information on probability estimates that appear in the variance of the test statistic under alternative hypothesis, and the values of these probabilities are generally unknown. We study the advantages and disadvantages of different sample size formulas with simulations. Sample sizes are calculated for the two-dimensional ordinal outcomes of efficacy and safety, for which bivariate Wilcoxon-Mann-Whitney test is appropriate than the multivariate parametric test.

• The Korean Journal of Applied Statistics
• /
• v.20 no.1
• /
• pp.183-194
• /
• 2007

The problem of calculating the sample sizes for comparing two independent binomial proportions is studied, when one of two probabilities or both are smaller than 0.05. The use of Whittemore(1981)'s corrected sample size formula for small response probability, which is derived based oB multiple logistic regression, demonstrates much larger sample sizes compared to those by the asymptotic normal method, which is derived for the comparison of response probabilities belonging to the normal range. Therefore, applied statisticians need to be careful in sample size determination with small response probability to ensure intended power during a planning stage of clinical trials. The results of this study describe that the use of the sample size formula in the textbooks might sometimes be risky.

• The Sea:JOURNAL OF THE KOREAN SOCIETY OF OCEANOGRAPHY
• /
• v.9 no.1
• /
• pp.20-29
• /
• 2004

Macrobenthos species-area relationship was investigated and empirical models were estimated to enable comparisons among species numbers of different sample size. The study aims to choose a way to predict cumulative relative species number (CRSN) in a given sample size Saemangeum, located in the west coast of South Korea, were visited in Apr., May and Aug.,2002 and a total of 261 biological samples from 87 stations were obtained by employing a quantitative sediment sampler, Smith-McIntyre grab and design of 3 replicates at each station. Relative species numbers (%) were baselined at sample size of 1000 $\textrm{cm}^2$ and the patterns of CRSN along the axis of sample size were measured and observed. In correlation analysis performed on a set of abiotic and biotic variables, log-transformed CRSN showed the only significant relationship with log-transformed density. Based on the result, three models, Log CRSN 2000, Log CRSN 3000 and Log CRSN were produced. The former two were devised to predict CRSN at 2000 and 3000 $\textrm{cm}^2$ respectively, and the latter at various sample sizes and samplers (all p-values were <0.001). Database from other studies (intertidal or subtidal macrofaunal samples from Kyonggi Bay and Saemangeum) were used to evaluate validity of the models. Observed CRSN below sample size of 3000 $\textrm{cm}^2$ fell under the range of 95% prediction interval and this was appeared to provide reliability of the models below that sample size.

• Fibers and Polymers
• /
• v.3 no.3
• /
• pp.97-102
• /
• 2002

A series of waterbome poly(urethane-urea) anionomers were prepared from isophorone diisocyanate (IPDI), polycaprolactone diol (PCL), dimethylol propionic acid (DMPA), ethylene diamine (EDA), and triethylamine (TEA), NaOH, or Cu($(COOCH_3)_2$) as neutralizing agent. This study was performed to decide the effect of neutralizing agent type on the particle size viscosity, hydrogen bonding index, adhesive strength, antistaticity, antibacterial and mechanical properties. The particle size of the dispersions decreased in the following order: TEA based samples (T-sample), NaOH based samples (N-sample), and Cu($(COOCH_3)_2$) based sample (C-sample). The viscosity of the dispersions increased in the order of C-sample, N-sample, and T-sample. Metal salt based film samples Of and C-sample) had much higher antistaticity than TEA based sample. By infrared spectroscopy, it was found that the hydrogen bonding index (or fraction) of samples decreased in the order of T-sam-pie, N-sample, and C-sample. The adhesive strength and tensile modulus/strength decreased in the order of T-sample, N-sam-pie, and C-sample. The C-sample had strong antibacterial halo, however, T- and N-samples did not

• Communications for Statistical Applications and Methods
• /
• v.22 no.5
• /
• pp.519-530
• /
• 2015

The geometric chart has proven more effective than Shewhart p or np charts to monitor the proportion nonconforming in high-quality processes. Implementing a geometric chart commonly requires the assumption that the in-control proportion nonconforming is known or accurately estimated. However, accurate parameter estimation is very difficult and may require a larger sample size than that available in practice in high-quality process where the proportion of nonconforming items is very small. Thus, the error in the parameter estimation increases and may lead to deterioration in the performance of the control chart if a sample size is inadequate. We suggest adjusting the control limits in order to improve the performance when a sample size is insufficient to estimate the parameter. We propose a linear function for the adjustment constant, which is a function of the sample size, the number of nonconforming items in a sample, and the false alarm rate. We also compare the performance of the geometric charts without and with adjustment using the expected value of the average run length (ARL) and the standard deviation of the ARL (SDARL).

• Journal of Animal Science and Technology
• /
• v.55 no.2
• /
• pp.103-107
• /
• 2013

A crucial first step in the planning of any scientific experiment is to evaluate an appropriate sample size to permit sufficient statistical power to detect the desired effect. In this study, we investigated the optimal sample size of quantitative trait locus (QTL) linkage analysis for simple random sibship samples in pedigreed chicken populations, under the variance component framework implemented in the genetic power calculator program. Using the program, we could compute the statistical power required to achieve given sample sizes in variance component linkage analysis in random sibship data. For simplicity, an additive model was taken into account. Power calculations were performed to relate sample size to heritability attributable to a QTL. Under the various assumptions, comparative power curves indicated that the power to detect QTL with the variance component method is highly affected by a function of the effect size of QTL. Hence, more power can be achievable for QTL with a larger effect. In addition, a marked improvement in power could be obtained by increasing the sibship size. Thus, the use of chickens is advantageous regarding the sampling unit issue, since desirable sibship size can be easily obtained compared with other domestic species.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.42 no.1
• /
• pp.151-157
• /
• 2019

This study is about the process capability index (PCI). In this study, we introduce several indices including the index $C_{PR}$ and present the characteristics of the $C_{PR}$ as well as its validity. The difference between the other indices and the $C_{PR}$ is the way we use to estimate the standard deviation. Calculating the index, most indices use sample standard deviation while the index $C_{PR}$ uses range R. The sample standard deviation is generally a better estimator than the range R. But in the case of the panel process, the $C_{PR}$ has more consistency than the other indices at the point of non-conforming ratio which is an important term in quality control. The reason why the $C_{PR}$ using the range has better consistency is explained by introducing the concept of 'flatness ratio'. At least one million cells are present in one panel, so we can't inspect all of them. In estimating the PCI, it is necessary to consider the inspection cost together with the consistency. Even though we want smaller sample size at the point of inspection cost, the small sample size makes the PCI unreliable. There is 'trade off' between the inspection cost and the accuracy of the PCI. Therefore, we should obtain as large a sample size as possible under the allowed inspection cost. In order for $C_{PR}$ to be used throughout the industry, it is necessary to analyze the characteristics of the $C_{PR}$. Because the $C_{PR}$ is a kind of index including subgroup concept, the analysis should be done at the point of sample size of the subgroup. We present numerical analysis results of $C_{PR}$ by the data from the random number generating method. In this study, we also show the difference between the $C_{PR}$ using the range and the $C_P$ which is a representative index using the sample standard deviation. Regression analysis was used for the numerical analysis of the sample data. In addition, residual analysis and equal variance analysis was also conducted.

• Korean Journal of Environmental Biology
• /
• v.32 no.3
• /
• pp.211-215
• /
• 2014

In order to accurately estimate the litter mass, we evaluated the required sample sizes across 13 chronosequence stands for five years (1994~1996, 2003~2004) in northern hardwood forests in New Hampshire, USA. It was found that the number of required litter traps in our stands (0.25~0.5 ha) within ${\pm}10%$ of the sample mean was appeared to be similar or higher than the 15 litter traps installed in this study. Notably, in 1994 and 1995, the number of required litter trap was twice higher than the 15 litter traps. Further, within ${\pm}20%$ of the sample mean, the number of required litter traps was less than 10 across all 13 stands for five years, which indicates that we can reduce the sample size. Precisely, the number of sample size had increased in stands with steep and high elevation, but no relations with stand age across 13 stands were observed. Based on these results, we suggest that it is important to sample litter mass for several years, in order to determine the number of appropriate sample size, and stands with steep and high elevation may need more litter traps.

• Health Policy and Management
• /
• v.23 no.2
• /
• pp.152-161
• /
• 2013

Korea has a single National Health Insurance program and all citizens are covered under this program, accounting 97% of the population, approximately 50 million people. Claims submitted by Health care providers are reviewed by Health Insurance Review and Assessment (HIRA) for the reimbursement. HIRA database contains not only individual beneficiary's information, but also healthcare service information such as diagnosis, procedures, prescriptions and tests for them. HRA database has gained attention as importance source for research due to its rich healthcare information and the demand of HIRA database has increased. Due to its tremendous size, however, researchers have had problems in accessing the database to conduct research. To meet this demand, we conducted a study to develop the inpatient sample data from HIRA database for research. This study has two purposes: 1) to determine a needed sample size; 2) to test reliability and validity of the sample data. We determined an adequate sample size to ensure representativeness and generality with additional consideration for convenience of calculation. The minimum sample size was 729,904 for the generality, and 488,861 for representativeness. After considering the convenience of calculation, our final sample size was 13% of the population, which was about 7.7 million beneficiaries. Age (5 years interval) and gender were used as stratification variables for sampling. In order to examine whether this sample data appropriately reflect population, we tested the reliability and validity of the sample data. From the sample data, we computed average expenditure of total claims per inpatient for 2011, frequency of top 30 disease, estimation of the number of stroke patients from the sample data, and then compared them to those from the population. Results confirmed reliability and validity of the sample data.