In this paper, we consider the problem of analysis of Balanced Incomplete Block Design (BIBD) in which yields of two plots get mixed-up. Estimates of mixed-up values, explicit expression for bias and test statistic for testing the hypothesis on homogeneity of treatment effects are developed. Also an illustrative example is given. Keywords: Mixed-up plots, Balanced Incomplete Block Design.

  Optimum return from the investment on inputs and minimum environmental pollution are the two major issues to be addressed while prescribing soil test based nutrient recommendations. A comprehensive knowledge of the soil resource is of fundamental importance for efficient land use planning. Green revolution by using high-yielding varieties and improved management technology has increased crop production at the cost of productivity of soil and possible risk of soil degradation. Decrease in the soil fertility and imbalanced use of nutrients are important factors responsible for stagnation or decrease in the crop yields over the years. Thus, it should be firmly understood that further increase in food production must be attained by judicious use of soil as a resource. Fertilizer being the costliest input, the scientific approach towards profitable agriculture would imply, use of plant nutrients according to the actual needs of the soil-crop situation. In this study, soil fertility maps were prepared using point estimates at different geographical locations using stratified multistage stratified random sampling. An appropriate kriging method was used for different nutrients using Akaike?s information criterion. The prepared soil fertility maps were validated using soil test values estimated in the next two consecutive years after mapping. There was a significant agreement between estimated soil test obtained through kriging and those determined in the following years by actual sampling. Keywords: Soil fertility maps, Geographic Information System (GIS), Global Positioning System (GPS), Sampling, Kriging.

  The Self-exciting threshold autoregressive (SETAR) nonlinear time-series model is thoroughly studied. This model is capable of describing cyclical data. The Genetic algorithm (GA), which is a powerful stochastic search and optimization procedure motivated by the principles of genetics and natural selection, is employed for estimation of parameters of the model. As an illustration, India?s lac export annual data for the period 1901-2001 is considered for building the model. Forecasting for hold-out data for the years 2002-?08 is also carried out. Superiority of GA-methodology over Search algorithm procedure is demonstrated for the data under consideration. Keywords: Akaike information criterion, Genetic algorithm, India?s annual lac export data, Forecasting for hold-out data, SETAR nonlinear time-series model.

  N.A

  The present work is an attempt to study the estimation of the population mean on the current occasion in two-occasion successive (rotation) sampling under a super-population model. An estimator of the population mean for the current occasion in two-occasion successive (rotation) sampling has been proposed. Optimum replacement policy and performance of the proposed estimator have been discussed. Results are interpreted via empirical studies. Keywords: Successive sampling, Super-population, Linear model, Auxiliary variable, Variance, Optimum replacement policy.

  N.A

  The present work intends to emphasize the role of imputation methods to deal with non-response when it occurs on both the occasions in two-occasion successive (rotation) sampling. Under imputation two difference-type estimators have been proposed for estimating the population mean on current occasion. Estimators for the current occasion are derived as a particular case when non-response occurs either on the first occasion or on the second occasion. Behaviors of the proposed estimators are examined and the optimum replacement policies are discussed. Effectiveness of the suggested imputation methods are discussed in two different situations: with and without non-response. Results are analyzed with the help of empirical studies. Keywords: Non-response, Imputation, Successive Sampling, Chain-type, Optimum replacement policy.

  N.A

  N.A

  N.A

  N.A

  The problem of domain estimation in the context of non-response arising out of mail surveys has been considered. Expressions for the unbiased domain estimator, variance of the estimator and unbiased variance estimators are obtained. Optimum values of the sample sizes have been derived under a cost function. The theoretical results are numerically illustrated. Keywords : Domain estimation, Mails surveys, Non-response.

  Estimates of proportion for small areas are often required for many policy planning and resource allocation. This paper examines an application of linear assumption based model-based direct (MBD) approach (Chandra and Chambers 2005, 2009) of small area estimation (SAE) to estimate small area proportions. The empirical performance of the MBD approach is compared with indirect method of SAE, the empirical best predictor (EBP) under a generalised linear mixed model (Rao 2003, chapter 5 and Saei and Chambers 2003). The empirical results based on two real data show that both the MBD and EBP perform well. The EBP is a computation intensive method, in contrast, MBD is easy to implement. In case of model misspecifications, the MBD appears to be more robust. Keywords: Binary variable, Small area estimation, MBD approach, GLMM, Empirical best predictor.

  In Dutta et al. (2009b) methods of construction of optimum covariate designs (OCDs) in partially balanced incomplete block design (PBIBD) set-ups have been proposed. Applying these methods actual designs can be obtained. Considering the suitable PBIBDs from Clatworthy?s table (1973) lists of OCDs, categorised for different classes of designs, have been prepared here. The constructional methods have also been indicated for ready reference. Also some new results and constructional methods in the set-up of balanced factorial designs with orthogonal factorial structure (OFS) with the list of OCDs has been given. Keywords: Block designs, Covariates, Optimal designs, Orthogonal arrays, Hadamard matrices, Khatri-Rao product, Kronecker product.