On the estimation of finite population mean with suitable adjustments of non response error

Abstract

The collection of necessary statistical information in all kinds of investigations in everyday life, through well-developed sample surveys, is the first and foremost objective which is absolutely necessary for any researcher and particularly to any statistician either to seek answers of questions coming up in mind about the undiscovered empirical word or to prepare plans and policies for the administration of a country to implement these for quick economic and social development. For the purpose of gathering information about a finite population, generally a sample of pre-determined size is taken from the population with any of the existing sampling techniques as per the need of the survey. But, it is generally observed in practice that in most of the cases, the population under consideration possesses some of the units on which the relevant information might not be collected, even after utmost efforts, due to a number of reasons and hence, a good number of units selected in the sample also does not respond, being the hard core ones, or are not accessible due to many reasons for collecting relevant information. This situation yields a special kind of non-sampling errors, popularly known as NON- RESPONSE, in the language of survey sampling theory. The problem of non-response in finite population sample surveys is one serious kind of non-sampling errors which occur in all kinds of sample surveys, particularly when the sample size is quite large or where a number of sensitive and confidential type questions have been included in the questionnaire. It is seen that whatever be the nature of the population to be sampled and whatever be the sampling strategy to be adopted for getting a representative and complete sample, non-response do occur and it cannot be completely eliminated. In other sense, non-response is an inherent characteristic of any type of population and, therefore, by no means it can be ignored while estimating population characteristics. newline