Estimation of standard error of the parameter of change using simulations

Djordje Petkovic
Statistical Office of the Republic of Serbia


The main objective of this paper is to present the procedure for estimating standard error of parameter of change (index) of turnover in R software (R core team, 2014) when samples are coordinated. The problem of estimating standard error is dealt with in the statistical literature by various types of approximations. In my paper I start from the method presented at the Consultation on Survey Methodology between Statistics Sweden and Statistical Office of the Republic of Serbia (SERSTAT 2013:22), make simulations and calculate estimate of the correlation and true value of standard error of change between turnovers from two years. I use two consecutive sampling frames of quarterly Structural Business Survey (SBS). These frames are updated with turnover from corresponding balance sheets. Important assumption is that annual turnover is highly correlated with quarterly turnover and that computed correlation can be referred to when comparing methods of estimation of correlation on the sample data.

[Full Text]

Romanian Statistical Review 2/2015