The effect of losing one view of the independent variableAnd its location in simple regression analysis

Abstract

The objective of the research , is to shed light on the most important treatment of the problem of missing values of time series data and its influence in simple linear regression. This research deals with the effect of the missing values in independent variable only. This was carried out by proposing missing value from time series data which is complete originally and testing the influence of the missing value on simple regression analysis of data of an experiment related with the effect of the quantity of consumed ration on broilers weight for 15 weeks. The results showed that the missing value had not a significant effect as the estimated model after missing value was consistent and significant statistically. The results also showed that the estimated missing value was larger than the original value when the missing value situated either in the middle or at the end of the series while the sign was negative or the estimated value was less than the original value when the missing value situated in the beginning of the time series. All of that would affect the estimated values outside the time series data according to estimated value of missing value. The research recommended to work on the analysis of the effect of missing more than one value and also when the missing is in the dependent variable only and in both dependent and independent variables.