A study on the role of interviewer in surveys

Non-sampling errors have many different sources, can occur at different stages of the survey and are a result of error contributions of different personnel or agencies Involved in executing the survey. In Interview based surveys,interviewer effectively influences both the response error as well as the response rate. In the present investigation an estimation procedure has been proposed to study the contribution of variance due to response errors as well as non-response errors introduced because of the interviewer.     

The power of two exact tests for structural change in the presence of heteroskedasticity

This paper compares analytically the power of the Jayatissa (1977) and Tsurumi (1984) tests for structural change in a heteroskedastic normal linear regression model with two equal-sized independent sub-samples. It is shown that the non-centrality parameter of the Tsurumi test cannot exceed that of the Jayatissa test and some conditions are obtained under which one test dominates the other.