Abstract: All the statistical analysis assumes that all the measurements are obtained correctly. In practice, the observations can be correctly obtained only in controlled environment and this assumption is violated in many real life data. Instead of the true value of the observations, they are obtained with some error 'measurement error' and the true value of the observation is unknown. The usual statistical tools become invalid when the data has measurement error. What are the consequences of presence of measurement error on the linear regression analysis and how to obtain the correct statistical inferences by using the erroneous data in linear regression analysis are the issues to be addressed in this talk.
School of Mathematical Sciences
NISER, PO- Bhimpur-Padanpur, Via- Jatni, District- Khurda, Odisha, India, PIN- 752050