stat53p q10

# stat53p q10 - ON SOME ASPECTS OF MEASUREMENT ERRORS IN...

This preview shows pages 1–3. Sign up to view the full content.

ON SOME ASPECTS OF MEASUREMENT ERRORS IN SAMPLE SURVEY PRASENJIT PAL M.Sc. (Agricultural Statistics), Roll No. 4406 I.A.S.R.I., Library Avenue, New Delhi-110012 Chairperson: Dr. U.C. Sud Abstract : Measurement error is defined as the difference between the finally recorded value from its true value. It occurs during data collection phase. The interviewer, the respondent, the questionnaire and the mode of the interview are the main sources of it. It is very serious in the sense that it increases the variance of the estimators. Ratio and regression method of estimators are affected in the presence of it. Both in correlated and uncorrelated data it underestimates the variance. For the unbiased estimation of variance we use repeated measurement techniques and interpenetrating sub samples procedures. If proper questionnaire design and appropriate sample frame work is used measurement error can be minimized. Key words : Measurement Error , Interpenetrating Sub Samples , Repeated Measurement , Ratio Estimators , Regression Estimators . 1. Introduction A survey comprises of number of survey operations. Broadly speaking there are five phases of different survey operation. These are as follows: i) Sample Selection ii) Data Collection iii) Data Processing iv) Estimation and Analysis v) Presentation of Results and Survey evaluation. Errors in survey estimates can be divided into two categories namely sampling errors and non sampling errors. The two principal categories of non sampling errors are: a) Errors due to non observation. b) Error in observation. The error in observation arises when elements are selected and observed but the finally recorded value differs from the true value. This difference may be due to measurement error. The term measurement error is used to denote those errors in individual data that occur during the data collection stage. Errors in individual data may also occur in the data processing stage of a survey. These errors are called processing errors and may be generated during such operations as coding, editing and data capture. Some authors use the term measurement errors in a wider sense that includes both measurement errors and the processing errors.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
On Some Aspects of Measurement Errors in Sample Survey The interviewer, respondent, questionnaire and mode of interview are generally the possible four sources of measurement errors. For example, an incorrect answer on the part of the respondent may be due to misunderstanding the question, not knowing the answer, or unwillingness to provide the true answer. Also, the interviewer may misunderstand the question, put it badly, and influence the respondent to give erroneous answer, or record the information incorrectly. Such errors can never be completely avoided in practice. It is the purpose of questionnaire design and interviewer training to keep these errors under control, so that they could not seriously damage the survey results. Besides this, an important source of measurement errors in survey data is the nature of variables. The nature of the variable is such that exact measurement on it is not available. This may happen chiefly due
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

### Page1 / 15

stat53p q10 - ON SOME ASPECTS OF MEASUREMENT ERRORS IN...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online