Thursday, March 29, 2012

ERRORS IN DEMOGRAPHIC DATA


Despite the care taken to ensure the quality of the data collected by enumeration and registration, it will some time give obvious indications of errors in basic information. Usually errors in censuses and vital statistic registration (VSR) are of two types:
1.       Coverage.
2.       Content.
    A third type of error is sampling error which is introduced when sampling is used.

1.    ERROR OF COVERAGE:
                                        It is the common observation that a conduct or survey some households are missed and not counted, others may be counted two times. In VSR (vital statistic registration) there are cases which skip in registration (birth, date, marriages) these situation errors in demographic data offer must be made to cover all units and events of all observations which fall in our domain.

2.    ERRORS IN CONTENT:
                                      The term error of content refers to instance where the characteristic of a person counted in a census enumeration or in registration or in a survey are in correctly reported, recorded and tabulated or sometime they are completely missing.
                                      Besides from the errors due to carelessness or mistakes in judgment in a census or VSR, errors of coverage or content or both can arise at any step from initiation of an original record publication to the final tabulation or final record. The five principles are:
                                 i.            The respondent.
                               ii.            The recorder/ enumerator /investigator.
                              iii.            The coding process.
                             iv.            The editing process.
                               v.            Compiling and printing process.
                                                    The respondent error may be occur due to ignorance, illiteracy or will full wrong answer. For example: a person may not remember his/her correct age or he/she may will fully give wrong age or income etc. sometimes the respondent does not understand the question and give wrong information, this type of errors known as respondent error.
                                                   The second source of error is one part of enumerator who may not listen correctly or may made wrong entries. The errors at the stage of coding, editing and classification are quiet common. Placing of code at the right place is very much necessary. Editors have to give extra care in comparing logical answer of similar and related question. Similarly at the time of classification, tabulation and printing certain error do crept.
DETECTION OF ERRORS:
                                          In VSR errors may be detected by comparison of either tabulated data or individual records. Comparisons of tabulated data may yield or may be made either by internal comparison with previous census results or comparison with independent collateral data. Internal comparison may revealed inconsistency between related data in the tabulated result ratio to male or female for ages at some stages of life may appear to be out of line. Comparison can be made with previous census and vital statistics. For example: a census may shoe fewer or more persons in an age group than expected keeping in mind mortality and migration patterns.
         i.            COMPARISON WITH COLLATERAL DATA means to look at related independent data that is number of children in schools, number of buses, housing unit, electric meters etc.
       ii.            COMPARISON OF INDIVIDUAL RECORDS census return of individual is critically examined with reference to itself and some other co-lateral records’. Registration of individual tax return etc.
      iii.             RE-ENUMERATION SURVEY/ POST ENUMERATION SURVEY
     iv.            In order to check the quality of census returns a re-enumeration survey or post enumeration survey is conducted on selected units. This survey is conducted by senior professionals and all efforts are being made to obtain accurate records. Schedules of these surveys are than compared with census returns. To measure the error of content, the post enumeration survey is sometimes known as content evaluation survey (CES) and it is supposed to be accurate among all exercise.


No comments:

Post a Comment