1. Check the questionnaire identification numbers entered in the database to
make sure each case has an identification number and no identification
number has been repeated in the database. If any identification number is
missing, go back to the paper questionnaires to find the correct number and
enter this into the database. If any identification number is repeated, check to
see if these cases are duplicates and delete one of the duplicated cases. If these
cases are not duplicates, check for the correct identification numbers in the
paper questionnaires.
Refer to the paper questionnaires as often as needed during the data
cleaning process. Every issue raised during data entry should be checked
against the paper questionnaires whenever possible.
2. Run the frequencies and means of numerical variables. Is there anything that
is unexpected? Are there any outliers that are greatly above or greatly below
the average value? Check any questionable data against the questionnaires
and correct any errors in data entry.
3. Look for missing data and check to make sure these are not a result of data
entry error.
