GCSE Information Technology Project

Error Checking

Before searching a database and drawing conclusions it is very important that you are using accurate data. Perhaps a few errors will have a minor effect, if any, on the final results, but this should not be assumed. Imagine a graph involving ages, where one of the points plotted is for a person aged 999 instead of 9. The average age quoted would also become nonsense.

Two factors are involved when checking for errors:

Verification - to visually check and verify that data as shown on a questionnaire (or other data capture form) has been transferred into the computer correctly - someone could read out the data to the keyboard operator and at the same time check that it is being transferred accurately, by watching the screen.

Common causes of data entry errors at a keyboard might be: typing a character twice; putting data in the wrong place; misreading (or mishearing) the original data; spelling and punctuation slipups. The use of simple codes for the data collected can greatly reduce the risk of many causes of error.

Validation - checks performed by the computer to ensure that a database can contain only entries that are valid (within sensible values).

Good database software will contain facilities to prevent most errors being accepted by the computer - multiple-choice selection, preset maximum and minimum values, maximum and minimum length (number of characters), numbers only, no blank entries allowed. The use of simple codes for the data collected can greatly reduce the risk of a typing error and also allow these automatic checks to be more fully implemented.

There are various examples using input devices other than a keyboard. A bar code representing say the numbers 2 1 3 2 might have a fifth bar representing the number for the total of 8 (it is actually a more complicated calculation). This final number is called a 'check digit'. If the scanner at the supermarket checkout misreads one of the bars, then the final figure will not be true and a warning is sounded for the operator to try it again. The recent rise of digital communication such as TV, telephones and the Internet, hangs on a similar principal - no transmission errors means top quality reception. (Digital data requires lots of memory and processing speed and had to wait for technology to reach that capability.)

When a database is complete, queries could be made to look for possible anomalies: Query Qu6 = 0 and Qu7 = 1, could be searching two fields where the data has been entered correctly but actually means that the person who filled in the questionnare is saying that they do not have a computer and they use it once a week! It would therefore be necessary to remove the whole record for that person from the database.


Page last updated 23/05/2003 © Copyright, I.D.Lee, Didcot Girls' School.
All rights reserved. The original material provided on this site may not be copied or redistributed without written consent from the school. But please feel free to add a link to this page on your own website if you find it useful.


Exam Projects Menu Back: Planning your survey | Planning a database
Next: Presenting your survey
first pageprevious pagetop of pagenext page