DATA INTEGRITY.

  • Accuracy:
  • Timeliness:
  • Relevance:

DATA CONTROL.

The quality of Input data is important to the accuracy of output.  Control must be instituted as early as possible in the system & everything possible must be done to ensure that data is complete and accurate before being input to the computer.

Objectives of Data Control.

The objectives of Control are:

  • To detect, correct and re-process all errors.
  • To ensure that all data is processed.
  • To preserve the integrity/reliability of maintained data.
  • To prevent and detect fraud/deception.

Note.  Control must be designed into the system & thoroughly tested.  Failure to build in adequate control may cause expensive systems to fail.  In addition, all users must be fully consulted to ensure that adequate controls are implemented.

Types of Data Controls.

The following are controls that can be used to ensure data accuracy:

  • Verification:

This is the process of checking & ensuring that data has been transcribed/ written out correctly.

Verification is whereby several computer users are given data to enter into the computer and the results are compared.  Or else, a second transcription is compared with the first one.  If the results are different, then there is inaccuracy in that data.

This method is mostly used to verify password changes.

Note.  Verification calls for manual intervention, hence errors are possible.  Note that some copying/transcription mistakes that bypass the verification stage are difficult to isolate during verification, e.g. the confusion of l (letter l) and 1 (one).  In this case, l might be input instead of 1 and vice versa, hence such mistakes go undetected.

The main types of errors, which might occur are: –

  • Missing data.
  • Duplicating of data.
  • Use of outdated records.
  • Incorrect batches of input data.
  • Incorrect recording at the source.
  • Incorrect data preparation.
  • Manual controls.

This involves considerable checking of the source documents.

Such checks may be:

  • Inspecting the source documents to detect missing entries, illegible entries, illogical or unlikely entries.
    • Comparing the document against stored data to verify entries.
    • Re-calculating to check calculations made on the document.

VALIDATION CHECKS.

A Computer cannot notice errors in the data being processed in the way that a Clerk or Machine operator does.

Data validation is the process of preventing wrong data from being processed.  It involves checking whether the results generated by the computer are valid or applicable.  During input or data preparation, the data must be checked for transcription errors, through a process known as Verification.

Once the data is brought into the computer memory directly from an input device, immediately before processing, the data is again subjected to checks built in the program described as validation checks, to check the data integrity or the conformity of the data to the processing requirements. 

Data validation includes testing for the following:

  • Test for reasonableness.

The computer program checks whether the data is reasonable, e.g., number of people should not be represented in decimals, i.e. 9½ children.

  • Test for numbers.

E.g., numbers should not be given as alphabets.

  • Test for alphabets.

E.g., alphabets should not be represented as numbers.

These checks can be made at 2 stages:

  • Input stage: When data is first input to the computer, different checks can be applied to prevent errors going forward for processing.  For this reason, the first computer run is often referred to as Validation or Data vet.
  • Updating stage: Further checking is possible during data processing (or when the data input are being processed).

The program checks the consistency of the input data with existing stored data.  This check is possible during the input run if the stored data is on-line at the time.

Note.  Validation is an online process (i.e. validation checks are build into the computer programs using the input data, so that incorrect data items are detected and reported).  Since the checks are under the influence of the computer, they are not prone to errors.

Exercise I.

  1. Distinguish between Data verification and Validation as used in the context of data collection.