Lesson Notes

Verification and Validation

Sources of errors

● Data Collection error – Bad handwriting, entering wrong information onto source
document

● Data preparation error – transcription/transposition errors – That data on source

documents are copied incorrectly.

● Data input errors – typing errors

A common problem with computer system is that it is very easy to input incorrect data.
Examples include users mistyping a name or address into a database, so that letters are sent to the
wrong persons. However, it is also possible for a device to misread a barcode and therefore not
provide any information to the user.

Errors are part of any computer information system. There are several approaches to dealing with
the problem that errors cause, specifically detecting when errors occur (so that they can be
corrected) and preventing errors (so that they do not occur). Here a few approaches are
considered.

Data Entry

Sometimes data cannot be automatically scanned into the system for storage, so a common way
is to type the data directly into the computer system. Data entry errors would then occur when an
operator enters data using a keyboard and mixes up digits and /or letters. For example, the
number 32 may be entered as 23. Other errors occur when data entry requirements are not clear.
Do you interpret 08/03/04 as 8th March 2004 or the 3rd August 2004? These data entry errors can
be either accidental or deliberate.

● Accidental errors occur unintentionally: data is entered or a command is issued by

mistake, in good faith, but in error.

● Deliberate: If errors are made deliberately then the user probably knows enough to get
round any validation checks. Examples include a disgruntled employee entering fictitious
data into a company’s database. Possible solutions to this problem include file access
permission imposed by the operating system, vigilances of the administrator and cross
referencing data to other supposedly reliable sources.

Software and hardware errors

There are two types of errors here. Software might malfunction, erasing or corrupting
previously entered data. Hardware may develop a fault (often intermittent) that corrupts data;
examples of such faults are bad sectors on a hard disk or bad memory, or a power surge. The
application may appear to accept data correctly but when that data is retrieved it is corrupt.

Transmission Errors

Transmission errors occur when data is sent from one computer to another. The data received by
one system is not the same as that transmitted by the sender.

Describe methods of verification and validation of data

1
Verification

Data verification is the checking for mistakes such as transcription errors when data is copied
from one medium or device to another. Verification checks do not guarantee that the entered data
is correct. You should be aware that if the original form was completed incorrectly, then the
entered data may pass some verification checks despite being incorrect.

One way to guarantee that the data is entered accurately is to use a procedure called double
(data) entry. Data is entered twice using a program that checks each second entry against the
first. The program identifies the pairs that do not match, so that the correct values can be re-
entered.

Visual checks

One verification method is the use on-screen prompts. After a set of data is entered, it is
redisplayed on the screen. The user is prompted to read it and confirm that it has entered
correctly. If the user has entered any data incorrectly, it is re-entered.

This method of verification is not very reliable since many data entry personnel will not read the
information that is redisplayed or check it carefully against the source document. However, the
method is relatively cheap to use.

Accidental errors can also be prevented by verification in asking for confirmation of instructions.
For example, if you try to delete files the computer may ask ‘Are you sure [y/n]? Or it may
display a summary of input data and ask for verification of its validity before the data is stored.

Validation

If any data input into a data processing system is wrong, the output it produces will be wrong
(Garbage in, Garbage out). In extreme cases, wrong input data can cause a computer program to
fail, for example if it has to divide by zero. For this reason, it is essential that all input data is
thoroughly checked or validated. Data validation is computerised checking of input data for
errors (data that may be unreasonable or incomplete) before it is processed. There are many
different methods of validation. The most appropriate method(s) to use will depend upon what
data is being entered. The most common methods are listed

● Range check

● Reasonableness check

● Data type check

● Inconsistency check

● Presence Check

Range Check

Range Check ensures that the data entered is within a certain range. For example, when you enter
a number corresponding to a month of the year, the range of acceptable numbers is 1 to 12. Here
are some other examples:

● A person cannot be older than the calculation of the present year minus their date of birth.

2
 ● Days in the month cannot be less than or greater than 31. Other calculations would
restrict certain month to 28 or 30 days.

Reasonableness check

Reasonableness ensures that the data entered does not deviate from an established norm (the data
should realistic). An example of data that is not realistic is the age of a high school student that
might be entered as 75.

Data type check

When a database is created each field in the database is given a type. Whenever data is entered
into a field the database will check that it is of the correct type, e.g. text, number, date, currency
etc. If it is not, then an error message will be displayed and the data will have to be re-entered.

Field Name Type Valid data Invalid Data

DOB Date 26/08/1992 26/18/76

Ryan

Sex Text Male, Female 123, WA2

Age Number 23 Twenty Three

Inconsistency /Consistency check

A consistency check compares the contents of two or more fields to make sure that they make
sense when the data in the fields are compared. This check compares new data with previous
data (for example, the age entered corresponds to the age calculation from the date of birth).
Consider the following employee record.

Surname Rogers

Marital Status Married

Number of Children 0

Claim for childcare Yes

allowance

It would be useful to have a consistency check to crosscheck the information in the ‘Number of
children’ and ‘Claim for childcare allowance’ fields, since you should not claim for child
allowance if you have no children. In this example, therefore, either an error has been made on
input or this is a deliberate attempt to claim for child allowance when there are no children.
Other examples of consistency checks are:

● Only employees of a certain category are allowed a car allowance. Check the category
against the car allowance.

● Only full time employees are paid overtime. Check that these fields correspond.

● Citizens over 75 years of age who are alive receive a pension allowance. Check the age
and status on whether the citizen is alive or dead against allowance field.

Presence Check
3
This type of check ensures that the data is actually entered and can be used on any field in a
database. For example, every person in a particular database must be given a job title. Therefore
a presence check could be carried out on the form field in each employee’s record to ensure that
a job title is entered on the form.

Difference between validation and verification

Verification Validation

● Is usually done by a user at the data input ● This is usually done after data is entered.
stage to ensure that data entered is the
same as that which was on a source ● Validation is done automatically by a
document. computer by using validation program to
detect errors.
● Different users type the same data to
double check that data are correctly
transcribe.

Ensure that you are able to answer the following questions in the future.

1. What is data validation?

2. What are some validation checks?
3. Explain how data verification is different from data validation.
4. Copy and complete the table below by giving an example of valid data for each

Field name Data Acceptable values

Item Code A0001 to Z9999
Quantity ordered >0
Date of order =today or <today
Customer ID 00000001 to 99999999
Employee Code 1 to 599

5. State an appropriate validation check for the following field names.

a) Month of the year b) gender

c) Height of a person d) surname

Which checks can be applied for each point in the picture above?