Equivalence partitioning and boundary value analysis are most suitable when we have specific inputs that are treated independently.

When the outcome of software execution depends on multiple inputs that are interrelated, the above-mentioned two techniques are not effective.

Decision table testing is a suitable method of analysing combinations of interrelated input values. This technique is also known as “cause-effect” table.

A decision table is designed as follows:

• Each input parameter has its own row in the table
• Each column represents a unique combination of input parameters. The result of each combination is determined by a (business) rule
• The result of combining the input values according to business rules is recorded on the last row of the table. An X represents and invalid combination.

Example 1:

An insurance software calculates premium based on the following rules:

• If the insured is male, then premium is $1500 / year
• If the insured is female, then premium is $1000 / year
• If the insured’s car has alarm, premium is reduced by 10%

The following table shows possible valid and invalid combinations of the three input parameters and the appropriate output values (premium) for each combination.

Example 2:

An e-commerce application calculates shipping fee as follows:

• Ground shipping = $10
• Air shipping = $20
• If the order value is over $30 then the shipping fee is reduced by 30%

The following table shows possible valid and invalid combinations of the three input parameters and the appropriate shipping fee for each combination.