Cause-effect analysis as a test design technique

Let's start with the concept of test design. In short, it is one of the stages of the software development process. It means the stage of designing, creating, and organizing tests into sets based on the analysis of requirements and product functionality. Test design could be compared to requirements analysis, but it is rather the result of this analysis. Problems occur when one does not know and understand the functionality for which test documentation needs to be created. When analyzing requirements and test design, it is very important to understand how even the simplest functionality is organized, as it may cause problems later on. A superficial approach at the stage of analysis and test design is unacceptable. It is necessary to understand the system thoroughly in order to understand what works and what is connected to what in the program. One of the test-design techniques that will be discussed in this article allows not just writing tests, but trying to find errors as a result of the operation or interaction of a particular functionality.

Cause/effect is a mathematical relationship that can be shown in testing practice by entering combinations of conditions (causes) in order to get a response from the system (effect).

Cause-effect testing is a test design technique that uses a graphical display of input data (causes) and output data (effects) for design. Different combinations of causes can lead to different results. The method requires an understanding of Boolean logic (logical operators –  and, or, not). Based on the analysis of the established relationships, it is possible to comprehensively select high-performance tests. Thanks to this method, it is possible to identify defects in requirements or specifications at an early stage. 

Tests are built in several stages:

• the specification is divided into work areas;
• the specification lists multiple causes and effects. A cause is a separate input condition or equivalence class. An effect is an output condition or a transformation of the system. Here, each cause and effect is assigned a number;
• a truth table is created based on the analysis of the semantic content of the specification, in which all possible combinations of causes are sequentially processed and the results (consequences) for each combination of causes are determined.

The table is provided with notes that set constraints and describe impossible combinations. The disadvantage of this approach is the poor study of boundary conditions.

Using the cause-effect testing method

Test templates using this technique should be written according to specifications or use cases. This document contains the steps and expected results of the test, but no specific data. They will be presented at the next stage of test case development.

Example:

This is a technique that helps to create functional cases, taking into account all the elements that interact and influence each other.

To apply this technique, you need to know:

• what are the input parameters (what does this functionality operate with);
• what is the output data (what is this functionality for, what does it do);
• how the input parameters affect the result.

Each preceding combination can be numbered and recorded in separate truth tables, in which the truth is marked as «1» and the false as «0». For uncertain states, the «X» symbol is used, which can be either «1» or «0».

All rows of the truth table are turned into tests. In this case:

• if possible, combine tests from independent tables using pairwise testing;
• for equivalence classes, you need to additionally specify input conditions.

This method allows starting from the causes of what needs to be done within the test scenario and smoothly moving to the consequences. Like all test design techniques, it helps determine the smallest number of tests to run in order to find more defects. By determining all the reasons and results, one can make sure that the system will respond to any manipulations. Also, this technique allows you to find flaws in the logic of the application description, which will help improve the documentation in the future.

Example from an online store:

Functionality: Registration.

Input parameters:

• input fields in the registration form;

Output parameters:

• user is registered, there is a record in the database.

Based on this cause and effect, one can make different variations. For example, we can try to make sure that nothing is added to the database or we can repeat the record. Different scenarios can be designed at the initial stage.

The next example is from a social network:

Functionality: user's news feed.

Input parameters:

• current user;
• friends of the user;
• groups to which the user is subscribed;
• blacklist.

Output parameters: 

• user's news feed.

If a friend or community posts something, we can see it in the feed, but we won't see anything that was added to the blacklist.

To sum up: cause and effect analysis is a testing technique that displays the input data and the system response in combination. By using this technique, you can significantly reduce the number of tests and pay more attention to more important and vulnerable areas of functionality.