A subdivision of fundamental attribution error is the tendency of people to attribute positive attributions of a job well done to themselves, while failure is assigned to someone else. As is mentioned in the introduction of the paper, it's common for a student to believe he or she passed a test do to their studying and hard work. Should they fail a test, they will attribute it to the professor and bad grading or an unfair test. In this particular experiment, MS Office users were tested to see how common this tendency really is.
In this particular experiment, the interface agents played a role similar to that of the operator on older telephone lines. The user issued commands, the agent would try and learn and understand their needs and personality and attempt to do their best to fulfill them. In this case, however, they were computerized, machines designed to try and understand the needs of the user. Prior research had confirmed the presence of self serving bias in many things and the examples given are economic behaviour, decision making, management, corporate governance, organizational behaviour and information systems and an earlier study revealed the same in human-computer interactions since people respond to computers socially, even though it is truly inanimate.
All students in the study were knowledgeable about the usage of the Microsoft applications provided as well as usage of agent. At some points, the agent became more autonomous and sometimes less. The conclusion drawn from the study was that users did not always assign positive qualities to themselves in an instance of success. However, the more autonomous an agent becomes, the more the user assigns both positive and negative qualities to the agent, in both success and failure.