clinical DSS DxPlain [1], which helps doctors perform diagnosis,
the user can select information from a patient’s medical
profile, including current symptoms and general information
such as age, and generate a list of possible diagnoses. The
user can also select some information to receive special focus,
which affects how the system generates the list. This
strategy of customization allows users to integrate their own
knowledge or expertise into the recommendations.
Customization is a design feature found not only in DSS but
in many types of human-machine systems. In this paper, I
make a theoretical contribution to understanding the effect
that customization has on user decision making in an HCI
context. I also offer a practical contribution to DSS design
and policy. I show that customization can create a decisionmaking
bias for users. This bias can lead users to erroneously
follow bad recommendations generated by an algorithm, although
it can also make users better rec