The paper addresses the issue of measuring web accessibility
in such a way that differences in measurements reflect differences in the
effectiveness experienced by disabled users. The paper presents the steps
upon which a measuring methodology called MAMBO is based, and the
data that are needed to compute the indexes, in addition to its concep-
tual rationale. An experimentation of MAMBO is then described, based
on analysis of 14 accessibility reports; results are shown and discussed,
including the effects that different severity judgments may have on the
metric, how to estimate confidence intervals on the values, and how the
metric can be used to estimate accessibility with respect to specific user
groups.
