Outliers from faulty distributed
assumptions: incorrect assumptions
about the distribution of the data can
also lead to the presence of suspected
outliers Iglewieze and Hoaglin,
(1993). Blood sugar levels,
disciplinary referrals, scores on
classroom tests where students are
well-prepared, and self-reports
of low-frequency behaviours (e.g.
number of times a student has been
suspended or held back a grade) may
give rise to bimodal, skewed,
asymptotic or flat distributions,
depending upon the sampling design.
The data may have a different
structure than the researcher originally
assumed, and long or short-term
trends may affect the data in
unanticipated ways. For example, a
study of college library usage rates
during the month of September may
find outlying values at the beginning
and end of the month, with
exceptionally low rates at the
beginning of the month when students
have just returned to campus or are on
break for labour weekend in (Nigeria)
and exceptionally high rates at the end
of the month, when mid-term
examinations have begun.