We have already mentioned the chief limitation of this
study, which is the presence of the University of Colorado
in one of Boulder’s ZIP codes. We also note the use of
secondary data as a limitation, especially as these data do
not permit reporting or testing of null hypotheses. As the
restaurant failure data obtained from the Health
Department of Boulder were combined with aggregated
Census data, each restaurant in a given ZIP code received
the same value from the census data. Hence, it was, for
example, not possible to compute an “SSE (within sum of
squares)” for an analysis of variance (ANOVA) test. Due
to this unfortunate limitation, the analysis was restricted
to observations. Similarly, from the current data, it was
not possible to differentiate between various restaurant
segments from a marketing perspective, or to test the
potential skew from the proportion of college students in
the town of Boulder.