results from this estimation are robust. While the coefficient fluctuates
somewhat, it remains statistically significant and negative in
the different specifications. This finding shows that the result is not
driven by the idiosyncratic features of any given ward.
Placebo test
We now introduce a simple placebo test to further test our
result. One possible concern is that dissatisfaction with waste
management is associated with demand for any kind of action. To
test for this possibility, we refit linear regressions with robust
standard errors on a new dependent variable on the perceived
importance of community organizations. It ranges from 0 to 4 or,
substantively, “not important at all” to “very important.” The
regression estimation is identical to that in the previous section
with the perceived importance of government action as the
dependent variable on a 0e4 scale.
From the placebo test, we find that being satisfied with the
current system does not reduce the importance ascribed to waste
management by community organizations. Instead, as shown in
Table 4, there is no relationship between waste management
satisfaction and the importance ascribed to community organization.
The coefficient is small and statistically insignificant.
Bolstering the plausibility of the placebo test, we also find that trust
in community organization is strongly associated with their
perceived relevance in waste management.
To summarize, we have discovered that greater satisfaction is
negatively correlated with the respondent's preference for government
leadership. In light of existing research, our findings
suggest that urban dwellers perceive waste management in Dar es
Salaam as a public good that requires effective action on part of the