Supplementary exercise 6.46 of IPS7e
------------------------------------

Stating null and alternative hypotheses for mu (population mean).

(a) The description indicates that the researcher (tenant group) is 
only interested in whether the apartments are smaller than advertised. 
Therefore it is valid to use a one-sided alternative hypothesis. 
  H0: mu=1250
  Ha: mu<1250
Note that the hypothesis representing the desired conclusion is given as
the alternative hypothesis, not the null hypothesis. This is because the
strongest test conclusion is a rejection of H0 and hence an adoption of
Ha.

(b) Again the description invites a one-sided alternative hypothesis,
  H0: mu=32
  Ha: mu>32
The description does not specify what data should be used to assess the
claims. If a single value for mileage is computed after 3000 miles, it is
not clear how this can be used for inference, because we would not know 
what its standard deviation (or standard error, if we're thinking about 
this value as our estimate) would be. If instead mileage was computed after 
each trip to the gas station to fill up gas, there would be a long series 
of values on which inference could be based.

(c) Here the description explicitly mentions that there is interest in both
directions of deviations (larger or smaller) from the optimal value. 
  H0: mu=5
  Ha: mu<>5