 
  
  
   
From the point of view of variance reduction, the adjustment of the
  's solely by
 's solely by   seems very little different to that
obtained by adjusting fully by
  seems very little different to that
obtained by adjusting fully by   and
  and   . We can examine
this suggestion, and the determine the effects upon expectation, by
considering separately the effects of adjusting by the two-hour
measurement,
 . We can examine
this suggestion, and the determine the effects upon expectation, by
considering separately the effects of adjusting by the two-hour
measurement,   , in addition to
 , in addition to   .
 .