Re: Comparing groups of different sample sizes

stef wrote:
I am wanting to compare the difference in hospital length of stay and
associated cost savings between a group of 167 patients whose
physicians refused a particular hospital consult service vs. 800
patients whose physicians did not refuse this consult service - in
order to evaluate the impact of this service on inpatient LOS and
cost-savings. Do the N's need to be equal? - do i need to take a
random sample of 167 patients from the 800 and only compare those - or
can i compare the entire 800 as long as variances on key variables are
equal?

They don't need to be equal - some common procedures assume equal
variances, but there are solutions to this procedures such as the t
test. Excluding data will result in less precise estimates and tests.

Thom

.

Relevant Pages

  • Comparing groups of different sample sizes
    ... I am wanting to compare the difference in hospital length of stay and ... associated cost savings between a group of 167 patients whose ...
    (sci.stat.edu)
  • remicade and polymyalgia rheumatica - no improvement
    ... Title: Infliximab in the Treatment of Polymyalgia Rheumatica: A Double ... To compare the efficacy and safety of prednisone plus ... infliximab and prednisone alone in patients with newly diagnosed ... 49 patients with newly diagnosed PMR. ...
    (alt.support.arthritis)
  • Re: Gynecologist Accused of Mutilating, Abusing Hundreds of Women
    ... patients, many other patients, have now come forward with horror stories, ... If the 'pride' merely involved removal of the clitoral hood, ... you simply can't compare an adult ...
    (alt.true-crime)
  • Purevision vs. Focus Night & Day?
    ... patients who have tried both, to hear how they compare. ... prescribed with monovision. ... with a multifocal silicone hydrogel in the near future? ...
    (sci.med.vision)