Re: confidence interval for ratio of correlated proportions

robert_dodier_at_yahoo.com
Date: 03/17/05 

Date: 17 Mar 2005 07:23:43 -0800

> The key requirements are that Pr[Y <= 0] be
> effectively zero and that SD[Y]/E[Y] be small.

These requirements are likely not met by the scenario described
by the original poster. As described, the probabilities in question
are probabilities of vehicular collision under different scenarios,
so I would guess the mean value is close to 0, and the
probability of 0 itself is substantial. Under the circumstances, the
approximation given is probably not very accurate, perhaps even
very inaccurate.

To the original poster -- it would be unfortunate to go to the trouble
of setting up a complicated simulation, only to throw away the results
with an inappropriate approximation.

For what it's worth,
Robert Dodier

Relevant Pages

  • Re: ACS-Law and the Copyright Speculative Invoicing
    ... Scenario" that you dreamed up is proof that your arguments all revolve ... tell me why the above is 'flawed use of statistics'. ... But it's still a hypothetical scenario, and it's not to say I'm necessarily right. ... If you were a judge in the case, you would have to decide one way or the other, and you'd have to decide 'on the balance of probabilities'. ...
    (uk.legal)
  • Re: Nationwide also to introduce "Card Reader Security"
    ... basis that the probabilities of guess i being the first successful ... the exact probabilities) no better or worse than 123123123. ... I was using the Si as a basis for speculating what ... mostly a little smaller in scenario D, ...
    (uk.finance)
  • Re: Nationwide also to introduce "Card Reader Security"
    ... the exact probabilities) no better or worse than 123123123. ... "The different order of cards visited might change the ... order in which any successes occur, ... were mostly a little smaller in scenario D, ...
    (uk.finance)
  • experts in small probabilities
    ... for a research project I have to work with estimations of very small ... probabilities. ... Experts will be asked how likely it is, that a scenario happens in T years. ...
    (sci.stat.consult)
  • Re: Is it possible for a probability to be unknown?
    ... to estimate such probabilities. ... How about a scenario like this: ... That is at least a repeatable experiment. ... I would guess a hit rate closer to 20-25%, ...
    (sci.math)