margin of error

Thread View: sci.math.stat

3 messages

[T] Back to Threads [V] Tree View [G] Group Articles

3 total messages Started by Kris Popat Sat, 03 Oct 1992 17:53

#631

Author: Kris Popat
Date: Sat, 03 Oct 1992 17:53

14 lines
973 bytes

Consider the statement,

         "The margin of error for this poll was +/- 3 percent."

I'd like to know what such a statement means in precise terms.  The
Chebyshev inequality allows us to make statements like, "the
probability that the measured value differs from the "true" value by
more than D, is less than P."  So does the "margin of error" here
correspond to D?  If so, what is the corresponding value of P?  Are we
supposed to assume it is some standard value, like 0.95?

Kris Popat
MIT Rm. E15-391  Cambridge, MA  02139

Re: margin of error

#632

Author: r hardin
Date: Sun, 04 Oct 1992 00:56

14 lines
877 bytes


>         "The margin of error for this poll was +/- 3 percent."

>I'd like to know what such a statement means in precise terms.  The
>Chebyshev inequality allows us to make statements like, "the
>probability that the measured value differs from the "true" value by
>more than D, is less than P."

It's worse that that, you can only say ``if thus and such has a
so and so distribution, then the probability ... etc''

You can't get a probability without a leading if, no matter how
much sampling you do.

Re: margin of error

#635

Author: Charles Packer
Date: Sun, 04 Oct 1992 10:49

11 lines
1111 bytes

In article <POPAT.92Oct3125308@image.mit.edu>, popat@image.mit.edu (Kris Popat) writes...
>correspond to D?  If so, what is the corresponding value of P?  Are we
>supposed to assume it is some standard value, like 0.95?


Yes, it's safe to assume that they are referring to a 95% confidence
interval, along with all the assumptions that go with it in survey
sampling. At any rate, when the NY Times reports in detail on on of
the surveys they do as a joint effort with CBS, they include a 
footnote that describes the "margin of error" in layman's terms,
and it sounds like the 95% confidence interval to me.

Thread Navigation

This is a paginated view of messages in the thread with full content displayed inline.

Messages are displayed in chronological order, with the original post highlighted in green.

Use pagination controls to navigate through all messages in large threads.

Back to All Threads