# In descriptive statistics, what is the "confidence level"?

Discussion in 'Microsoft Excel Charting' started by Dora Smith, Feb 11, 2007.

1. ### Dora SmithGuest

Here are teh descriptive statistics for one of my columns.

Total Records Collected

Mean 617,416.82
Standard Error 38,027.85
Median 629,507.50
Mode #N/A
Standard Deviation 201,224.45
Sample Variance 40,491,281,082.30
Kurtosis -0.26
Skewness 0.21
Range 829,194.00
Minimum 248,690.00
Maximum 1,077,884.00
Sum 17,287,671.00
Count 28.00
Confidence Level(95.0%) 78,026.70

I thought it would be great to get the confidence interval, but instead it
gives me a "confidence level" that doesn't even make sense, and the manual
does not discuss what the "confidence level" is. Frankly the number is
far too small in relation to the mean and the standard deviation to appear
to make sense.

I want to know the level one can have 95% confidence of achieving. That's
the only confidence level that has any meaning to me. That figure would
have to be a couple of standard deviations above 617,000, it seems to me.
Not, 78,026. Since the SD is 201,000, what does the number 78,026 even
mean?

What exactly is Excel's "confidence level"?

--
Yours,
Dora Smith
Austin, TX

Dora Smith, Feb 11, 2007

2. ### GuestGuest

the same question to multiple groups.

Jerry

"Dora Smith" wrote:

> Here are teh descriptive statistics for one of my columns.
>
> Total Records Collected
>
> Mean 617,416.82
> Standard Error 38,027.85
> Median 629,507.50
> Mode #N/A
> Standard Deviation 201,224.45
> Sample Variance 40,491,281,082.30
> Kurtosis -0.26
> Skewness 0.21
> Range 829,194.00
> Minimum 248,690.00
> Maximum 1,077,884.00
> Sum 17,287,671.00
> Count 28.00
> Confidence Level(95.0%) 78,026.70
>
>
>
> I thought it would be great to get the confidence interval, but instead it
> gives me a "confidence level" that doesn't even make sense, and the manual
> does not discuss what the "confidence level" is. Frankly the number is
> far too small in relation to the mean and the standard deviation to appear
> to make sense.
>
> I want to know the level one can have 95% confidence of achieving. That's
> the only confidence level that has any meaning to me. That figure would
> have to be a couple of standard deviations above 617,000, it seems to me.
> Not, 78,026. Since the SD is 201,000, what does the number 78,026 even
> mean?
>
> What exactly is Excel's "confidence level"?
>
> --
> Yours,
> Dora Smith
> Austin, TX
>
>
>
>

Guest, Feb 11, 2007

3. ### David BiddulphGuest

If you look at the figures, the confidence level looks more related to the
standard error, rather than to the standard deviation. It therefore looked
as if it related to confidence in estimating the population mean from this
sample.

This seems to be confirmed at
http://cameron.econ.ucdavis.edu/excel/exstat.html
--
David Biddulph

"Dora Smith" <> wrote in message
news:45cf182d\$0\$1360\$...
> Here are teh descriptive statistics for one of my columns.
>
> Total Records Collected
>
> Mean 617,416.82
> Standard Error 38,027.85
> Median 629,507.50
> Mode #N/A
> Standard Deviation 201,224.45
> Sample Variance 40,491,281,082.30
> Kurtosis -0.26
> Skewness 0.21
> Range 829,194.00
> Minimum 248,690.00
> Maximum 1,077,884.00
> Sum 17,287,671.00
> Count 28.00
> Confidence Level(95.0%) 78,026.70
>
>
>
> I thought it would be great to get the confidence interval, but instead it
> gives me a "confidence level" that doesn't even make sense, and the manual
> does not discuss what the "confidence level" is. Frankly the number is
> far too small in relation to the mean and the standard deviation to appear
> to make sense.
>
> I want to know the level one can have 95% confidence of achieving.
> That's the only confidence level that has any meaning to me. That figure
> would have to be a couple of standard deviations above 617,000, it seems
> to me. Not, 78,026. Since the SD is 201,000, what does the number
> 78,026 even mean?
>
> What exactly is Excel's "confidence level"?
>
> --
> Yours,
> Dora Smith
> Austin, TX
>
>

David Biddulph, Feb 11, 2007