AdminShade
March 25 2006 11:13 AM EST
As some of you might know I am working on figuring out some spells.
I now have 2 ways of calculating something I believe is the same, and they get a bit different results.
The first forrmula to is:
2 *Standard Deviation
-------------------------
square root of the number of data points
The second formula is one used by excel:
CONFIDENCE(alpha,standard_dev,size)
the calculation behind it i don't know exactly but it's a calculation in which the alpha determines what % of confidence is calculated:
an alpha of 0.05 makes the calculation predict conficence that 95% of the measurement points fall within the average + or - the calculated number.
explanation in numbers:
CONFIDENCE(0.05,2.5,50) equals 0.692951.
In other words, the average equals 30 ± 0.692951. or average = 29,31 - 30.69 with 95% confidence.
I think the formula excel uses would be more accurate, however I am not sure.
Some data to look at:
Average Damage: 156,8
Margin of Error: 5,01
Confidence Level: 4,91
Standard Deviation: 30,69
Amount of measurements: 150
So the Confidence level is smaller, if only a bit, than the mean error.
Shot in the dark guess, because I don't have Excel, but often time sqrt(n-1) is used instead of sqrt(n). Often it gives an unbiased estimator when taking measurements from a sample. Perhaps that explains the small discrepancy.
Sacredpeanut
March 25 2006 7:09 PM EST
n-1 is used to calculate Standard Deviation of a sample and n is used to calculate the Standard Deviation of an entire population. Since we are trying to find the Standard Deviation of a sample here n-1 is probably more appropriate.
Also strictly speaking 1.96 should be used not 2 for a 95% confidence interval.
Try comparing the formula 1.96* (SD / sqr (n-1)) to the Excel number.
AdminShade
March 25 2006 8:03 PM EST
Try comparing the formula 1.96* (SD / sqr (n-1)) to the Excel number.
new formula: 4,927835085
excel number: 4,91
AdminShade
March 27 2006 10:42 AM EST
Bump!
Some information on how Excel calculates it:
CONFIDENCE
Returns the confidence interval for a population mean. The confidence interval is a range on either side of a sample mean. For example, if you order a product through the mail, you can determine, with a particular level of confidence, the earliest and latest the product will arrive.
Syntax
CONFIDENCE(alpha,standard_dev,size)
Alpha is the significance level used to compute the confidence level. The confidence level equals 100*(1 - alpha)%, or in other words, an alpha of 0.05 indicates a 95 percent confidence level.
Standard_dev is the population standard deviation for the data range and is assumed to be known.
Size is the sample size.
Remarks:
If any argument is nonnumeric, CONFIDENCE returns the #VALUE! error value.
If alpha < 0 or alpha > 1, CONFIDENCE returns the #NUM! error value.
If standard_dev < 0, CONFIDENCE returns the #NUM! error value.
If size is not an integer, it is truncated.
If size < 1, CONFIDENCE returns the #NUM! error value.
If we assume alpha equals 0.05, we need to calculate the area under the standard normal curve that equals (1 - alpha), or 95 percent. This value is 1.96. The confidence interval is therefore:
X(average) +- 1.96 (sigma / root of n)
Example
Suppose we observe that, in our sample of 50 commuters, the average length of travel to work is 30 minutes with a population standard deviation of 2.5. We can be 95 percent confident that the population mean is in the interval:
30 +- 1.96 (2.5 / root of 50)
or:
CONFIDENCE(0.05,2.5,50) equals 0.692951. In other words, the average length of travel to work equals 30 +- 0.692951 minutes, or 29.3 to 30.7 minutes.
So I think Excel uses the right calculation, except for the N -1 part which for measurements over 150 shouldn't make much of a difference anyway.
This thread is closed to new posts.
However, you are welcome to reference it
from a new thread; link this with the html
<a href="/bboard/q-and-a-fetch-msg.tcl?msg_id=001ko1">Anybody into statistics? I have a question for you then!</a>