Statistics and outliers question?

Question

I dont remember statistics that much, but I want to know the formula for this problem.

Basically, I want to average numbers out. But in doing so, I want to omit the outliers so they don't contribute to the average. Get it? So What would be that formula? Thanks in advance!

Answer 1

I've done this many times, and here's the procedure I usually follow.

Sometimes, by ordering the data or doing a scatter diagram, you just see obvious outliers and throw them out.

Then, whether you did the preceding or not, enter the data using a computer, a statistical calculator, or, in the absence of those, calculate the sum of the x's and x-squares, and get the mean and standard deviation. (It's preferable to use n-1 in the calculation to account for the loss of a degree of freedom, using the sample variance,, but for what you're doing, that's not absolutely necessary.)

Then take the mean, plus or minus standard deviations. Anything outside that range can be considered an outlier.

Throw out the outliers, and use all of the remaining data.

If you follow this procedure, you'll be okay.

P.S. A good formula to use for the sample variance is this:

s^2 = (n sumsq - sumx^2) / [n(n-1)]

where sumx is the sum of your x-values, and sumsq is the sum of the x^2 values.

Answer 2

Add up the values and divide by the number of values.

For example, if you drop 3 outliers out of 25, then divide by 22.

Be aware that this isn't exactly "kosher."

An alternative would be to use the median instead of a mean. That is to line up all the scores in order and select the one that 1/2 of the values are above and 1/2 are below. This measure of central tendency reduces the impact of outliers and is statistically valid when identified as the median rather than the mean.

Answer 3

Take all the numbers drop whatever highs and lows you want. Find out how many number remain. Add up all the remaining numbers, divide by the number of terms that you added up. Here's an example
1 2 3 4 5 6 7 8 9
drop highs and lows- 2 3 4 5 6 7 8
add up all the numbers - 35
divide by the number of terms - 35/ 7 = 5
average = 5

Answer 4

There is no formula for that. If you have outliers you need new data. Having outliers will taint your results.

Looks like new survey time, bro.