Reply #20: No--I am not saying that those who make

[CaTeacher]

less than 200,000 are failures. (where in the world did you get that?)It is a completely false allegation to say that I was trying to say that lots of people make more than $200,000 as well. All I am trying to explain is that an "average" is an inappropriate way to look at this data and I tried to explain why using an "average" is inappropriate.

I am saying that when you take a straight "average" of a set of numbers, that the numbers that are outliers will have a great effect on the average (please read the example on student grades very carefully--I have used this example to successfully teach these simple statistical concepts to many young children--let's say that a student brought home an A in PE and an F in statistics--Do you think that their parents would say--oh--she has a C average--we are happy? Sometimes a straight average just doesn't cut it.).

I will reiterate--hopefully it will be clearer this time:

Assume that I am comparing several businesses--which are all "successful" in that they have been functioning for several years and they are bringing in a profit. Let's assume that the profit from these successful businesses ranges from 30,000 to 3 million per year. The average income for a successful business may very well be several hundred thousand dollars per year--because the outliers at the high ends will skew the data in that direction. You could have many many businesses earning 30K,40K,50K,60K--that are successful businesses and only one or two that brought in over a million dollars. But--the data at the high end is going to skew the numbers to be artificially high if you take a simple average--even though only a handful of the businesses in your sample make this high amount. (An average of 200,000 doesn't necessarily mean that even a single business in the sample actually made that exact amount.)

Using an average is simply not a good statistical model in this case. There is too much variability (deviation) in the data and the outliers will have a huge impact on the average--and thus an average is not an accurate representation of what an average business earns. (A more appropriate statistic would be the MEDIAN)

When you include the failures (bankruptcies etc) all you are accomplishing is effectively skewing the data in the opposite direction.

Unfortunately, it is a harsh reality that the majority of personal businesses fail within a few years. So--if instead of only using the data from profitable business in our calculations (as * does in his rosy calculations)--now we include the businesses that are on the margin--and the ones that actually failed this year. If you include in your average a few businesses that failed with significant debt--you will dramatically skew the data in the other direction--and effectively make the average profit appear very low.

Try averaging just 2 businesses--for example one was -500,000 in debt and the other had 500,000 in profit! You would say--oh the average business makes ZERO dollars! Well that is not very accurate!

My point is simply that using an average is not an appropriate way to look at this data. When you throw out the failures--the high numbers will skew the average high--and when you put in the numbers for the huge failures--it brings the average down significantly.

For this data, it would be more appropriate to look at the MEDIAN of the data. Using a simple average is an inappropriate way to analyze this type of data.