How many sexual partners have you had

[Nikolai]

The median can't be five. Nobody has voted five.

EDIT: If not someone has voted above four after I voted.

EDIT2:The median was four when I checked a few minutes ago.

[Ben Kenobi]

Obviosly the average is much higher but it is actually impossible to calculate on the provided data-- (assume for a moment that one of the responsders is Wilt Chamberlin)

Yeah, I had mean and median mixed up before. I listed the floor of the mean (ie, the lowest possible).

[KrazyHorse]

The median can't be five. Nobody has voted five

The sample median is defined, in the case that the sample contains an even number (N) responses, as the mean of the N/2 lowest response and the (N/2 + 1) lowest response.

[Flubber]

Question: Is medians often used in surverys like this? In school, get got told "this is what median is, this is what an average is." But we never used median in maths, only averages.

For a data set like this a median is a far more descriptive and indicative number. Consider a data set regarding numbers of sexual partners

there are ninty nine "ones" and a single "10,000" so the median is one while the mean is what ?100.99-- Which means that the average is 101 partners??? Means can give undue weight to extreme outliers, particularly in cases where one side of the distribution in bounded as here with the impossibility of a response less than zero. I have seen some types of analysis that completely rejects the data from some outliers but I am not a stats guy so have no further comment on the methodology


But median numbers ARE used-- I see it frequently in local real estate reports in the paper about house sales for exampe . It seems to be viewed as a far better indicator than the average (which can be unduly swayed by ONE 10 million dollar sale) but the mean is also sometimes reported

[KrazyHorse]

There are other medians, by the way. I'm not sure how to report medians from a sample as an estimator of population medians when it comes to discrete variables, or how to report an error on this median. I'm used to dealing with continuous distributions where things are much more intuitively defined.

[Jon Miller]

How is the average a very useful quantity? We aren't looking for the average number of partners, we are looking for the number of partners the 'average' person has. This is going to be the median number of partners, not the average number of partners.

JM

[Flubber]

After 40 votes the median is still 4 (I am disregarding banana votes)-- So after 40 votes there are 38 valid ones and both the 19th and 20th person in the progression said 4

If you count banana as zero as well then we have 40 valid votes and the 20th value is 3 while the 21st value is 4.

For 10 points and the win what is the median?

[Ben Kenobi]

Isn't the error simply: 1/(SQRT (40)), which would be about 0.15 or so?

[Ben Kenobi]

The median would be 3 and 4.

[Flubber]

AS an aside I find the median of this group and the general distribution to be around what I would expect intuitively. I suspect that most people were honest in this poll

[KrazyHorse]

I suspect that most of the 6 votes for 31+ are banana votes

[OzzyKP]

Or are dirty homosexuals.

[notyoueither]

After 40 votes the median is still 4 (I am disregarding banana votes)-- So after 40 votes there are 38 valid ones and both the 19th and 20th person in the progression said 4

If you count banana as zero as well then we have 40 valid votes and the 20th value is 3 while the 21st value is 4.

For 10 points and the win what is the median?

What if we count relations with entities in fur?

[ricketyclik]

I put 31+ but it might be slightly less, I can't remember. 31+ before my current darling who I met at age 24, only her since.

Not that I haven't thought about others of course...

[Dis]

what qualifies as a sexual partner exactly?

To be honest, I'm not even sure.