Morning, Guys!
Let's do the easy ones first. Steve! Still wasting your time?

Sax, no need. As you will see in a moment.
DMF! Thanks for the lesson. My brother is a Maths teacher and specialises in Pure Maths and Statistics. (He doesn't like Applied Maths much) I'll be sure to email him with your post. Yes indeed, the mean, the median (Medium?) and the mode. Not many of us know about those different averages.
But now let me tell you about the voting system. I conducted my poll along the lines of how a British parliamentary poll would have been conducted. On polling day, which in this country is always a Thursday, the polling stations open at 7am and close at 10pm. School buildings are often used as polling stations which is great for the kids because they get the day off. The senior person in charge at the polling station (PS) is known as the Returning Officer (RO). Each person coming into the PS to vote is given instructions on how to vote for their chosen candidate. These instructions are to vote for one and
only one candidate by marking a X in the space next to the name of the candidate of their choice. This is done in the privacy of a voting booth - no-one looking over the shoulder etc. The ballot paper is then placed in the ballot box for the ballots to be counted later.
Typically, there will be three candiates representing the three major parties - Labour/Conservative/Liberal Democrat (Lab./Con.LD). There may be additional parties such as the Raving Monster Loony Party - I'm being serious. So there could be a few more than three, but there need not be seven.
Some points to note:
- If a voter votes for more than one candidate, or marks the ballot paper in any other way other than a single X in ONE of the designated boxes, the vote is void and is deemed to be a "spoilt ballot paper". So, for example, whereas I might be tempted to write "Piss off, Blair" on my own ballot paper at the next election, I must show restraint or else my vote will be discounted.
- The polls are open only for those 15 hours. The RO does not have the option of keeping the PS open if voter turnout has been low, in order to allow a few more people to come along.
- The incumbent party is not empowered to instruct the RO to close the poll when they feel enough votes have been cast and/or that the votes collected thus far are the ones they need to get the result they want.
- When counting the votes, the RO is not empowered to sift through the spoilt ballot papers, making his own interpretation of "what the voter really meant" if more than one vote had been cast, or if the X spanned two boxes. (And that is precisely what your stance has been throughout this entire debate - tinkering with dubious votes, and making your own interpretation as to where to place those votes according to your agenda. Thus it also becomes clear WHY the RO is not empowered to do this.)
So as you can see, DMF, those 2/3 votes equate to an X which spans more than one box in the ballot paper. And 2.5/3.5 would equate to marking Lab./LibDem or maybe Lab./Con. on the paper. Which is it?! How can the RO be sure which
single vote the person would have made? They can't. And the only fair way to deal with spoilt ballot papers is to discount them. And that's what I did with those 2/3, 2.5 etc. votes. Maybe I should have said I would close the poll after 24 hours, but the voting had trickled off by then anyway.
Basically, we may measure variables at four different levels -- dichotomous, nominal, ordinal (or continuous categorical as I called it, an apt if incomplete description), and interval/ratio. You erroneously believe that the rules that apply to dichotomous variables (yes and no, for example) likewise apply to ordinal variables for polling purposes when they do not. The example you gave of the beers is nominal -- that is, it is categorical without degrees of difference between categories. Measures such as religion or race fall in this measurement level. Your poll employed an ordinal variable with values ranging from "all the time" to "most of the time" to "none of the time." Interval and ratio measures are truly continuous and consist of things like income and age, though we may turn interval measures into ordinal ones for polling purposes.
Yes, very impressive. But don't you think that's way over the top for a guy who just wants to get a broad feel for what a relative handful of his playmates do in a freaking *game*?


<-- sometimes one is not enough.