Any statisticians on here?!

[psychojuice]

I want to work out the probability for a legal case I am working on. Can anyone help?

There are 34 candidates for interview. There are 8 interviewers, working in pairs:

Interviewers 1, 2, 3, 4 and 5 conduct 12 interviews (each, but working in pairs).

Interviewers 6 and 7 conduct 3 interviews. (each, but working in pairs).

Interviewer 8 only conducts 2 interviews. (each, but working in pairs).

The maximum score a candidate can get is 55 and the lowest score is negative 44.

Question was is the probability that interviewer 8, who only interviews two candidates, both scoring 34?

[paulajayne]

So you want us to do your homework?

If we do what do you learn?

[psychojuice]

So you want us to do your homework?

If we do what do you learn?

Homework? I stopped doing homework 20 years ago!

[Crazy council]

I will have a go later or tomorrow, am not bad at numbers..

Looking at it, i think you would need the scores from the others to predict a probility, The confidence in accuracy will be very low becuase of the low sample rate. How far off the mean is 34,

was is the probability that interviewer 8, who only interviews two candidates, both scoring 34?

same probeblity as randem, if you have further data, i will re-run, if you need me to put down the workings ( math ) i will do later
Also, if the averidge scores run between 20 and 44, then its more likley that both would be scored the same.

Updated . 34 seem to be the statistical likleyhood if the questions are considered generic by all applicats. from a score of 1-99 ( 55 to minus 44 ) they have score 68, as closes as damit to 2 thirds, again,


Just my 2p,

[davenewbergr]

How many other candidates are involved and how many interviewers there are is wholly irrelevant. This is not a conditional probability issue: the scores given by one interviewer are completely independent of the scores given by other interviewers, particularly so in regard to different candidates.
There isn't enough information in your question. One interviewer, 2 interviews, and you want to know how likely it is that he would award the same score - it's like watching Wimbledon and seeing two consecutive 230kph aces by Federer and asking 'what was the probability that he would do that?' It's not calculable. (Edit: and of course it's completely irrelevant what other tennis players did for their serves.)
It may be possible to calculate the probability that someone will carry out two unrelated actions. But to do so you would have to assume an enormous number of variables. 1. make a cup of tea 2. switch on the TV. Is he English? Is he blind? Is he a child? Is he a she? ... and so on.
Assume the person makes a cup of tea, and switches on the TV. How do you subsequently calculate the probability that he would have done that? In one very meaningful sense, the probability is 1. The only way you could retrospectively calculate the probability is, again, to make assumptions about the person involved, and then of course the probability data is useless because if any one assumption is incorrect, the probability is wrong and has provided you with no new information.
Even if you could calculate some sort of probability the data set (1 interviewer and 2 interviewees - 2 data points) is too small to draw meaningful conclusions. Even all the data points (68) is hardly enough for meaningful conclusions.
Crazy Council's question 'How far off the mean is 34' is misleading; by definition some data points must be away from the mean, and the probability that any two data points will be a given distance from the mean does not provide meaningful conclusions concerning the likelihood that two data points would be the same.
It might be possible to draw some conclusions where all 8 interviewers interviewed the same 2 candidates, especially if one interviewer's scores were markedly different - but even then the data set remains too small for meaningful results. It's late and if I'd written this at 4pm it would have been easier. If I was a judge you would get absolutely nowhere if you invited me to reach conclusions with regard to unlawfulness from evidence of 1 person carrying out 2 interviews.

[Crazy council]

lol, if i designed the tests, i would make the score out of 120127.5 , lol, becuase i think the statistical averidge ( =*2 /3 ) is my favourate number on a hot sunny day.
But then again, i dont think they would let me ,lol

[davenewbergr]

Imagine trying to prove that a die was loaded from only two rolls of the dice. The thesis is that one of the dice - you don't know which - is tweaked to land on 6. The first throw is a 3 and a 6 and the second throw is a 4 and a 6.