I'm designing an experiment that involves a subject attempting to determine various attributes of a playing card drawn at random from a standard 52 card pack.
The experiment is in four phases:
I'm planning to do several tests of n presentation for each phase with each test card being drawn from a full pack of 52 cards.
The probabilities for success for each phase for a random quess being correct are 1 in 2, 1 in 4, 1 in 13 (4 in 52) and 1 in 52.
It's been a long time since I did much with statistics and, through not using it, I've lost it.
What I want to do is find the 95% and 99% confidence limits for n tests for each of the phases. Once I have these values I can get some idea of what a reasonable number of tests for each trial is. Too few and it's going to be hard to separate a possibly real effect from random. Too high a value and the subjects are going to find the process too tedious.
What I'd like are the formulae for calculating the 95% and 99% confidence limits for n tests for each of the phases. I can probably handle the first case using the binomial distribution but the other three are defeating me at the moment.
I appreciate any help or pointers that anyone is able to give me. Thanks in advance.
By the way, this is not an ESP experiment. Those have been done to death for the moment.
The experiment is in four phases:
- Attempt to determine the colour, red or black.
- Attempt to determine the suit, hearts, clubs, diamonds or spades.
- Attempt to determine the face value ace-10, jack, queen or king
- Attempt to determine the full details, e.g. 5 of clubs.
I'm planning to do several tests of n presentation for each phase with each test card being drawn from a full pack of 52 cards.
The probabilities for success for each phase for a random quess being correct are 1 in 2, 1 in 4, 1 in 13 (4 in 52) and 1 in 52.
It's been a long time since I did much with statistics and, through not using it, I've lost it.
What I want to do is find the 95% and 99% confidence limits for n tests for each of the phases. Once I have these values I can get some idea of what a reasonable number of tests for each trial is. Too few and it's going to be hard to separate a possibly real effect from random. Too high a value and the subjects are going to find the process too tedious.
What I'd like are the formulae for calculating the 95% and 99% confidence limits for n tests for each of the phases. I can probably handle the first case using the binomial distribution but the other three are defeating me at the moment.
I appreciate any help or pointers that anyone is able to give me. Thanks in advance.
By the way, this is not an ESP experiment. Those have been done to death for the moment.