[HALP] Stats Problem, t-distribution?

T

Teratos

It's been awhile since I took a statistics course.

The goal is to determine if N = 10 is a large enough sample size.

The problem
Code: 
N = [( t sub alpha  *  sigma) / d ]^2

WHERE:
N     = sample size = 10
sigma = standard deviation = +- 3
d     = limit of acceptable error (+- 2 to +- 4)
alpha = significance level = 5% = 0.05

t sub alpha = value of student's t-distribution with (1 - alpha/2)
                confidence level and (N-1) degrees of freedom

I'm confused by the t.

Is the tail (0.975, 9) where (1-alpha/2) = 0.975 and N -1 = 9?
This seems incorrect since tail values are usually much much smaller (like .025 or .0025).

Help me please, I'm tired and confused.
 
Last edited:
Amazing isn't it? I was thinking when I posted this "TI will probably be awake and get a good laugh out of it". I guess I got what I deserved.
 
your tail is just your alpha, and it depends on the question if it's two tail test you would have to add 1/2 the alpha to either end. so if I had an alpha of .1 on a two tail test, I would check both .05 and .95.

In this case, if it is a two tail test. So you would need to check .05/2 or .025 and .975 with n-1 degrees a freedom to get your t-scores to check against what ever your checking (variances, etc). Since you're dividing the alpha by two...I think this a two tail test.

t-test can be .975, .95 if you are doing an upper tail test. Since there are essentially three type of tests: <, <=>, >
 
Last edited:
I-C-E said:
your tail is just your alpha, and it depends on the question if it's two tail test you would have to add 1/2 the alpha to either end. so if I had an alpha of .1 on a two tail test, I would check both .05 and .95.

In this case, if it is a two tail test. So you would need to check .05/2 or .025 and .975 with n-1 degrees a freedom to get your t-scores to check against what ever your checking (variances, etc). Since you're dividing the alpha by two...I think this a two tail test.

t-test can be .975, .95 if you are doing an upper tail test. Since there are essentially three type of tests: <, <=>, >
Click to expand...

Thank you. That clears it up.
 
You must log in or register to reply here.
Back
Top