[Marauder]

Killim then, thank you for the info.
Edit: Well, I could also create a thread in this very subforum.

[Wolfman]

For a sample of n = 45, find the probability of a sample mean being greater then 551 if μ is 550 and σ is 3.7.

My problem is mainly that I don;t know what formula I need to be using. For whatever it matters, the book claims the answer is 0.0351.

[Kaiserschmarrn]

The sample mean has a Normal distribution - in your case N(550, 3.7) or N(550, 3.7^2) depending on which formula your class uses. You need to standardise the sample mean and look up the probability in the Normal table.

[Nets]

For a sample of n = 45, find the probability of a sample mean being greater then 551 if μ is 550 and σ is 3.7.

What is the underlying distribution from which the sample is drawn?

[Wolfman]

That is literally all of the information I was given to work with.

[Nets]

Well then who knows!

The sample mean is not generally normally distributed, but I guess for n = 45 it is an okay approximation. Listen to Kaiser.

[Wolfman]

I finished the rest of the section and that problem didn't come up again and it wasn't on the test, so I'm thinking I'm just going to ignore it.

[Kaiserschmarrn]

The sample mean is not generally normally distributed

True. My bad for not mentioning that it's an approximation.

I finished the rest of the section and that problem didn't come up again and it wasn't on the test, so I'm thinking I'm just going to ignore it.

What class/subject are you taking? That's actually a pretty simple problem and it's probably a good idea to understand it.

[Wolfman]

A basic stats class. I can do the work for this normally, I'm just missing information, or something.

[Kaiserschmarrn]

Maybe my earlier post wasn't too helpful.

The standardised sample mean is



If you look up 1.81 (the z-value) in the Normal table you find the probability that the sample mean is less than 551 which turns out to be 0.9649. Since you need the probability that the sample mean is greater than 551 - i.e. the complement - you have to subtract the result from 1.

1 - 0.9649 = 0.0351.

[Wolfman]

Hmmm... Thanks. Not sure what was throwing me off, seems pretty obvious now.

[Wolfman]

Alright, I give up. For the love of god, someone explain to me what the hell I'm doing wrong.

[Kaiserschmarrn]

Well, I'm a bit late for this, but your mistake was to use the level of confidence given to you - i.e. C=99% - as your alpha. When you calculate and look up the chi^2 variable you have to use alpha=0.01 instead of C=.99. Also, if I remember correctly, the chi^2 distribution is not symmetric and you therefore cannot simply use the two formulas (1-alpha)/2 and (1+alpha)/2.

Penn State has a very good statistics intro website here. The STAT415 class (bottom left of the page) goes through estimation and hypothesis testing in detail and offers good explanations and examples if you are interested.

[Wolfman]

Well, I'm a bit late for this, but your mistake was to use the level of confidence given to you - i.e. C=99% - as your alpha. When you calculate and look up the chi^2 variable you have to use alpha=0.01 instead of C=.99. Also, if I remember correctly, the chi^2 distribution is not symmetric and you therefore cannot simply use the two formulas (1-alpha)/2 and (1+alpha)/2.

Thanks. Sadly, I realized what I was doing wrong on Thursday in class, a week after the test on chi^2

Penn State has a very good statistics intro website here. The STAT415 class (bottom left of the page) goes through estimation and hypothesis testing in detail and offers good explanations and examples if you are interested.

I think the Penn State online classes were one of the ones posted in the Free Learning Resources section. I might go through it, thanks.

[Kaiserschmarrn]

Thanks. Sadly, I realized what I was doing wrong on Thursday in class, a week after the test on chi^2

If it's any consolation, the definition of significance levels and construction of confidence intervals seems convoluted and unintuitive to almost everybody at first. Most people memorise it, scratch their heads and forget it immediately after the test. If they ever need to do statistical testing later they generally use statistical software as a black box. I can't really fault them for that since one intro class is not enough to solidify the concepts.

[Wolfman]

I'll be taking the advanced stats classes too, probably.

[Kaiserschmarrn]

^ Good luck.

After the initial confusion, statistics can actually be quite enjoyable. It's definitely worth the effort, in my opinion, and since data analysis is used extensively in more and more fields nowadays, you will be way ahead of the majority of students.

[Moonchild]

Good luck with taking an advanced course in statistics, I just had a test in Stochastic Processes today, which is a continuation of Probability and Statistics course, and it was hell.

[Wolfman]

I'm a math major, who is not actually good at math, and I will get through this degree with a combination of determination and stupidity.

[Seeker8]

Found a thread for my silly questions.

Is: (-27)^(-1/3) = -1/3 ?
That's the answer i got by working it out but google says the answer is:
0.166666667 - 0.288675135 i (don't know what the "i" means)
but my calculator says -1/3.