Support NeoGAF

[Escape Goat]

I'm trying to figure out this statistics problem that deals with confidence intervals.

Two large employers are considering jointly offering child care for their employees. As a part of the feasability study, they wish to estimate the mean weekly child care cost of their employees. A sample of 10 employees who use child care reveals the following amounts spent last last: 107, 92, 97, 95, 105, 101, 91, 99, 95 and 104.

Develop a 90 percent confidence interval for the population mean. Interpret the result.

Since the standard deviation isn't given I have to use the Sample divided by the square root of the population right? Gaaah! I hope someone out there knows something about stats!!

 

[Mr Gump]

Fucking HATED that subject.

 

[ManDudeChild]

The answer is 42.

Spoiler

don't know answer

 

[Ferrio]

If you asked me 2 years ago I could of told ya. CI's were the easiest thing in the whole class to me.

 

[Foreign Jackass]

I remember getting excellent marks in this class. That's all I remember from it. :lol

 

[Escape Goat]

At least I know I'm not alone in my hatred of the subject.

 

[darscot]

Sadistics is not math. Math has clear rules 1+1=2. Statistics is more like this 1+1= 47 based on our sample this could possible be correct 19 times out of 20.

 

[fallout]

Sadistics is not math. Math has clear rules 1+1=2. Statistics is more like this 1+1= 47 based on our sample this could possible be correct 19 times out of 20.

Well, in reality, there is no set answer to any kind of measurement. You're forced to take into account errors because they're there, no matter what you do. Tell me that your computer desk is 106.5 cm long and I'll ask you how many atoms that is. Give me an answer there and I'll ask you how big an atom is. You can only be so accurate.

In fact, how people judge your data will be based on how well you calculated your error. Stats does take it some extremes, but it's absolutely necessary in any scientific field.

 

[darscot]

LOL, You sounds exactly like my Statistics Prof. I guess it's all part of the whole statisitcs deal. It's about math and numbers no one really cares about. Most people ask for a desk 162cm wide and they could care less how many atoms are in it.

 

[fallout]

Heh, well clearly that's an extreme example, but what I said about the error calculations of your measurements being one of the most important things of a piece of research is true. If your levels of accuracy are satisfactory, then we'll talk. My desk example isn't very good, but if the size of a desk really mattered, wouldn't you care if I could measure it within ± 50 cm or ± 0.5 cm? Also, what if I told you that I could measure it to within ± 0.5 cm, but in actuality, my error calculations are wrong and I can only measure it to within ± 50 cm? Such things are usually debated about and with good reason. A measurement is meaningless without error in a scientific paper.

As for your comment about 1 + 1 = 47, that's a little different. Based on our definitions of 1, +, = and 47, I would have to say that 1 + 1 != 47. Unless of course, you have some fucked up version of one or more of those characters, then I'd just have to call you a weirdo.

Anyway, I know what you're getting at, but I'm just trying to illustrate that it does serve a purpose. Any theoretical physicist or pure mathematician would probably just smile at your post and chuckle a little, so it's not like your alone (although, for different reasons I assume).

 

[darscot]

I was hoping everybody that read my post would chuckle a little it was meant as humor in the form of massive exageration. If yours was meant as humor it's far too dry for me to get. As this is clearly your field why don't you help Teh Hamburglar with his homework. I'm just here for stupidity and moral support for those of us that did not enjoy the subject.

 

[fallout]

Heh, sorry. Like I said, I do understand what you're getting at, I just wanted to illustrate the other side a little. It wasn't meant to be rude or anything, so forgive me for that. I don't think many people really enjoys stats (I certainly don't), but then, most people don't particularily enjoy doing laundry. I'm sure you could argue that laundry is slightly necessary to some people.

And honestly, I'd like to help, but I really don't know much about stats other than what I learned from my light physics background. So the line Develop a 90 percent confidence interval for the population mean. Interpret the result. throws me for a loop since I know those words, but not in the order they're in.

 

[SoVos20]

You use the sample standard error (estimate of the standard error) which is the sample standard diviation (estimate of the standard diviation) divided by the root of the number of samples. Remember since it is only 10 to use the t test statistic.

 

[SoVos20]

Solution

95.3875, 101.8125

98.6 (+/-) 1.383*1.7525

 

[Escape Goat]

Thanks, SoVos!!

Thanks for the input, fallout!