statistics

[loesch]

I was going through the Idiot's Guide to Statistics and I noticed that they gave a different standard deviation formula than did some of the other books I was using. Did anyone else notice the same thing? I am scheduled to take this next Wednesday. Also if anyone has anyother information on the exam, that would be greatly appreciated.
John

[snazzlefrag]

I was going through the Idiot's Guide to Statistics and I noticed that they gave a different standard deviation formula than did some of the other books I was using. Did anyone else notice the same thing? I am scheduled to take this next Wednesday. Also if anyone has anyother information on the exam, that would be greatly appreciated.
John

Hi John,

There can be several different ways to calculate Standard Deviation depending on what data you are working with (just as there could be several different ways to calculate "12", depending on what type of data you are using....Eg, 2*6, 3*4, 2+10, 5+7, 15-3, etc.).

For example, if you know the variance, the formula for SD is as simple as "the square root of the variance". If you don't know the variance, then the formula for SD will have to be the formula for variance, with a square root symbol in front of it. Make sense?

The "square root of the variance" is the formula you should MEMORIZE for the exam. Seriously, ignore any formulas that make your nose bleed, okay?

In fact, you will actually earn more points if you are just familiar with the basic QUALITIES of SD (ie, the properties of a normal bell curve). There will most likely be quite a few questions which drill you on your knowledge of this information.


Memorize this:
(Properties of Normal Distribution/Bell Curve)

[0.15%|2.35%|13.5%|34%|<---0--->|34%|13.5%|2.35%|0.15%]

Notice that...
1SD=68% (34%+34%)
2SD=95% (13.5%+34%+34%+13.5%)
3SD=99.7% (2.35%+13.5%+34%+34%+13.5%+2.35%)

Know that as the size of the sample 'n' increases, the SD will get smaller and smaller. Or obviously, as the size of the sample 'n' decreases, the SD will be larger and larger (This is known as the "Central Limits Theory").

Recognize that if the SD is small, the bell curve will look narrow. If the SD is large, the bell curve will look wide. Eg, they might show you several bell curves and ask, "which of the following bell curves reflects the smallest Standard Deviation?" Your answer will be the bell curve that looks really narrow.

If you can memorize everything I just wrote, you will be prepared to answer a lot of possible questions that may appear on the exam.

I hope this info will encourage you, rather than discourage you. It's hard to explain in writing, and I know how intimidating stats can be until you grasp it. But hopefully you understand what I am saying, and realise that the formula for SD is very simple, and knowing the CONCEPT of Standard Deviation (and the properties of a normal distribution/bell curve) is much more important for this exam than being able to crunch lots of numbers.

Do Great,
Snazzlefrag

[loesch]

The information was very helpful. I feel like I know the variance, SD and the normal bell curve pretty well. I was wondering how in depth should I study? Should I go over the chi-square test and variance analysis or am I just wasting my time. Thanks for all your input.

[snazzlefrag]

The information was very helpful. I feel like I know the variance, SD and the normal bell curve pretty well. I was wondering how in depth should I study? Should I go over the chi-square test and variance analysis or am I just wasting my time. Thanks for all your input.

While you will need to know some of the more complex concepts, and be familiar with various formulas, it is a thorough knowledge of the BASICS which will see you through this exam. I can't stress this enough.

If you can master all the basic concepts, terms, formulas, etc., you CAN pass this exam even if you don't fully understand all the advanced concepts. You may not score in the 70's, but you will pass.

With that in mind, any of the more complex material you can become comfortable with, will no doubt help your score. Just make sure you don't invest too much study time on the complex stuff until you have fully grasped all the basic stuff. Also, don't let the harder concepts scare you, panic you, or freak you out, to the extent that you end up getting confused or overwhelmed by even the basic stuff.

--
You should learn how to calculate variance and, more importantly, understand the concept of what variance is. Chi-square wasn't included on my version of the exam at all. Don't sweat it!

Hope that helps,
Snazzlefrag

[Old Rusty Pipe]

Just a quick note on Chi-square. On my version of the test I did have one or two questions concerning it. I'm pretty sure that I got both of them right, and the only thing I reviewed on Chi-square was the 1 page summary of the topic in the Complete Idiot's Guide. Its not essential, but it wouldn't hurt to skim over it just a little bit. As was already pointed out, variance and SD are a big chunk of the exam. Probability is also another important topic you want to be comfortable on. Finally, know your Null Hypothesis. Know it well, I had something like 10 to 15 questions all concerning it.