Good resource for learning statistics?

[egFace]

While I was an undergraduate, I took a statistics class. However, the professor was absent half the time due to illness, and she did not teach us anything beyond t-tests. I am now a first-year PhD student, and I am currently taking an intermediate stats class that is supposed to cover everything up to multiple regression. But my professor is not qualified to teach, and the book does not cover any of the theory or math behind the statistical tests. I desperately want to achieve a deep understanding of statistics. What is a good book that will dive into the theoretical aspects of stats?

[geographyrocks]

I have An Introduction to Statistical Problem Solving.  I guess it's an okay book.  It definitely covers the math behind the tests as I learned how to do stats by hand before being handed a computer program that will do all the hard work for you. 

I really get more out of reading papers that use the methods I'm interested in or googling methods than reading books on the subject though.  It helps me to see how they're applied in research rather than just learning the method itself. 

And I sympathize with you on the poor stats classes.  My first class was taught by a professor who just didn't give a crap (and was always extremely surprised when I knew all the answers.  Me. A girl. But that's another story.)  I had a very competent professor for my second class except he wasn't very good at teaching the material.  He also had a problem with girls knowing the answers.  Of course, that's what you can encounter when you go to school in the Midwest. 

[spunky]

 the book does not cover any of the theory or math behind the statistical tests. I desperately want to achieve a deep understanding of statistics. What is a good book that will dive into the theoretical aspects of stats?

 

may i interest you in one of the classics, Casella & Berger's Statistical Inference?

 

it's a pretty awesome book

Edited November 27, 2014 by spunky

[juilletmercredi]

Why would you say that your professor is not qualified to teach your statistics class?  There's a difference between not being a good teacher and being actually not qualified.

 

If you are interested in going deep into the theory and math behind statistics, you should investigate whether your university will allow you to complete a concurrent MA in statistics or a graduate certificate.  Many universities do allow this, but they don't widely advertise it - it's something you have to ask about.  Even if they don't, you might be able to cobble together your own, although if you haven't taken multivariable calculus yet you will have to do that.

 

Psychological statistics classes are not really going to go into depth about the theory because most psychologists aren't interested in that; they want to know the applied stuff.  (Quantitative psychologists being the exception, of course, spunky )  If you really want to know the math and theory behind stats, you may be interested in taking calculus I-III (which is through multivariable) and a linear algebra class if you haven't already.  I haven't taken calc III, but I took calc I and II and linear algebra and that's what helps you understand the math.  

[Applemiu]

Why would you say that your professor is not qualified to teach your statistics class?  There's a difference between not being a good teacher and being actually not qualified.

 

If you are interested in going deep into the theory and math behind statistics, you should investigate whether your university will allow you to complete a concurrent MA in statistics or a graduate certificate.  Many universities do allow this, but they don't widely advertise it - it's something you have to ask about.  Even if they don't, you might be able to cobble together your own, although if you haven't taken multivariable calculus yet you will have to do that.

 

Psychological statistics classes are not really going to go into depth about the theory because most psychologists aren't interested in that; they want to know the applied stuff.  (Quantitative psychologists being the exception, of course, spunky )  If you really want to know the math and theory behind stats, you may be interested in taking calculus I-III (which is through multivariable) and a linear algebra class if you haven't already.  I haven't taken calc III, but I took calc I and II and linear algebra and that's what helps you understand the math.  

 

I was wondering if a Times Series statistics class could be useful as well for a future psychology researcher. What do you think, from your experience? Usually Times Series is more targeted for finance, but I thought maybe it could be useful for neuroimaging.

[bakalamba]

I had a great graduate stats instructor, but the textbook was also very helpful, Statistics: Unlocking the Power of Data by the Lock family statisticians. Online Stat Book is pretty great, covers most of the material, although it doesn't always go into the depth necessary.