Statistics training...

[Ollie123]

---

Members don't see this ad.

Hi everyone,
For those of you currently in programs, how comfortable were you with statistics going into your doctoral programs? Did you feel your undergrad training was adequate to prepare you or were you overwhelmed at first by grad-level stats? If so, how hard was it to adjust?

I ask because I feel I am seriously behind-the-curve on stats. The undergrad stats course I took was quite honestly, just terrible. I taught most of it to myself because the professor just confused me(and most of the class it seemed). Most of it was learning to do calculations by hand, and not actually learning what the statistics mean. In other words, I could generate P values, R values, etc. if I was given a table, but I'd have no idea what any of it meant.

Anyways, I'm REALLY interested in stats (I know, I know, I've embraced my weirdness though😉 ), and I feel like I need to "catch up" a bit before I start grad school. I'm working with a grad student right now sitting in on her dissertation stats, but she is using HLM so this is some pretty advanced stuff and I feel like I'm missing most of the information in between the BASICS and this stuff.

Does anyone have any good texts to recommend that might just help me develop a good overview of stats? Mainly, what the different commonly used tests are, how to interpret them, things like that. Info on performing these tests in SPSS or SAS would be good too. I'm a pretty mathematically-oriented person, but I have not had as many opportunities to learn this stuff as I'd like so I've decided to take the initiative and teach myself some before I go to grad school so I'm not walking in completely blind. Any suggestions at all you have on what might be good books or resources to look at would be appreciated.

 

[Alicorn]

Sounds like you were peeking into my head! While you did request for doctoral students' responses and I'm just a lowly post-bacc'er, I couldn't help but empathize with your situation. My undergrad stats class was horrendous too and I ended up teaching myself and my friends the basics to get through. Now, I really want to learn stats as well as SPSS and SAS.

I had bought two books about SPSS that seem rather thorough though I am looking to get some straight stats books or even take some additional classes. The books that I nabbed were:

Discovering Statistics Using SPSS (Introducing Statistical Methods S.) (2nd Edition) (Paperback) by Andy Field

and

SPSS Survival Manual (Spiral-bound) by Julie Pallant

The former actually has humor and some rather hilarious examples. Hope that these will be useful!

 

[Ollie123]

Thanks! I will definitely look those up.

I normally don't complain about professors, but my stats class was just ridiculous. Attendance was less than 10%. When there are 15 people showing up on a regular basis for a 300 person lecture you know something is wrong. He had the gall to complain about attendance too, despite the fact that he actually HINDERED my learning of statistics. I had a C until I stopped going to class and looking at his online notes and just derived my own damn formulas and taught myself. I ended up with an A at the end, though I still don't feel I learned a thing. He was "Encouraged into retirement" after that semester though, so at least he isn't teaching anymore.

I used to hold review sessions in the dorms for my friends in the class

That's really sad, because I didn't feel like I knew the material at all either, but I was at least able to clarify things somewhat for the people that made the mistake of listening to the prof.

 

[paramour]

Seems to be stats night--I just finished ordering some SPSS programming books. I've not used SPSS in many, many years, particularly since our program taught SAS for the last two. Unfortunately, however, I'm using SPSS for my thesis analyses since I could easily buy SPSS for personal use where SAS is practically impossible to get ahold of . . .

Anyway, for SAS, try Step-by-Step Basic Statistics Using SAS: Student Guide or The Little SAS Book: A Primer.

SPSS really varies on if you're looking for something to do the "point and click" drop down menus or if you're interested in the programming language (i.e., learning the actual syntax commands) behind it. SPSS Reference Guide (ISBN 0131778587) is an excellent manual for this, but it may be a bit difficult to find. You may want to try places specializing in used or out-of-print books.

G'luck!

 

[Ollie123]

Thanks!
At this point I'm looking to learn both. I assume point and click is probably an easier way to get started, but eventually you need to get into syntax to do more advanced work. Does that seem true?

I have a slight background in computer programming as I was a comp sci major before psychology. I'm not great, but computer code doesn't scare me as it does some psych people I know so hopefully that will work to my advantage.

 

[psychanon]

You will probably be fine. You probably know more stats than you think. Grad stats courses don't assume any more than the basics that you learn in undergrad. They even review the basics (starting at the bottom with mean, SD, etc.), although they do it pretty fast. You'll probably find that most in your cohort don't remember much from undergrad either (esp. those who have been out for a while), so you can work together.

As for books, there's a book called something like the Cartoon Book of Statistics which explains things in a kind of fun way.

I'd find out what program your future lab tends to use before you spend too much time learning it. I'm in my 3rd year, and I've never used SAS in grad school. I mainly use SPSS, and when it can't do what I want to do there's always mplus, amos, hlm, etc. (but definitely don't worry about those yet).

You can do pretty much anything in SPSS without using syntax, but using syntax is much, much easier, especially for doing repetitive analyses. Syntax has the huge advantage of being savable, so you can re-run analyses on different datasets, look back at how you constructed variables, etc. The good news is that learning syntax is extremely easy. Just press "paste" in the data analysis windows. The syntax for the analysis will be pasted into a syntax window, and then to run it just highlight it and click on the triangle icon that looks like a play button. It's easy to figure out how to modify the syntax to use different variables, etc.

 

[Sorg1123]

wow, there seem to be quite a few of us in this situation. (what are the odds of that?) My undergrad stats courses weren't too pretty either. I did fine in them, but I really didn't learn much. At least I took in enough to know that I didn't have a chance in hell of winning that jackpot last night. I bought a ticket anyway.

SDSU/UCSD requires students to take a "statistics screening" exam. It's been a while since my last stats test (4 years), so I'm going to buy a stats book on half.com and study up from time to time before the big test. There are some good web sites to check out:

Statsoft offers an electronic textbook:
http://www.statsoft.com/textbook/stathome.html

Hyperstat is pretty useful too:
http://davidmlane.com/hyperstat/

 

[Therapist4Chnge]

You will probably be fine. You probably know more stats than you think. Grad stats courses don't assume any more than the basics that you learn in undergrad. They even review the basics (starting at the bottom with mean, SD, etc.), although they do it pretty fast. You'll probably find that most in your cohort don't remember much from undergrad either (esp. those who have been out for a while), so you can work together.

Pretty much, though the review will probably just be a day. Stats seems to be one of those things that takes awhile to sink in...so review it a few weeks before your classes, and the adjustment will be much more manageable. I learned on SPSS 6 or 7 IIRC in undergrad...and I think they were up to 11-12 when I started my stats.

So yeah....start early, and don't stress too much about it.



-t

 

[Ollie123]

Thanks all.
I just dropped quite a chunk of money on some stats and a few emotion textbooks I wanted to buy anyways (one written by what is likely to be my future mentor! ). Figure if nothing else, anything I can add to my reference library is a good thing.

Thanks for alleviating some of my fears about it too....sounds like I'm not the only one coming in with a poor background in stats. Hopefully some practice over the summer will put me in good shape for next year.

 

[paramour]

If anyone is interested in another SPSS/stats book, try Fundamental Statistics for Social Research: Step-by-step calculations and computer techniques using SPSS for Windows, by Duncan Cramer. I typically do not care for most SPSS books I've reviewed, but this one is a nice review of statistics in general while also providing one with the instructions for how to use SPSS for the procedure in question.

 

[Lunabin]

Reading and Understanding Multivariate Statistics (Paperback)
by Laurence G. Grimm (Author), Paul R. Yarnold (Editor)

Reading and Understanding More Multivariate Statistics (Paperback)
by Laurence G. Grimm (Editor), Paul R. Yarnold (Editor)

Discovering Statistics Using SPSS (Introducing Statistical Methods S.) (2nd Edition) (Paperback) by Andy Field (Author)

These 3 got me through grad school and dissertation (along with some good ol' fashioned consultation). Just before I started my Ph.D. program, I read through an undergrad stats book. My school gave us a 'stats proficiency' test to be sure we knew enough to start the class (anyone who didn't pass took a short refresher course)... very helpful!

My professor really started at the beginning and spent a couple of weeks going over the basics (mean, median, mode, z-score, types of errors, basic research designs, t-tests, etc.).

Don't panic too much - you're going to take a class, so hopefully the prof. will actually teach you the material. My classmates and I also did lots of study groups throughout all of our stats courses - - this was a lot of help, and was the start of a lot of friendships!

P.S. - Stats go well with wine and cheese on a Friday night! Good luck and have fun!