There comes a time in every young woman’s life when she must confront her destiny and actually write her damn thesis.
Unfortunately, mathematics and I parted company years ago and my understanding of statistics is beyond dismal. I’ll be sitting on a huge pile of raw data and no clue of what to do with it.
So, mathematically inclined Dopers (which I assume is everyone but me), please recommend a good basic primer of statistics that will make everything clear, or at least clear enough that I won’t make an ass of myself to my thesis committee.
Thanks!
Robin
(Mods, please move if you deem appropriate. It’s in CS because I’m asking for book recommendations.)
If you’re not mathematically inclined, I’m not sure if this is your best bet. But I’ve worn the binding on my copy of Phiip R. Bevington’s Data Reduction and Error Analysis for the Physical Sciences. Not complete, by any means (as I’ve learned over the years), but plenty of useful basic stuff and a good intro for anyone having to reduce data and analyze results.
" How to lie with statistics" is a classic. You read it in half a day. It is well written, a fun read, and it teaches all the basics of applied statistics.
The Schuam’s Outline of Statistics is actually pretty good with the added advantage of being cheap. I don’t know how it would be if you were really unfamiliar with statistics, though. There are a ton of stats textbooks out there that are more or less equal. Many of them are Statistics for _______ where the blank is, say, business applications, medical applications, social sciences, physical sciences, biology, etc. They all contain the same content but the example problems tend to be taylored to a particular audience, which can be helpful for someone struggling to understand the concepts.
On preview, I see someone mentioned the ‘How To Lie…’ book. That would also be good - especially if you’re trying to figure out how to best summarize your data. I don’t have my copy here, but I don’t think it has much about how to actually go about computing things.
Is there a community college near you that maybe you could audit an Intro to Statistics class, or if you talk to the prof. they might even let you sit in on?
I’m actually pretty good at math and have taken a lot of math courses, but statistics and I didn’t get along! I would have had a hard time trying to teach myself statistics from reading a book. Other people I’ve talked to found it easy, so maybe it was just me.
Maybe if you post what your trying to do, someone here can help you get started. I have a couple statistics books and I’d be willing to give you some help if I can.
Someone on another message board I post at used to give me little tutoring sessions when I was pulling my hair out and it really helped me.
Most of the Schaum’s Outlines are pretty good, at least for fundamentals and problem solving. I haven’t reviewed The Cartoon Guide To Statistics but the ones for physics and chemistry were actually pretty good.
Sorry about the delay in responding; I’m crazy busy right now.
I’m writing a thesis on communication using content analysis. I know how to gather the data, but I’m not sure what to do with it once I have it. I’m not sure which tests will give me which information. I’m looking for a basic reference guide that will help me figure that out.
In general, statistics textbooks are good references, but if you need to work in a specific field, you’re far better off getting a text in that field. If you don’t already have a content analysis book, I’ve seen this one recommended in reviews for a couple other books, so it might be worth checking out.
That’s not to say that having a statistics book is a bad idea. I’ll second the Schaum’s series, as they’re very cheap and walk you through solving a lot of problems. And if you have any specific questions about how to get information out of data, GQ is a great place to have those answered.
As a Prof who mentors master’s level students, I am curious if you had to propose your thesis yet. If so, didn’t you ahve to propose your analyses? If you did not yet propose, how do you have data to analyze.
Stats are generally simple and straight-forward. Certain stats are appropriate for for cartain types of data (nominal/ordinal/interval/ratio) and for different designs (within-subjects/between-groups).
I recommend you first define your variables, and operationalize them, then meet with your committee members and discuss what it is you are hypothesizing. The stats should flow from each hypothesis.
For example:
Hyp: I hypothesize that males will show higher math scores than females.
IV: Gender - 2 levels, Male, Female
DV: Math scores, as measured by SAT math subscale scores
Analysis: Independent samples t-test
I know this is a basic example, but every hypothesis you have should be supported with operationally defined variables, and the stats are simply a matter of what is appropriate.
Simplified:
Multiple DVs = MANOVA
Multiple IVs = Factorial Design ANOVA
Non-equivalent groups = Possible use of covariate
Predicting one variable from others = multiple regression
Predicting group membership = discriminant analysis
etc…
phungi, I’m writing my research proposal right now. Part of what my committee wants is pilot data which needs to be analyzed. I’ve got the hypothesis, and it’s operationalized.
A big part of my problem is that the research course I took was kinda light on statistics and analysis. We’re using SPSS for the actual math and stuff, so I don’t need to know how to use the formulae. I just don’t know which tests to run for what.
Best of luck with your analysis. I’m working on my analysis, myself.
My field is very different from yours, and I’d make a pretty big bet that my analysis will use completely different techniques (I’ve got an ANOVA, a biodiversity index, and an ordination method to get done.)
I don’t use SPSS. I’m using SAS and Canoco. Still, when I was looking for a good text to help me out, I was recommended this one many, many times. Maybe it would be a good resource for you.
Bumping because I have basically the same question that MsRobyn had. I too have a thesis of sorts to write, and I expect it to need some statistical analysis work. I don’t have the time or inclination to read a complete statistics textbook, and I know I can use something like R to do the heavy lifting, but a basic primer on statistics would help me understand what questions to ask R.
I’ll check out the references above; anything else good (book or website) come out in the last 12 years?
Hands down, the best practical guide to scientific statistical analysis is “The Statistical Analysis of Experimental Data”. It has tons of practical advice on how to analyze different data sets for a wide variety of purposes.
MsRobyn, what did you end up doing with your huge pile of raw data? Did any of the books that anyone recommended, or any other books or resources, turn out to be useful?
When I was in grad school, the statistics department had a person (or persons) whose job was to consult with non-statisticians on their data analysis issues. You may want to check with your statistics department (assuming your school has one) and see if they have such a resource.