Good starting points for learning how to do empirical research?

June 21, 2012 11:21 AM

Good starting points for learning how to do empirical research?

I am part way through a PhD program in law (in the UK, if it matters). I have no training in empirical/social science research methods but my supervisor wants me to do quite a lot of both qualitative and quantitative research. I agree that it would be useful but am finding it difficult to get started because I really don't know where to start. My supervisor thinks I can figure it out from books and isn't a lot of help at the moment. I think I need to take a proper course but nothing will be available until September and I may not be able to attend in any case (or may annoy my supervisor by doing so).

Are there online courses that could help me get a proper understanding of "real-world" research methods, so I don't end up with data that is unusable or flawed in some way? Are there good beginner's books that stand alone, without a teacher?

I'm asking anonymously in case my supervisor comes across the question or connects it with me.

posted by anonymous to education (9 answers total)

I had a similar thing happen midway through a history dissertation.

Do you have statistical software available through your university? Here's a free decent guide to SPSS, which is one of the most widely used.
posted by mareli at 11:44 AM on June 21, 2012

I find L. Michele Issel's textbook "Health Program Planning and Evaluation" to be a pretty good resource for qualitative and quantitative research. Mostly health stuff but it applies to other fields, too.
posted by entropone at 11:57 AM on June 21, 2012 [1 favorite]

0. Make sure you have a reasonably clear question of interest before you continue.

1. Read similar research, and try to understand the logic behind it. There must be researchers publishing research that is similar (or shares certain features) with yours.

2. Consult with the statisticians/methodologists at your university. Some universities have special programs set up so that students or faculty members can meet with someone who helps design and analyze experiments. Brainstorm with them about how to answer your question. If they have experience in general social science research, they'll be able to clarify your ideas and shoot down ones that won't work.

3. Get a book on research methods that is used in your field - perhaps the text used in your department - and read relevant parts of it. Look up unfamiliar ideas that you found in the similar research, and try to see how the book explains them. If you can find someone (maybe same person as in 2?) to answer your questions as you go, that will help tremendously.

Make sure that if you're consulting with someone, it is someone with plenty of experience. Sometimes people with very little experience (who have taken the classes, etc) have an rough understanding of things, but may not be very good at explaining the ideas, since they ideas aren't solid for them yet. Good luck!
posted by Philosopher Dirtbike at 2:06 PM on June 21, 2012

Oh, and don't go analyzing data for your first time (whether it is in SPSS or otherwise) without a copilot. You need someone to explain things to you as you go, or you'll end up doing silly things. You want someone that is able to say "Stop that. What you're doing makes no sense," because software will be more than happy to do whatever you ask, sensible or not.
posted by Philosopher Dirtbike at 2:09 PM on June 21, 2012

You might find a book like How to Think Straight About Psychology helpful - it's an introduction to what it is that psychologists are doing when they do research, with some general chapters on things like how to tell if you have a testable, falsifiable hypothesis. Psychological methods are fairly different from say, sociology, but I think it's a type of introduction to empirical research that could be good for getting your head around how to think like this kind of a researcher.
posted by heyforfour at 2:12 PM on June 21, 2012

Psychological methods are fairly different from say, sociology, but I think it's a type of introduction to empirical research that could be good for getting your head around how to think like this kind of a researcher.

I was going to suggest the opposite, that the OP pick up a basic intro text to social research methods, if she/he really has absolutely no background in empirical research. But it's a bit hard to say, start telling the OP to consult with a statistician, or talk about experiments, since that may not at all be appropriate. There is much more to social research than just experiments.

The OP is in a law PhD program - will you even be studying people (individuals? groups? communities?) Or are you looking at texts (laws, regulations, policies, speeches, decisions)? It's easier to suggest a starting point if we know what the general area of inquiry will be.

Feel free to memail me if you want.
posted by DiscourseMarker at 3:32 PM on June 21, 2012

I like Creswell's Designing Mixed Methods book. Covers qual and quant.
posted by k8t at 8:51 PM on June 21, 2012

This question is too big to answer. There are thousands of research methods out there, and the ones you pick depend on your research question. If you told us more about your research question, we could help you find a method appropriate for it.

Strauss and Corbin is a very detailed and very readable qualitative methods book. Open coding, closed coding, and microanalysis are all frequently used in the social sciences. Note that although it's called Basics of Qualitative Research, Strauss and Corbin is actually more of a mixed methods book. They cover the transition from Qualitative to Quantitative, rather than something purely Qualitative like a case study, where you're basically just telling a story. But if you need a framework, something that tells you "what to do" just to get started, Strauss and Corbin is good at that. Just keep in mind that it's the beginning, not the end.

But the best way to learn to do research is to do research. Courses are generally about right answers. Research is about not knowing the answers. You can't really learn to deal with the uncertainty of research without practicing it. Doing real research is not about memorizing and following methods, it's about making judgement calls. You have to try things out, and you have to make some mistakes before you learn how to get it right. Even if you have studied particular methods in detail, you will never understand the need for the quirks of particular research methods until you learn from your own mistakes.

For example: You can study how to design an interview protocol, and even design one, but you will never learn to judge how much flexibility and options you need to build into your protocol until you have an interview tank on you because the interviewee said something completely unexpected.

So no, you're not going to figure it out from books. A course will help (as long as its a course that provides feedback on your mistakes, not an online slide show), but you're not going to figure research out from a course either. The only way to truly learn research is to get your hands dirty. Fundamentally, that's what graduate school is.

This is what your supervisor is trying to do: get you out there. Maybe he/she is not the best way for you in particular, but it does need to happen, so give it a try*. Design something and show it to your supervisor for feedback. Don't worry about making mistakes. Your supervisor won't let you do anything truly important to begin with*. Remember that perfect is the enemy of good, and more importantly (for a Ph.D.) perfect is the enemy of done.

So the basic method of learning to do research is to A) have a question, and B) try to answer it. A informs B. If you're really stuck on not knowing what to do, it's probably because you haven't focused on A enough. Really spend some time on defining what you want to learn from the research. That's Philosopher Dirtbike's step 0, and it's incredibly important. This is also why you're not going to get very useful responses in this AskMeFi. Without knowing what your research question is, we can't tell you where to focus your efforts.

tl;dr: just read the bold parts.

*note: DO NOT actually do any research involving human subjects without clearing the method (and specific protocol) you intend to use through your supervisor first. I don't know what the UK laws on human subjects are, but I suspect there are some.
posted by yeolcoatl at 5:16 AM on June 22, 2012

Take an epidemiology course
posted by tiburon at 9:52 PM on June 23, 2012