27.10.2014 Views

Russel-Research-Method-in-Anthropology

Russel-Research-Method-in-Anthropology

Russel-Research-Method-in-Anthropology

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Multivariate Analysis 695<br />

just because you th<strong>in</strong>k they might come <strong>in</strong> handy. Of course, you can overdo it.<br />

There is noth<strong>in</strong>g more tedious than an <strong>in</strong>terview that drones on for hours without<br />

any obvious po<strong>in</strong>t other than that the researcher is gather<strong>in</strong>g data on as many<br />

variables as possible.<br />

3. The source of ideas has no necessary effect on their usefulness. You can get ideas<br />

from an exist<strong>in</strong>g theory or from brows<strong>in</strong>g through data tables—or from talk<strong>in</strong>g<br />

about research problems with your friends. The important th<strong>in</strong>g is not how you<br />

get a hunch, it’s whether you can test your hunches and create plausible explanations<br />

for whatever f<strong>in</strong>d<strong>in</strong>gs come out of those tests. If others disagree with your<br />

explanations, then let them demonstrate that you are wrong, either by reanalyz<strong>in</strong>g<br />

your data or by produc<strong>in</strong>g new data. But stumbl<strong>in</strong>g onto a significant relation<br />

between some variables does noth<strong>in</strong>g to <strong>in</strong>validate the relation.<br />

So, when you design your research, try to th<strong>in</strong>k about the k<strong>in</strong>ds of variables<br />

that might be useful <strong>in</strong> test<strong>in</strong>g your hunches. Use the pr<strong>in</strong>ciples <strong>in</strong> chapter 5<br />

and consider <strong>in</strong>ternal state variables (e.g., attitudes, values, beliefs); external<br />

state variables (e.g., age, height, gender, race, health status, occupation, wealth<br />

status); physical and cultural environmental variables (e.g., ra<strong>in</strong>fall, socioeconomic<br />

class of a neighborhood); and time or space variables (Have attitudes<br />

changed over time? Do the people <strong>in</strong> one community behave differently from<br />

those <strong>in</strong> another otherwise similar community?).<br />

In applied research, important variables are the ones that let you target a<br />

policy—that is, focus <strong>in</strong>tervention efforts on subpopulations of <strong>in</strong>terest (the<br />

rural elderly, victims of violent crime, overachiev<strong>in</strong>g third graders, etc.)—or<br />

that are more amenable to policy manipulation (knowledge is far more manipulable<br />

than attitudes or behavior, for example). No matter what the purposes<br />

of your research, or how you design it, the two pr<strong>in</strong>ciple rules of data analysis<br />

are:<br />

1. If you have an idea, test it.<br />

2. You can’t test it if you don’t have data on it.<br />

So, don’t be afraid to play and have a good time with data analysis. If you<br />

hang around people who use complex statistical tools <strong>in</strong> their research, you’ll<br />

hear them talk<strong>in</strong>g about ‘‘massag<strong>in</strong>g’’ their data, ‘‘teas<strong>in</strong>g out signals’’ from<br />

their data, and ‘‘separat<strong>in</strong>g the signals from the noise.’’ These are not the sorts<br />

of phrases used by people who are bored with what they’re do<strong>in</strong>g.<br />

Enjoy.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!