Relationship between two variables help!

[Charles Crabtree]

I'm brand new to STATA and statistics and have been left a little on my own in terms of figuring out how to figure both out.

I have two variables ranked 1-5 in order of importance and am interested in figuring out whether a high ranking (5) of one variable correlates with a low ranking of another. My hypothesis is that a high ranking on variable 1 will correlate with a a low ranking on variable 2 and vice versa.

How would I run this through Stata. Regress v1 v2? Any pointers on interpreting the output would also be appreciated.

[Dick Campbell]

You can correlate the two variables. If you run the regression, ask for standardized coefficients in addition to the usual output. The standardized coefficient in bivariate regression equals the correlation. You also should run a simple cross tabulation of the two variables.

[Charles Crabtree]

So something like this?

[Charles Crabtree]

And the .28 beta would suggest that it's not that predictive right? Or that it's not statistically significant at least?

[Dick Campbell]

The significance test on the unstandardized regression coefficient tells you that it is statistically significant. Whether the relationship is substantively meaningful or not is up to your judgement. I'm not going to interpret your results for you, but I'd suggest that you look carefully at the cross tab. It might useful to ask for row and/or column percents. .