The regression coefficient is simply very small. If you're having trouble wrapping your mind around the scientific notation (e.g., -5.29e-08), just copy and paste the numbers into Microsoft Excel, right-click on the cells and then on "Format Cells". Under "Category", choose "Number" and select 10 or so decimal places. Then you can see what the numbers look like in "regular" notation (the regression coefficient point estimate is equal to -0.0000000529). The coefficient is very small, indicating that your DV is expected to change very little with increasing X scores. This makes sense given that your X variable is not significant at the .05 level and R squared is very small also.

Note that for a nested design (observations nested within countries), multilevel regression analysis is likely more appropriate given that nesting affects the standard errors and test statistics and can lead to Type-1 error inflation (i.e., incorrectly rejecting the null hypothesis of no effect more often than expected based on the nominal alpha level).

or divide x by 100,000. btw, if you have questions about whether the approach is correct, you'll probably want to start another thread with more info about what you mean with the country groups.

Just one thing I might add. Given the DV is an index between two values least squares is probably not appropriate to use. See the attached search for details. Best wishes David Booth

Why would you think the sample is too small? The most important question is: what kind of relationship do you logically expect between your x and y, and does the regression conform this?

If you can share the raw data, we could have a look...