Do you have Excel?. It does a fine job of doing regressions. This is a power regression. You could use an initial population of 9.5 with the year 1960 corresponding to t=0.

Then, 1970 would be t=10 and so forth. For the western population:

Just running it through my TI, I get

In your case, with the 2:

Or, in exponential :

This one has an R^2 of .9972, which is very accurate. The closer to 1 the better. This is close to 1 for sure.

Their statement of "when will it double?". From where?. The initial population or the latest one?.

Try both. When will it double from the latest of 2*23.9. Plug in and solve for t.

We get t=43.53. That would about half way through the year 2003.