That would be correct; increasing flavor concentration (percentage) would/will have a lesser impact:
In this case, it calls for adding only 1.702mL of flavor, to get the flavor percentage from 2% to 6%. This is a 4.26% increase (1.702mL divided by 40mL = 0.04255 or 4.255%) in total volume. A 4.26% dilution of 12mg/mL nic is only 0.5112mg/mL (4.26% of 12mg/mL = 0.5112mg.mL). So, this leaves a final nic level of slightly less than 11.5mg/mL (12mg/mL minus 0.5112 = 11.4888mg/mL). Maybe some folks would detect that small of a loss; but I would not be able to.
However, because @KEn Nicklowicz' example was fixing an over-flavored recipe, it has a much greater impact. Because, instead of a 4.26% dilution, there is a 66.66% dilution (12mg/mL -->11.49mg/mL vs. 12mg/mL --> 4mg/mL). That I would notice!
This example provides a good visual why it can be better to start with low flavor percentages... and tweak upward; vs. starting high and having to tweak downward. It results in a lot more volume (potentially wasted material).
It also demonstrates why small test batches are beneficial. If one starts with a 40mL sample size... and makes an over-flavoring error (as in the example), all of a sudden you can have 4 ounces (120mL) of something that still may not be enjoyed, even though it is now "correct." I personally find 10-15mL samples to be plenty.
