Ouch,

If the pipet is not graduated the error could be enormas. So I will assume graduation: you would then assume linarity in the error, and work 7.5 * 0.6%. Next ask what terms the error is expressed in in terms of statistics. The confidence levels are generally 67%, 95 or 99.7% confidence levels (+/- 1 sig, +/- 2 sig or +/- 3 sigma).

As a retired quality control manager I find that most undereducated people instinctively give +/- 2 sigma numbers when asked about variation. So off the top of my head, I would say you have been given insufficent information to give an accurate answer in terms of probability or confidence level. If the 0.6% is 1 sigma, you would say that at the 95% confidence level the error would be less than or equal to +/- (7.5 * 0.006 * 2 sigma) or 0.09 ml which is 1.2%. That seems awlful small to me for the pipets I have used, so let me go back to a non-graduated pipet solution and say that the error cannot be determined, and you should use the correct size pipet for the experiment.

David