The calculated lifetimes are different from what I expected
Are you using the right fluorescence filter cube?
Are you using the right reference solution and reference lifetime for calibrating the system? Did you change anything in the optical setup between taking a Reference and a Sample? Please note that each optical setup requires its own reference calibration; i.e. switching dichroics, filters, objectives, light source, magnifications, or power light levels changes the path lengths inside the system. The same holds when using a different MCP gain for the intensified CCD camera. Changing ND filters or the exposure time does not require a different reference calibration.
Is there reason to believe that the (photophysics of) your sample is as expected?
Is there auto-fluorescence influencing the fluorescence lifetime distribution?
In case your application is FRET; is there leak-through from your acceptor in the donor channel?